A stereoselective synthesis via a 5-exo-trig cyclisation of trans-2-oxohexahydro-2H-furo[3,2-b]pyrroles (pyrrolidine-trans-lactones) - Potent, novel elastase inhibitors

Article abstract of DOI:10.1039/a807540i

trans-2-Oxohexahydro-2H-furo[3,2-b]pyrroles (such as 2) are conformationally strained 5,5-fused ring systems and are potent human neutrophil elastase (HNE) inhibitors. A stereoselective synthesis is described based on intramolecular 5-exo-trig cyclisations of aldehyde acrylates 4 and 5 mediated by samarium(II) iodide to give predominantly trans-products 9 and 15. The n-propyl group in 15 is also generated with stereoselectivity for the desired β-isomer.

Full text of DOI:10.1039/a807540i

View Article Online / Journal Homepage / Table of Contents for this issue
A stereoselective synthesis via a 5-exo-trig cyclisation of trans-2-
oxohexahydro-2H-furo[3,2-b]pyrroles (pyrrolidine-trans-lactones)—
potent, novel elastase inhibitors
Simon J. F. Macdonald,* Keith Mills,a Julie E. Spooner, Richard J. Uptonb and
Michael D. Dowle
Enzyme Chemistry 2, Development Chemistry 3,a Physical Sciences,b
GlaxoWellcome Medicines Research Centre, Gunnels Wood Road, Stevenage, UK SG1 2NY
Received (in Cambridge) 28th September 1998, Accepted 9th October 1998
trans-2-Oxohexahydro-2H-furo[3,2-b]pyrroles (such as 2) are conformationally strained 5,5-fused ring systems
and are potent human neutrophil elastase (HNE) inhibitors. A stereoselective synthesis is described based on
intramolecular 5-exo-trig cyclisations of aldehyde acrylates 4 and 5 mediated by samarium() iodide to give
predominantly trans-products 9 and 15. The n-propyl group in 15 is also generated with stereoselectivity for
the desired β-isomer.
O
Introduction
CO2H
R
CO2Et
Prn
Prn
Human neutrophil elastase (HNE)1 is a serine protease stored
in the azurophilic granules of neutrophils. It is released in
response to inflammatory stimuli and has a major role in pro-
tein digestion following phagocytosis. Excessive elastase release
has been implicated in respiratory diseases such as acute
respiratory distress syndrome, cystic fibrosis, emphysema1 and
chronic bronchitis.2 Inhibitors of HNE are actively being
investigated3 as potential therapies for these diseases.
O
OH
CBZN
CBZN
CBZN
O
4 R = Prn
2
3
5 R = H
Scheme 1
As part of our programme to develop low molecular weight
non-peptide inhibitors of HNE for the treatment of chronic
bronchitis, we are pursuing compounds developed from a lead
triterpene identified by high throughput screening.4 Our initial
development of this lead led us to strained cyclopentane-trans-
lactones typified by 1 which are micromolar inhibitors of
potential reactivity of the strained trans-lactone led us to post-
pone lactone formation to the end of the synthesis: hence the
initial disconnection to the hydroxy acid 3. This in turn is dis-
connected via the 5-exo-trig cyclisation to the aldehyde 4 which
we envisaged being available readily from ethyl propiolate and
3-aminopropanal diethyl acetal. From previous structure–
activity relationships,5,6 it was clear that the β-stereochemistry
of the propyl group (i.e. drawn up as for 2) gave increased
inhibitory activity against HNE when compared with the
α-propyl group. Thus stereocontrol in the cyclisation was
critical to the success of this approach.
O
O
The literature contains examples of intramolecular reductive
cyclisations of aldehydes with electron deficient olefins for the
preparation of cyclopentanes by electro-reduction,7 treatment
1
9,10
11
with Bu3SnH,8 SmI2
or VCl2 and for the preparation of
HNE.5 These compounds inhibit HNE by acylating the nucleo-
philic hydroxy of serine-195† in the active site of the enzyme.
To enhance the potency of these inhibitors we designed, with
the aid of computer modelling, a second generation of inhibi-
tors based on trans-oxohexahydro-2H-furo[3,2-b]pyrroles,
called colloquially pyrrolidine-trans-lactones 2. We believe such
compounds were unknown until our studies;6 they now repre-
sent an important new class of heterocycles characterised by
their inhibitory activity against serine proteases including
HNE, chymotrypsin, cathepsin G and thrombin. We attribute
their potency to a combination of conformational ring strain
(from the trans-fused ring system) and molecular fit in the
active site of the enzyme.6 We describe here a synthetic
approach which gives the desired stereocontrol over the three
contiguous stereocentres via a samarium iodide mediated 5-
exo-trig cyclisation.
oxacycles10,12,13 and pyrrolidines.10,14–16 However at the time of
this work we were unaware of any precedent for reductive cyclis-
ations of aldehydes onto urethane acrylates (as in 4 or 5) and
only a single example (Scheme 2)11 with substitution α to the
Me
CO2Et
OH
Me
CO2Et
ref. 11
77%
CHO
isomer ratio 2.7:1
Scheme 2
ester (as in 4). Since this work was completed, an example of
this type of cyclisation has been described for the formation of
a piperidine.14
The retrosynthesis is shown in Scheme 1. Concern for the
Results
We explored initially the des-propyl series (Scheme 3). The
† Numbered according to chymotrypsin.
J. Chem. Soc., Perkin Trans. 1, 1998, 3931–3936
3931

View Article Online
1'
CO2Et
CO2Et
2
3
Prn
R1
CO2Et
Prn
CO2Et
RN
OEt
OEt
a
c
CBZN
e
a
H2N(CH2)2CH(OEt)2
CO2Et
10
1''
R2
6
7 R = H, 2:1 Z:E
8 R = CBZ E only
14 R2 = CH(OEt)2
11 R1 = H
12 R1 = Br
b
d
b
4
R2 = CHO
CO2Et
OH
CO2Et
O
O
CO2R3
OH
c
Prn
Prn
Prn
d
CBZN
O
O
3
3
CBZN
+
3
6a
6a
3a
3a
O
g
9 trans only
CBZN
CBZN
CBZN
5
Prn β:α 3.5:1
15:16 4.3:1
16
2
᎐
Scheme 3 Reagents and conditions: (a) HC᎐CCO2Et, Et2O, rt, 3 days,
᎐
Prn β:α 3.1:1
100% crude; (b) n-BuLi, THF, Ϫ78 ЊC, 30 min, then CBZCl, warm to rt,
45 min, 100% crude; (c) TsOHؒpy, H2O, Me2CO, reflux, 7 h, 30%;
(d) SmI2, THF, HMPA, MeOH, 0 ЊC, 90 min, 53%.
15 R3 = Et
R3 = H
f
3
Scheme 4 (all compounds racemic) Reagents and conditions: (a) KOH,
EtOH, rt, 17 h; c HCl; (HCHO)n, pyridine, piperidine, reflux, 90 min,
56%; (b) Br2, CHCl3, AcOH, 5 ЊC, 2 h, 100% crude; then DBU, THF,
reflux, 75 min, 60%; (c) CBZNH(CH2)2CH(OEt)2 13, NaH, DMF, rt,
2.5 h, 15%; (d) TsOHؒpy, H2O, Me2CO, reflux, 3 h, 100% crude; (e)
SmI2, THF, MeOH, HMPA, 5 ЊC, 90 min, 48%; ( f ) LiOH, EtOH, H2O,
reflux, 6 h, 100% crude; (g) 2,4,6-trichlorobenzoyl chloride, Et3N,
CH2Cl2 rt, 1 h; then DMAP, PhMe, reflux 3.5 h, 35%.
commercially available 3-aminopropanal diethyl acetal 6 was
readily converted into the protected aldehyde 8 by 1,4-addition
with ethyl propiolate17 giving a 2:1 E:Z ratio of olefins 7 which
upon deprotonation with n-butyllithium‡ and treatment with
benzyl chloroformate gave only the (E)-product 8 with an
olefinic coupling constant J = 14.5 Hz. Deprotection of
the acetal 8 with pyridinium toluene-p-sulfonate in aqueous
acetone gave the stable crystalline aldehyde 5. Reductive cyclis-
ation with excess samarium iodide in the presence of HMPA
and using methanol as the proton source, gave a 53% yield of
racemic trans-hydroxy ester 9.§ Neither the cis-isomer of
hydroxy ester 9 nor the anticipated more stable product the cis-
lactone, were observed. Other conditions such as Bu3SnH,8
VCl2,11 TiCl3 and a Zn–Cu couple,18 zinc dust and TMSCl,19 or
Mg in MeOH13 failed to give significant, if any, trans-hydroxy
ester 9.
1
chemistries of the β- and α-isomers were determined by H-
NMR, and by correlation of the corresponding β- and α-allyl
isomers which were separable and available via a different
route.6 We believe there is minimal epimerisation of the carbon
bearing the n-propyl in transforming 15 to the trans-lactone 2.
1
The H-NMR spectra of 16, 2β and 2α are complicated by
peak broadening due to cis–trans isomerisation of the benzyl
carbamates. However warming to 50 ЊC (or higher) sharpens
many of the resonances and allowed structural assignments
based on NOE, COSY and HMQC experiments. Table 1 shows
clearly the difference in the diagnostic coupling constant
between H6a and H3a in the cis-lactone 16 (5 Hz) and in the
trans-lactones 2β and 2α (10 Hz). These values are in good
agreement with those predicted by Macromodel.23 Additional
evidence for a difference in ring fusion stereochemistry is that
a strong NOE is observed between H6a and H3a in the cis-
lactone 16 where none is observed between H6a and H3a in 2β
or 2α (Table 1).
We next investigated cyclisation of the propyl aldehyde 4
(Scheme 4). Diethyl propylmalonate 10 was converted into the
propylacrylate 11 under standard conditions.20 Dibromination
followed by dehydrobromination with DBU in THF gave the
(E)-bromoacrylate 12 in 60% yield after distillation.¶ This was
subjected to an addition–elimination reaction with the sodium
salt of benzyl carbamate 13 to give a 15% yield of the unsatur-
ated acetal 14.¶ Alternative conditions [PdCl2(MeCN)2, CuCl,
CH2C(Me)CO2Me,13 DME] for a similar transformation failed
to give any identifiable products.21 This acetal 14 was de-
protected smoothly to afford the crude (E)-acrylate aldehyde 4.
The (E)-geometry was determined by an NOE between H1Ј
and H1Љ. No NOE was observed between H1Ј and H3. This
acrylate aldehyde 4 underwent a SmI2 mediated cyclisation to
give a 48% yield of trans-hydroxy ester 15 (Prn β:α = 3.8:1) and
an 11% yield of the β-propyl cis-lactone 16 (all racemic prod-
ucts). A repeat cyclisation on a larger scale gave a 62% yield
The coupling constants between H3a and H3 in these
systems are also of diagnostic use in differentiating between the
β and α n-propyl groups (12 and 7 Hz respectively). These are
Table 1 Key NOE’s and observed (obs.) and predicteda (pred.) 1H
coupling constants (Hz) for the cis- and trans-lactones 16, 2β and 2α
O
O
O
H
H
1
3
of 15 with similar stereoselectivity. The H-NMR spectra of
O
O
O
3
3a
3
the ester 15 correlate with those of a sample (Prn β:α = 1:1.5)
prepared by a different route (unpublished results) giving con-
fidence in the stereochemical assignment of this material.
Saponification of the ethyl ester 15 under standard condi-
tions, followed by a Yamaguchi cyclisation22 of the resulting
hydroxy acid 3 under dilute conditions, gave the desired trans-
lactone 2 (35% yield) as a 3.1:1 β:α mixture of isomers which
were not separated readily by flash chromatography. The stereo-
3a
6a
H
H
H
H
H
H
6a
6a
3a
H
H
CBZN
CBZN
CBZN
H
H
16
2β
2α
16
2β
2α
Obs.
Pred.
Obs.
Pred.
Obs.
Pred.
J3a–6a
J3a–3
5
5.3
0.6
10
12
9.9
11.9
10
7
9.9
6.7
‡ The sodium salt of 7 is highly insoluble in DMF. The lithium salt is
preferred due to its solubility in THF.
<2b
a Predicted values were calculated using Macromodel (ref. 23). b Line
broadening due to cis–trans isomerisation of the carbamate prevented
accurate determination of this coupling constant even after warming.
§ This material 9 is identical to that prepared by other routes (ref. 6 and
unpublished results).
¶ The olefin geometry is supported by the lack of an NOE between the
olefinic proton and the n-propyl protons.
3932
J. Chem. Soc., Perkin Trans. 1, 1998, 3931–3936

View Article Online
CO2Et
OH
R
OEt
O•
OEt
I2SmO
N
I2SmO
CBZ
2
2
1'
O
CBZ
3
2
R
N
R
•
CBZN
3
H
3
H
β
A
B
R
I2SmO
N
CO2Et
2
1'
R
OH
O•
CBZ
OEt
2
3
3
H
CBZN
C
α
OEt
H
OEt
H
2
1'
2
O•
CBZ
CO2Et
OH
O
CBZ
R
N
I2SmO
R
N
R
3
•
3
I2SmO
D
2
3
E
CBZN
intramolecular
protonation
α
CO2Et
H
R
2
3
R
OH
CBZ
1'
O•
N
2
3
OEt
I2SmO
CBZN
β
F
Fig. 1 Possible transition states for the 5-exo-trig cyclisations of 4 and 5. Unless shown otherwise protonation of B, C, E and F is assumed to occur
from the most open (exo) face (from the back face as drawn). The cis-products from E and F are shown as hydroxy esters, although in fact only the cis-
lactone 16 was isolated.
also in good agreement with those predicted by Macromodel.23
Again, NOE’s between H3 and H6a for 2β and between H3 and
H3a for 2α confirm the propyl stereochemistries. The difference
in coupling between H3a and H3 in 16 (<2 Hz) and 2β (12 Hz)
illustrates some of the structural differences between cis and
trans-lactones.
transition state A (rather than D) may be explained in terms
of minimising electronic repulsion between the developing
methylene radical centre and the negatively charged aldehyde
oxygen.9b,11,26,27 Our results show the preference for A over D is
less pronounced for R = propyl than R = H which is not easily
rationalised.
Following cyclisation, the resultant radical species is then
proposed to undergo further reduction by SmI2 to give an
anionic species (not shown) which is then rapidly protonated.27
We assume that the stereochemistry of the n-propyl group is
either determined by the facial preference for O to C proton
transfer in the ketene acetal species or the facial preference for
protonation of C1Ј. The results suggest that B is preferred to C.
If one assumes that protonation occurs from the more open
face (exo), this can be rationalised as minimising interactions
between the OEt and H3 (as shown in C) leading to a preference
for the β-propyl product. However application of this argument
for preferring F over E is less clear cut. The F intermediate
leads to the thermodynamically more stable β-propyl product
(the product that is actually isolated) but is the less preferred
conformation because the OEt is underneath the pyrrolidine
ring. If however C3 alkoxide protonation occurs first, then an
intramolecular protonation from the C3 hydroxy may occur.
This would appear to be more facile in E than F and would lead
to the desired β-propyl product (Fig. 1).
Discussion
Although there have been considerable advances in the use of
samarium() iodide mediated transformations, there is still only
a hazy understanding of the mechanistic details involved.24
However some suggestions and comments can be made. The
formation of 5-membered carbocycles by intramolecular reduc-
tive cyclisation of aldehydes onto electron deficient (E)-olefins
are known to give predominantly trans products and the olefin
geometry (E or Z) has been noted as playing a critical role
in affecting the cis- or trans-geometry of the products.11,13 As
mentioned previously however, there is little precedent for
stereocontrol of substituents α to the ester (Scheme 2).11
The various transition states for the SmI2 cyclisations of 4
and 5 are depicted in Fig. 1. These have been drawn as radical
intermediates although they could be anionic intermediates.
However as these radicals behave as nucleophiles, similar steric
or electronic effects should apply.25 The preference for the
formation of trans-products (across C2–C3) 9 and 15 from
There are distinct physicochemical and biochemical differ-
J. Chem. Soc., Perkin Trans. 1, 1998, 3931–3936
3933

View Article Online
ences6 between the trans-lactone 2 and the cis-lactone 16 (vide
supra). The pyrrolidine trans-lactones 2 are characterised by
the infra-red lactone carbonyl stretching frequency at νmax
1795 cmϪ1 in contrast to that of the cis-lactone 16 at 1778
cmϪ1.6 This may indicate increased ring strain. Additionally, the
β-propylpyrrolidine trans-lactone of 2β is a potent inhibitor of
HNE with an IC50 of 45 nM (after a 15 min preincubation) in
contrast to 16 which is inactive (up to 100 µM).6 This is entirely
consistent with the different molecular shapes of 2 and 16; 2
exists in a ‘pseudo-planar’ conformation and 16 in an ‘envelope’
conformation. These molecular shapes explain the lack of
‘molecular fit’ of 16 in the active site of HNE. The difference in
ring strain is also believed to contribute to the acylating power
of the trans-fused molecule over the cis-fused analogue.
We have briefly explored these 5-exo-trig cyclisations for
the preparation of the analogous pyrrolidine trans-lactams.6
Pyrrolidine products were not obtained although some recent
protocols from Aurrecoechea and Fernandez–Acebes28 have
not been investigated.
hexane 1:1 (8 cm3) at 5 ЊC followed by filtration and washing of
the filter cake with Et2O (5 cm3) gave the aldehyde 5 (0.795 g,
30%) as a white solid. (Further material was available from the
residue by chromatography.) (Found: C, 63.2; H, 6.4; N, 4.5.
C18H19NO5 requires C, 62.9; H, 6.3; N, 4.6%); TLC Rf = 0.24
(3:1 hexane–Et2O); δH(CDCl3, 250 MHz) 9.78 (1 H, s, CHO),
8.18 (1 H, d, J 14.5, NCHCH), 7.37 (5 H, br s, Ph), 5.26 (2 H, s,
PhCH2), 5.20 (1 H, d, J 14.5, NCHCH), 4.19 (2 H, q, J 6.5,
CO2CH2CH3), 3.93 (2 H, t, J 7, NCH2), 2.76 (2 H, t, J 7, 2H,
CH2CHO), 1.28 (3 H, t, J 6.5, CO2CH2CH3); νmax (Nujol mull)/
cmϪ1 2924, 2853, 1741, 1716, 1689, 1634, 1454 and 1411.
trans-3-Hydroxy-2-ethoxycarbonylmethylpyrrolidine-1-carb-
oxylic acid benzyl ester 9
To samarium() iodide in THF (0.1 M) (10 cm3, 1 mmol) and
methanol (2 drops) at Ϫ5 ЊC was added with stirring the alde-
hyde 5 (50 mg, 0.164 mmol) in THF (1 cm3) over 5 min. After 30
min, TLC showed no reaction. HMPA (CARCINOGEN) (0.2
cm3) was then added causing the deep blue–green colour to
rapidly change to purple and then to colourless. After a further
30 min TLC showed almost complete consumption of starting
material. After addition of 2 M HCl (10 cm3), the crude prod-
uct was extracted with EtOAc (3 × 10 cm3). The combined
extracts were washed with brine (10 cm3), 5% sodium thio-
sulfate solution (10 cm3), brine (2 × 10 cm3) and dried (MgSO4).
Solvent removal in vacuo followed by flash chromatography on
silica (9385 Merck) eluting with 80:20 ether–hexane afforded
recovered starting material 5 (0.009 g) and the desired product 9
(0.022 g, 44%; 53% based on consumed starting material). This
material was identical to samples prepared by an alternative
route;6 TLC Rf = 0.40 (Et2O); δH(CDCl3, 250 MHz) 7.40–7.24
(5 H, m, Ph), 5.18–5.10 (2 H, m, PhCH2), 4.26 (1 H, br s,
NCHCHOH), 4.20–3.99 (3 H, m, CHOH, OCH2CH3), 3.78–
3.58 (1 H, m, NHCH), 3.51–3.40 (1 H, m, NHCH), 3.12 (0.6 H,
dd, J 16 and 3, HCHCO2Et), 2.87 (0.4 H, dd, J 16 and 3,
HCHCO2Et), 2.76 (0.6 H, s, OH), 2.55 (0.4 H, s, OH), 2.36
(0.6 H, d, J 16, HCHCO2Et), 2.20 (0.4 H, d, J 16, HCHCO2Et),
2.06 (1 H, m, NCH2HCH), 1.92 (1 H, m, NCH2HCH), 1.26
(3 H, t, J 7, CO2CH2CH3); νmax(thin film)/cmϪ1 3443, 2978,
1731, 1701 and 1415.
Experimental
General methods
Unless otherwise indicated, δH NMR spectra were recorded
on a 250 MHz Bruker spectrometer. 400 MHz and 750 MHz
spectra were recorded on a Varian VXR 400 and Inova 750
MHz Varian spectrometers respectively. Spectra are referenced
internally to residual proton solvent signals. Data for 1H NMR
are reported as follows: chemical shift (δ ppm), integration,
multiplicity (s = singlet, b = broad, m = multiplet, d = doublet,
t = triplet, q = quartet), coupling constant (J in Hz), and
assignment. Proton assignments were aided by COSY, HMQC,
NOE and ROESY experiments. Infra-red spectra were recorded
on a Perkin-Elmer 1700 series spectrometer. Thermospray mass
spectra were recorded on an HP5989B Engine, using ammonia
as carrier gas, with the filament on and in positive ion mode.
Electrospray mass spectra were recorded on a Micromass
Autospec Voltage Spec in positive ion mode using PEG/PEG-
NH3 conditions. Thin layer chromatography (TLC) was per-
formed on Polygram Sil G/UV254 pre-coated plastic sheets and
visualised by UV or potassium permanganate in dilute aqueous
sodium carbonate. Flash chromatography was carried out on
Merck 9385 silica gel 60.
(E)-3-Bromo-2-propylacrylic acid ethyl ester 12
A solution of bromine (2.26 cm3, 43.75 mmol) in chloroform
(20 cm3) was added over 2 h to a solution of the ethyl 2-
methylenepentanoate 11 (6.0 g, 42.0 mmol) (prepared according
to ref. 20) in chloroform (15 cm3) and acetic acid (3 cm3) at 5 ЊC.
After a further 2 h, the mixture was diluted with 2 M aqueous
sodium carbonate (200 cm3) and sufficient solid sodium metabi-
sulfite to decolourise the mixture. This was then extracted with
chloroform (2 × 40 cm3). After drying (MgSO4), concentration
in vacuo gave a colourless oil (13.0 g). The oil (13.0 g, 43 mmol)
was dissolved in dry THF (45 cm3) and DBU (11.7 cm3, 78
mmol) in THF (30 cm3) was added over 5 min. The mixture was
then heated at reflux for 75 min, cooled, and then diluted with
water (500 cm3) and 2 M hydrochloric acid (25 cm3). After
extraction with hexane (2 × 50 cm3), the combined extracts
were washed with brine (2 × 100 cm3), dried (MgSO4) and
evaporated in vacuo to afford a yellow oil (7.8 g). This was
distilled in vacuo to give the bromoacrylate 12 (5.75 g, 60%),
bp 58.5–61 ЊC/4 mm Hg (Found: C, 43.7; H, 6.1. C8H13BrO2
requires C, 43.45; H, 5.9%); δH(CDCl3, 250 MHz) 7.51 (1 H, s,
BrCH), 4.21 (2 H, q, J 6.5, CO2CH2CH3), 2.46 (2 H, t, J 8,
CH2CH2CH3), 1.50 (2 H, sextet, J 8, CH2CH2CH3), 1.31 (3 H, t,
J 6.5, CO2CH2CH3), 0.96 (3 H, t, J 8, CH2CH2CH3); νmax(thin
film)/cmϪ1 2964, 1719, 1607 and 1464.
(E)-3-[Benzyloxycarbonyl(3-oxopropyl)amino]acrylic acid ethyl
ester 5
The enamine 7 (2:1 Z:E) was prepared according to ref. 15.
The enamine 7 (24.17 g, 98.5 mmol) was dissolved in THF (200
cm3) under N2 and cooled to Ϫ70 ЊC. BunLi (1.6 M in hexanes)
(62 cm3, 99.3 mmol) was added dropwise over 90 min to give a
yellow–green solution. After 30 min, benzyl chloroformate
(14.2 cm3, 99.5 mmol) in THF (20 cm3) was added over 60 min
and the cooling bath removed immediately and the reaction
allowed to warm to room temperature. After 45 min TLC
showed absence of starting material. The reaction was
quenched with saturated NH4Cl solution (200 cm3) and
extracted with ethyl acetate (3 × 150 cm3). The combined
extracts were washed with brine (100 cm3), dried (MgSO4) and
concentrated in vacuo to afford (E)-3-[benzyloxycarbonyl(3,3-
diethoxypropyl)amino]acrylic acid ethyl ester 8 as a yellow oil
(41.68 g). A portion of this crude carbamate 8 (3.27 g, 8.62
mmol), pyridinium toluene-p-sulfonate (2.17 g, 8.62 mmol),
water (10 cm3) and acetone (40 cm3) were heated at reflux for
7 h. The acetone was removed in vacuo and the residue diluted
with EtOAc (50 cm3) and water (40 cm3). The aqueous layer was
separated and extracted with EtOAc (2 × 25 cm3). The com-
bined organic extracts were washed with brine (25 cm3), dried
(MgSO4) and the solvent removed in vacuo to afford an orange
oil which solidified on standing. Trituration with EtOAc–
(E)-3-[Benzyloxycarbonyl(3,3-diethoxypropyl)amino]-2-propyl-
acrylic acid ethyl ester 14
To the benzyl carbamate 13 (0.560 g, 2 mmol) and bromo-
3934
J. Chem. Soc., Perkin Trans. 1, 1998, 3931–3936

View Article Online
acrylate 12 (0.480 g, 2.2 mmol) in dry DMF (7 cm3) under N2
was added sodium hydride (60% in oil) (0.090 g, 2.2 mmol).
After 40 min, evolution of hydrogen was complete and the solu-
tion was orange–brown. After a further 2 h the mixture was
diluted with water (100 cm3) and brine (50 cm3). Citric acid (2 g)
was added. Extraction with EtOAc (2 × 60 cm3) followed by
washing of the combined extracts with brine (3 × 50 cm3) and
drying (MgSO4) and concentration in vacuo gave a pale yellow
oil (0.9 g). This was purified by flash column chromatography
on silica gel eluting with hexane–ether 2:1 to afford the desired
product 14 (0.130 g, 15%); TLC Rf = 0.48 (1:1 hexane–Et2O);
δH(CDCl3, 400 MHz) 7.64 (1 H, s, NCH), 7.40–7.30 (5 H, m,
Ph), 5.21 (2 H, s, PhCH2), 4.49 (1 H, t, J 5, CH(OEt)2), 4.19
(2 H, q, J 7, CO2CH2CH3), 3.68 (2 H, m, NCH2CH2), 3.60 (2 H,
m, 2H, CH(OCH2CH3)2), 3.43 (2 H, m, CH(OCH2CH3)2), 2.28
(2 H, m, CH2CH2CH3), 1.92 (2 H, m, NCH2CH2), 1.45 (2 H, m,
CH2CH2CH3), 1.28 (3 H, t, J 7, CO2CH2CH3), 1.08 (6 H, t, J 7,
CH(OCH2CH3)2), 0.91 (3H, t, J 7, CH2CH2CH3); νmax(KBr
diffuse reflectance)/cmϪ1 2963, 2934, 2882, 1724, 1709 and
Data for the â-propyl cis-lactone 16. TLC Rf = 0.61 (Et2O);
the resonances for the two rotamers are described separately for
clarity; δH(CDCl3, 400 MHz, 30 ЊC) rotamer 1: 7.40–7.30 (5 H,
br s, Ph), 5.18 (1 H, d, J 12, PhHCH), 5.10 (1 H, d, J 12,
PhHCH), 5.04 (1 H, ddd, J 5, 5 and 1, NCH2CH2CH), 4.24
(1 H, dd, J 5 and <2, NCH), 3.80 (1 H, m, NHCH), 3.41 (1 H,
m, NHCH), 2.86 (1 H, ddd, J 7, 7 and <2, PrCHCO), 2.31
(1 H, ddd, J 14, 6 and 1, NCH2HCH), 2.01(1 H, br s, NCH2-
HCH), 1.80–1.43 (2 H, m, CH2CH2CH3), 1.38–1.25 (2 H, m,
CH2CH2CH3), 0.82 (3 H, br t, CH2CH2CH3); rotamer 2: 7.40–
7.30 (5 H, s, Ph), 5.18 (1 H, d, J 12, PhHCH), 5.10 (1 H, d, J 12,
PhHCH), 5.04 (1 H, ddd, J 5, 5 and 1, NCH2CH2CH), 4.24
(1 H, dd, J 5 and <2, NCH), 3.89 (1 H, m, NHCH), 3.41 (1 H,
m, NHCH), 2.67 (1 H, ddd, J 7, 5 and <2, PrCHCO), 2.31
(1 H, ddd, J 14, 6 and 1, NCH2HCH), 2.01 (1 H, br s, NCH2-
HCH), 1.80–1.43 (2 H, m, CH2CH2CH3), 1.38–1.25 (2 H, m,
CH2CH2CH3), 0.99 (3 H, br t, CH2CH2CH3); νmax(KBr diffuse
reflectance)/cmϪ1 2961, 2940, 2865, 1778, 1704 and 1416; m/z
ϩ
(thermospray, NH3) MNH4 at 321 (30%); m/z (thermospray,
ϩ
1637; m/z (thermospray NH3) MNH4 439 (20%), M Ϫ OEt
NH3) Found: MHϩ: 304.154. C17H22NO4 requires 304.154 (0.2
ppm error).
371 (20); m/z (thermospray, NH3) [Found: M Ϫ OEtϩ, 376.211.
C21H30NO5 requires 376.212 (1.3 ppm error)].
Data for the propyl ester 15. TLC Rf = 0.52 (Et2O); δH(CDCl3,
400 MHz, 30 ЊC) 7.32–7.20 (5 H, m, Ph), 5.20–4.98 (2 H, br m,
PhCH2) 4.42–4.24 (1 H, br d), 4.04 (2 H, q, J 7), 3.96–3.84 (1 H,
br m), 3.74–3.54 (1 H, br m), 3.38–3.24 (1 H, br m), 3.17–2.43
(1 H, br m), 2.07–1.90 (1 H, br m), 1.83–1.74 (1 H, br m), 1.74–
1.43 (3 H, br m), 1.38–1.05 (5 H, br m), 0.87–0.66 (3 H, br m);
(E)-3-[Benzyloxycarbonyl(3-oxopropyl)amino]-2-propylacrylic
acid ethyl ester 4
A mixture of the acetal 14 (0.048 g, 0.12 mmol), pyridinium
toluene-p-sulfonate (0.050 g, 0.20 mmol), water (0.5 cm3) and
acetone (5 cm3) were stirred at reflux for 3 h under nitrogen.
After removal of most of the acetone in vacuo, the mixture was
diluted with brine (25 cm3) and 2 M sulfuric acid (2 cm3) and
extracted with EtOAc (2 × 20 cm3). The combined extracts were
washed with 8% NaHCO3 solution (10 cm3) and brine (15 cm3),
dried (MgSO4) and the solvent removed in vacuo to afford the
aldehyde 4 as a colourless oil (0.040 g, 100%). A small sample
for analysis was purified by flash chromatography on silica gel
9385 eluting with 30:70 ether–hexane; TLC Rf = 0.24 (1:1
ether–hexane); δH(CDCl3, 400 MHz) 9.77 (1 H, t, J 1.5, CHO),
7.54 (1 H, s, NCH), 7.38–7.31 (5 H, m, Ph), 5.20 (2 H, s,
PhCH2), 4.20 (2 H, q, J 7, CO2CH2CH3), 3.90 (2 H, t, J 7,
NCH2), 2.79 (2 H, m, CH2CHO), 2.21 (2 H, m, CH2CH2CH3)
1.44 (2 H, m, CH2CH2CH3), 1.28 (3 H, t, J 7, CO2CH2CH3),
0.90 (3 H, t, J 7.5, CH2CH2CH3); νmax(KBr diffuse reflectance)/
cmϪ1 2975, 1730, 1715 and 1632; m/z (thermospray, NH3)
Found MHϩ: 348.180. C19H26NO5ؒHϩ requires 348.181 (2 ppm
error).
ϩ
m/z (thermospray, NH3) MNH4 at 350 (95%); analytical
HPLC (system as described above) 15β and 15α co-eluted with
material of a different α:β ratio prepared by a different route
(unpublished results).
trans-2-(1-Carboxybutyl)-3-hydroxypyrrolidine-1-carboxylic
acid benzyl ester 3
To the ester 15 (0.0153 g, 0.04 mmol) in THF (2 cm3) was added
lithium hydroxide monohydrate (0.018 g, 0.4 mmol) in water
(0.2 cm3) and the mixture heated at reflux for 18 h. TLC showed
limited reaction. The THF was removed in vacuo and replaced
with EtOH (2 cm3) and water (1 cm3). Further lithium hydrox-
ide monohydrate (0.010 g) was added. After heating at reflux
for 6 h, TLC showed absence of starting material. The ethanol
was then removed in vacuo. The residue was diluted with 1 M
hydrochloric acid (10 cm3) and extracted with EtOAc (3 × 5
cm3). The combined extracts were washed with brine (5 cm3)
and dried (MgSO4). Solvent removal in vacuo afforded a colour-
less oil (0.015 g, >100%); TLC Rf = 0.29 (streak) (Et2O);
δH(CDCl3, 250 MHz) 7.40–7.22 (5 H, br s, Ph), 5.82–5.30 (1 H,
br s), 5.30–4.97 (2 H, br m), 4.42–4.25 (1 H, br m), 4.18–3.94
(1H, br m), 3.81–3.60 (1H, br m), 3.52–3.30 (1 H, br m), 3.07–
2.84 (0.3 H, br m), 2.59–2.39 (0.7 H, br m), 2.15–1.82 (2 H, br
m), 1.80–1.08 (5 H, br m), 0.97–0.72 (3 H, br m); m/z (thermo-
spray, NH3) MHϩ at 322 (100%).
rel-(2R,3S)-3-Hydroxy-2-[(1S)-1-ethoxycarbonylbutyl]-
pyrrolidine-1-carboxylic acid benzyl ester 15â, rel-(2R,3S)-3-
hydroxy-2-[(1R)-1-ethoxycarbonylbutyl]pyrrolidine-1-carb-
oxylic acid benzyl ester 15á and rel-(3R,3aR,6aR)-3-propyl-2-
oxohexahydro-2H-furo[3,2-b]pyrrole-4-carboxylic acid benzyl
ester 16
To aldehyde 4 (0.038 g, 0.11 mmol) in dry THF (2 cm3) was
added dropwise over 2 min samarium iodide in THF (0.1 M, 8
cm3, 0.8 mmol) at 5 ЊC. After 20 min HMPA (0.05 cm3) was
added and after 90 min TLC showed absence of starting
material. The reaction was quenched with brine (100 cm3), 2 M
HCl (5 cm3) and water (50 cm3) and extracted with EtOAc
(2 × 60 cm3). The combined extracts were washed with brine
(MgSO4). Solvent removal in vacuo gave a yellow oil (56
mg); analytical HPLC, column: Inertsil M at 215 nm, eluting at
1.0 cm3 minϪ1 with 100% “A” and 0% “B” for 2 min, to 100%
“B” for 40 min and 100% “B” for 10 min. Solvent “A” =
H2O ϩ 0.1% H3PO4 and solvent “B” = 95% MeCN–H2O ϩ
0.1% H3PO4), 15β (27.58 min), 15α (27.35 min) and 16 (27.17
min) in a ratio of 3.9:1.1:1.0 plus unidentified material. Flash
chromatography on silica gel 9385 eluting with hexane–Et2O
1:1 then 1:2 afforded the β-propyl cis-lactone 16 (4.3 mg, 11%)
and a mixture of the β- and α-propyl esters 15 (18.5 mg, 48%).
rel-(3R,3aR,6aS)-3-Propyl-2-oxohexahydro-2H-furo[3,2-b]-
pyrrole-4-carboxylic acid benzyl ester 2â and rel-(3S,3aR,6aS)-
3-propyl-2-oxohexahydro-2H-furo[3,2-b]pyrrole-4-carboxylic
acid benzyl ester 2á
The hydroxy acid 3 (0.015 g, 0.047 mmol), triethylamine (10 µl,
0.071 mmol), 2,4,6-trichlorobenzoyl chloride (10 µl, 0.061
mmol) and DCM (5 cm3) were stirred at room temperature in a
stoppered flask. After 1 h, TLC showed absence of starting
material. The reaction mixture was then diluted with PhMe (30
cm3) and added over 1 h to a refluxing solution of DMAP
(0.034 g, 0.28 mmol) in PhMe (25 cm3). After refluxing for a
further 2.5 h, the solvent was removed in vacuo and the residue
treated with 1 M HCl (5 cm3). The mixture was extracted with
EtOAc (3 × 5 cm3). The combined extracts were washed with
brine (2 × 5 cm3), dried (MgSO4) and concentrated in vacuo to
J. Chem. Soc., Perkin Trans. 1, 1998, 3931–3936
3935

View Article Online
E. P. Vacek, J. C. Williams, D. J. Wolanin and S. Woolson, J. Med.
Chem., 1997, 40, 3173; (e) Sterling Winthrop: D. J. Hlasta, J. H.
Ackermann, J. J. Court, R. P. Farrell, J. A. Johnson, J. L. Kofron,
D. T. Robinson, T. G. Talomie, R. P. Dunlap and C. A. Franke,
J. Med. Chem., 1995, 38, 4687.
4 (a) M. P. Weir, S. S. Bethell, A. Cleasby, C. J. Campbell, R. J.
Dennis, C. J. Dix, H. Finch, H. Jhoti, C. J. Mooney, S. Patel, C. M.
Tang, M. Ward, A. Wonacott and C. W. Wharton, Biochem., 1998,
37, 6645; (b) M. J. O’Neill, J. A. Lewis, H. M. Noble, S. Holland,
C. Mansat, J. E. Farthing, G. Foster, D. Noble, S. J. Lane, P. J.
Sidebottom, S. M. Lynn, M. V. Hayes and C. J. Dix, J. Nat. Prod.,
submitted for publication.
5 S. J. F. Macdonald, M. D. Dowle, D. M. Buckley, J. G. Montana,
P. J. Knox, R. A. Smith, M. Bottwood, H. Jhoti, O. M. P. Singh
and A. J. Wonacott, manuscript in preparation.
6 S. J. F. Macdonald, D. J. Belton, D. M. Buckley, J. E. Spooner,
M. S. Anson, L. A. Harrison, K. Mills, R. J. Upton, M. D. Dowle,
R. A. Smith, C. Risley and C. J. Molloy, J. Med. Chem., 1998, 41,
3919.
7 C. G. Sowell, R. L. Wolin and R. D. Little, Tetrahedron Lett., 1990,
31, 485 and references therein.
8 E. J. Enholm and G. Prasad, Tetrahedron Lett., 1989, 30, 4939.
9 (a) S. Fukuzawa, M. Iida, A. Nakanishi, T. Fujinama and S. Sakai,
J. Chem. Soc., Chem. Commun., 1987, 920; (b) E. J. Enholm and
A. Trivellas, Tetrahedron Lett., 1989, 30, 1063; (c) E. J. Enholm and
A. Trivellas, J. Am. Chem. Soc., 1989, 111, 6463; (d) K. Tadano,
Y. Isshiki, M. Minami and S. Ogawa, Tetrahedron Lett., 1992, 33,
7899.
10 S. C. Shim, J. T. Hwang, H. Y. Kang and M. H. Chang, Tetrahedron
Lett., 1990, 31, 4765.
11 T. Inokuchi, H. Kawafuchi and S. Torii, J. Org. Chem., 1991, 56,
4983.
12 E. Lee, J. S. Tae, Y. H. Chang, Y. C. Park, M. Yun and S. Kim,
Tetrahedron Lett., 1994, 35, 129.
13 G. L. Lee, E. B. Choi, E. Lee and C. S. Pak, J. Org. Chem., 1994, 59,
1428.
14 E. Lee, T. S. Kang, B. J. Joo, J. S. Tae, K. S. Li and C. K. Chung,
Tetrahedron Lett., 1995, 36, 417.
15 E. Lee, K. S. Li and J. Lim, Tetrahedron Lett., 1996, 37, 1445.
16 (a) J. E. Baldwin, S. C. M. Turner and M. G. Moloney, Tetrahedron
Lett., 1992, 33, 1517; (b) J. E. Baldwin, S. C. M. Turner and M. G.
Moloney, Tetrahedron, 1994, 50, 9411.
give the crude product (0.0243 g). This material was purified by
preparative TLC on silca eluting three times with 62.5:62.5:1
hexane–EtOAc–AcOH to afford the desired product as a
colourless oil (0.005 g, 35%); TLC Rf = 0.24 (1:1 hexane–
Et2O ϩ 2 drops AcOH); 2β and 2α were not readily separable
by flash chromatography but their δH data is described separ-
ately for clarity (the ratio by integration is β:α = 3:1); 2β: δH
(d6-DMSO, 750 MHz, 75 ЊC) 7.39–7.32 (5H, m, Ph), 5.10 (1H,
d, J 12.5, PhHCH), 5.08 (1H, d, J 12.5, PhHCH), 4.02 (1H,
ddd, J 12, 10 and 5.5, NCH2CH2CH), 3.83 (1H, dd, J 11 and
10, NHCH), 3.63 (1H, m, NHCH), 3.33 (1H, dd, J 12 and 10,
NCH), 2.95 (1H, ddd, J 12, 7 and 3, PrCHCO), 2.27 (1H, m,
NCH2HCH), 2.01 (1H, m, NCH2HCH), 1.84 (1H, br m,
HCHCH2CH3), 1.71 (1H, br m, HCHCH2CH3), 1.46 (1H, br
m, CH2HCHCH3), 1.18 (1H, br m, CH2HCHCH3), 0.76 (3H,
br t, CH3); 2α: δH (d6-DMSO, 750 MHz, 75 ЊC) 7.39–7.32 (5H,
m, Ph), 5.14 (1H, d, J 12.5, PhHCH), 5.04 (1H, d, J 12.5,
PhHCH), 4.31 (1H, ddd, J 11, 10.5 and 5.5, NCH2CH2CH),
3.83 (1H, dd, J 11 and 10, NHCH), 3.66 (1H, dd, 10.5 and 7,
NCH), 3.57 (1H, m, NHCH), 2.84 (1H, br m, PrCHCO), 2.27
(1H, m, NCH2HCH), 2.04 (1H, m, NCH2HCH), 1.52 (2H, br
m, CH2CH2CH3), 1.41 (1H, br m, CH2HCHCH3), 1.28 (1H, br
m, CH2HCHCH3), 0.81 (3H, br t, CH3); comparison of this
spectrum with that of a 2:1 α:β propyl sample prepared by a
different route (unpublished results) gives good agreement;
comparison with the α- and β-allyl-trans-lactone analogues6
(which are separable) gives excellent agreement of the relevant
coupling constants; νmax(KBr diffuse reflectance)/cmϪ1 2960,
2930, 2865, 1796, 1715, 1454, 1433, 1390, 1356, 1324, 1109
and 1026; m/z (thermospray, NH3), MHϩ at 304 (100%); m/z
(electrospray) Found Mϩ: 303.147. C17H21NO4 requires Mϩ:
303.147; analytical HPLC (system as described above) 3α
(28.83 min), 3β (29.17 min) in a ratio of 1:3.1; this material
co-eluted with a 2:1 α:β sample prepared by a different route
(unpublished results).
17 R. Huisgen, K. Herbig, A. Siegl and H. Huber, Chem. Ber., 1966,
99, 2526.
18 J. E. McMurry, T. Lecta and J. G. Rico, J. Org. Chem., 1989, 54,
3748.
19 E. J. Corey and S. G. Pyne, Tetrahedron Lett., 1983, 24, 2821.
20 (a) H. Stetter and H. Kuhlmann, Synthesis, 1979, 29; (b) H. Stetter
and K. Marten, Liebigs Ann. Chem., 1982, 240.
21 T. Hosokawa, M. Takano, Y. Kuroki and S.-I. Murahashi,
Tetrahedron Lett., 1992, 33, 6643.
22 J. Inanaga, S. Katsuki, S. Takimoto, S. Ouchida, K. Inone,
A. Nakano, N. Okukuda and M. Yamaguchi, Chem. Lett., 1979,
1021.
23 Macromodel 6.0, Columbia University, 1996; F. Mohamadi, N. G. J.
Richards, W. C. Guida, R. Liskamp, M. Lipton, C. Caufield,
G. Chang, T. Hendrickson and W. C. Still, J. Comput. Chem., 1990,
11, 440.
Acknowledgements
We would like to thank Ian Davidson for mass spectral meas-
urements, Tim Underwood for analytical HPLC experiments,
Darren Green and Gianpaolo Bravi for generating coupling
constants with Macromodel, and Lee A. Harrison and Profes-
sor Tim Gallagher for many helpful comments and discussions
during the preparation of this manuscript.
References
1 S. Sinha, W. Watorek, S. Karr, J. Giles, W. Bode and J. Travis, Proc.
Natl. Acad. Sci. USA, 1987, 84, 2228.
2 R. L. Vender, J. Invest. Med., 1996, 44, 531.
3 For reviews: (a) D. J. Hlasta and E. D. Pagani, Annu. Rep. Med.
Chem., 1994, 195; (b) P. D. Edwards and P. R. Bernstein, Med. Res.
Rev., 1994, 14, 127; for leading references see: (c) Merck: P. E. Finke,
S. K. Shah, D. S. Fletcher, B. M. Ashe, K. A. Brause, G. O.
Chandler, P. S. Dellea, K. M. Hand, A. L. Maycock, D. G. Osinga,
D. J. Underwood, H. Weston, P. Davies and J. B. Doherty, J. Med.
Chem., 1995, 38, 2449; (d) Zeneca: C. A. Veale, P. R. Bernstein,
C. M. Bohnert, F. J. Brown, C. Bryant, J. R. Damewood, R. Earley,
S. W. Feeney, P. D. Edwards, B. Gomes, J. M. Hulsizer, B. J.
Kosmider, R. D. Krell, G. Moore, T. W. Salcedo, A. Shaw, D. S.
Silberstein, G. B. Steelman, M. Stein, A. Strimpler, R. M. Thomas,
24 D. P. Curran, T. L. Fevig, C. P. Jasperse and M. J. Totleben, Synlett,
1992, 943.
25 D. P. Curran, N. A. Porter and B. Giese, Stereochemistry of Radical
Reactions, VCH, Weinheim, 1995, pp. 16–19.
26 Ref. 25, p. 41.
27 G. A. Molander and C. Kenny, J. Am. Chem. Soc., 1989, 111, 8236.
28 J. M. Aurrecoechea and A. Fernandez-Acebes, Tetrahedron Lett.,
1992, 33, 4763; 1993, 34, 549.
Paper 8/07540I
3936
J. Chem. Soc., Perkin Trans. 1, 1998, 3931–3936