Preparation of (+)-nemorensic acid and approaches to nemorensine using the partial reduction of electron deficient furans

Article abstract of DOI:10.1039/b202514k

Starting from a commercially available furoic acid, the synthesis of (+)-nemorensic acid is described in nine steps, and in 32% overall yield. Key steps in our sequence are a chiral auxiliary controlled, stereoselective, Birch reduction of 3-methyl-2-furo

Full text of DOI:10.1039/b202514k

View Article Online / Journal Homepage / Table of Contents for this issue
Preparation of (؉)-nemorensic acid and approaches to
nemorensine using the partial reduction of electron deficient furans
Timothy J. Donohoe,*a,b Jean-Baptiste Guillerminb and Daryl S. Walterc
1
a Dyson Perrins Laboratory, South Parks Road, Oxford, UK OX1 3QY.
E-mail: timothy.donohoe@chem.ox.ac.uk; Fax: (01865) 275674; Tel: (01865) 275649
b Department of Chemistry, University of Manchester, Oxford Road, Manchester,
UK M13 9PL
c Department of Chemistry, Roche Discovery Welwyn, Broadwater Road, Welwyn Garden City,
Herts, UK AL7 3AY
Received (in Cambridge, UK) 12th March 2002, Accepted 15th April 2002
First published as an Advance Article on the web 2nd May 2002
Starting from a commercially available furoic acid, the synthesis of (ϩ)-nemorensic acid is described in nine steps,
and in 32% overall yield. Key steps in our sequence are a chiral auxiliary controlled, stereoselective, Birch reduction
of 3-methyl-2-furoic acid and the stereoselective reaction of an oxonium ion generated within a tetrahydrofuran ring.
Attempts to complete the synthesis of nemorensine did not succeed because of the low nucleophilicity of platynecine,
the alkaloid base portion of the natural product.
The first synthesis of nemorensic acid was reported by Klein
Introduction
in 19856 and since that time five further syntheses have
Pyrrolizidine alkaloids are a broad class of natural product,
appeared,7 including our own.8
isolated from a variety of different plant sources. In 1994 the
In this paper we wish to report, in full, our attempts at
existence of some 6000 plants containing a variety of more
preparing nemorensic acid using the (stereoselective) partial
than 180 known pyrrolizidine alkaloids was noted.1 These
compounds are important because of their biological activity,
which ranges from carcinogenicity to hepatotoxicity and
reduction of substituted furan 1 as a key step.8 We also wish to
disclose our work on the preparation of platynecine and initial
efforts at coupling the two halves of nemorensine itself.
pyrrolizidine alkaloids have been reported to be responsible for
severe cases of livestock poisoning.2
The toxic effect of pyrrolizidine alkaloids is related to P450
Results and discussion
mediated oxidation of the 4-azabicyclo[3.3.0] ring system to
Our synthesis began with the coupling of (commercially avail-
give a pyrrole unit: this then has the ability to act as a double
able) 3-methyl-2-furoic acid with (R,R)-bis(methoxymethyl)-
electrophile and is eminently capable of interstrand cross
pyrrolidine; this was accomplished by reaction of the
linking of DNA.3
corresponding acyl chloride with the amine in a mixture of
We are interested in synthesising macrocyclic bislactone
CH2Cl2 and 2 M NaOH and afforded the desired amide 3
members of this class of compound and our attention was
in 91% yield, Scheme 1. This amide was then subjected to a
drawn to one such compound, nemorensine, that has not yet
Birch reductive alkylation using sodium in liquid ammonia
been synthesised.
and subsequent addition of methyl iodide quenched the
The isolation of nemorensine was reported in 1973 from
reaction and yielded compound 4 in 87% and as a 30 : 1 mixture
three different varieties of Senecio Nemorensis L.4 After some
of diastereoisomers (this compound was fully characterised
confusion regarding the stereochemistry of the compound, the
and the ratio of diastereoisomers assessed by comparison to an
issue was proven unequovically by X-ray crystal structure
authentic 1 : 1 standard).9 † The 1H NMR spectrum of 4
in 1995.5 Hydrolysis of the lactone units of nemorensine gives
showed the appearance of a new methyl peak at 1.4 ppm and an
two components, platynecine (a necine base) and nemorensic
absence of any peaks between 6.0 and 7.0 ppm corresponding
acid, Fig. 1.4
to the furan nucleus.
The dihydrofuran obtained was then submitted to an allylic
oxidation using Jones’ reagent in a mixture of water–acetone
to afford the lactone 5 in excellent yield. The IR spectrum of
5 contained two bands at 1764 and 1626 cmϪ1 characteristic of
the carbonyl stretches. This lactone was then hydrogenated
using palladium hydroxide on carbon to afford the
compounds 6 and 7 in 88% yield and with a respective ratio of
92 : 8: these could be easily separated by chromatography on
silica gel. The cis relationship between the two methyl groups
† We have previously reported methodology for the stereoselective
reduction and are confident in both the sense and level of diastereo-
selectivity. We also know that the auxiliary can be removed without
Fig. 1
scrambling the stereochemistry at C-2. See ref. 9.
DOI: 10.1039/b202514k
J. Chem. Soc., Perkin Trans. 1, 2002, 1369–1375
1369
This journal is © The Royal Society of Chemistry 2002

View Article Online
a methoxy peak in the proton NMR around 3.3 ppm, and both
were fully characterised.
The allylation reaction of the methyl acetal mixture 9 proved
to be challenging. Different Lewis acids and conditions were
used to generate compound 10 Scheme 3. This reaction works in
Scheme 3 Reagents: i. TiCl4, allyltrimethylsilane, DCM.
two steps, the first being formation of an oxonium ion and the
second being attack on the oxonium by the allyltrimethylsilane
nucleophile. Strong Lewis acids worked best and nucleophilic
attack was encouraged by using a large excess (20 eq.) of
allyltrimethylsilane in dichloromethane at room temperature.12
Under these conditions, the reaction was reproducible and
high yielding for formation of 10: moreover, examination of
the crude NMR spectrum of the reaction showed only a
single diastereoisomer of product. In the absence of the other
diastereoisomer for comparative purposes, the diastereo-
selectivity was assigned conservatively as being у10 : 1.
The structure of allylated product 10, could be clearly
assigned by examination of its NMR data. A multiplet inte-
grating for one proton at 5.75 ppm, and another, integrating for
two protons, at 5.05 ppm confirmed the presence of a terminal
double bond in the molecule. The sense of the diastereoselec-
tivity was proved by an NOE experiment which showed strong,
and reciprocal, NOE enhancements between the methine at C-3
and the allylic methylene attached to C-5.8
Scheme 1 Reagents: i. SOCl2, then (R,R)-bis(methoxymethyl)pyrrol-
idine, (aq.) NaOH; ii. Na, NH3, MeI; iii. CrO3, H2SO4, (aq.) acetone;
iv, H2, Pd(OH)2/C, EtOH.
on the major lactone 6 was proven unambiguously by X-ray
crystallography.10
The lactone 6 was then treated with the Petasis reagent
(Cp2TiMe2) to afford the enol ether 8, Scheme 2. Petasis’ reagent
This sense of diastereoselectivity (addition of a nucleophile
trans to the C-3 methyl group) is in accord with the selectivity
displayed by the parent 3-methyltetrahydrofuran derived
oxonium ion.12 It may also be the case that geminal substitution
at C-2 aids the stereoselectivity, as has been noted by Woerpel13
in related systems. Assuming that the intermediate oxonium ion
reacts in an envelope conformation Fig. 2, the bulky amide
Scheme 2 Reagents: i. Cp2TiMe2, THF; ii. MeOH, HCl; iii. AcOH.
was prepared by treating bis(cyclopentadienyl)titanium
dichloride with 2.2 equivalents of MeLi in ether.11 The orange
needles, obtained after recrystallisation from pentane, could be
kept in the freezer in the dark and re-used for up to two weeks.
The methylenation reaction was followed by IR (disappearance
of the band at 1787 cmϪ1 characteristic of the lactone carbonyl
and appearance of a band at 1673 cmϪ1 characteristic of form-
ation of the enol ether). After addition of petrol to the reaction,
a solid precipitated and the mixture was filtered through a
cotton plug. It was, however, difficult to remove all of the
titanium by-products by filtration and the enol ether 8 was not
stable to purification by column chromatography. Therefore, we
used crude material directly in the next step.
Fig. 2
With the aim of making an acetal, the enol ether 8 was
treated with acetic acid. However, a multi-component mixture
was obtained each time which showed evidence of formation of
the corresponding lactol and ring opened isomers. It was
decided to replace the acetate by a more stable group and so the
reaction of enol ether 8 with methanol was performed under
acidic conditions, and this gave acetal 9 in 86% from the lactone
6. The acetal product obtained was a 1 : 1 mixture of two
diastereoisomers separable by column chromatography; this
separation was only performed for characterisation. Each of
the acetal diastereoisomers, which were not assigned, possessed
group at C-2, sits in a pseudo-equatorial position forcing the
methyl group to adopt a pseudo-axial orientation where it may
block the top face of the oxonium ion as shown.
More recently, Woerpel14 has proposed a second model based
on torsional strain which is also relevant here. In this model,
the oxonium ion sits opposite the ‘flap’ of the envelope and
the preferred attack is from inside, rather than outside, the
envelope. Nucleophilic attack on an oxonium ion in this
conformation from outside the envelope would result in the
development of a disfavoured eclipsing interaction between
substituents at C-1 and C-2 of the resulting tetrahydrofuran
1370
J. Chem. Soc., Perkin Trans. 1, 2002, 1369–1375

View Article Online
ring, see Fig. 2. Conversely, nucleophilic attack from inside
the envelope gives the tetrahydrofuran ring in a staggered
conformation and is preferred.
Note that both arguments predict the same outcome (one of
only two that are possible) and without further evidence it is
difficult to separate the relative contributions that each makes
to the stereoselectivity shown.
Ruthenium tetraoxide catalysed oxidation of the terminal
double bond within 10 to give the carboxylic acid 11 was
realised following Sharpless’ methodology,15 Scheme 4;
Scheme 5 Reagents: i. Ba(OH)2, H2O, ∆; ii. PtO2/C, H2, MeOH; iii.
Rh/C, H2, THF.
one atmosphere of hydrogen. However, several attempts to
repeat this work all met with failure in our hands. The only
compound we obtained was the product of hydrogenolysis of
the allylic alcohol, followed by hydrogenation of the remaining
1
double bond. A H NMR spectrum of the compound 12 so
obtained revealed the presence of a (3H) doublet at 1.32 ppm
corresponding to a methyl group while the alkene region
(5.00–6.00 ppm) was free from any signals.
Changing the catalyst from platinium to palladium and the
solvent from methanol to ethanol made little difference and
hydrogenolysis was still the principal reaction. Furthermore,
reduction under homogeneous conditions, using Wilkinson’s
catalyst, was attempted but gave only recovered starting
material.
Eventually, hydrogenation was achieved by using H2 and
Rh/C (5%) in THF. Under these conditions, and after stirring
overnight at room temperature under a balloon of hydrogen,
platynecine was isolated in 72% yield Scheme 5. The curved
shape of retronecine means that hydrogenation would be
expected to occur from the lower face of the molecule, as
shown, to furnish platynecine. This was proved by matching
of the spectroscopic data (1H, 13C NMR) of the compound so
obtained with data from authentic platynecine.17
Scheme 4 Reagents: i. RuCl3ؒxH2O, NaIO4; ii. (6 M) HCl.
compound 11 was used directly in the next step without
purification. Cleavage of the chiral auxiliary was achieved
by hydrolysis using a 6 M HCl solution heated at reflux and
nemorensic acid was obtained as a colourless solid in 83% yield
from 10.
All data (1H, 13C NMR, ms, [α]D) of synthetic (ϩ)-
nemorensic acid were in accordance with those reported in the
literature.7 Moreover, the structure of the synthetic material
was proved by X-ray crystallography.8
In total, this synthesis was realised in 32% overall yield over
nine steps. Moreover, it is amenable to the synthesis of substan-
tial amounts of enantiopure nemorensic acid and, using this
route, hundreds of milligrams of this molecule have been
synthesised.
Having completed the synthesis of nemorensic acid and
made significant quantities of material, it was decided to
attempt the first synthesis of nemorensine.
Attempted coupling of nemorensic acid with platynecine
Following encouraging precedent by Niwa,18 we prepared the
cyclic anhydride of nemorensic acid in quantitative yield by
coupling with one equivalent of DCC in dichloromethane,
Scheme 6. The urea by-product was removed by filtration. The
Fig. 1 reveals the main challenge in this coupling, namely
that the primary hydroxy group on platynecine has to be
coupled to the most hindered carboxylic acid on nemorensic
acid, whereas the less reactive alcohol (secondary hydroxy
group) has to couple to the most accessible acid. A successful
coupling should therefore involve differentiation of either the
hydroxys of platynecine and/or the carboxylic acids of nemo-
rensic acid. Without this differentiation between groups, the
coupling could conceivably occur in two ways to form either
nemorensine or its unnatural regioisomer.
The necine base of nemorensine, platynecine, can be
obtained from the naturally occurring pyrrolizidine alkaloid
monocrotaline, which is commercially available. Double
hydrolysis of this molecule was effected following a literature
procedure Scheme 5 which consisted of reacting monocrotaline
with 2.2 eq. of barium hydroxide octahydrate in water, followed
by heating at reflux.1 This reaction worked well and 98% yield
of retronecine could be isolated. All data (1H, 13C NMR) were
in accordance with that reported in the literature.1
Scheme 6 Reagents: i. DCC, CH2Cl2.
anhydride 13 could not be purified by column chromatography
and was therefore used straight away in the next step. While 1H
and 13C NMR and IR indicated complete formation of 13, it
proved to be extremely sensitive to hydrolysis.
However, we were not able to address the regioselectivity of
anhydride coupling with platynecine because of the extreme
sensitivity of the anhydride and low nucleophilicity of the diol.
Little or no coupling was observed between the two under
a variety of conditions (neutral, basic, etc.). In fact, opening
the anhydride with a range of unhindered alcohols and
alkoxides did not give any ester products but always returned
nemorensic acid.
The next step, stereoselective hydrogenation of retronecine,
proved more difficult to accomplish. This reaction has been
reported in the literature,16 when Drewes and co-workers
described successful hydrogenation using PtO2 on carbon under
Being unable to open the anhydride successfully, we decided
to mono-protect nemorensic acid and attempt more con-
ventional ester coupling reactions. Several conditions were
J. Chem. Soc., Perkin Trans. 1, 2002, 1369–1375
1371

View Article Online
investigated for placement of a single benzyl group on nemo-
rensic acid. Firstly, we showed that the diacid could indeed be
activated and then coupled with a nucleophile by preparing
the dibenzylester 14 in 35% yield, Scheme 7. Attempts to
In total, we have developed an extremely efficient route to
nemorensic acid and protected derivatives thereof; coupling
strategies which involve platynecine as a nucleophile were not
successful in our hands and other ways of completing the
synthesis must be pursued.
Experimental
General experimental
All reactions, except aqueous reactions, were carried out under
an atmosphere of argon.
Proton nuclear magnetic resonance spectra (NMR) were
recorded on a Gemini 200 at 200 MHz, a Varian Unity Inova
300 at 300 MHz, a Varian Unity Inova 400 at 400 MHz and
a Varian Unity Inova 500 at 500 MHz. 13C nuclear magnetic
resonance spectra were recorded on a Varian Unity Inova 300
at 75 MHz, a Varian Unity Inova 400 at 100 MHz and a Varian
Unity Inova 500 at 125 MHz. Chemical shifts (δ) are quoted
in parts per million (ppm), downfield from tetramethylsilane.
Coupling constants (J) are quoted in Hz. Infrared spectra (IR)
were recorded on an ATI Mattson Genesis FTIR as evaporated
films (EF). Optical rotation were recorded on an Optical
Activity Ltd. AA-100 polarimeter and [α]D values (all reported
at 21 ЊC) are given in units of 10Ϫ1 deg cm2 gϪ1; concentration c
in units of g 100 mlϪ1. Mass spectra were recorded on a Kratos
Concept or a Fisson VG Trio 2000. Melting points were
obtained on a Kofler block and are uncorrected.
Scheme 7 Reagents: i. DCC, DMAP, BnOH (excess); ii. Et3N (1.2 eq.),
BnBr, CH2Cl2; iii. (EtO)2POCl, Et3N; then platynecine, NaH, THF.
mono-protect by reducing the amount of coupling reagent and
benzyl alcohol gave inseparable mixtures of the two possible
regioisomers, together with doubly protected nemorensic acid.
Eventually, the best conditions found for this mono-
protection were those involving benzyl bromide in CH2Cl2 with
1.2 eq. of Et3N. Under these conditions, 81% of a 7 : 1 ratio of
regio-isomers 15 was obtained. The ratio was measured by
proton NMR by comparison of the integration of the methylene
benzyl peak and compared to the mixtures obtained earlier.
The connectivity (major compound) within compound 15
was shown by a HMBC experiment which revealed that the
benzyl singlet at δ 5.23 was coupled to the carbonyl at δ 177.1
and that the AB signals at δ 2.32 and 2.65 were both coupled
to a different carbonyl at δ 171.4. This two and three
bond coupling array can arise after benzylation at the most
hindered acid group, which is presumably favoured because of
the enhanced acidity of the α-oxy carboxylic acid relative to the
other one.
Tetrahydrofuran (THF) was dried and distilled from sodium
metal using benzophenone as indicator, under an atmosphere
of argon. Dichloromethane (DCM) was distilled over calcium
hydride. Petroleum ether (boiling range 40–60 ЊC) was distilled
prior to use and ammonia was distilled from sodium metal and
ferric chloride.
[(2ЈR,5ЈR)-2Ј,5Ј-Bis(methoxymethyl)tetrahydro-1ЈH-pyrrol-1Ј-
yl](3-methyl-2-furyl)methanone 3
Thionyl chloride (12 ml, 0.16 mol) was added to 3-methyl-2-
furoic acid (2.39 g, 19.0 mmol). The resulting dark mixture was
heated at reflux for 3–4 h and the excess thionyl chloride was
then removed under pressure (azeotroped with toluene). The
brown liquid obtained was mixed with dichloromethane (10 ml)
and was then added dropwise to a stirring mixture of (R,R)-
bis(methoxymethyl)pyrrolidine (3.32 g, 21.0 mmol), 2.0 M aq.
sodium hydroxide (40 ml, 80 mmol) and dichloromethane
(40 ml) at 0 ЊC. The solution was allowed to warm to room
temperature and stirred overnight. The dichloromethane was
removed in vacuo and the residue extracted with ethyl acetate
(3 × 20 ml). The combined organics were then dried over
magnesium sulfate and evaporated to dryness in vacuo to yield
a brown oil. The crude product was purified by column chrom-
atography (25 : 75 ethyl acetate–petrol) to give the title
compound as a pale oil (4.6 g, 17 mmol, 91%). (Found M ϩ Hϩ
268.1557, C14H21NO4 requires 268.1549, deviation 2.98 ppm),
[α]D ϩ130.1 (c 2.17 in CHCl3); νmax/cmϪ1 2927, 1615; δH (300
MHz, CDCl3) 1.79–2.20 (4H, m), 2.25 (3H, s), 2.88–3.04 (2H,
m), 3.12 (3H, s), 3.26 (3H, s), 3.23–3.36 (1H, m), 3.48 (1H, dd,
J 9.0 and 3.0), 4.37–4.48 (1H, m), 4.64–4.74 (1H, m), 6.25 (1H,
d, J 1.5), 7.25 (1H, d, J 1.5); δC (75 MHz, CDCl3) 11.60, 23.98,
27.30, 57.10, 57.50, 58.90, 59.00, 71.86, 73.77, 114.9, 128.6,
Having successfully selectively protected one of the carb-
oxylic acid groups, we now decided to probe the reactivity of
the other towards platynecine.
Coupling of acid 15 was attempted under a variety of differ-
ent conditions and was successful with non-hindered alcohols
using diethylchlorophosphate and a sodium alkoxide19 (details
not shown). However, the poor nucleophilicity of platynecine
thwarted its coupling under conditions that we knew to work
with simpler alcohols. In fact, we were only ever able to isolate
just 1–2% of a coupled product 16 which was incompletely
characterised (and therefore its structure only tentatively
assigned) because of the small amount available. The 1H NMR
spectrum of 16 shows an ABX system at 4.35 and 4.54 ppm,
which corresponds to the methylene group attached to oxygen
of the newly formed ester linkage. In contrast, the methylene
attached to the primary alcohol in platynecine resonates at 3.8
ppm which points to coupling at the primary alcohol end (C8)
of platynecine. If nemorensine were ever to be made via this
route, then the acid 15 would have to be coupled to the more
hindered secondary alcohol of platynecine, which is clearly a
formidable proposition.
141.6, 143.0, 159.6 (C᎐O); m/z (CI) 268 (M ϩ 1, 100%).
᎐
(2R,2ЈR,5ЈR)-[2Ј,5Ј-Bis(methoxymethyl)tetrahydro-1ЈH-pyrrol-
1Ј-yl][2,3-dimethyl-2,5-dihydrofuran-2-yl]methanone 4
So, our attempts to complete the synthesis of nemorensine
were thwarted by the low nucleophilicity of platynecine; the
concave nature of this compound means that both primary and
secondary hydroxy groups are placed in the cleft and are
poor nucleophiles. A successful approach to nemorensine will
probably involve coupling of acid 15 with a protected form of
retronecine followed by stereoselective hydrogenation.
To a blue solution of sodium (0.44 g, 19 mmol) in liquid
ammonia (75 ml) at Ϫ78 ЊC was added the substrate 3 (1.25 g,
4.67 mmol) in 1 ml of THF. After 10 min, isoprene was added
(until the blue colour dispersed) followed by iodomethane
(1 ml, 16 mmol) to give a yellow solution. After 15 min, the
1372
J. Chem. Soc., Perkin Trans. 1, 2002, 1369–1375

View Article Online
solution was quenched with a solution of saturated aq.
ammonium chloride (10 ml). The ammonia was then allowed to
evaporate and the residual solution was extracted with diethyl
ether (3 × 20 ml). The combined organics were dried over
sodium sulfate, filtered and evaporated in vacuo. The crude
material was purified by column chromatography (20 : 80
diethyl ether–petrol) to afford the title compound as a
colourless oil (1.15 g, 4.06 mmol, 87%). (Found M ϩ Hϩ
284.1861, C15H25NO4 requires 284.1862, deviation 0.35 ppm),
[α]D ϩ130.9 (c 1.28 in CHCl3); νmax(film)/cmϪ1 2924, 1621;
δH (300 MHz, CDCl3) 1.44 (3H, s), 1.74–2.04 (7H, m), 2.97
(1H, t, J 9.0), 3.10–3.30 (2H, m), 3.21 (3H, s), 3.27 (3H, s), 3.46
(1H, dd, J 9.0 and 3.0), 4.19–4.27 (1H, m), 4.56 (2H, br s),
4.64–4.72 (1H, m), 5.46 (1H, dd, J 3.5 and 1.5); δC (75 MHz,
CDCl3) 12.83, 23.30, 27.30, 27.95, 57.63, 57.92, 58.71, 58.78,
71.39, 73.07, 74.01, 93.42, 120.4, 141.0, 172.2; m/z (CI) 284
(M ϩ 1, 100%), 97 (25).
was extracted with diethyl ether (3 × 20 ml). The resulting
organics were dried over magnesium sulfate, filtered and
evaporated in vacuo to afford the product as an orange solid
(1.12 g, 5.39 mmol, 88%). δH (200 MHz, CDCl3) Ϫ0.18 (6H, s,
CH3), 6.05 (10H, s, CH ).
[(2R,5R)-2,5-Bis(methoxymethyl)tetrahydro-1H-pyrrol-1-yl]-
[(2R,3R)-2,3-dimethyl-5-methylenetetrahydrofuran-2-yl]-
methanone 8
To a solution of compound 6 (315 mg, 1.05 mmol) in THF
(15 ml) was added dimethyldicyclopentadienyltitanium (877 mg,
4.20 mmol) in the dark at room temperature. The reaction was
then stirred for 4 hours at 70 ЊC. The reaction was followed by
IR (disappearance of the band at 1783 cmϪ1 and formation of a
band at 1673 cmϪ1). Petrol was added and the yellow precipitate
removed by filtration over Celite®. The filtrate was then evap-
orated to dryness to give the crude product as a red oil. This
compound was not particularly stable and therefore was used
immediately in the next step. νmax/cmϪ1 2952, 1673, 1622;
δH (200 MHz, CDCl3), 1.14 (3H, d, J 7.5), 1.35 (3H, s),
1.80–2.30 (5H, m), 2.50–2.80 (2H, m), 3.10 (1H, t, J 9.0),
3.15–3.55 (2H, m), 3.30 (6H, br s), 3.50 (1H, dd, J 9.0 and 3.0),
3.82 (1H, br s), 4.15–4.30 (2H, m), 4.50–4.65 (1H, m).
(5R,2ЈR,5ЈR)-5-[2Ј,5Ј-Bis(methoxymethyl)tetrahydro-1ЈH-
pyrrol-1Ј-ylcarbonyl]-4,5-dimethyl-2,5-dihydrofuran-2-one 5
To a solution of substrate 4 (1.15 g, 4.06 mmol) in acetone
(15 ml), was added at 0 ЊC a solution of chromium trioxide
(2.7 g, 27 mmol) in water (6.5 ml) and 5 drops of conc. sulfuric
acid. The reaction was stirred at 50 ЊC for 7 hours before being
quenched by addition of brine. The solution was then extracted
with dichloromethane (7 × 20 ml). The combined organics were
dried over magnesium sulfate, filtered and evaporated in vacuo
to give the crude product. Flash chromatography (50 : 50
diethyl ether–hexane) afforded the title compound as a colour-
less oil (1.20 g, 4.04 mmol, 98%). (Found M ϩ Hϩ 298.1652,
C15H23NO5 requires 298.1654, deviation 0.67 ppm), [α]D ϩ169.8
(c 1.63 in CHCl3); νmax/cmϪ1 2926, 1764, 1626; δH (300 MHz,
CDCl3) 1.64 (3H, s), 1.70–2.08 (4H, m), 2.20 (3H, d, J 1.5), 2.90
(1H, dd, J 9.0 and 6.0), 3.04–3.37 (2H, m), 3.14 (3H, s),
3.27 (3H, s), 3.42 (1H, dd, J 9.0 and 2.5), 4.09–4.14 (1H, m),
4.73 (1H, q, J 6.0), 5.84 (1H, d, J 1.5); δC (75 MHz, CDCl3)
14.78, 24.06, 26.65, 27.26, 57.66, 58.46 (× 2), 58.92, 71.42,
73.94, 90.03, 116.15, 168.4, 171.3, 171.6; m/z (CI) 298 (M ϩ 1,
100%).
[(2R,5R)-2,5-Bis(methoxymethyl)tetrahydro-1H-pyrrol-1-yl]-
[(2R,3R)-5-(methoxymethyl)-2,3,5-trimethyltetrahydrofuran-
2-yl]methanone 9
To a solution of crude compound 8 in 2,2Ј-dimethoxypropane
(2 ml) stirred at room temperature under argon was added
4 ml of methanol followed by a few drops of concentrated
hydrochloric acid. After 2 hours, the reaction was quenched by
addition of a 2.0 M sodium hydroxide solution and extracted
with diethyl ether (3 × 15 ml). The combined organics were then
dried over magnesium sulfate and evaporated in vacuo to give
the crude product as a red oil. Purification by flash chrom-
atography (60 : 40 diethyl ether–petrol) afforded the title
compound as a yellow oil (299 mg, 0.91 mmol, 86% over two
steps).
For the mixture of diastereoisomers: (Found M ϩ Hϩ
330.2276, C17H31NO3 requires 330.2280, deviation 1.12 ppm);
νmax/cmϪ1 2934, 2826, 1630; m/z (CI) 330 (M ϩ 1, 95), 298
(M Ϫ 31, 100%).
(5R,4R,2ЈR,5ЈR) and (5R,4S,2ЈR,5ЈR)-5-[2Ј,5Ј-Bis(methoxy-
methyl)tetrahydro-1ЈHpyrrol-1Ј-ylcarbonyl]-4,5-dimethyltetra-
hydrofuran-2-one (6 and 7)
Diastereoisomer 1. δH (300 MHz, CDCl3) 1.20 (3H, d, J 6.5),
1.35 (3H, s), 1.42 (3H, s), 1.78 (1H, dd, J 13.0 and 11.0),
1.84–2.05 (4H, m), 2.16 (1H, dd, J 13.0 and 9.0), 2.62 (1H, qdd,
J 7.0, 2.5 and 2.0), 3.04–3.12 (1H, dd, J 9.0 and 8.0), 3.21–3.30
(1H, m), 3.28 (3H, s), 3.33 (6H, s), 3.51–3.58 (2H, m), 4.16–4.24
(1H, m), 4.47–4.55 (1H, m); δC (75 MHz, CDCl3) 15.7, 20.8,
23.2, 23.8, 26.8, 40.5, 45.6 (× 2), 49.3, 57.4, 58.0, 58.7, 71.0,
75.0, 87.9, 107.2, 174.1.
To a solution of compound 5 (1.20 g, 1.68 mmol) in ethanol
(15 ml), was added acetic acid (2.0 ml) and palladium (10% on
carbon, 50 mg). The solution was stirred under 1 atm of
hydrogen overnight. The catalyst was removed by filtration
through Celite®, washed with ethanol, and the filtrate evap-
orated in vacuo. The two diastereoisomers were then separated
by column chromatography (55 : 45 diethyl ether–hexane) to
afford 6 and 7 as colourless oils (1.05 g, 3.51 mmol, 88%).
6: (850 mg, 71%) (Found M ϩ Hϩ 300.1808 C15H25NO5
requires 308.1811, deviation 0.97 ppm), [α]D ϩ22.9 (c 0.21 in
CH2Cl2); νmax(film)/cmϪ1 2977, 1787, 1623; δH (300 MHz,
CDCl3) 1.27 (3H, d, J 7.0), 1.48 (3H, s), 1.78–2.10 (4H, m), 2.19
(1H, dd, J 17.5 and 10.0), 2.61 (1H, dd, J 17.5 and 9.0), 2.95–
3.10 (1H, m), 3.16 (1H, dd, J 9.0 and 6.5), 3.24–3.40 (2H, m),
3.30 and 3.32 (6H, 2s), 3.50 (1H, dd, J 9.0 and 3.0), 4.24–4.32
(1H, m), 4.72 (1H, q, J 7.0); δC (75 MHz, CDCl3) 15.67, 21.43,
24.17, 27.45, 35.63, 36.63, 57.44, 58.25, 58.67, 58.83, 71.31,
74.56, 88.39, 172.3, 174.7; m/z (CI) 300 (M ϩ 1, 100%).
Diastereoisomer 2. δH (300 MHz, CDCl3) 1.14 (3H, d, J 7.0),
1.21 (3H, s), 1.44 (3H, s), 1.79 (1H, t, J 13.0), 1.88–2.03 (4H, m),
2.20 (1H, dd, J 13.0 and 7.0), 2.85–2.99 (1H, m), 3.15 (1H, t,
J 9.0), 3.19–3.26 (1H, m), 3.22 (3H, s), 3.32 and 3.35 (6H, s),
3.56 (1H, dd, J 9.0 and 3.0), 3.81 (1H, dd, J 9.0 and 4.0), 4.15–
4.24 (1H, m), 4.43–4.46 (1H, m); δC (75 MHz, CDCl3) 15.4,
21.3, 23.9, 25.7, 26.5, 30.2, 38.9, 42.3, 50.3, 57.6, 58.5, 58.7,
70.9, 73.6, 87.7, 107.3, 173.8.
[(2R,3R,5R)-5-Allyl-2,3,5-trimethyltetrahydrofuran-2-yl]-
[(2R,5R)-2,5-bis(methoxymethyl)tetrahydro-1H-pyrrol-1-yl]-
methanone 10
Dimethyldicyclopentadienyltitanium11
To a solution of dichlorodicyclopentadienyltitanium (1.52 g,
6.13 mmol) in diethyl ether (15 ml) was added, at 10–15 ЊC, a
1.6 M solution of methyllithium (8.4 ml, 13.40 mmol) in diethyl
ether. The solution was stirred for one hour at room temper-
ature before being quenched with ice water. The aqueous layer
Allyltrimethylsilane (2.18 g, 19.1 mmol) was added to a solu-
tion of compound 9 (315 mg, 0.96 mmol) in CH2Cl2 (5 ml) at
room temperature, followed by the addition of a 1 M solution
of titanium() chloride (2.9 ml, 2.87 mmol). The red–brown
J. Chem. Soc., Perkin Trans. 1, 2002, 1369–1375
1373

View Article Online
solution was stirred for one hour and quenched by addition of
sodium hydroxide (2 M). The solution was then extracted with
diethyl ether (3 × 20 ml), dried over magnesium sulfate and
evaporated in vacuo to give the crude product as a yellow oil.
Purification by flash chromatography (20 : 80 diethyl ether–
petrol) afforded the title compound as a colourless oil (265 mg,
0.78 mmol, 81%). (Found M ϩ Hϩ 340.2492, C19H33NO4
requires 340.2488, deviation 1.17 ppm), [α]D ϩ46.06 (c 6.0,
CH2Cl2); νmax/cmϪ1 2974, 1632; δH (300 MHz, CDCl3) 1.14 (3H,
d, J 6.5), 1.22 (3H, s), 1.28 (3H, s), 1.41 (1H, t, J 13.0), 1.86–
2.00 (4H, m), 2.04 (1H, dd, J 13.0 and 7.0), 2.25 (2H, 2 dd,
J 13.0 and 8.0), 2.68–2.82 (1H, m), 3.09–3.16 (1H, t, J 7.0), 3.27
(1H, t, J 7.0), 3.32 (3H, s), 3.34 (3H, s), 3.55 (1H, dd, J 9.0 and
3.0), 3.69 (1H, dd, J 9.0 and 3.0), 4.12–4.20 (1H, m), 4.40–4.51
(1H, m), 5.05 (2H, m), 5.68–5.82 (1H, m); δC (75 MHz, CDCl3)
15.46, 21.84, 24.00, 26.34, 29.35, 40.18, 42.49, 46.69, 57.20,
57.45, 58.53, 58.71, 70.90, 74.03, 82.16, 86.92, 117.7, 134.7,
174.2; m/z (CI) 340 (M ϩ 1, 100%).
0.70 mmol) and barium hydroxide octahydrate (457 mg,
1.45 mmol). The resulting mixture was stirred at 60 ЊC for
45 min and heated at reflux for a further 3 hours. The solution
was allowed to cool to room temperature and three pieces of
CO2 solid were carefully added. The reaction was stirred for
30 min, filtered and washed with water and chloroform. The
solvents and aqueous were evaporated in vacuo. The resulting
oil was purified by flash column chromatography (13 : 5 : 2
CHCl3–MeOH–Et3N) to afford the title compound as a pale
beige solid. After recrystallisation from hot acetone, the com-
pound was obtained as a colourless crystalline solid (93 mg,
0.60 mmol, 86%). δH (300 MHz, D2O) 1.76–1.88 (2H, m),
2.52–2.64 (1H, m), 3.06–3.14 (1H, m), 3.30 (1H, ddd, J 15.5, 5.0
and 2.0), 3.70 (1H, dt, J 15 and 2.0), 4.04–4.12 (3H, m), 4.26–
4.28 (1H, m), 5.60 (1H, d, J 1.5); δC (75 MHz, D2O) 35.26,
52.61, 58.33, 61.18, 70.44, 76.32, 124.9, 137.2. All data were in
accordance with that reported in the literature.1
Platynecine17
2-((2R,4R,5R)-5-[(2R,5R)-2,5-Bis(methoxymethyl)tetrahydro-
1H-pyrrol-1-ylcarbonyl]-2,4,5-trimethyltetrahydrofuran-2-yl)-
acetic acid 11
A solution of retronecine (50 mg, 0.32 mmol) in THF (10 ml)
containing a catalytic amount of rhodium on carbon (5%) was
stirred under an atmospheric pressure of hydrogen for 12 hours.
The solution was then filtered through Celite®, washed with
methanol and the solvents evaporated in vacuo. Purification by
column chromatography (10 : 5 : 1 CHCl3–MeOH–NH4OHsol)
afforded the title compound as a pale yellow oil, which crystal-
lised at room temperature (38 mg, 0.24 mmol, 75%). (Mϩ
157.1104, C8H15NO2 requires 157.1103, deviation 0.64 ppm),
δH (300 MHz, CD3OD) 1.54–2.00 (4H, m), 2.26–2.40 (1H, m),
2.62–2.82 (2H, m), 2.95–3.24 (3H, m), 3.84 (2H, d, J 5.5),
4.13–4.17 (1H, m); δC (75 MHz, CD3OD) 27.40, 35.84, 43.55,
53.36, 55.09, 60.17, 71.25, 71.58; m/z (CI) 158 (M ϩ 1,
100%). All data were in accordance with that reported in the
literature.17
To a solution of compound 10 (410 mg, 1.20 mmol) in a 2 : 2 : 3
mixture of carbon tetrachloride (3 ml), acetonitrile (3 ml) and
water (4.5 ml) was added sodium periodate (1.19 g, 5.60 mmol).
Ruthenium trichloride hydrate (catalytic) was then added
to this biphasic solution. The reaction was allowed to stir for
2 hours and quenched with a 2 M solution of sodium
hydroxide. The solution was extracted with dichloromethane
(1 × 20 ml) and the aqueous layer acidified with conc. hydro-
chloric acid to pH < 1. The aqueous layer was then extracted
again with dichloromethane (5 × 20 ml). The organic layer was
dried over magnesium sulfate and evaporated in vacuo to afford
the crude compound as a colourless oil. This compound was
used in the next step without further purification. (Found M ϩ
Hϩ 358.2232, C18H31NO6 requires 358.2229, deviation 0.84
ppm), νmax/cmϪ1 3167, 2976, 1730, 1629; δH (300 MHz, CDCl3)
1.09 (3H, d, J 6.5), 1.22 (3H, s), 1.36 (3H, s), 1.51 (1H, t, J 12.5),
1.76–1.98 (4H, m), 2.25 (1H, dd, J 13.0 and 7.5), 2.48 (2H, s),
2.60–2.72 (1H, m), 3.04–3.36 (2H, m), 3.26 (6H, s), 3.38–3.50
(2H, m), 4.04–4.16 (1H, m), 4.22–4.36 (1H, m); δC (75 MHz,
CDCl3) 15.29, 21.79, 24.29, 26.73, 29.02, 40.47, 43.80, 46.12,
57.69 (× 2), 58.79 (× 2), 70.98, 74.29, 80.67, 87.44, 174.3, 174.7;
m/z (CI) 358 (M ϩ 1, 100%).
(1R,6S,8R)-1,6,8-Trimethyl-3,9-dioxabicyclo[4.2.1]nonane-2,4-
dione 13
To a solution of nemorensic acid (20 mg, 0.09 mmol) in
dichloroethane (1 ml) was added a solution of 1,3-dicyclo-
carbodiimide (19 mg, 0.09 mmol) in dichloroethane (1 ml). The
reaction was stirred at 80 ЊC for 7 hours. The solvents were then
evaporated in vacuo. The residue was taken up in acetone and
filtered through a cotton plug. The acetone was evaporated to
afford the title compound as a clear oil (18 mg, 0.09 mmol,
99%). νmax(film)/cmϪ1 1805, 1758; δH (400 MHz, CDCl3) 1.17
(3H, d, J 7.0), 1.45 (1H, dd, J 12.5 and 8.0), 1.49 (3H, s), 1.52
(3H, s), 2.24 (1H, dd, J 12.5 and 8.0), 2.78–2.88 (1H, m), 2.86
(1H, d, J 13.5), 2.92 (1H, d, J 13.5); δC (100 MHz, CDCl3)
16.23, 20.64, 26.44, 40.18, 47.66, 47.75, 79.23, 86.18, 162.7,
174.6. This compound was not stable enough to obtain any
mass spectrometry measurements.
(2R,3R,5S)-5-Carboxymethyl-2,3,5-trimethyltetrahydrofuran-
2-carboxylic acid ((؉)-nemorensic acid)
A 6 M solution of hydrochloric acid (10 ml) was added to the
crude compound 11. The reaction was heated at reflux for
5 hours and then quenched by addition of a 2 M solution
of NaOH until pH = 14. The solution was extracted with
dichloromethane (1 × 20 ml). The aqueous layer acidified until
pH < 1 by addition of HCl concentrated and extracted with
dichloromethane (5 × 20 ml). The remaining organic layer was
then dried over magnesium sulfate and evaporated in vacuo to
afford the title compound as a colourless solid (217 mg,
1.00 mmol, 83% over two steps from 10). (Found M ϩ Hϩ
217.1073, C10H16O5 requires 217.1076, deviation 1.38 ppm), mp
171–173 ЊC (lit.5 174–175 ЊC); [α]D ϩ84.4 (c 0.18, EtOH) (lit.5
ϩ87.2 (c 0.24 EtOH)); νmax(film)/cmϪ1 3451, 3120, 2974, 2934,
1712; δH (400 MHz, CD3OD) 1.12 (3H, d, J 6.8), 1.28 (3H, s),
1.39 (3H, s), 1.68 (1H, t, J 12.4), 2.25 (1H, dd, J 12.4 and 6.8),
2.47 (1H, d, J 13.6), 2.56–2.69 (1H, m), 2.60 (1H, d, J 13.6);
δC (100 MHz, CD3OD) 14.22, 20.62, 27.73, 41.66, 46.03, 46.31,
82.48, 86.81, 175.1, 179.7; m/z (CI) 234 (M ϩ 18, 100%), 217
(10), 171 (20).
(2R,3R,5S)-5-Benzyloxycarbonylmethyl-2,3,5-trimethyltetra-
hydrofuran-2-carboxylic acid benzyl ester 14
To a solution of nemorensic acid (20 mg, 0.09 mmol) in dichloro-
methane (2 ml) at Ϫ20 ЊC were successively added benzyl
alcohol (10 mg, 0.09 mmol) followed by 1,3-dicyclohexyl-
carbodiimide (28 mg, 0.14 mmol) and a catalytic amount of
4-(dimethylamino)pyridine (2 mg, 0.02 mmol). The reaction
was slowly allowed to warm to room temperature and was
stirred for 12 hours before addition of a saturated solution
of ammonium chloride. The solution was extracted with
dichloromethane (3 × 20 ml), dried over magnesium sulfate and
evaporated in vacuo. The oil obtained was purified by column
chromatography (40 : 60 diethyl ether–petrol) to give the title
compound 14 (12 mg, 0.030 mmol, 35%). (Found M ϩ Hϩ
397.2016, C24H28O5 requires 397.2015, deviation 0.25 ppm),
[α]D ϩ12.6 (c 0.70, CHCl3); νmax(film)/cmϪ1 1728; δH (300 MHz,
Retronecine1
Water (7 ml) was added to a mixture of monocrotaline (229 mg,
1374
J. Chem. Soc., Perkin Trans. 1, 2002, 1369–1375

View Article Online
CDCl3) 0.98 (3H, d, J 7.0), 1.22 (3H, s), 1.34 (3H, s), 1.46 (1H,
t, J 13.0), 2.32 (1H, dd, J 13.0 and 7.0), 2.53 (1H, d, J 13.5),
2.53–2.65 (1H, m), 2.62 (1H, d, J 13.5), 5.00 (2H, s), 5.60 (2H,
d, J 7.0), 7.22–7.30 (10H, m); δC (75 MHz, CDCl3) 14.13, 20.45,
28.04, 40.02, 43.90, 45.23, 66.06, 66.57, 81.47, 85.30, 127.9,
128.1, 128.1, 128.1, 128.4, 135.8, 170.7, 174.6; m/z (CI) 397
(M ϩ 1, 90), 94 (M Ϫ 302, 100%).
4.35 (1H, dd, J 11 and 6.2), 4.54 (1H, dd, J 11 and 10), 4.80–
4.90 (1H, m), 5.08 (2H, s), 7.23–7.38 (5H, m); m/z (ES)
C25H36NO6 requires 446, obtained 446 (100%). It was not
possible to obtain further characterisation on this compound
because of the small mass available.
References
(2R,3R,5S)-5-Carboxymethyl-2,3,5-trimethyltetrahydrofuran-
2-carboxylic acid benzyl ester 15
1 J. T. Hovermale, F. F. Fleming, J. D. White and A. M. Craig,
Heterocycles, 1994, 38, 135 and references therein.
2 A. R. Mattocks, Chemistry and Toxicology of Pyrrolizidine
Alkaloids, Academic Press, Inc., London, 1986; See also D. J.
Robins, Nat. Prod. Rep., 1995, 12, 413 and references therein.
3 D. B. Hagan and D. J. Robins, J. Chem. Soc., Perkin Trans. 1, 1988,
1165; J. J. Tepe and R. M. Williams, J. Am. Chem. Soc., 1999, 121,
2951; J. J. Tepe and R. M. Williams, Angew. Chem., Int. Ed., 1999,
38, 3501.
4 A. Klasek, P. Sedmera, A. Boeva and F. Santavy, Collect. Czech.
Chem. Commum., 1973, 38, 2504; N. T. Nghia, P. Sedmera,
A. Klasek, A. Boeva, S. Dvorakova and F. Santavy, Collect. Czech.
Chem. Commum., 1976, 41, 2952; A. Klasek, P. Sedmera, F. Vokun,
A. Boeva, S. Dvorakova and F. Santavy, Collect. Czech. Chem.
Commum., 1980, 45, 548.
5 M. P. Dillon, N. C. Lee, F. Stappenbeck and J. D. White, J. Chem.
Soc., Chem. Commun., 1995, 1645.
6 L. L. Klein, J. Am. Chem. Soc., 1985, 107, 257.
7 See ref. 5 J. R. Rodríguez, A. Rumbo, L. Castedo and J. L.
Mascareñas, J. Org. Chem., 1999, 64, 4560; T. Honda and
F. Ishikawa, J. Org. Chem., 1999, 64, 5542; B. Liu and K. D. Moeller,
Tetrahedron Lett., 2001, 42, 7163; F. López, L. Castedo and J. L.
Mascareñas, Chem. Eur. J., 2002, 8, 884.
To a solution of nemorensic acid (33 mg, 0.15 mmol) in
dichloromethane (1 ml) at room temperature were successively
added triethylamine (18.5 mg, 0.18 mmol) followed by benzyl
bromide (27 mg, 0.16 mmol). The reaction was stirred at room
temperature for 12 hours. The solvents were then evaporated
and the residue dry loaded on silica to allow purification by
flash column chromatography (EtOAc). The title compound
was isolated as a colourless oil (7 : 1 mixture of regioisomers, 38
mg, 0.12 mmol, 81%). (Found M ϩ Hϩ 307.1540, C17H22O5
requires 307.1545, deviation 1.63 ppm), νmax(film)/cmϪ1 3257,
1735, 1720; δH (400 MHz, CDCl3) 1.12 (3H, d, J 7.0), 1.36 (3H,
s), 1.43 (3H, s), 1.71 (1H, t, J 12.5), 2.02 (1H, dd, J 12.5 and
7.0), 2.32 (1H, d, J 13.5), 2.46–2.58 (1H, m), 2.65 (1H, d,
J 13.5), 5.23 (2H, s), 7.32–7.41 (5H, m); δC (125 MHz, CDCl3)
14.28, 20.46, 26.99, 40.93, 46.29, 46.52, 68.15, 81.41, 86.39,
128.6, 128.9, 129.0, 135.2, 171.4, 177.1; m/z (CI) 324 (M ϩ 18,
100%), 307 (35).
8 T. J. Donohoe, J.-B. Guillermin, C. Frampton and D. S. Walter,
J. Chem. Soc., Chem. Commun., 2000, 465.
9 T. J. Donohoe, M. Helliwell, C. A. Stevenson and T. Ladduwahetty,
Tetrahedron Lett., 1998, 39, 3071; T. J. Donohoe, A. A. Calabrese,
C. A. Stevenson and T. Ladduwahetty, J. Chem. Soc. Perkin Trans. 1,
2000, 3724.
10 T. J. Donohoe, C. A. Stevenson, M. Helliwell, R. Irshad and
T. Ladduwahetty, Tetrahedron: Asymmetry, 1999, 10, 1315.
11 N. A. Petasis and E. I. Bzowej, J. Am. Chem. Soc., 1990, 112,
6392.
12 A. Schmitt and H.-U. Reissig, Synlett, 1990, 40; C. Brückner,
H. Lorey and H.-U. Reissig, Angew. Chem., Int. Ed. Engl., 1986, 25,
556.
13 J. T. Shaw and K. A. Woerpel, Tetrahedron, 1999, 55, 8747.
14 C. H. Larsen, B. H. Ridagway, J. T. Shaw and K. A. Woerpel, J. Am.
Chem. Soc., 1999, 121, 12208.
15 P. H. J. Carlsen, T. Katsuki, V. S. Martin and K. B. Sharpless, J. Org.
Chem., 1981, 46, 3936.
16 S. E. Drewes, I. Antonowitz, P. T. Kaye and P. C. Coleman, J. Chem.
Soc., Perkin Trans. 1, 1981, 287.
17 S. E. Denmark, D. L. Parker Jr. and J. A. Dixon, J. Org. Chem.,
1997, 62, 435.
(2R,3R,5S)-5-(8-Platynecinylcarboxymethyl)-2,3,5-trimethyl-
tetrahydrofuran-2-carboxylic acid benzyl ester 16
To a cold (0 ЊC) solution of compound 15 (19 mg, 0.062 mmol)
in THF (1 ml) was added triethylamine (22 mg, 0.22 mmol)
followed by diethylchlorophosphate (16 mg, 0.093 mmol). The
solution was stirred at 0 ЊC for 1 hour and at room temperature
for a further 30 min. The mixture was then filtered through a
cotton plug and added to a cold (0 ЊC) solution of platynecine-
alkoxide in THF [this solution was previously prepared by
stirring, at 0 ЊC, a mixture of NaH (60% in mineral oil) (8 mg,
0.20 mmol) and platynecine (24 mg, 0.155 mmol) in THF (1 ml)
for 20 min]. The resulting solution was stirred at room temper-
ature for 12 hours before addition of a saturated solution of
ammonium chloride. After extraction with dichloromethane
(3 × 5 ml), the organics were dried using magnesium sulfate,
filtered and evaporated in vacuo. Purification by flash chrom-
atography (1 : 20 : 79 NH4OHsol–MeOH–CHCl3) afforded the
title compound as a colourless oil (ca. 1 mg, 0.002 mmol, 1%).
δH (300 MHz, CDCl3) 1.12 (3H, d, J 7.0), 1.34 (3H, s), 1.49 (5H,
s), 1.96–2.25 (3H, m), 2.34 (1H, dd, J 13.0 and 7.0), 2.58–2.62
(2H, m), 2.66–2.73 (1H, m), 2.78–2.84 (1H, m), 3.00–3.15 (2H,
m), 3.55–3.65 (1H, m), 3.84–3.92 (1H, m), 4.03 (1H, br d, J 8.0),
18 H. Niwa, T. Sakata and K. Yamada, Bull. Chem. Soc. Jpn., 1994, 67,
1990.
19 E. Vedejs, S. Ahmad, S. D. Larsen and S. Westwood, J. Org. Chem.,
1987, 52, 3937; J. D. White, J. C. Amedio Jr., S. Gut, S. Ohira and
L. R. Jayasinghe, J. Org. Chem., 1992, 57, 2270.
J. Chem. Soc., Perkin Trans. 1, 2002, 1369–1375
1375