Synthesis of 7-deoxypancratistatin from carbohydrates by the use of olefin metathesis

Article abstract of DOI:10.1002/chem.200501429

The stereocontrolled synthesis of (+)-7-deoxypancratistatin is described. The convergent synthesis has been achieved by two different strategies, both of which commence from a pentose and piperonal. The latter is converted into allylic bromide 7, which is then coupled with a protected methyl 5-deoxy-5-iodo-D-ribofuranoside in the presence of zinc metal. The first strategy involves a total of only 13 steps from D-ribose and piperonal, but suffers from a low yield in the zinc-mediated reaction between ribofuranoside 9, benzylamine, and bromide 7.The second strategy involves a total of 18 steps from D-xylose and piperonal. The former is converted into ribofuranoside 28, which is coupled with bromide 7 in the presence of zinc, and this is followed by ring-closing olefin metathesis. Subsequent Overman rearrangement, dihydroxylation. and deprotection then affords the natural product.

Full text of DOI:10.1002/chem.200501429

FULL PAPER
DOI: 10.1002/chem.200501429
Synthesis of 7-Deoxypancratistatin from Carbohydrates by the Use of Olefin
Metathesis
Anders E. Håkansson, Anders Palmelund, Henriette Holm, and Robert Madsen*[a]
Abstract: The stereocontrolled synthe-
sis of (+)-7-deoxypancratistatin is de-
scribed. The convergent synthesis has
been achieved by two different strat-
egies, both of which commence from a
pentose and piperonal. The latter is
converted into allylic bromide 7, which
first strategy involves a total of only 13
steps from d-ribose and piperonal, but
suffers from a low yield in the zinc-
mediated reaction between ribofurano-
side 9, benzylamine, and bromide 7.
The second strategy involves a total of
18 steps from d-xylose and piperonal.
The former is converted into ribofura-
noside 28, which is coupled with bro-
mide 7 in the presence of zinc, and this
is followed by ring-closing olefin meta-
thesis. Subsequent Overman rearrange-
ment, dihydroxylation, and deprotec-
tion then affords the natural product.
Keywords: antitumour agents · car-
bohydrates · natural products · ole-
fin metathesis · total synthesis · zinc
is then coupled with
a protected
methyl 5-deoxy-5-iodo-d-ribofurano-
side in the presence of zinc metal. The
Introduction
two of which set up the trans-fused BC ring junction. The
only difference between the two compounds is the lack of a
hydroxy group at the 7-position of 2 compared with that of
1. The interest in 1 and 2 was greatly intensified by the
promising results obtained when the compounds were tested
against the NCI human tumour cell line panel. Both com-
pounds showed activities against a variety of different
cancer types, with a mean panel GI50 value of 91 nm for 1
and with 2 being slightly less active.[3] Their mechanism of
action, however, is still not known in any great detail.[4]
Potent activities against a set of diverse viruses have also
been reported for both compounds.[5] Unfortunately, 1 and 2
are only available in minute quantities by isolation from the
natural sources, and the further clinical development of
these compounds has been hampered by the lack of materi-
al. As a result, it has been an important task to develop
short and asymmetric syntheses of both compounds.[6]
The first total synthesis of racemic 1 was reported by
Danishefsky and Lee in 1989,[7] and six years later Hudlicky
and co-workers described the first enantioselective synthe-
sis.[8] At present, eight different groups have reported syn-
thetic strategies leading to 1.[7–14] Most remarkable is the
synthesis by Trost and Pulley, whereby 1 is prepared in 11%
overall yield from conduritol A acetonide after only 12
steps.[9] Racemic 2 was first synthesised by Ohta and Kimoto
in 1976 as an advanced intermediate in the total synthesis of
another hydroxylated phenanthridone, lycoricidine.[15] A few
years later, enantiopure 2 was prepared for the first time by
Paulsen and Stubbe as an intermediate in their synthesis of
Pancratistatin (1) and the related 7-deoxypancratistatin (2)
are both hydroxylated phenanthridones isolated from the
plant family Amaryllidaceae. Pancratistatin was first isolated
in 1984 by Pettit and co-workers from the Hawaiian plant
Hymenocallis littorale[1] and later 7-deoxypancratistatin was
isolated from Haemanthus kalbreyeri by Ghosal and co-
workers.[2] Both compounds possess six contiguous stereo-
genic centers in the C ring of the phenanthridone skeleton,
[a] Dr. A. E. Håkansson, Dr. A. Palmelund, H. Holm,
Prof. Dr. R. Madsen
Center for Sustainable and Green Chemistry
Department of Chemistry, Building 201
Technical University of Denmark
2800 Kgs. Lyngby (Denmark)
Fax : (+45)4593-3968
E-mail: rm@kemi.dtu.dk
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Chem. Eur. J. 2006, 12, 3243 – 3253
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3243

(+)-lycoricidine.[16] Thus, the first two syntheses of 2 were
actually achieved before the compound was known to be a
natural product. Following its isolation in 1989, the first
total synthesis was reported by Keck and co-workers in
1995.[17] At present, six groups have published syntheses of
2.[15–21] Most notable are the synthesis by Hudlickyꢁs group
and the second-generation synthesis by Keck and co-work-
ers. Hudlickyꢁs group used a whole-cell biooxidation of bro-
mobenzene as the first step to afford (1S,2S)-3-bromocyclo-
hexa-3,5-diene-1,2-diol, which was then converted into 2 by
way of a total of ten linear steps in 3.0% overall yield.[18]
Keck and co-workers employed a radical cyclisation as the
key step in a synthesis that involved a total of 19 steps start-
ing from d-gulono-1,4-lactone and piperonal.[20] The longest
linear sequence consisted of 16 steps from piperonal, and af-
forded 2 in 14% overall yield.
A series of deoxy analogues at the aminocyclitol part
(ring C) of 2 have been synthesised and tested against the
NCI human tumour cell line panel.[3c,22] The results indicated
that at least two hydroxy groups on this ring are essential
for the activities of 1 and 2. Another necessary feature is
the trans-fused BC ring junction formed by the lactam.
These findings have led to the conclusion that enhanced ac-
tivities in pancratistatin-like compounds are not expected
unless they have the fully functionalised aminocyclitol
moiety of ring C in 1 and 2.[3c]
We have recently described a new strategy for the synthe-
sis of enantiopure cyclitols and aminocyclitols.[23] The
method is based on three consecutive organometallic reac-
tions starting from a carbohydrate. Firstly, a methyl iodogly-
coside is fragmented with zinc metal to produce an unsatu-
rated aldehyde, which is then alkylated with a vinyl or an
allyl organometallic reagent in the same pot. The product is
a diene, which is subjected to ring-closing olefin metathesis
to afford a cyclitol. If an amine is added during the zinc-
mediated fragmentation, the intermediate aldehyde is con-
verted into the corresponding imine. In this case, the alkyla-
tion reaction generates an aminodiene, which is transformed
into an aminocyclitol by ring-closing metathesis.[23] We have
applied this method in the synthesis of several cyclitols and
aminocyclitols.[24]
Scheme 1. Retrosynthesis: first-generation synthesis.
O-nucleophile.[18] Alkene 11 can be prepared by ring-closing
metathesis of the corresponding diene, which could be ob-
tained by a zinc-mediated fragmentation of iodofuranoside
9 followed by imine formation with benzylamine and allyla-
tion with bromide 7. The latter could be derived from cin-
namic acid 5, which, in turn, could be formed by a Heck
coupling between bromide 4 and acrylic acid. Bromide 4 is a
known compound and is easily available from piperonal.
Allylic bromide 7: Treatment of piperonal with bromine and
iron filings in glacial acetic acid gave 6-bromopiperonal 3 in
84% yield (Scheme 2).[25] The aldehyde was then oxidised to
Herein, we report the chemical synthesis of 7-deoxypan-
cratistatin (2) by using a zinc-mediated tandem reaction fol-
lowed by ring-closing metathesis as the key steps. The syn-
thesis has been achieved by two different strategies, both of
which start from a carbohydrate and a common substituted
cinnamyl bromide.
Scheme 2. a) Ref. [25]; b) NaClO2, NaH2PO4, acetone, H2O, then MeOH,
H2SO4, 658C; c) acrylic acid, Pd(OAc)2, Ph3P, Bu3N, toluene, 1108C;
d) ClCO2Et, Et3N, THF, 08C, then NaBH4, THF, H2O, 0 8C; e) Ms2O,
Et3N, LiBr, THF, ꢀ408C ! RT; f) MsCl, Et3N, LiCl, THF, ꢀ308C ! RT.
Results and Discussion
Retrosynthesis: first-generation synthesis: Close inspection
of the aminocyclitol moiety revealed that it could be assem-
bled by a metal-mediated fragmentation/allylation reaction
followed by ring-closing metathesis (Scheme 1). The trans-
diol moiety in 2 could be installed from alkene 11 by means
of an epoxidation and a trans-diaxial ring-opening with an
the corresponding carboxylic acid with sodium chlorite, and
this was followed by ester formation in acidic methanol. The
Heck coupling with acrylic acid was carried out with a cata-
lytic amount of palladium(ii) acetate and triphenylphosphine
in the presence of tributylamine to afford cinnamic acid 5 in
3244
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3243 – 3253

Synthesis of 7-Deoxypancratistatin
FULL PAPER
84% yield. The carboxylic acid was selectively reduced in
the presence of the methyl ester by anhydride formation
with ethyl chloroformate followed by treatment with sodium
borohydride. No reduction of the ester functionality was ob-
served and the allylic alcohol 6 was isolated in 77% yield.
At this stage, two different allylic halides, 7 and 8, were pre-
pared to investigate the influence of the halide in the frag-
mentation/allylation reaction. Bromide 7 was obtained by
mesylating the free hydroxy group with methanesulfonic an-
hydride in the presence of lithium bromide, while chloride 8
was prepared in a similar manner by treatment with mesyl
chloride and lithium chloride.
was taken up in dry THF containing benzylamine and a
metal. The mixture was sonicated at 408C while allylic bro-
mide 7 was added over a period of 3 h. Again, indium, zinc,
and magnesium metals were tested in this second reaction.
This procedure gave trace amounts of 10 when indium was
used as the metal, while none of the desired product was ob-
tained with magnesium. When zinc was employed in the al-
lylation reaction, an encouraging 18% of the desired com-
pound was isolated as a single diastereomer. The product
was completely converted into the lactam during the reac-
tion. Under these conditions, the imine is formed in situ in
the allylation reaction. Another procedure was also investi-
gated, in which the imine was formed in a separate opera-
tion with 3 Š molecular sieves in dry THF. Unfortunately,
zinc-mediated allylation of this preformed imine afforded
only an 8% yield of lactam 10. The major by-product in all
of these cases seems to arise from a metal-mediated homo-
coupling of the allylic bromide 7. To determine the stereo-
chemical outcome, 10 was submitted to metathesis with
Grubbsꢁ first-generation catalyst[28] to afford cyclohexene 11
Zinc-mediated reaction with 9: Methyl iodofuranoside 9 is
easily available from d-ribose in two steps[26] and was there-
fore used as a convenient substrate in our earlier methodol-
ogy studies.[23,24c] To identify the optimum conditions for the
fragmentation/allylation reaction, a series of experiments
was performed (Scheme 3). Initial attempts to form 10 by a
1
in 83% yield (Scheme 3). The H NMR coupling constants
in 11 and additional NOE experiments showed that the
lactam was formed with the desired stereochemistry at the
new stereocenters.
A series of experiments was conducted in an attempt to
optimise the yield of lactam 10. Firstly, the allylic bromide 7
was replaced with chloride 8. The zinc-mediated allylation
with this compound gave lactam 10 in 7% yield as a single
diastereomer. Next, the isopropylidene group in 9 was re-
placed by two benzyl groups and by two triethylsilyl (TES)
groups to form 12[29] and 13, respectively. In both cases, the
allylic bromide 7 was used for the allylation. However, the
reaction with 13 gave only a trace amount of the desired
product, while the transformation with 12 gave the corre-
sponding lactam in 12% yield as a 3:1 mixture of two dia-
stereomers. Finally, the reaction was performed with p-anisi-
dine instead of benzylamine. This afforded the allylation
product in 10% yield as a 1:1 mixture of two diastereomers,
and the product was found not to undergo lactamisation
under the reaction conditions. No further attempts were
made to optimise the formation of 10. These results indicate
that the optimum procedure for this transformation involves
fragmentation of the iodoglycoside under aqueous condi-
tions followed by transfer of the formed unsaturated alde-
hyde to an anhydrous solution containing benzylamine and
zinc. The allylic bromide is then slowly added to this solu-
tion.
Lactam 11 was submitted to acidic conditions to hydrolyse
the acetonide and thereby liberate the two hydroxy groups
(Scheme 4). An attempt was made to epoxidise alkene 14 by
the Sharpless procedure with [VO(acac)2]/tBuOOH[30]
(acac=acetylacetonate) to give syn epoxy alcohol 15. Un-
fortunately, this resulted only in a complicated mixture ac-
cording to TLC. Instead, allylic alcohol 14 was epoxidised
with m-chloroperbenzoic acid (m-CPBA) to give epoxy al-
cohol 15 as a single diastereomer. When using commercially
available m-CPBA (50–60% pure), a somewhat sluggish re-
Scheme 3. a) 1) Zn, THF, H2O, 40 8C, ultrasound, 2) Zn, BnNH2, THF,
408C, ultrasound; b) [Ru(=CHPh)(PCy3)2(Cl)2], CH2Cl2. TES=triethyl-
silyl.
tandem procedure were inspired by our previous aminocy-
clitol syntheses.[23] A suspension of 9, benzylamine, and the
appropriate metal in THF was sonicated at 408C while allyl-
ic bromide 7 was added over a period of 3 h. The slow addi-
tion was necessary due to the instability of the allylic bro-
mide under the reaction conditions. When this procedure
was tested with indium, zinc, and magnesium, only degrada-
tion products were observed. One explanation for this dis-
couraging result could be the somewhat slower rate of the
reductive fragmentation of 9 in the absence of water.[23,27] To
address this problem, the fragmentation/allylation sequence
was divided into two separate reactions. Initially, glycoside 9
was fragmented with zinc in a THF/water mixture. Excess
zinc was removed by filtration and the aldehyde was isolat-
ed by extraction. Without further purification, the aldehyde
Chem. Eur. J. 2006, 12, 3243 – 3253
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3245

R. Madsen et al.
action was observed, but this was circumvented by switching
to freshly recrystallised m-CPBA.[31] The relative stereo-
chemistry in the epoxy alcohol 15 was not determined since
both diastereomers would undergo the trans-diaxial ring-
opening to produce the desired product. However, it has
previously been shown that these m-CPBA epoxidations are
directed by the allylic alcohol to give the syn epoxy alco-
hol.[32] Epoxide 15 was then refluxed in an aqueous medium
containing a catalytic amount of sodium benzoate for three
Scheme 5. Retrosynthesis: second-generation synthesis.
same starting materials as the first-generation synthesis, the
only difference being the presence or absence of benzyl-
amine.
Diol 21: In the first experiment, iodofuranoside 9 was al-
lowed to react with allylic bromide 7 in the presence of zinc
(Scheme 6). This afforded a 5:2 mixture of two diastereo-
Scheme 4. a) H+-resin, MeOH, 658C; b) m-CPBA, CH2Cl2; c) BzONa,
H2O, 1008C; d) H2, Pd/C, AcOH.
days[18] to afford tetrol 16. Debenzylation of the benzyl-
amide was achieved by prolonged treatment with hydrogen
in the presence of palladium on activated charcoal in glacial
acetic acid. 7-Deoxypancratistatin (2) was isolated in 22%
overall yield from 11, m.p. 308–3098C, [a]2D3 = +73.8 (c =
0.9, DMF), with physical and spectral data in excellent
agreement with those reported for the natural sub-
stance.[2,17,18] No attempt has been made to optimise the last
four reactions in the synthesis.
This first-generation synthesis employs a total of 13 steps
starting from d-ribose and piperonal, and thus represents
one of the shortest syntheses of (+)-7-deoxypancratistatin
from commercially available starting materials. The longest
linear sequence is 11 steps from piperonal, but unfortunately
the overall yield is only 1.4%. The main reason for the low
yield is the difficult tandem reaction between furanoside 9,
benzylamine, and bromide 7. As a result, it was decided to
develop a second-generation synthesis, in which a different
zinc-mediated tandem reaction would serve as the key step.
Scheme 6. a) 7, Zn, THF, H2O, ultrasound; b) 22, CH2Cl2, 408C; c) (imi-
dazole)2CO, toluene, 508C, then phthalimide, [Pd(PPh3)4], THF, 658C.
meric coupling products, 17 and 18, in 69% yield. It was en-
couraging to observe that the yield of this tandem reaction
improved significantly when it was performed in the absence
of benzylamine. The stereochemical outcome was establish-
ed by NMR after the metathesis reaction. Two other iodo-
furanosides were also investigated. TES-protected furano-
side 13 gave a 1:1 mixture of the two diastereomeric dienes
19 and 20 after removal of the TES groups in the work-up.
The yield in this case increased to an impressive 97%. Un-
protected methyl 5-deoxy-5-iodo-d-ribofuranoside[29] was
Retrosynthesis: second-generation synthesis: In this case,
the target molecule was envisaged as arising from alkene 32
(Scheme 5). Here, the amino group is to be introduced at a
late stage in the synthesis, for example, by an Overman rear-
rangement from diol 21.[33] This molecule can be prepared
by metathesis from the corresponding diene, which, in turn,
may be obtained by a zinc-mediated tandem reaction be-
tween iodofuranoside 9 and bromide 7. Thus, the crucial
step in the second-generation synthesis may use some of the
3246
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3243 – 3253

Synthesis of 7-Deoxypancratistatin
FULL PAPER
also tested and gave a 68% yield of the same two diaster-
eomers in a 3:2 ratio. Diene 19 was converted into the cor-
responding cyclohexene 21 in 75% yield with Hoveydaꢁs
catalyst 22.[34] The metathesis reaction could also be per-
formed with Grubbsꢁ first-generation[28] and second-genera-
tion catalysts,[35] but the conversion was more sluggish in
these cases.
Several experiments were then carried out with a view to
introducing the nitrogen functionality. Firstly, it was at-
tempted to selectively protect the homoallylic alcohol in
diol 21 to make the allylic alcohol available for Overman re-
arrangement.[36] Unfortunately, silylation with one equiva-
lent of TBSOTf or benzylation with BnBr/Bu2SnO predomi-
nately protected the allylic alcohol. At this point, three pro-
tecting group manipulations could be envisaged for blocking
of the homoallylic alcohol. However, this approach was not
desirable at this stage in the synthesis. Instead, it was at-
tempted to invert the allylic alcohol in 21 by either a Mitsu-
Scheme 7. a) 3 steps, ref. [42]; b) AcOH, H2O, 60 8C; c) I2, Ph3P, imida-
zole, THF, 658C; d) Tf2O, pyridine, CH2Cl2, ꢀ208C ! RT, then NaNO2,
DMF; e) TESCl, pyridine; f) TBSOTf, 2,6-lutidine, CH2Cl2, 08C.
isopropylidene group was then removed under acidic condi-
tions to afford diol 25 in 97% yield. Selective iodination at
the 5-position was performed with iodine and triphenylphos-
phine[29] to produce 26 in 88% yield. To invert the configu-
ration at the 3-position, two methods were tested. The free
alcohol was converted into the corresponding triflate by
treatment with triflic anhydride and reacted with either
sodium nitrite[43] or sodium trifluoroacetate.[44] Both reac-
tions gave a clean displacement of the triflate while leaving
the primary iodide unaffected. In the latter case, the formed
trifluoroacetate was hydrolysed in the work-up by treatment
with a 3:1 mixture of acetic acid and methanol. Ribofurano-
side 27 was isolated in 59% yield from 26 by using sodium
nitrite and in 47% yield by using sodium trifluoroacetate.
Finally, a silyl group was introduced at the 3-position to
form 28 and 29, respectively.
Ribofuranosides 27–29 were submitted to the zinc-mediat-
ed tandem reaction to investigate the influence of a bulky
substituent in the 3-position. The dienes were then subjected
to ring-closing metathesis and converted into the corre-
sponding cyclohexenes. After these reactions, furanoside 27
furnished a 37% yield of 30:31 in a 1:3 ratio in favour of the
undesired isomer (Scheme 8). Introducing a bulky protect-
ing group at the 3-position changed the ratio of 30 and 31 to
2:1 in favour of the desired compound. The silyl groups
were removed in the work-up following the fragmentation/
allylation reaction. TES-protected furanoside 28 gave 30:31
in 65% overall yield, while the yield with TBS-protected 29
was 61%. In all cases, traces of the last two isomers were
also observed in isolated yields below 10%. The yields and
the selectivities were not improved by using allylic chloride
8 in the tandem reaction. Catalyst 22[34] was chosen for the
metathesis reaction since it gave a better yield than the
Grubbsꢁ first- and second-generation catalysts.[28,35] The ster-
eochemical outcome of the tandem reaction is the same as
that observed in our earlier methodology studies with iodo-
furanoside 9 and cinnamyl bromide.[23]
nobu reaction[37] or
a selective oxidation/reduction se-
quence.[38] The nitrogen functionality could then potentially
be introduced by epoxidation and ring-opening of the epox-
ide with an appropriate nitrogen nucleophile. However,
none of the experiments aimed at inverting the allylic alco-
hol were successful. Finally, it was attempted to use a palla-
dium-catalysed allylic substitution reaction with a soft nucle-
ophile. Diol 21 was converted into the corresponding cyclic
carbonate, which was then subjected to palladium(0) and
phthalimide. Interestingly, but also very disappointingly, this
reaction only afforded the arene product 23. In the light of
these results, it was decided not to continue the experiments
with diol 21.
New ribofuranoside: Instead, a slightly modified synthesis
was envisaged, employing a ribofuranoside with a benzyl
group in the 2-position. After the fragmentation/allylation
reaction and ring-closing metathesis, this benzyl group will
block the homoallylic alcohol in diol 21. Hence, this modi-
fied strategy will save some protecting group manipulations
late in the synthesis at the expense of some additional steps
to prepare the ribofuranoside starting material.
To the best of our knowledge, there is no efficient litera-
ture method for the preparation of 2-O-benzyl ribofurano-
sides. Kim and co-workers reacted methyl 5-O-tosyl-b-d-ri-
bofuranoside with Bu2SnO and benzyl bromide.[39] This af-
forded a 1:1 mixture of the 2- and 3-benzylated furanosides
in 68% overall yield. We attempted to use this procedure
on d-ribono-1,4-lactone, in which the 2-position often dis-
plays a higher reactivity than the 3-position.[40] Treatment of
5-bromo-5-deoxy-d-ribono-1,4-lactone[41] with Bu2SnO and
benzyl bromide did indeed afford the 2-benzylated lactone
as the major product. Unfortunately, the yield never exceed-
ed 30% and this procedure was therefore abandoned. In-
stead, it was decided to start from d-xylose, which is also a
cheaper carbohydrate than d-ribose, but requires an inver-
sion at C-3. Fischer glycosylation of d-xylose with methanol
followed by isopropylidene protection and benzylation gave
fully protected 24 in 85% overall yield (Scheme 7).[42] The
Completion of the synthesis: With allylic alcohol 30 in hand,
the focus then turned to the Overman rearrangement to in-
troduce the nitrogen functionality.[33] Initially, it was at-
tempted to install the trichloroacetimidate by treating 30
with trichloroacetonitrile in the presence of a catalytic
amount of sodium hydride. This reaction proceeded very
Chem. Eur. J. 2006, 12, 3243 – 3253
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3247

R. Madsen et al.
lective dihydroxylation with osmium tetroxide/NMO[47] to
give the desired diol 33 in 94% yield. Treatment of diol 33
with potassium carbonate in refluxing methanol promoted
methanolysis of the trichloroacetamide and a lactone-to-
lactam transformation yielding 2-O-benzyl-7-deoxypancra-
tistatin (34) in 81% yield. Finally, the benzyl group was re-
moved by hydrogenolysis over Pearlmanꢁs catalyst to afford
7-deoxypancratistatin (2) in 79% yield, m.p. 308–3108C,
[a]2D3 = +72.7 (c = 2.3, DMF), with the same spectral data
as those of the previous product.
Conclusion
In summary, we have described two synthetic strategies for
the preparation of 7-deoxypancratistatin. Both strategies are
based on a zinc-mediated tandem reaction between a ribo-
furanoside and allylic bromide 7, followed by ring-closing
olefin metathesis. The first strategy requires only a total of
13 steps from commercially available starting materials, but
unfortunately it suffers from a low yield in the zinc-mediat-
ed reaction. The second strategy requires a total of 18 steps,
and the longest linear sequence involves 13 steps from d-
xylose giving an overall yield of 4.3%. The two syntheses
are relatively short compared to previous syntheses of 7-de-
oxypancratistatin. They highlight the utility of the zinc-
mediated tandem reaction combined with ring-closing meta-
thesis for converting carbohydrates into important carbocy-
clic structures in relatively few steps.
Scheme 8. a) 7, Zn, THF, H2O, 40 8C, ultrasound, then H+-resin, MeOH,
508C; b) 22, CH2Cl2, 408C; c) CCl3CN, DBU, CH2Cl2, ꢀ458C ! ꢀ208C,
then 1 mmHg, neat, 1208C; d) OsO4, NMO, THF; e) K2CO3, MeOH,
658C; f) H2, Pd(OH)2/C, EtOAc.
slowly and so a different approach with DBU as the base
was chosen. Under these conditions, the desired imidate was
produced very cleanly. To facilitate the Overman rearrange-
ment, heat is normally applied, but a number of additives
have also proven beneficial for the [3,3]-sigmatropic rear-
rangement. Mercury(ii) and palladium(ii) have been shown
to catalyse the rearrangement at or slightly above room
temperature.[33] Other authors have seen great improve-
ments by adding potassium carbonate.[45] Danishefsky and
Lee performed a closely related reaction in their total syn-
thesis of pancratistatin, whereby an allylic trichloroacetimi-
date was rearranged under pyrolysis conditions.[7] Applying
these conditions to our system gave the desired trichloroacet-
amide 32 in an overall yield of 41% from the allylic alcohol
30. Further heating of the reaction mixture led to decompo-
sition. Several experiments were carried out with the aim of
improving the yield, such as by adding palladium(ii) com-
plexes or potassium carbonate, but only decomposition was
observed. The use of a trifluoroacetimidate instead of the
chloro compound was investigated in a final attempt to im-
prove the yield. This modification has been shown to give
an improved yield in the allylic rearrangement in some
cases.[46] The trifluoroacetimidate of 30 was generated at
ꢀ788C with butyllithium as the base and then rearranged in
refluxing toluene to afford the corresponding trifluoroacet-
amide in 30% yield. No attempt was made to further opti-
mise this result due to the precautions necessary when han-
dling trifluoroacetonitrile, which is a highly toxic gas.
Experimental Section
General: Thin-layer chromatography was performed on aluminium plates
precoated with silica gel 60. Compounds were visualised by heating after
dipping in a solution of Ce(SO4)2 (2.5 g) and (NH4)6Mo7O24 (6.25 g) in
10% aqueous H2SO4 (250 mL). Flash column chromatography was per-
formed on silica gel 60 (0.035–0.063 mm), while dry column chromatogra-
phy[48] was carried out with silica gel 60 (0.015–0.040 mm). Optical rota-
tions were measured with a Perkin Elmer 241 polarimeter. IR spectra
were recorded on a Perkin Elmer 1720 Infrared Fourier Transform spec-
trometer. NMR spectra were recorded on a Varian Unity Inova 500 or a
Varian Mercury 300 spectrometer. Mass spectrometry was carried out at
the Department of Chemistry, University of Copenhagen. Microanalyses
were conducted at the Department of Chemistry, University of Copenha-
gen or at the Institute of Physical Chemistry, University of Vienna.
6-Bromo-5-methoxycarbonyl-benzo[1,3]dioxole (4):
NaClO2 (59.4 g, 0.657 mol) and NaH2PO4·H2O (160 g, 1.16 mol) in H2O
(1000 mL) was added dropwise to
suspension of 3[25] (44.83 g,
A
solution of
a
0.196 mol) in acetone (1000 mL) containing 2-methyl-2-butene (250 mL,
2.36 mol), keeping the temperature below 308C. The mixture was stirred
vigorously for 2 h, then adjusted to pH 1 with concentrated aqueous HCl
and extracted with EtOAc (3”250 mL). The combined organic layers
were concentrated in vacuo and the residue was taken up in MeOH
(400 mL) containing concentrated H2SO4 (3 mL). The mixture was stirred
under reflux for 2 days. After the mixture had been cooled to room tem-
perature, H2O (100 mL) was added, which led to the deposition of a pre-
cipitate. This precipitate was collected by filtration and the mother liquor
was partly concentrated in vacuo to produce a second crop of precipitate,
which was also collected by filtration. The combined solids were dis-
solved in CH2Cl2 (300 mL), and this solution was washed with H2O
(250 mL) and saturated aqueous NaHCO3 solution (250 mL), dried
With a feasible route to trichloroacetamide 32 at hand, at-
tention was then turned to the introduction of the cis-diol at
the cyclohexane moiety. This was achieved by a diastereose-
3248
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3243 – 3253

Synthesis of 7-Deoxypancratistatin
FULL PAPER
(MgSO4), and concentrated in vacuo to afford 4 (47.09 g, 93%) as off-
found: C 48.24, H 3.78, Br 26.56; HRMS (FAB) calcd for C12H12BrO4:
white crystals. Rf
=
0.81 (heptane/EtOAc, 1:1); m.p. 89.0–90.08C
298.9919 [M+H]+; found: 298.9899.
(MeOH) (lit.:[49] 88–88.58C); IR (KBr): n˜ = 3056, 2956, 1718, 1615, 1485,
6-((E)-3-Chloro-1-propenyl)-5-methoxycarbonyl-benzo[1,3]dioxole (8): A
solution of 6 (4.01 g, 17.0 mmol) in THF (45 mL) at ꢀ308C under N2 was
treated first with Et3N (3.6 mL, 25.9 mmol) and then LiCl (2.16 g,
51.1 mmol) was added. MsCl (1.5 mL, 19.4 mmol) was added dropwise
over a period of 10 min and the reaction mixture was allowed to warm
slowly to room temperature. After 3 h, the reaction had reached comple-
tion and the mixture was quenched by adding 1m HCl (25 mL). The re-
sulting mixture was extracted with EtOAc (3”40 mL), and the combined
organic layers were dried (Na2SO4) and concentrated to dryness under
reduced pressure. The residue was purified by flash column chromatogra-
1244, 1034 cmꢀ1 1H NMR (300 MHz, CDCl3): d = 7.33 (s, 1H), 7.10 (s,
;
1H), 6.05 (s, 2H), 3.90 ppm (s, 3H); 13C NMR (75 MHz, CDCl3): d =
165.8, 151.0, 147.2, 124.7, 115.0, 114.5, 111.1, 102.6, 52.4 ppm; elemental
analysis calcd (%) for C9H7BrO4: C 41.73, H 2.72, Br 30.84; found: C
41.67, H 2.65, Br 30.85.
6-((E)-2-Carboxyvinyl)-5-methoxycarbonyl-benzo[1,3]dioxole (5): A de-
gassed solution of Bu3N (46 mL, 0.19 mol), acrylic acid (5.5 mL,
0.080 mol), and Ph3P (629 mg, 2.40 mmol) in toluene (35 mL) under N2
atmosphere was heated to reflux and then Pd(OAc)2 (229 mg, 1.02 mmol)
and a solution of 4 (10.03 g, 0.0387 mol) in toluene (20 mL) were added
sequentially. The reaction mixture was stirred under reflux overnight and
then cooled to room temperature, whereupon 2m HCl (200 mL) was
added. The resulting mixture was added to refluxing EtOAc (2000 mL),
which was stirred until all the solids had dissolved. The solution obtained
was washed with brine (200 mL) and dried (MgSO4). Most of the solvent
was removed in vacuo to leave an off-white solid, which was collected by
filtration. The solid was washed with ice-cooled diethyl ether (2”
100 mL) to give 5 (8.09 g, 84%) as a white powder. Rf = 0.31 (hexane/
EtOAc/AcOH, 1:1:0.01); m.p. 262–2648C (EtOAc); IR (KBr): n˜ = 2962,
phy (hexane/EtOAc, 9:1) to afford 8 (3.60 g, 83%) as a white solid. Rf
0.53 (hexane/EtOAc, 3:1); m.p. 76.0–76.58C (hexane/EtOAc); IR (KBr):
n˜
2953, 1714, 1505, 1237, 1117, 1037 cmꢀ1 1H NMR (300 MHz,
=
=
;
CDCl3): d = 7.53 (dt, J = 15.5, 0.9 Hz, 1H), 7.37 (s, 1H), 7.00 (s, 1H),
6.09 (dt, J = 15.5, 7.2 Hz, 1H), 6.03 (s, 2H), 4.25 (dd, J = 7.2, 0.9 Hz,
2H), 3.87 ppm (s, 3H); 13C NMR (75 MHz, CDCl3): d = 166.8, 151.2,
147.5, 134.6, 133.0, 126.8, 122.5, 110.4, 107.4, 102.2, 52.3, 45.6 ppm; ele-
mental analysis calcd (%) for C12H11ClO4: C 56.60, H 4.35; found: C
56.53, H 4.29.
(7R,8R)-6-Benzyl-7-((4S,5R)-2,2-dimethyl-5-vinyl-[1,3]dioxolan-4-yl)-8-
vinyl-7,8-dihydro-6H-[1,3]dioxolo[4,5-g]isoquinolin-5-one (10): Freshly
activated Zn[23] (2.06 g, 31.5 mmol) was added to a solution of 9 (1.40 g,
4.46 mmol) in THF/H2O (3:1, 60 mL) and the mixture was sonicated at
408C until complete conversion of the starting material, as monitored by
TLC. The mixture was then filtered through a pad of Celite, which was
successively rinsed with CH2Cl2 and H2O. The combined filtrate was ex-
tracted with CH2Cl2 (3”100 mL), and the combined organic layers were
dried (Na2SO4) and concentrated in vacuo. The crude product was redis-
solved in THF (40 mL) and then BnNH2 (0.58 mL, 5.31 mmol) and fresh-
ly activated Zn (3.30 g, 50.5 mmol) were added under N2. Bromide 7
(3.36 g, 11.2 mmol) was dissolved in THF (7 mL) and this solution was
added in small portions to the mixture over 1 h with sonication at 408C.
The mixture was sonicated for an additional 4.5 h at 408C and then left
overnight at room temperature. It was then quenched with H2O and fil-
tered through a pad of Celite, and the pad was successively rinsed with
CH2Cl2 and H2O. The organic phase was washed with H2O (2”100 mL)
and saturated aqueous NaHCO3 solution (100 mL), dried (Na2SO4), and
concentrated in vacuo. The residue was purified twice by column chroma-
tography (hexane/EtOAc, 5:1, followed by CH2Cl2/EtOAc, 19:1) to
2568, 1731, 1681, 1506, 1237, 1121, 1037 cmꢀ1 1H NMR (300 MHz,
;
[D6]DMSO): d = 12.36 (brs, 1H), 8.22 (d, J = 15.8 Hz, 1H), 7.45 (s,
1H), 7.34 (s, 1H), 6.40 (d, J = 16.2 Hz, 1H), 6.17 (s, 2H), 3.82 ppm (s,
3H); 13C NMR (75 MHz, [D6]DMSO): d = 168.7, 167.2, 152.0, 149.7,
142.6, 132.2, 125.5, 121.8, 110.6, 108.2, 103.7, 53.5 ppm; elemental analysis
calcd (%) for C12H10O6: C 57.60, H 4.03; found: C 57.46, H 3.97.
6-((E)-3-Hydroxy-1-propenyl)-5-methoxycarbonyl-benzo[1,3]dioxole (6):
A suspension of 5 (29.37 g, 0.124 mol) in THF (400 mL) at 08C under N2
was first treated with Et3N (22 mL, 0.158 mol) and then ethyl chlorofor-
mate (14.0 mL, 0.146 mol) was added dropwise. The mixture was stirred
at 08C for 1.5 h and then filtered into ice-cold H2O (75 mL). NaBH4
(12.0 g, 0.317 mol) was added in small portions to the resulting solution,
keeping the temperature at 08C. The mixture was stirred for 3 h at 08C
and then the reaction was quenched by slowly adding 2m HCl (250 mL).
The resulting mixture was extracted with EtOAc (3”200 mL) and the
combined organic layers were washed with saturated aqueous NaHCO3
solution (2”200 mL), dried (Na2SO4), and concentrated in vacuo. The
residue was recrystallised from heptane/EtOAc and the mother liquor
was concentrated in vacuo and purified by flash column chromatography
(heptane/EtOAc, 9:1) to give a combined yield of 6 of 22.45 g (77%) as
afford 10 (350 mg, 18%) as a clear oil that solidified on standing. Rf
=
a
white solid. Rf
(hexane/EtOAc); IR (KBr): n˜
1260, 1120, 1038 cmꢀ1 1H NMR (300 MHz, CDCl3): d = 7.39 (dt, J =
=
0.41 (hexane/EtOAc, 1:1); m.p. 105.0–106.08C
0.38 (hexane/EtOAc, 3:1); m.p. 46.5–48.08C (hexane); [a]2D3 = ꢀ131.2 (c
= 1.0, CHCl3); 1H NMR (300 MHz, CDCl3): d = 7.54 (s, 1H), 7.24–7.21
(m, 5H), 6.41 (s, 1H), 5.95–5.93 (m, 2H), 5.78–5.66 (m, 1H), 5.66 (d, J =
14.4 Hz, 1H), 5.30–5.19 (m, 3H), 4.71 (dt, J = 10.3, 1.2 Hz, 1H), 4.50 (dt,
J = 17.0, 1.3 Hz, 1H), 4.29 (dd, J = 8.9, 6.0 Hz, 1H), 4.15 (dd, J = 10.0,
5.9 Hz, 1H), 4.03 (d, J = 14.5 Hz, 1H), 3.39 (dd, J = 10.2, 1.3 Hz, 1H),
3.01 (dd, J = 6.2, 1.2 Hz, 1H), 1.48 (s, 3H), 1.18 ppm (s, 3H); 13C NMR
=
3524, 3434, 2900, 1713, 1610, 1488,
;
15.7, 1.3 Hz, 1H), 7.35 (s, 1H), 6.99 (s, 1H), 6.15 (dt, J = 15.9, 5.8 Hz,
1H), 6.02 (s, 2H), 4.32 (dd, J = 5.7, 1.4 Hz, 2H), 3.85 (s, 3H), 1.93 ppm
(brs, 1H); 13C NMR (75 MHz, CDCl3): d = 167.1, 151.1, 147.0, 135.6,
130.6, 130.0, 122.0, 110.1, 107.3, 102.0, 63.9, 52.2 ppm; elemental analysis
calcd (%) for C12H12O5: C 61.01, H 5.12; found: C 60.88, H 5.11.
(75 MHz, CDCl3): d
= 162.8, 150.9, 147.6, 138.1, 137.4, 133.8, 132.3,
129.7, 128.4, 127.5, 123.2, 120.2, 116.8, 109.5, 108.8, 107.9, 101.8, 79.5,
79.4, 58.3, 50.5, 42.9, 28.4, 25.6 ppm; elemental analysis calcd (%) for
C26H27NO5: C 72.04, H 6.28, N 3.23; found: C 71.81, H 6.23, N 3.19.
6-((E)-3-Bromo-1-propenyl)-5-methoxycarbonyl-benzo[1,3]dioxole (7): A
solution of 6 (4.34 g, 18.4 mmol) in THF (70 mL) at ꢀ408C under Ar was
treated first with Et3N (4.0 mL, 28.3 mmol) and then LiBr (5.15 g,
59.3 mmol) was added. Ms2O (4.82 g, 27.7 mmol) was added in small por-
tions and the reaction mixture was allowed to warm slowly to room tem-
perature. After the reaction mixture had been stirred for 4 h, the reaction
had gone to completion and the mixture was quenched by adding 1m
HBr (50 mL). The resulting mixture was extracted with EtOAc (3”
50 mL) and the combined organic layers were dried (MgSO4) and con-
centrated in vacuo. The residue was purified by flash column chromato-
graphy (hexane/EtOAc, 9:1) to afford 7 (4.94 g, 90%) as a white solid.
Rf = 0.40 (heptane/EtOAc, 3:1); m.p. 65.5–66.08C (heptane); IR (KBr):
(3R,4S,4aR,11bR)-5-Benzyl-3,4-isopropylidenedioxy-3,4a,5,11b-tetrahy-
dro-4H-[1,3]dioxolo[4,5-j]phenanthridin-6-one
(11):
[Ru(=CHPh)-
(PCy3)2(Cl)2][28] (34 mg, 0.041 mmol) was added to a degassed solution of
10 (149 mg, 0.344 mmol) in CH2Cl2 (40 mL) under N2 and then the
system was degassed once more with N2. The mixture was protected from
light and was left to stir for 3 days at room temperature. Thereafter, a
1.5m solution of P(CH2OH)3 in 2-propanol (1 mL) was added and the
mixture was stirred overnight at 408C. It was then cooled and washed
with H2O (3”20 mL) and the organic layer was dried (Na2SO4) and con-
centrated in vacuo. The residue was purified by dry column chromatogra-
phy (EtOAc/hexane, 0:1 ! 1:0) to give 11 (116 mg, 83%) as white crys-
n˜
= ;
2954, 1711, 1608, 1500, 1486, 1230, 1115, 1034 cmꢀ1 1H NMR
(500 MHz, CDCl3): d = 7.51 (brd, J = 15.6 Hz, 1H), 7.38 (s, 1H), 7.01
(s, 1H), 6.18 (dt, J = 15.5, 7.8 Hz, 1H), 6.04 (s, 2H), 4.17 (dd, J = 7.6,
0.9 Hz, 2H), 3.87 ppm (s, 3H); 13C NMR (75 MHz, CDCl3): d = 166.8,
151.1, 147.5, 134.4, 133.1, 127.0, 122.4, 110.3, 107.2, 102.1, 52.3, 33.6 ppm;
elemental analysis calcd (%) for C12H11BrO4: C 48.18, H 3.71, Br 26.71;
tals. Rf
= 0.51 (hexane/EtOAc, 2:1); m.p. 213.0–214.58C (hexane/
EtOAc); [a]2D3 = +24.5 (c = 0.6, CHCl3); 1H NMR (300 MHz, CDCl3):
d = 7.59 (s, 1H), 7.21–7.10 (m, 5H), 6.84 (s, 1H), 6.27 (dt, J = 10.0,
1.7 Hz, 1H), 6.02 (dt, J = 9.8, 3.1 Hz, 1H), 5.96 (s, 2H), 5.35 (d, J =
Chem. Eur. J. 2006, 12, 3243 – 3253
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3249

R. Madsen et al.
15.8 Hz, 1H), 5.01 (d, J = 15.9 Hz, 1H), 4.55–4.51 (m, 1H), 4.27 (dd, J
= 9.2, 6.9 Hz, 1H), 3.59 (dd, J = 12.0, 9.2 Hz, 1H), 3.44 (dd, J = 11.9,
1.7 Hz, 1H), 1.26 (s, 3H), 1.21 ppm (s, 3H); 13C NMR (75 MHz, CDCl3):
d = 165.6, 151.2, 147.0, 140.3, 133.8, 128.4, 128.1, 126.9, 126.6, 126.1,
123.7, 109.4, 109.3, 103.9, 101.8, 75.0, 71.9, 60.8, 46.6, 38.7, 27.6, 25.4 ppm;
elemental analysis calcd (%) for C24H23NO5: C 71.10, H 5.72, N 3.45;
found: C 70.79, H 5.67, N 3.40.
ited amount of 25 was purified by flash column chromatography (hep-
tane/EtOAc, 1:1 ! 7:10) in order to obtain full characterisation data for
the two anomers. a-25: The analytical data are in accordance with litera-
ture values.[50] b-25: Rf = 0.16 (heptane/EtOAc, 1:1); [a]2D3 = ꢀ39.1 (c =
1.4, CHCl3); IR (film): n˜ = 3435, 2930, 1641, 1455, 1195, 1103, 1039 cmꢀ1
;
1H NMR (300 MHz, CDCl3): d = 7.42–7.27 (m, 5H), 4.92 (s, 1H), 4.66
(d, J = 12.0 Hz, 1H), 4.60 (d, J = 11.9 Hz, 1H), 4.43–4.31 (m, 2H), 3.96
(d, J = 1.7 Hz, 1H), 3.88 (d, J = 4.1 Hz, 2H), 3.41 ppm (s, 3H); 13C
NMR (75 MHz, CDCl3): d = 137.4, 128.7, 128.1, 127.9, 107.3, 88.4, 82.4,
76.1, 72.1, 62.4, 55.6 ppm; HRMS (FAB) calcd for C13H18O5Na: 277.1052
[M+Na]+; found: 277.1047.
Methyl
5-deoxy-5-iodo-2,3-bis-O-triethylsilyl-d-ribofuranoside
(13):
TESCl (2.7 mL, 16.1 mmol) was added to a solution of methyl 5-deoxy-5-
iodo-d-ribofuranoside[29] (1.464 g, 5.34 mmol) in pyridine (40 mL) under
N2 and the mixture was stirred at room temperature for 3 h. It was then
diluted with hexane (50 mL), washed with H2O (3”30 mL), dried
(Na2SO4), and concentrated in vacuo. The residue was purified by flash
column chromatography (hexane/EtOAc/Et3N, 99:0:1 ! 97:2:1) to give a
4:1 anomeric mixture of 13 (2.236 g, 83%) as a colourless oil. Rf = 0.93
(hexane/EtOAc, 3:1); 1H NMR (300 MHz, CDCl3): d = 4.82 (d, J =
3.8 Hz, 0.2H), 4.69 (s, 0.8H), 4.06–3.92 (m, 1.8H), 3.89–3.72 (m, 1.2H),
3.51–3.20 (m, 2H), 3.41 (s, 0.6H), 3.39 (s, 2.4H), 0.97 (t, J = 8.1 Hz, 9H),
0.96 (d, J = 8.1 Hz, 9H), 0.70–0.57 ppm (m, 12H); 13C NMR (75 MHz,
CDCl3, b-isomer): d = 108.2, 80.6, 77.2, 75.9, 55.6, 9.6, 7.0, 6.9, 5.0 ppm
(2”); elemental analysis calcd (%) for C18H39IO4Si2: C 43.02, H 7.82;
found: C 42.61, H 7.46.
Methyl 2-O-benzyl-5-deoxy-5-iodo-d-xylofuranoside (26): A solution of
25 (9.115 g, 35.8 mmol), Ph3P (14.17 g, 54.0 mmol), and imidazole
(4.942 g, 72.6 mmol) in THF (150 mL) was heated to reflux and a solu-
tion of I2 (13.72 g, 54.1 mmol) in THF (40 mL) was added slowly. The sol-
ution was refluxed for 45 min, then cooled to room temperature, filtered,
and concentrated in vacuo. The residue was purified by flash column
chromatography (heptane/EtOAc, 3:1 ! 2:1) to give a combined yield of
a-26 and b-26 of 11.50 g (88%) as colourless oils. Pure samples of both
anomers were isolated. a-26: Rf = 0.50 (heptane/EtOAc, 1:1); [a]D23
+75.7 (c = 3.4, CHCl3); IR (KBr): n˜ = 3470, 2910, 1198, 1118, 1033,
992 cmꢀ1 1H NMR (300 MHz, CDCl3): d = 7.41–7.29 (m, 5H), 4.84 (d,
=
;
J = 4.4 Hz, 1H), 4.70 (d, J = 12.0 Hz, 1H), 4.64 (d, J = 11.9 Hz, 1H),
4.47 (dd, J = 6.5, 5.3 Hz, 1H), 4.37 (q, J = 6.6 Hz, 1H), 3.92 (dd, J =
5.0, 4.5 Hz, 1H), 3.41 (s, 3H), 3.30 (dd, J = 10.0, 7.1 Hz, 1H), 3.19 (dd,
J = 10.0, 6.2 Hz, 1H), 1.97 ppm (brs, 1H); 13C NMR (75 MHz, CDCl3):
d = 137.7, 128.8, 128.5, 128.3, 100.8, 85.3, 77.6, 75.3, 73.2, 55.7, 3.0 ppm;
HRMS (FAB) calcd for C13H17IO4: 364.0172 [M]+; found: 364.0168. b-26:
Rf = 0.64 (heptane/EtOAc, 1:1); [a]2D3 = ꢀ56.8 (c = 4.3, CHCl3); IR
(7R,8S)-7-((1R,2R)-1,2-Dihydroxy-but-3-enyl)-8-vinyl-7,8-dihydro-1,3-
dioxolo[4,5-g,2]benzopyran-5-one (19): Freshly activated Zn[23] (140 mg,
2.14 mmol) was added to a solution of 13 (102 mg, 0.203 mmol) in THF/
H2O (4:1, 4 mL) under N2. The slurry was sonicated at 408C while a solu-
tion of 7 (201 mg, 0.672 mmol) in THF (2 mL) was added by means of a
syringe pump over a period of 4 h. After an additional 2 h at 408C under
sonication, the reaction mixture was filtered through a pad of Celite,
which was subsequently rinsed with EtOAc and H2O. The filtrate was ex-
tracted with EtOAc (3”25 mL) and the combined organic layers were
dried (Na2SO4) and concentrated in vacuo. The residue was redissolved
in MeOH (10 mL) and heated at 408C for 2 h in the presence of an
acidic ion-exchange resin (Amberlite IR-120, 4 mL). The mixture was
then filtered and concentrated in vacuo. The residue was purified by
flash column chromatography (hexane/EtOAc, 3:2) to give 19 (31 mg,
50%) as a white solid. Rf = 0.59 (heptane/EtOAc, 2:1); m.p. 50–528C
(CH2Cl2); [a]2D3 = +214.3 (c = 1.7, CHCl3); IR (KBr): n˜ = 3414, 2912,
(KBr): n˜
= ;
3496, 2927, 1455, 1367, 1103, 1025, 930 cmꢀ1 1H NMR
(300 MHz, CDCl3): d = 7.41–7.30 (m, 5H), 5.01 (s, 1H), 4.61 (s, 2H),
4.55 (dt, J = 7.6, 3.8 Hz, 1H), 4.25 (brd, J = 3.2 Hz, 1H), 4.01 (s, 1H),
3.38 (s, 3H), 3.32 (d, J = 7.7 Hz, 2H), 2.74 ppm (brs, 1H); 13C NMR
(75 MHz, CDCl3): d = 137.2, 128.7, 128.2, 127.8, 107.2, 86.2, 83.9, 73.7,
72.1, 55.4, 1.9 ppm; HRMS (FAB) calcd for C13H17IO4: 364.0172 [M]+;
found: 364.0175.
Methyl 2-O-benzyl-5-deoxy-5-iodo-d-ribofuranoside (27): A solution of
26 (501 mg, 1.38 mmol) in CH2Cl2 (10 mL) was cooled to ꢀ208C under
Ar. Pyridine (0.50 mL, 6.21 mmol) was added, and then Tf2O (0.45 mL,
2.73 mmol) was added dropwise. The mixture was stirred for 2.5 h at
room temperature and then the reaction was quenched with ice-cold 2m
HCl (20 mL). The mixture was extracted with CH2Cl2 (2”30 mL) and
the combined organic layers were washed with saturated aqueous
NaHCO3 solution (40 mL), dried (MgSO4), and concentrated in vacuo.
The residue was redissolved in DMF (5 mL) under Ar and NaNO2
(392 mg, 5.68 mmol) was added. The resulting mixture was stirred for 4 h
at room temperature, then quenched with H2O (50 mL) and extracted
with diethyl ether (5”30 mL). The combined organic layers were dried
(MgSO4) and concentrated in vacuo. The residue was purified by flash
column chromatography (heptane/EtOAc, 4:1) to give a combined yield
of a-27 and b-27 of 298 mg (59%) as a colourless oil. Pure samples of
both isomers were isolated. a-27: Rf = 0.18 (heptane/EtOAc, 2:1); [a]D23
= +24.4 (c = 1.5, CHCl3); IR (KBr): n˜ = 3526, 2930, 1454, 1197, 1127,
1
1708, 1618, 1482, 1270, 1035 cmꢀ1; H NMR (300 MHz, CDCl3): d = 7.45
(s, 1H), 6.62 (s, 1H), 6.10–5.90 (m, 4H), 5.41 (dt, J = 17.4, 0.9 Hz, 1H),
5.30 (dt, J = 10.2, 0.9 Hz, 1H), 5.20 (brd, J = 10.5 Hz, 1H), 5.10 (brd, J
= 17.1 Hz, 1H), 4.55 (dd, J = 6.9, 3.0 Hz, 1H), 4.29 (brt, J = 5.1 Hz,
1H), 3.98 (dd, J = 6.9, 4.5 Hz, 1H), 3.48 (dd, J = 8.4, 3.0 Hz, 1H), 3.05
(brs, 1H), 2.70 ppm (brs, 1H); 13C NMR (75 MHz, CDCl3): d = 164.5,
153.1, 148.0, 138.5, 136.2, 133.8, 119.3, 118.4, 118.0, 109.5, 107.4, 102.4,
80.2, 73.4, 72.6, 44.9 ppm; elemental analysis calcd (%) for C16H16O6: C
63.15, H 5.30; found: C 62.84, H 5.32.
(3R,4R,4aR,11bS)-3,4-Dihydroxy-3,4,4a,11b-tetrahydro-6H-
[1,3]benzodioxolo[5,6-c,1]benzopyran-6-one (21): Catalyst 22[34] (71 mg,
0.113 mmol) was added to a solution of 19 (690 mg, 2.28 mmol) in
CH2Cl2 (12 mL) under Ar and the mixture was stirred under reflux for
5 h. The solvent was then removed in vacuo and the residue was crystal-
lised from absolute EtOH to give 21 (473 mg, 75%) as a white solid. Rf
= 0.40 (hexane/EtOAc, 1:4); m.p. 231.5–233.08C (EtOH); [a]2D3 = ꢀ20.0
(c = 0.7, DMSO); IR (KBr): n˜ = 3448, 2908, 1707, 1615, 1484, 1285,
1088, 1027, 699 cmꢀ1 1H NMR (500 MHz, CDCl3): d = 7.42–7.30 (m,
;
5H), 4.93 (d, J = 4.2 Hz, 1H), 4.74 (d, J = 11.7 Hz, 1H), 4.64 (d, J =
11.8 Hz, 1H), 4.07 (dt, J = 5.0, 2.3 Hz, 1H), 3.93 (dd, J = 6.0, 4.0 Hz,
1H), 3.90 (dd, J = 6.4, 2.1 Hz, 1H), 3.43 (s, 3H), 3.31 (d, J = 5.0 Hz,
2H), 2.96 ppm (brs, 1H); 13C NMR (75 MHz, CDCl3): d = 137.0, 128.7,
128.4 (2”), 102.9, 84.3, 77.3, 72.7 (2”), 55.3, 7.6 ppm; HRMS (FAB)
calcd for C13H16IO4: 363.0093 [MꢀH]+; found: 363.0107. b-27: Rf = 0.41
1034 cmꢀ1 1H NMR (500 MHz, [D6]DMSO): d = 7.30 (s, 1H), 7.14 (s,
;
1H), 6.15 (s, 1H), 6.13 (s, 1H), 5.57 (dd, J = 10.2, 0.8 Hz, 1H), 5.43 (dd,
J = 10.2, 1.7 Hz, 1H), 5.20 (dd, J = 3.8, 0.9 Hz, 1H), 4.83 (t, J = 4.0 Hz,
1H), 4.80 (d, J = 7.7 Hz, 1H), 4.21–4.15 (m, 1H), 4.02 (dd, J = 8.5,
3.8 Hz, 1H), 3.71–3.66 ppm (m, 1H); 13C NMR (125 MHz, [D6]DMSO):
d = 162.9, 152.3, 147.1, 138.3, 129.6, 125.9, 117.4, 107.8, 107.7, 102.1, 77.0,
66.8, 64.1, 33.5 ppm; elemental analysis calcd (%) for C14H12O6: C 60.87,
H 4.38; found: C 60.68, H 4.24.
Methyl 2-O-benzyl-d-xylofuranoside (25): Furanoside 24[42] (17.0 g,
57.8 mmol) was dissolved in 70% aqueous AcOH (75 mL) and the solu-
tion was heated at 608C for 1 h. It was then concentrated in vacuo and
the remaining volatiles were coevaporated with absolute EtOH (2”
50 mL) and toluene (50 mL) to give 25 (14.25 g, 97%) as a syrup. A lim-
(heptane/EtOAc, 2:1); [a]D23
=
+0.2 (c = 4.1, CHCl3); IR (KBr): n˜ =
1H NMR
3420, 2954, 2906, 1455, 1243, 1125, 1076, 953, 736 cmꢀ1
;
(500 MHz, CDCl3): d = 7.41–7.31 (m, 5H), 4.91 (s, 1H), 4.74 (d, J =
11.8 Hz, 1H), 4.63 (d, J = 11.8 Hz, 1H), 4.17–4.11 (m, 1H), 3.99 (brq, J
= 5.8 Hz, 1H), 3.93 (d, J = 5.4 Hz, 1H), 3.38 (s, 3H), 3.37 (dd, J = 10.6,
5.6 Hz, 1H), 3.28 (dd, J = 10.6, 6.6 Hz, 1H), 2.65 ppm (brd, J = 8.8 Hz,
1H); 13C NMR (75 MHz, CDCl3): d = 136.7, 128.6, 128.3, 128.0, 105.6,
83.7, 82.4, 74.7, 73.0, 55.3, 7.7 ppm; HRMS (FAB) calcd for C13H16IO4:
363.0093 [MꢀH]+; found: 363.0106.
3250
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3243 – 3253

Synthesis of 7-Deoxypancratistatin
FULL PAPER
Methyl
2-O-benzyl-5-deoxy-5-iodo-3-O-triethylsilyl-d-ribofuranoside
rified by flash column chromatography (heptane/EtOAc, 4:1) to give a
clear oil (538 mg, Rf = 0.08 (heptane/EtOAc, 3:1)). This oil was dis-
solved in CH2Cl2 (20 mL) and the solution was degassed under Ar. Cata-
lyst 22[34] (41 mg, 0.065 mmol) was added and the mixture was stirred
under reflux for 3.5 h. The solvent was then removed in vacuo and the
residue was purified by flash column chromatography (heptane/EtOAc,
2:1) to give 30 (330 mg, 42%) and 31 (184 mg, 23%), both as white
foams. 30: Rf = 0.43 (hexane/EtOAc, 1:1); m.p. 70–728C (CH2Cl2); [a]D23
= ꢀ140.9 (c = 1.5, CHCl3); IR (KBr): n˜ = 3419, 2911, 1713, 1617, 1481,
(28): TESCl (0.80 mL, 4.77 mmol) was added to a solution of 27 (719 mg,
1.97 mmol) in pyridine (10 mL) under Ar. The reaction mixture was stir-
red at room temperature for 2 h, and was then diluted with hexane
(30 mL) and washed with H2O (30 mL) and brine (30 mL). The organic
layer was dried (Na2SO4) and concentrated in vacuo, and the residue was
purified by flash column chromatography (hexane/EtOAc, 15:1) to give a
combined yield of a-28 and b-28 of 916 mg (97%) as a colourless oil.
Pure samples of both isomers were isolated. a-28: Rf = 0.49 (heptane/
EtOAc, 3:1); [a]2D3 = +76.8 (c = 1.4, CHCl3); IR (film): n˜ = 2953, 2875,
1402, 1280, 1117, 1035 cmꢀ1 1H NMR (500 MHz, CDCl3): d = 7.47 (s,
;
1455, 1240, 1175, 1018, 842 cmꢀ1; H NMR (500 MHz, CDCl3): d = 7.41–
1H), 7.39–7.30 (m, 5H), 6.73 (s, 1H), 6.06 (d, J = 1.0 Hz, 1H), 6.05 (d, J
= 1.0 Hz, 1H), 5.73 (brd, J = 10.5 Hz, 1H), 5.47 (ddd, J = 10.3, 3.9,
2.0 Hz, 1H), 4.85 (brt, J = 4.2 Hz, 1H), 4.76 (d, J = 11.6 Hz, 1H), 4.72
(d, J = 11.6 Hz, 1H), 4.53–4.49 (m, 1H), 4.14 (brt, J = 4.4 Hz, 1H),
3.61–3.58 (m, 1H), 2.42 ppm (brs, 1H); 13C NMR (75 MHz, CDCl3): d =
163.6, 153.0, 147.9, 137.7, 137.3, 129.7, 128.9, 128.6, 128.2, 126.3, 118.0,
109.6, 107.2, 102.2, 75.1, 75.0, 74.4, 64.7, 34.6 ppm; HRMS (FAB) calcd
for C21H19O6: 367.1181 [M+H]+; found: 367.1191. 31: Rf = 0.20 (hexane/
EtOAc, 1:1); 1H NMR (300 MHz, CDCl3): d = 7.48 (s, 1H), 7.40–7.29
(m, 5H), 6.70 (s, 1H), 6.06 (d, J = 1.0 Hz, 1H), 6.05 (d, J = 1.0 Hz,
1H), 6.04–6.01 (m, 1H), 5.49 (brd, J = 9.9 Hz, 1H), 4.97 (m, 1H), 4.89
(d, J = 12.0 Hz, 1H), 4.69 (d, J = 12.0 Hz, 1H), 4.44–4.38 (m, 1H), 3.72
(dd, J = 5.1, 1.8 Hz, 1H), 3.45–3.42 (m, 1H), 2.82 ppm (brd, J = 9.9 Hz,
1H); 13C NMR (75 MHz, CDCl3): d = 163.1, 153.0, 147.9, 137.5, 136.9,
129.5, 128.7, 128.1 (2”), 125.4, 118.2, 109.8, 106.6, 102.3, 76.7, 74.5, 70.5,
63.5, 39.0 ppm.
1
7.26 (m, 5H), 4.91 (d, J = 4.0 Hz, 1H), 4.69 (d, J = 11.9 Hz, 1H), 4.65
(d, J = 12.0 Hz, 1H), 3.93 (dd, J = 6.5, 5.3 Hz, 1H), 3.80–3.75 (m, 2H),
3.49–3.43 (m, 1H), 3.45 (s, 3H), 3.30 (dd, J = 11.0, 4.1 Hz, 1H), 0.96 (t,
J = 7.9 Hz, 9H), 0.69–0.60 ppm (m, 6H); 13C NMR (75 MHz, CDCl3): d
= 138.0, 128.4, 128.3, 127.9, 103.2, 81.1, 77.9, 74.3, 73.0, 55.9, 9.2, 7.0,
5.0 ppm; HRMS (FAB) calcd for C19H30IO4Si: 477.0958 [MꢀH]+; found:
477.0973. b-28: Rf = 0.53 (heptane/EtOAc, 3:1); [a]2D3 = +11.2 (c = 2.1,
CHCl3); IR (film): n˜ = 2954, 2875, 1455, 1241, 1147, 1039, 846 cmꢀ1 1H
;
NMR (300 MHz, CDCl3): d = 7.38–7.27 (m, 5H), 4.87 (s, 1H), 4.72 (d, J
= 11.8 Hz, 1H), 4.64 (d, J = 11.8 Hz, 1H), 4.16 (dd, J = 7.4, 4.5 Hz,
1H), 3.94–3.87 (m, 1H), 3.76 (d, J = 4.7 Hz, 1H), 3.49 (dd, J = 10.9,
3.7 Hz, 1H), 3.39 (s, 3H), 3.28 (dd, J = 11.0, 5.2 Hz, 1H), 0.97 (t, J =
7.9 Hz, 9H), 0.66 ppm (d, J = 7.9 Hz, 6H); 13C NMR (75 MHz, CDCl3):
d = 137.9, 128.5, 128.0, 127.9, 105.8, 82.8, 81.0, 75.7, 72.7, 55.5, 9.4, 6.9,
4.9 ppm; HRMS (FAB) calcd for C19H30IO4Si: 477.0958 [MꢀH]+; found:
477.0956.
(1S,4R,4aR,11bR)-4-Benzyloxy-1,4,4a,11b-tetrahydro-1-(2,2,2-trichloro-
acetylamino)-6H-[1,3]benzodioxolo[5,6-c,1]benzopyran-6-one (32): DBU
(0.46 mL, 3.07 mmol) and CCl3CN (0.31 mL, 3.09 mmol) were successive-
ly added to a solution of 30 (780 mg, 2.13 mmol) in CH2Cl2 (50 mL)
under Ar at ꢀ458C. The mixture was stirred for 7 h at ꢀ208C, and then
the reaction was quenched with saturated aqueous NH4Cl solution
(35 mL). The resulting mixture was extracted with CH2Cl2 (3”25 mL)
and the combined organic layers were dried (MgSO4) and concentrated
in vacuo. The residue was passed through a short pad of silica gel, eluting
with heptane/EtOAc (4:1). After concentration of the eluate and co-
evaporation of the volatiles with toluene (3”5 mL), the residue was
heated to 1208C under reduced pressure (~1 mmHg) for 40 h. It was
then purified by flash column chromatography (heptane/EtOAc, 3:1 !
2:1) to give 32 (442 mg, 41%) as a white solid. Rf = 0.55 (heptane/
EtOAc, 1:1); m.p. 234–2368C (EtOAc/heptane); [a]2D3 = ꢀ33.5 (c = 1.1,
Methyl 2-O-benzyl-3-O-tert-butyldimethylsilyl-5-deoxy-5-iodo-d-ribofura-
noside (29): A solution of 27 (190 mg, 0.52 mmol) in CH2Cl2 (2 mL)
under Ar at 08C was treated with 2,6-lutidine (0.15 mL, 1.30 mmol) and
then TBSOTf (0.18 mL, 0.78 mmol) was added. The reaction mixture was
stirred at 08C for 1 h, then diluted with hexane (10 mL) and washed with
H2O (10 mL) and brine (10 mL). The organic layer was dried (Na2SO4)
and concentrated in vacuo, and the residue was purified by flash column
chromatography (heptane/EtOAc, 9:1) to give a combined yield of a-29
and b-29 of 230 mg (92%) as a colourless oil. Pure samples of both iso-
mers were isolated. a-29: Rf
+76.8 (c = 3.2, CHCl3); IR (film): n˜ = 2927, 2856, 1471, 1251, 1172,
1027 cmꢀ1 1H NMR (500 MHz, CDCl3): d = 7.40–7.27 (m, 5H), 4.94 (d,
= =
0.61 (heptane/EtOAc, 3:1); [a]D23
;
J = 4.3 Hz, 1H), 4.68 (d, J = 12.0 Hz, 1H), 4.64 (d, J = 12.1 Hz, 1H),
3.93 (dd, J = 6.4, 5.6 Hz, 1H), 3.81–3.76 (m, 2H), 3.46 (dd, J = 10.8,
3.8 Hz, 1H), 3.45 (s, 3H), 3.29 (dd, J = 11.0, 4.3 Hz, 1H), 0.90 (s, 9H),
0.11 (s, 3H), 0.08 ppm (s, 3H); 13C NMR (75 MHz, CDCl3): d = 138.0,
128.4, 128.3, 127.9, 103.3, 80.9, 77.8, 74.4, 73.0, 55.9, 26.0, 18.3, 9.1, ꢀ4.2,
ꢀ4.5 ppm; MS (FAB) calcd for C19H30IO4Si: 477.10 [MꢀH]+; found:
477.10. b-29: Rf = 0.65 (heptane/EtOAc, 3:1); [a]2D3 = +11.8 (c = 2.9,
1
CHCl3); IR (KBr): n˜ = 3421, 1712, 1505, 1482, 1263, 1059, 1036 cmꢀ1; H
NMR (300 MHz, CDCl3): d = 7.54 (s, 1H), 7.40–7.30 (m, 5H), 6.80 (d, J
= 9.5 Hz, 1H), 6.68 (s, 1H), 6.10–6.02 (m, 1H), 6.06 (s, 1H), 6.04 (s, 1H),
5.89 (dd, J = 10.5, 1.4 Hz, 1H), 4.82 (brs, 1H), 4.69 (d, J = 11.6 Hz,
1H), 4.64 (d, J = 11.4 Hz, 1H), 4.48 (brt, J = 9.3 Hz, 1H), 4.16–4.09 (m,
1H), 3.03 ppm (dd, J = 10.0, 2.3 Hz, 1H); 13C NMR (75 MHz, CDCl3): d
= 164.2, 161.5, 152.5, 148.4, 137.3, 137.2, 130.9, 128.8, 128.4, 128.1, 128.1,
126.9, 118.0, 110.7, 107.6, 102.4, 76.6, 72.5, 71.0, 50.7, 38.4 ppm; HRMS
(FAB) calcd for C23H19Cl3NO6: 510.0278 [M+H]+; found: 510.0260.
1
CHCl3); IR (film): n˜ = 2928, 2857, 1471, 1254, 1151, 1040 cmꢀ1; H NMR
(500 MHz, CDCl3): d = 7.37–7.27 (m, 5H), 4.88 (s, 1H), 4.71 (d, J =
11.6 Hz, 1H), 4.63 (d, J = 11.8 Hz, 1H), 4.16 (dd, J = 7.3, 4.3 Hz, 1H),
3.94–3.90 (m, 1H), 3.77 (d, J = 4.3 Hz, 1H), 3.47 (dd, J = 10.7, 4.7 Hz,
1H), 3.39 (s, 3H), 3.26 (dd, J = 10.8, 5.4 Hz, 1H), 0.91 (s, 9H), 0.13 (s,
3H), 0.10 ppm (s, 3H); 13C NMR (75 MHz, CDCl3): d = 137.9, 128.5,
127.9, 127.9, 106.0, 82.6, 81.1, 75.8, 72.7, 55.5, 25.9, 18.2, 9.2, ꢀ4.3,
ꢀ4.5 ppm; HRMS (FAB) calcd for C19H30IO4Si: 477.0958 [MꢀH]+;
found: 477.0959.
(1R,2S,3S,4R,4aR,11bR)-4-Benzyloxy-2,3-dihydroxy-1,2,3,4,4a,11b-hexa-
hydro-1-(2,2,2-trichloroacetylamino)-6H-[1,3]benzodioxolo[5,6-c,1]benzo-
pyran-6-one (33): N-Methylmorpholine N-oxide monohydrate (25 mg,
0.185 mmol) and OsO4 (4% in H2O, 0.1 mL, 0.016 mmol) were succes-
sively added to a solution of 32 (45 mg, 0.088 mmol) in THF (2 mL)
under Ar. The mixture was stirred for 5 days under Ar and then the reac-
tion was quenched by the addition of 10% aqueous NaHSO3 (20 mL).
The resulting mixture was extracted with EtOAc (3”15 mL) and the
combined organic layers were washed with brine (30 mL), dried
(MgSO4), and concentrated in vacuo. The residue was purified by flash
column chromatography (hexane/EtOAc, 1:1) to give 33 (45 mg, 94%) as
a white solid. Rf = 0.12 (heptane/EtOAc, 1:1); m.p. 213–2148C (hexane/
(3R,4R,4aR,11bS)-4-Benzyloxy-3-hydroxy-3,4,4a,11b-tetrahydro-6H-
[1,3]benzodioxolo[5,6-c,1]benzopyran-6-one (30): Freshly activated Zn[23]
(1.43 g, 21.8 mmol) was added to a solution of 28 (1.03 g, 2.15 mmol) in
THF/H2O (3:1, 40 mL) under Ar. The slurry was sonicated at 408C while
a solution of 7 (1.991 g, 6.66 mmol) in THF (10 mL) was added by means
of a syringe pump over a period of 4 h. After an additional 3 h at 408C
under sonication, the reaction was quenched by the addition of H2O
(30 mL). The mixture was filtered through a pad of Celite, which was
subsequently rinsed first with EtOAc and then with H2O. The filtrate was
extracted with EtOAc (3”30 mL) and the combined organic layers were
dried (MgSO4) and concentrated in vacuo. The residue was redissolved in
MeOH (60 mL) and this solution was heated at 508C for 3 h in the pres-
ence of an acidic ion-exchange resin (Amberlite IR-120). The mixture
was then filtered and concentrated in vacuo to give a residue that was pu-
EtOAc); [a]2D3
= +8.9 (c = 0.6, DMSO); IR (KBr): n˜ = 3423, 2905,
1708, 1483, 1259, 1075, 1036 cmꢀ1; H NMR (300 MHz, [D6]DMSO): d =
8.81 (d, J = 9.7 Hz, 1H), 7.40–7.30 (m, 6H), 6.84 (s, 1H), 6.09 (s, 1H),
6.08 (s, 1H), 5.00 (d, J = 3.6 Hz, 1H), 4.89–4.85 (m, 1H), 4.68 (d, J =
11.9 Hz, 1H), 4.67 (d, J = 7.5 Hz, 1H), 4.63 (d, J = 11.7 Hz, 1H), 4.13–
4.01 (m, 1H), 3.93–3.84 (m, 1H), 3.35–3.29 ppm (m, 1H); 13C NMR
(75 Hz, [D6]DMSO): d = 163.4, 161.1, 151.3, 147.2, 138.1, 136.4, 128.4,
1
Chem. Eur. J. 2006, 12, 3243 – 3253
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3251

R. Madsen et al.
127.8 (2”), 117.8, 109.0, 108.3, 102.2, 93.1, 76.8, 75.9, 71.7, 70.8, 69.2, 51.5,
38.0 ppm; HRMS (FAB) calcd for C23H21Cl3NO8: 544.0333 [M+H]+;
found: 544.0350.
Acknowledgements
Financial support from the Lundbeck Foundation is gratefully acknowl-
edged.
(1R,2S,3S,4S,4aR,11bR)-2-Benzyloxy-1,3,4-trihydroxy-1,3,4,4a,5,11b-
hexahydro-2H-[1,3]dioxolo[4,5-j]phenanthridin-6-one
(34):
K2CO3
(1.12 g, 8.10 mmol) was added to suspension of 33 (399 mg,
a
0.732 mmol) in MeOH (50 mL) under Ar and the mixture was refluxed
for 18 h. It was then cooled, neutralised with an acidic ion-exchange resin
(Amberlite IR-120), and concentrated in vacuo. The residue was purified
by flash column chromatography (CH2Cl2/MeOH, 100:0 ! 96:4) to give
34 (236 mg, 81%) as a white solid. Rf = 0.48 (CHCl3/MeOH, 17:3); m.p.
[1] a) G. R. Pettit, V. Gaddamidi, G. M. Cragg, D. L. Herald, Y. Sagawa,
J. Chem. Soc. Chem. Commun. 1984, 1693–1694; b) G. R. Pettit, V.
Gaddamidi, D. L. Herald, S. B. Singh, G. M. Cragg, J. M. Schmidt,
F. E. Boettner, M. Williams, Y. Sagawa, J. Nat. Prod. 1986, 49, 995–
1002.
164.0–165.08C (MeOH); [a]D23
n˜ = 3392, 2905, 1656, 1464, 1261, 1074, 1038 cmꢀ1
=
+31.0 (c = 1.5, DMSO); IR (KBr):
;
1H NMR (300 MHz,
[2] S. Ghosal, S. Singh, Y. Kumar, R. S. Srivastava, Phytochemistry 1989,
28, 611–613.
[D6]DMSO): d = 7.40–7.26 (m, 6H), 6.98 (s, 1H), 6.92 (s, 1H), 6.08 (s,
2H), 5.27 (brs, 1H), 5.17 (brs, 1H), 4.85 (m, 1H), 4.67 (d, J = 12.0 Hz,
1H), 4.63 (d, J = 12.0 Hz, 1H), 4.60 (brs, 1H), 4.03 (brs, 1H), 3.86–3.82
(m, 1H), 3.78–3.65 (m, 2H), 2.95 ppm (brd, J = 10.3 Hz, 1H); 13C NMR
(75 MHz, [D6]DMSO): d = 164.0, 150.5, 145.9, 138.3, 135.1, 128.4, 127.6,
127.6, 123.8, 106.8, 105.8, 101.6, 77.5, 71.1, 71.0, 70.6, 66.0, 50.2, 40.7 ppm;
HRMS (FAB) calcd for C21H22NO7: 400.1396 [M+H]+; found: 400.1400.
[3] a) G. R. Pettit, R. A. Backhaus, M. R. Boyd, A. W. Meerow, J. Nat.
Prod. 1993, 56, 1682–1687; b) G. R. Pettit, B. Orr, S. Ducki, Anti-
Cancer Drug Des. 2000, 15, 389–395; c) U. Rinner, H. L. Hillebren-
ner, D. R. Adams, T. Hudlicky, G. R. Pettit, Bioorg. Med. Chem.
Lett. 2004, 14, 2911–2915.
[4] a) N. Kekre, C. Griffin, J. McNulty, S. Pandey, Cancer Chemother.
Pharmacol. 2005, 56, 29–38; b) A. McLachlan, N. Kekre, J. McNul-
ty, S. Pandey, Apoptosis 2005, 10, 619–630.
[5] B. Gabrielsen, T. P. Monath, J. W. Huggins, D. F. Kefauver, G. R.
Pettit, G. Groszek, M. Hollingshead, J. J. Kirsi, W. M. Shannon,
E. M. Schubert, J. DaRe, B. Ugarkar, M. A. Ussery, M. J. Phelan, J.
Nat. Prod. 1992, 55, 1569–1581.
[6] For a recent review on the synthesis of hydroxylated phenanthri-
dones from Amaryllidaceae, see: U. Rinner, T. Hudlicky, Synlett
2005, 365–387.
[7] S. J. Danishefsky, J. Y. Lee, J. Am. Chem. Soc. 1989, 111, 4829–4837.
[8] X. Tian, T. Hudlicky, K. Kçnigsberger, J. Am. Chem. Soc. 1995, 117,
3643–3644.
[9] B. M. Trost, S. R. Pulley, J. Am. Chem. Soc. 1995, 117, 10143–10144.
[10] T. J. Doyle, M. Hendrix, D. VanDerveer, S. Javanmard, J. Haseltine,
Tetrahedron 1997, 53, 11153–11170.
[11] P. Magnus, I. K. Sebhat, Tetrahedron 1998, 54, 15509–15524.
[12] J. H. Rigby, U. S. M. Maharoof, M. E. Mateo, J. Am. Chem. Soc.
2000, 122, 6624–6628.
[13] G. R. Pettit, N. Melody, D. L. Herald, J. Org. Chem. 2001, 66, 2583–
2587.
7-Deoxypancratistatin (2): From 11: Acidic ion-exchange resin (Amber-
lite IR-120, 2 mL) was added to a solution of 11 (418 mg, 1.03 mmol) in
MeOH (5 mL) and the mixture was refluxed for 53 h. It was then filtered
and concentrated in vacuo, and the concentrate was passed through a
pad of silica gel to give a crude yield of 14 (405 mg, Rf = 0.08 (hexane/
EtOAc, 2:1)). This was taken up in CH2Cl2 (25 mL) under Ar and m-
CPBA (890 mg, 5.16 mmol) was added. The mixture was stirred for 24 h
at room temperature, and then the reaction was quenched by the addi-
tion of 10% aqueous Na2SO3 (50 mL). The resulting mixture was stirred
for 1.5 h, then extracted with CH2Cl2 (2”35 mL), and the combined or-
ganic layers were dried (Na2SO4) and concentrated in vacuo to give a
crude yield of 15 (325 mg, Rf = 0.24 (hexane/EtOAc, 1:2)). An aqueous
solution of sodium benzoate (5.2 mg in 20 mL) was added to the residue
and the slurry was refluxed for 3 days. The mixture was then cooled and
concentrated in vacuo. The residue of 16 (Rf = 0.19 (hexane/EtOAc,
1:4)) was redissolved in AcOH (7 mL) and the solution was degassed.
Pd/C (90 mg) was added and the system was evacuated twice by the ap-
plication of high vacuum and filled with H2. The system was fitted with a
balloon filled with H2 and the reaction mixture was stirred at room tem-
perature. Additional H2 was introduced as needed. After one week, the
mixture was filtered through a pad of Celite and concentrated in vacuo.
The residue was purified by flash column chromatography (CH2Cl2/
MeOH, 9:1) to give 2 (70 mg, 22%) as a white solid. Rf = 0.61 (CHCl3/
[14] H. Ko, E. Kim, J. E. Park, D. Kim, S. Kim, J. Org. Chem. 2004, 69,
112–121.
[15] S. Ohta, S. Kimoto, Chem. Pharm. Bull. 1976, 24, 2977–2984.
[16] H. Paulsen, M. Stubbe, Liebigs Ann. Chem. 1983, 535–556.
[17] a) G. E. Keck, S. F. McHardy, J. A. Murry, J. Am. Chem. Soc. 1995,
117, 7289–7290; b) G. E. Keck, S. F. McHardy, J. A. Murry, J. Org.
Chem. 1999, 64, 4465–4476.
[18] T. Hudlicky, X. Tian, K. Kçnigsberger, R. Maurya, J. Rouden, B.
Fan, J. Am. Chem. Soc. 1996, 118, 10752–10765.
[19] N. Chida, M. Jitsuoka, Y. Yamamoto, M. Ohtsuka, S. Ogawa, Het-
erocycles 1996, 43, 1385–1390.
[20] G. E. Keck, T. T. Wager, S. F. McHardy, J. Org. Chem. 1998, 63,
9164–9165.
[21] J. L. AceÇa, O. Arjona, M. L. León, J. Plumet, Org. Lett. 2000, 2,
3683–3686.
[22] a) J. McNulty, J. Mao, R. Gibe, R. Mo, S. Wolf, G. R. Pettit, D. L.
Herald, M. R. Boyd, Bioorg. Med. Chem. Lett. 2001, 11, 169–172;
b) T. Hudlicky, U. Rinner, D. Gonzalez, H. Akgun, S. Schilling, P.
Siengalewicz, T. A. Martinot, G. R. Pettit, J. Org. Chem. 2002, 67,
8726–8743.
[23] L. Hyldtoft, R. Madsen, J. Am. Chem. Soc. 2000, 122, 8444–8452.
[24] a) P. R. Skaanderup, R. Madsen, J. Org. Chem. 2003, 68, 2115–2122;
b) T. L. Andresen, D. M. Skytte, R. Madsen, Org. Biomol. Chem.
2004, 2, 2951–2957; c) L. Keinicke, R. Madsen, Org. Biomol. Chem.
2005, 3, 4124–4128; d) F. G. Hansen, E. Bundgaard, R. Madsen, J.
Org. Chem. 2005, 70, 10139–10142.
MeOH, 5:1); m.p. 308–3098C (MeOH) (lit.:[18] 306–3098C); [a]2D3
+73.8 (c = 0.9, DMF) (lit.:[17b] +74.6, c = 0.85, DMF).
=
From 34: Pd(OH)2/C (432 mg) was added to a solution of 34 (225 mg,
0.563 mmol) in EtOAc (30 mL). The system was twice evacuated by the
application of high vacuum and filled with H2. The system was fitted with
a balloon filled with H2 and the reaction mixture was stirred for 20 h at
room temperature. The mixture was then filtered through a pad of
Celite, which was subsequently rinsed with MeOH (100 mL). The com-
bined organic phases were concentrated in vacuo and the residue was pu-
rified by flash column chromatography (CH2Cl2/MeOH, 9:1) to give 2
(138 mg, 79%) as a white solid. Rf = 0.13 (CHCl3/MeOH, 17:1); m.p.
308–3108C (MeOH) (lit.:[18] 306–3098C); [a]2D3 = +72.7 (c = 2.3, DMF)
(lit.:[17b] +74.6, c = 0.85, DMF); IR (KBr): n˜ = 3401, 2902, 1656, 1610,
1466, 1264, 1039 cmꢀ1 1H NMR (300 MHz, [D6]DMSO): d = 7.31 (s,
;
1H), 6.91 (s, 1H), 6.87 (brs, 1H), 6.08 (s, 2H), 5.38 (d, J = 4.0 Hz, 1H),
5.10 (d, J = 5.7 Hz, 1H), 5.08 (d, J = 6.4 Hz, 1H), 4.80 (d, J = 7.6 Hz,
1H), 4.36–4.29 (m, 1H), 3.97 (dd, J = 6.8, 3.5 Hz, 1H), 3.88–3.82 (m,
1H), 3.77–3.64 (m, 2H), 2.98 ppm (brd, J = 10.0 Hz, 1H); 13C NMR
(75 MHz, [D6]DMSO): d = 164.0, 150.5, 145.9, 135.4, 123.8, 106.8, 105.5,
101.6, 73.4, 70.3, 70.2, 68.7, 50.4, 40.1 ppm; HRMS (FAB) calcd for
C14H16NO7: 310.0927 [M+H]+; found: 310.0926.
[25] P. C. Conrad, P. L. Kawiatkowski, P. L. Fuchs, J. Org. Chem. 1987,
52, 586–591.
[26] L. M. Lerner, Carbohydr. Res. 1977, 53, 177–185.
3252
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 3243 – 3253

Synthesis of 7-Deoxypancratistatin
FULL PAPER
[27] J.-L. Luche, L. A. Sarandeses in Organozinc Reagents (Eds.: P. Kno-
chel, P. Jones), Oxford University Press, New York, 1999, pp. 307–
323.
[28] P. Schwab, R. H. Grubbs, J. W. Ziller, J. Am. Chem. Soc. 1996, 118,
100–110.
[29] P. R. Skaanderup, C. S. Poulsen, L. Hyldtoft, M. R. Jørgensen, R.
Madsen, Synthesis 2002, 1721–1727.
[30] K. B. Sharpless, R. C. Michaelson, J. Am. Chem. Soc. 1973, 95,
6136–6137.
[31] O. Bortolini, S. Campestrini, F. Di Furia, G. Modena, J. Org. Chem.
1987, 52, 5093–5095.
[32] A. H. Hoveyda, D. A. Evans, G. C. Fu, Chem. Rev. 1993, 93, 1307–
1370.
[33] L. E. Overman, N. E. Carpenter, Org. React. 2005, 66, 1–107.
[34] S. B. Garber, J. S. Kingsbury, B. L. Gray, A. H. Hoveyda, J. Am.
Chem. Soc. 2000, 122, 8168–8179.
[35] M. Scholl, S. Ding, C. W. Lee, R. H. Grubbs, Org. Lett. 1999, 1, 953–
956.
[39] K. S. Kim, Y. H. Ahn, E. Y. Hurh, K. T. Kim, Y. H. Joo, Bull.
Korean Chem. Soc. 1999, 20, 829–831.
[40] a) I. Lundt, R. Madsen, Synthesis 1992, 1129–1132; b) M. Flasche,
H.-D. Scharf, Tetrahedron: Asymmetry 1995, 6, 1543–1546.
[41] H. Kold, I. Lundt, C. Pedersen, Acta Chem. Scand. 1994, 48, 675–
678.
[42] S. J. Yoo, H. O. Kim, Y. H. Lim, J. M. Kim, L. S. Jeong, Bioorg. Med.
Chem. 2002, 10, 215–226.
[43] R. Albert, K. Dax, R. W. Link, A. E. Stꢂtz, Carbohydr. Res. 1983,
118, C5–C6.
[44] S. J. Mantell, P. S. Ford, D. J. Watkin, G. W. J. Fleet, D. Brown, Tetra-
hedron 1993, 49, 3343–3358.
[45] T. Nishikawa, M. Asai, N. Ohyabu, M. Isobe, J. Org. Chem. 1998,
63, 188–192.
[46] I. Savage, E. J. Thomas, P. D. Wilson, J. Chem. Soc. Perkin Trans. 1
1999, 3291–3303.
[47] V. VanRheenen, R. C. Kelly, D. Y. Cha, Tetrahedron Lett. 1976, 17,
1973–1976.
[48] D. S. Pedersen, C. Rosenbohm, Synthesis 2001, 2431–2434.
[49] E. Brown, J.-P. Robin, R. Dhal, Tetrahedron 1982, 38, 2569–2579.
[50] J. M. J. Tronchet, J. F. Tronchet, F. Rachidzadeh, F. Barbalat-Rey, G.
Bernadinelli, Carbohydr. Res. 1988, 181, 97–114.
[36] A selective Overman rearrangement from diol 21 is not possible;
see ref. [7].
[37] O. Mitsunobu, Synthesis 1981, 1–28.
[38] R. Madsen, in Glycoscience: Chemistry and Chemical Biology (Eds.:
B. Fraser-Reid, K. Tatsuta, J. Thiem), Springer, Berlin, 2001,
pp. 195–229.
Received: November 18, 2005
Published online: February 9, 2006
Chem. Eur. J. 2006, 12, 3243 – 3253
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3253