Chiral building blocks from d-xylose and their application in synthesis of avocadotriol monoacetate

Article abstract of DOI:10.1016/j.tet.2008.07.103

Facile routes to several enantiomerically pure flexible chiral building blocks carrying a hidden syn or anti 1,3-diol motif were developed with the inexpensive and readily available carbohydrate d-xylose as the starting material. Application of the newly developed chiral building blocks in total synthesis is exemplified through a synthesis of (2S,4S)- and (2S,4R)-avocadotriol. Both triols were selectively acetylated on the primary hydroxyl group in high yields with Novozyme 435/vinyl acetate. On the basis of comparison of the ¹H NMR, optical rotation, and melting point data, the natural avocadotriol 1-monoacetate was assigned to be of (2R,4R) configuration.

Full text of DOI:10.1016/j.tet.2008.07.103

Tetrahedron 64 (2008) 9377–9383
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Chiral building blocks from
D-xylose and their application in synthesis
of avocadotriol monoacetate
*
Jian Gao, Zhi-Bin Zhen, Ya-Jun Jian, Yikang Wu
State Key Laboratory of Bio-organic and Natural Products Chemistry, Shanghai Institute of Organic Chemistry, Chinese Academy of Sciences, 354 Fenglin Road,
Shanghai 200032, China
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 15 June 2008
Received in revised form 19 July 2008
Accepted 25 July 2008
Available online 30 July 2008
Facile routes to several enantiomerically pure flexible chiral building blocks carrying a hidden syn or anti
1,3-diol motif were developed with the inexpensive and readily available carbohydrate D-xylose as the
starting material. Application of the newly developed chiral building blocks in total synthesis is exem-
plified through a synthesis of (2S,4S)- and (2S,4R)-avocadotriol. Both triols were selectively acetylated on
the primary hydroxyl group in high yields with Novozyme 435/vinyl acetate. On the basis of comparison
of the 1H NMR, optical rotation, and melting point data, the natural avocadotriol 1-monoacetate was
assigned to be of (2R,4R) configuration.
Keywords:
Chiron
Carbohydrates
Natural products
Enantioselective synthesis
Epoxides
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
operations were involved, explorations on more practical routes to
flexible chiral building blocks at lower cost are still warranted.
D
-xylose (1) is one of the carbohydrates frequently employed as
OAc
O
OH
chiral pools in synthesis. However, like most other carbohydrates, 1
rarely can be utilized directly because of co-existence of several
hydroxyl groups. In most cases, differentiation between the hy-
droxyl groups and deoxygenation at one or more positions are pre-
requisites for execution of chiron-based synthetic plans. It is hence
highly desirable to convert 1, through practical routes, into gener-
ally applicable chiral building blocks that are directly usable in
synthesis.
O
HO
D-Xylose
O
1
2
OH
OH
O
1
O
4
HO
2
3
3
OH
OAc
2. Results and discussions
A key issue in making use of diol 3 as a precursor for chiral
building blocks is to differentiate the two hydroxyl groups. Because
one of them is primary and the other secondary, on the basis of
common knowledge of organic chemistry, one usually takes it for
granted that the reaction would preferentially occur at the primary
OH. However, as will be shown in the following, this is not always
the case.
Another hidden problem that deserves to be mentioned is the
preparation of 3 from 2 (Scheme 1). According to the literature
cleavage of the acetyl groups can be readily achieved with KOH/
EtOH–H2O/rt/12 h followed by p-TsOH/THF/rt/24 h,3a,c KCN/
MeOH,4 NaOMe/MeOH then Amberlite IR-120 H resin,5a or 10%
NaOH/MeOH/rt/12 h then acidification with HCl (cleaving a mono-
acetate).2 Although none of the literature ever explicitly mentioned
any difficulties about 3, in repeating the documented procedures
we noticed that the outcomes of the different protocols did vary
considerably. Over hydrolysis led to ring-opening of the lactone,
One of the good starting points for such endeavors appears to be
the diacetate 2 reported1 by Okabe, which can be conveniently
prepared from 1 through a practical sequence consisting of (1) Br2
oxidation of 1 into a five-membered lactone, (2) acetylation of all
the hydroxyl groups, (3) Et3N-mediated elimination of the AcO at
the C-3, and (4) Ra-Ni catalyzed hydrogenation of the double bond
formed by the elimination to reintroduce a stereogenic center at
the C-2 in a highly stereoselective way. The two acetyl groups in
diacetate 2 can be readily removed by treatment with a base in
MeOH, yielding a known2 diol 3.
There have been some records in the literature on elaboration of
3 into various chiral building blocks.3 However, because relatively
expensive reagents (e.g., TBDPSCl and BH3) and inconvenient
* Corresponding author.
E-mail address: yikangwu@mail.sioc.ac.cn (Y. Wu).
0040-4020/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2008.07.103

9378
J. Gao et al. / Tetrahedron 64 (2008) 9377–9383
OAc
4
OH
O
OTs
p-TsCl
NaOMe
MeOH
rt/20 min
87%
O
O
2
HO
O
1
Na2CO3
MeOH
O
OH
ref. 1
Py
CO2Me
OH
O
O
O
5 d
60%
H
1
HO
OH
3
5
3
4a
OH
OH
OAc
2
D-Xylose
52%
TBSCl
Et3N
(cf also ref 5a)
See text
OH
OTBS
HO
OH
OR
O
NaBH4
O
O
TBSO
O
O
O
OH
+
MeOH
7
12b
12a
6a
OH
OR
OH
ca. 1:1 R = Allyl, Piv, Bz
p-TsCl
Et3N
n-Bu2SnO
p-TsCl
91%
HO
OH
TBSO
OTs
8
NaBH4
OTBS
0 °C/20 min
CH2Cl2-MeOH
HO
H
O
O
O
HO
H
MeOH
K2CO3
O
TBSO
83%
TBSO
85% from 6
11
10
OTs
9
0 °C/2 h
Scheme 1.
which tended to lower the yield. Acidification also made remark-
able difference. It appears that the acidic resin conditions of Var-
geese and Abushanab5a more often afforded higher yields of diol 3.
It is interesting to note that the IR spectrum of 3 also changed
with time. Freshly chromatographed 3 (neat) showed apparently
two carbonyl absorptions, with one at 1737 cmꢀ1 and the other at
1771 cmꢀ1. However, recording the spectrum after the sample was
allowed to stand at ambient temperature for an hour or so essen-
tially only one carbonyl absorption signal at 1771 cmꢀ1 could be
seen.
With 3 in hand, we attempted selective tosylation under the
literature3b conditions (p-TsCl/Py/rt/20 h). Somewhat unexpected,
after 20-h reaction (Scheme 1) only small amounts of 4a was
formed while most of the starting 3 remained unchanged. However,
with extension of the reaction time to 5 days, the desired 4a could
be formed in 60% yield.
attempts to use allyl, benzoyl (Bz) or pivalyol (Piv) group as the
protecting group always led to an almost 1:1 mixture of two po-
sitional isomers, indicating that the secondary OH was just as re-
active as the primary one in these cases. The selectivity appears to
depend very much on the protecting group involved in the reaction.
It is not clear yet why no product with a Ts or TBS on the secondary
OH was formed in these cases. However, it can be definitely ex-
cluded that under the given conditions any 4b or 6b (both prepared
from 13,1 Scheme 2) might be converted into 4a or 6a, respectively,
via intramolecular migration of the protecting group from the
secondary OH to the primary one (Scheme 2).
CO2Me
CO2Me
HO
RO
OH
O
O
O
4b R = Ts
O
In our hands tosylate 4a was isolated as a white solid with the
6b R = TBS
26
OR
O
optical rotation ([
a
]
D
þ21.8 (c 1.7, CHCl3)) slightly smaller than that
O
13
14
20
reported3b in the literature (an oil, [
a
]
þ28.6 (c 1.4, CHCl3)). In the
D
beginning we suspected that our sample was partially racemized
over the prolonged reaction. However, 1H and 13C NMR and HPLC3d
analyses all excluded the possibility of co-existence of the C-2
epimer or any other impurities. Therefore, the 4a was believed to be
pure and used as such in the next step. On treatment with NaOMe
in MeOH at ambient temperature for 20 min, tosylate 4a was
smoothly converted into epoxide 5 in 87% isolated yield.
Diol 3 could also be converted to mono-TBS ether 6a as de-
scribed by Vargeese and Abushanab,5a although in our hands the
yield for 6a (of larger optical rotation than that reported in the
literature) was only 52%, with substantial part of starting 3
unreacted. Extension of reaction time or using more reagents led to
a significantly increased content of di-TBS protected species in the
product mixture. Treatment of 6a with NaBH4 in MeOH gave an
intermediate triol 7, which was selectively tosylated under the
conditions of Martinelli6 to afford 8. On treatment with K2CO3/
MeOH the crude 8 underwent facile cyclization, producing epoxide
9 in 85% overall yield from 6a.
OH
OTs
Py
O
O
O
O
(Not detected)
rt/1 d
OTs
4a
4b
OH
Et3N
THF
OTBS
OH
O
O
O
O
(Not detected)
rt/1 d
6a
Scheme 2.
OTBS
OH
6b
The chiral building blocks derived from
D
-xylose such as 5, 9,
and 11 all contain a hidden 1,3-diol motif of either syn or anti
relationship, which may be utilized as precursors for the corre-
sponding fragments in those target compounds encapsulate 1,3-
diol(s) partial structures. Avocadotiol monoacetate (15),7 a potent
acetyl-CoA carboxylase inhibitor, is one of such compounds
(Scheme 3).
The relative configuration of natural 15 was reported to be 1,3-
syn on the basis of the relative configuration of the related triol.7a,c
However, the assignment is not conclusive because no structural
determination has ever been performed directly on 15 itself. To
demonstrate the utility of the chiral building blocks mentioned
Tosylation of 6a gave 10 in 91% yield. Careful treatment of 10
with NaBH4 in CH2Cl2/MeOH led to reduction of the lactone car-
bonyl group followed by cyclization, affording epoxide 11 with an
inverted configuration at the C-2.
It is noteworthy that although protection of either 4a or 6a was
relatively facile,5b the apparent preference for the primary OH did
not seem to exist with other protecting groups. For instance,

J. Gao et al. / Tetrahedron 64 (2008) 9377–9383
9379
SOCl2
CH Cl
O
Removal of the TBS protecting group in 21 with n-Bu4NF (TBAF) led
to known14 tiols (2S,4R)-22 or (2S,4S)-22, respectively.
CH2N2
CO2H
8
CO2H
9
2
2
Cl
DMF
8
17
Selective acetylation of the triols was expected to be a difficult
task if using conventional acetylation conditions. Therefore, we
opted to try those unconventional protocols developed in recent
years. Novozyme 435/vinyl acetate appeared rather mild and highly
selective,15 and was thus tested first.
The experimental outcome using this set of conditions was in-
deed very pleasing. Treatment of triol with Novozyme 435/vinyl
acetate led to clean formation of the corresponding 1-mono-
acetates within 30 min. The position of acetyl group was secured by
an unmistakable downfield shift of the methylene group (–CH2O–)
in addition to the coupling information from COSY and HMQC
experiments.
It is interesting to note that the two diastereomers, (2S,4R)-22
and (2S,4S)-22, showed remarkable difference in reactivity toward
vinyl acetate in the absence of Novozyme 435. Stirring of (2S,4S)-22
in vinyl acetate at ambient temperature for 5 days resulted in
(2S,4S)-15 cleanly. However, (2S,4R)-22 was essentially complete
inert under the same conditions.
The optical rotation for (2S,4S)-15 and (2S,4R)-15 synthesized in
this work was determined to be þ6.7 and ꢀ11.4, respectively
(Fig. 1). The former appears to be compatible with the values
reported by Kashman7b and Browse,7d respectively, suggesting that
the natural 15 should be of (2R,4R) configuration because of the
opposite signs. However, the rotation of (2S,4R)-15 is also rather
close to that of Kawabata,7a which implies a (2S,4R) configuration.
Apparently, comparison of the optical rotations alone in this case
cannot lead to any definite conclusion.
18
16
LiAlH4/Et2O 0 °C/2 h
I2/PPh3
imidazole
CH2Cl2
84%
O
I
20
OH
9
9
H
19
OH
78% from 16
t-BuLi
CuCN
MeLi
OTBS
t-BuLi
CuCN
MeLi
HO
OH
H
O
9
71%
11
TBSO
76%
syn
OH
anti
OH
OH
OTBS
(2S,4R)-21
OTBS
11
11
(2S,4S)-21
TBAF
TBAF
OH
OH
OH
11
OH
OH
OH
(2S,4R)-22
11
(2S,4S)-22
Novozyme 435
93%
Novozyme 435
91%
AcO
AcO
OH
OH
OH
OH
OAc
OAc
11
11
(2S,4S)-15
(2S,4R)-15
Scheme 3.
13C NMR is often considered as a sensitive tool for differentiation
between diastereomers. However, in the present case it turns out to
be useless because the spectrum for (2S,4S)-15 and (2S,4R)-15 is
identical to each other (and also identical to that of the natural 15).
The critical differences between the diastereomers were finally
above and to gain more direct evidence on the relative as well as
absolute configuration of the natural 15, we conducted the total
syntheses described below.
We were aware from the beginning that the optical rotations
documented in the literature over the years varied considerably
from each other, which implied that a definite conclusion on the
relative and absolute configurations might not be reachable through
simple comparisonof the rotationsof asynthetic(2S,4S)-15 (1,3-syn)
with those in the literature. To avoid this situation, we also syn-
thesized (2S,4R)-15 (1,3-anti) in parallel for comparison. As will be
seen below, this anti isomer indeed played a critical role in con-
firming the syn configuration for the natural 15 despite the in-
terference of the discrepancy between the literature rotation data.
Both isomers of 15 were derived via catenation of a 12-carbon
chain species with either 9 or 11 (Scheme 3). The syntheses started
with Arndt–Eistert homologation8 of the readily available 10-
undecenoic acid (16). The acid was first converted into acid chloride
17 by treatment with SOCl2 in CH2Cl2 in the presence of a catalytic
amount of DMF as described by Quinkert.9 The subsequent chain
extension and reduction of the carboxylic group were then per-
formed as reported by Kato and Mori,10 giving alcohol 19 in 78%
overall yield from 16.
found in 1H NMR, in the region of
d 4.22–4.11 (Table 1). The ob-
served relevant signals for (2S,4S)-15 were essentially the same as
those for the natural one, with all three protons (two H-10s and H-2)
appearing between
d 4.11–4.03. Two protons of (2S,4R)-15, how-
ever, appeared at 4.22–4.12, a blank region for natural and (2S,4S)-
d
15, providing a critical piece of evidence for rejecting the (2S,4R)
relative configuration (i.e., the 1,3-anti configuration) for the nat-
ural 15. It should also be noted that the melting point for (2S,4S)-15
is much closer to that for the natural 15 than the (2S,4R)-15 (Fig. 1).
Once the 1,3-anti configuration was excluded, the remaining
task became much easier. Although the three rotation data for the
natural 15 in the literature differ considerably from each other, they
all are levorotatory. This means the natural 15 must be an mirror
image to (2S,4S)-15, possessing a (2R,4R) configuration.
(This work)
OH
OH
2
OH
OH
16
3
1
11
17
11
Before combining the long chain moiety with either epoxide 9
or 11, the hydroxyl group terminal of 19 must be activated into
a proper carbanion species. This was most conveniently achieved
via an iodide, which might undergo iodine–lithium exchange to
afford the desired carbanion. In execution of this plan, the alcohol
19 was converted into iodide 20 as described11 by Bach and
Lemarchand. The exchange of the iodide into the corresponding
lithium species was realized by reaction with t-BuLi.12 Further ad-
dition of CuCN/MeLi13 to the organolithium generated a cuprate
species, which on reaction with epoxide 9 or 11 afforded diol
(2S,4R)-21 or (2S,4S)-21, respectively.
OAc
(2S,4S)-15
(2S,4R)-15
OAc
28
21
[α]D −11.4 (c 0.15, CHCl3)
M.p. 44-45 °C
[α]D +6.7 (c 0.35, CHCl3)
M.p. 52-53 °C
[α]D23 −14.0 (c 0.5, CHCl3) (ref. 7a)
[α]D22 −2.5 (c 0.89, CHCl3) (ref. 7d)
OH
11
OH
[α]D25 −4.6 (ref. 7b)
OAc
M.p. 58-59°C (ref. 7b)
(2R,4R)-15
(Structure deducted for the natural 15)
The (2S,4R)-21 was inseparable from the unreacted 9 on silica
gel. Therefore, it was directly desilylated without any purification.
Figure 1. The absolute configurations and physical data for synthetic (2S,4S)- and
(2S,4R)-15 as well as the natural 15.

9380
J. Gao et al. / Tetrahedron 64 (2008) 9377–9383
Table 1
Comparison of 1H NMR data of the synthetic and natural 15, with the unmistakable differences in the
d
4.22–4.03 region italicizeda
Natural 15 (Ref. 7a)
Natural 15 (Ref. 7d)
(2S,4S)-15 (this work)
(2S,4R)-15 (this work)
5.79 (ddt, J¼17.1, 10.2, 6.7 Hz, 1H, H-16)
4.97 (d, J¼17.1 Hz, 1H, one of H-17)
4.91 (d, J¼10.2 Hz, 1H, one of H-17)
4.11–4.07 (2H, m, one of H-1 and H-2)
3.98 (dd, J¼11.9, 4.5 Hz, 1H, one of H-1)
3.90–3.85 (m, 1H, H-4)
5.82 (ddt, J¼17, 10.5, 6.7 Hz)
4.99 (ddt, J¼17, 1.6, 1.2 Hz)
4.93 (ddt, J¼10.5, 2.1, 1.2 Hz)
4.11 (m)
5.81 (ddt, J¼17.0, 10.2, 6.6 Hz, 1H)
4.99 (dd, J¼17.0, 1.6 Hz, 1H)
4.93 (d, J¼10.4 Hz, 1H)
4.15–4.06 (m, 2H)
4.04–3.96 (m, 1H),
3.93–3.84 (m, 1H)
5.81 (ddt, J¼18.0, 10.4, 6.8 Hz, 1H)
4.99 (dd, J¼17.2, 1.6 Hz, 1H)
4.92 (dd, J¼10.4, 1.2 Hz, 1H)
4.22–4.12 (m, 2H)
3.89 (m)
4.03 (dd, J¼11.2, 7.6 Hz, 1H)
3.97–3.89 (m, 1H)
2.09 (s, Ac)
2.09 (s)
2.10 (s, 3H)
2.10 (s, 3H)
2.04–2.00 (m, 2H H-15)
1.58–1.50 (m, 2H, H-3), 1.438–1.43 (m, 2H, H-5), 1.24 (m, 18H)
2.04 (br q, J¼7 Hz)
2.03 (q, J¼6.8 Hz, 2H)
2.03 (q, J¼7.2 Hz, 2H)
1.47–1.32 (m)
1.65–1.18 (m, 22H)
1.72–1.22 (m, 22H)
a
The data from Ref. 7b are not included here because they were oversimplified (all designated as ‘m’ without giving the ranges of the multiplets) and hence not informative
enough for comparison. For clarity, the signals for OH groups are not shown here.
3. Conclusions
mixture was filtered off through Celite. The filtrate was concen-
trated to dryness on a rotary evaporator to afford the known diol 32
(crude) as a yellowish oil (4.450 g, 30.2 mmol), which was used as
such in the subsequent step.
Facile access to several flexible chiral building blocks has been
developed with inexpensive and readily available D-xylose as the
starting material. The newly synthesized chiral building blocks all
possess a hidden 1,3-diol motif and may find applications in syn-
thesis of various target molecules containing 1,3-diol partial
structures.16 Such a potential is illustrated through a synthesis of
(2S,4S)-15 and (2S,4R)-15, which served as reference samples in
direct data comparison in an effort to assign the absolute config-
uration of natural avocadotriol monoacetate (natural 15). On the
basis of careful comparison of the 1H NMR spectra, the relative
configuration of natural 15 was secured to be 1,3-syn. The absolute
configuration of the natural 15 was then deducted to be (2R,4R)
because its rotation is of opposite sign to that of (2S,4S)-15. The
rapid, mild, and highly selective acetylation of the 1,2,4-triols using
Novozyme 435/vinyl acetate at ambient temperature may be also of
preparative significance in synthesis.
4.3. Tosylation of diol 3 leading to 4a
A solution of diol 3 (528 mg, 4.0 mmol) and p-TsCl (839 mg,
4.401 mmol) in anhydrous pyridine (15 mL) was stirred at ambient
temperature for 5 days. Aq satd CuSO4 (20 mL) was added. The
mixture was extracted with EtOAc (2ꢂ50 mL). The combined EtOAc
phases were washed with aq satd CuSO4 until no more blue pre-
cipitates formed. The EtOAc phase was then washed in turn with
water (5.0 mL) and brine (5.0 mL), and dried over anhydrous
Na2SO4. Removal of the solvent by rotary evaporation and column
chromatography (1:1 EtOAc/PE) on silica gel gave the known
monotosylate 4a3b as a white solid (687 mg, 2.399 mmol, 60%). Mp
26
20
101–102 ꢁC;3e
[
a]
þ21.8 (c 1.7, CHCl3) (lit.3b
[
a]
þ28.6 (c 1.47,
D
D
CHCl3)); 1H NMR (300 MHz, CDCl3)
J¼8.2 Hz, 2H), 4.65–4.51 (m, 2H), 4.23 (dd, J¼11.4, 2.9 Hz, 1H), 4.11
(dd, J¼11.6, 5.5 Hz, 1H), 3.69 (d, J¼4.0 Hz, –OH), 2.63 (ddd, J¼13.1,
8.5, 6.1 Hz, 1H), 2.44 (s, 3H), 2.12–1.92 (m, 1H); 13C NMR (75 MHz,
d
7.77 (d, J¼8.3 Hz, 2H), 7.35 (d,
4. Experimental
4.1. General
CDCl3)
d 176.0, 145.5, 132.0, 130.1, 128.0, 73.5, 69.1, 67.8, 32.3, 21.7.
Unless otherwise stated, the 1H NMR and 13C NMR spectra were
recorded in deuterochloroform at ambient temperature using
a Varian Mercury 300 or a Bruker Avance 300 instrument (oper-
ating at 300 MHz for proton). The FTIR spectra were scanned with
a Nicolet Avatar 360 FTIR spectrometer. The ESIMS and ESIHRMS
were recorded with a PE Mariner API-TOF and an APEX III (7.0 Tesla)
FTMS mass spectrometer, respectively. The melting points were
uncorrected. Optical rotations were recorded on a Jasco P-1030
polarimeter. Dry THF and Et2O were distilled from Na/Ph2CO under
N2 prior to use. Dry DMF, CH2Cl2, and pyridine were distilled over
CaH2 under N2 prior to use. PE (chromatography eluent) stands for
petroleum ether (bp 60–90 ꢁC). DMF stands for N,N0-dime-
thylformamide. Molecular sieves (4 Å) were activated by heating
3ꢂ1 min (with the moisture briefly ventilated between the heating
sessions by opening the oven door) in a 700 W household
microwave oven using ca. 75% of the full power. Novozyme 435
(immobilized on macroporous acrylic resin) was purchased from
Sigma–Aldrich and used as received.
4.4. Conversion of 4a into epoxide 5
A freshly prepared solution of NaOMe in anhydrous MeOH
(2.0 N, 0.35 mL, 0.70 mmol) was added to a solution of 4a (100 mg,
0.346 mmol) in anhydrous MeOH (2.0 mL) stirred at ambient
temperature. The stirring was continued for 20 min, when TLC
showed completion of the reaction. The mixture was neutralized
with diluted HOAc. Solvent was removed by rotary evaporation. The
residue was diluted with EtOAc (20 mL), washed with water (5 mL)
and brine (5 mL), and dried over anhydrous Na2SO4. Removal of the
solvent by rotary evaporation and column chromatography (2:3
EtOAc/PE) on silica gel gave epoxide 5 as a yellowish sticky oil
(44 mg, 0.301 mmol, 87%). [
(400 MHz, CDCl3)
23
a
]
D
ꢀ8.0 (c 0.65, CHCl3); 1H NMR
d
4.37 (dd, J¼7.3, 5.1 Hz, 1H), 3.80 (s, 3H), 3.17–
3.12 (m, 1H), 2.82 (t, J¼4.5 Hz, 1H), 2.54 (dd, J¼5.0, 2.7 Hz, 1H), 1.97–
1.92 (m, 2H); 13C NMR (100 MHz, CDCl3)
d
174.9, 68.4, 52.7, 49.1,
47.2, 37.3; FTIR (film) 3367, 2924, 2854, 1742, 1461, 1261, 1100,
801 cmꢀ1
;
ESIMS m/z 169.1 ([MþNa]þ). ESIHRMS calcd for
C6H10O4Na ([MþNa]þ): 169.0471; found: 169.0463.
4.2. Deacetylation of 2 leading to crude diol 3
Finely powdered Na2CO3 (220 mg, 2.1 mmol) was added to
a solution of diacetate 2 (6.500 g, 30.0 mmol) in anhydrous MeOH
(60 mL) stirred at ambient temperature. The mixture was stirred for
1 h, when TLC showed full disappearance of the starting 2. The
white solids were filtered off through Celite. Amberlite IR-120 H
resin was carefully added to the filtrate with stirring until the
mixture became neutral (pH testing paper). The resin in the
4.5. TBS protection of diol 3 leading to 6a
A solution of diol 3 (0.821 g, 6.215 mmol), dry Et3N (1.0 mL,
7.125 mmol), and TBSCl (1.032 g, 6.857 mmol) in anhydrous THF
(30 mL) was stirred at ambient temperature for 3 days. Water
(10 mL) was added. The mixture was concentrated on a rotary
evaporator. The residue was dissolved in CH2Cl2 (100 mL), washed

J. Gao et al. / Tetrahedron 64 (2008) 9377–9383
9381
in turn with 0.3 N HCl (2ꢂ20 mL), water (20 mL), and brine (20 mL),
4.8. Conversion of 10 into 11
and dried over anhydrous Na2SO4. Removal of the solvent by rotary
evaporation and column chromatography (1:2 EtOAc/PE) on silica
gel afforded the known5a 6a as a colorless sticky oil (796 mg,
NaBH4 (20 mg, 0.512 mmol) was added in portions to a solution
of 10 (205 mg, 0.512 mmol) in CH2Cl2/MeOH (1:1 v/v, 3.0 mL) stir-
red in an ice-water bath. After completion of the addition, the
mixture was stirred at ambient temperature until TLC showed
complete disappearance of the starting 10 (ca. 20 min). Water
(3 mL) was added. The mixture was concentrated on a rotary
evaporator to remove MeOH and CH2Cl2. The residue was extracted
with EtOAc (2ꢂ25 mL). The combined organic layers were washed
brine (5 mL) and dried over anhydrous Na2SO4. Solvent was re-
moved by rotary evaporation. The residue was dissolved in MeOH
(2.5 mL). With cooling (ice-water bath) and stirring, finely pow-
dered anhydrous K2CO3 (46 mg, 0.623 mmol) was introduced. After
completion of the addition, the mixture was stirred at ambient
temperature for ca. 1 h (when TLC showed completion of the re-
action). Water (5 mL) was added. The mixture was concentrated on
a rotary evaporator to remove MeOH. The residue was extracted
with EtOAc (2ꢂ20 mL). The combined organic layers were washed
with water (5 mL) and brine (5 mL), and dried over anhydrous
Na2SO4. Removal of the solvent by rotary evaporation and column
25
25
3.231 mmol, 52%). [
a
]
þ11.1 (c 1.15, EtOH) (lit.5a
[a
]
þ8.57 (c 1.16,
D
D
EtOH)); 1H NMR (300 MHz, CDCl3)
d 4.58–4.40 (m, 2H), 3.94 (dd,
J¼11.6, 3.2 Hz, 1H), 3.74 (dd, J¼11.7, 3.8 Hz, 1H), 3.65 (br s, OH), 2.61
(ddd, J¼15.1, 8.2, 6.3 Hz, 1H), 2.15 (dt, J¼12.7, 8.7 Hz, 1H). 0.92 (s,
9H), 0.06 (s, 6H); 13C NMR (75 MHz, CDCl3)
d 177.3, 77.3, 68.0, 63.9,
32.3, 25.7, 18.2, ꢀ3.7, ꢀ5.5.
4.6. Conversion of 6a into epoxide 9
NaBH4 (454 mg, 12 mmol) was added in portions to a solution
of 6a (2.464 g, 10 mmol) in MeOH (50 mL) stirred in an ice-water
bath. After completion of the addition, the mixture was stirred at
ambient temperature until TLC showed complete disappearance of
the starting 6. Amberlite IR-120 H resin was added with stirring
until pH¼6 (testing paper). The resin was filtered off. The filtrate
was concentrated on a rotary evaporator to give the intermediate
triol as a yellowish sticky oil. To this oily residue were added in
turn dry THF (50 mL), n-Bu2SnO (72 mg, 0.3 mmol), activated 4 Å
molecular sieves (1.5 g), dry Et3N (1.54 mL, 11 mmol), and p-TsCl
(1.904 g, 10 mmol). The mixture was stirred at ambient tempera-
ture for 1 h, when TLC showed complete disappearance of the
triol. The mixture was filtered through Celite. The filtrate was
concentrated on a rotary evaporator to dryness. The residue (crude
tosylate) was dissolved in MeOH (40 mL). The solution was then
stirred in an ice-water bath. Finely powdered anhydrous K2CO3
(495 mg, 5.0 mmol) was introduced. After completion of the ad-
dition, the mixture was stirred at ambient temperature for ca. 1 h
(when TLC showed completion of the reaction). Water (10 mL) was
added. The mixture was concentrated on a rotary evaporator to
remove MeOH. The residue was extracted with EtOAc (3ꢂ50 mL).
The combined organic layers were washed with water (20 mL) and
brine (20 mL), and dried over anhydrous Na2SO4. Removal of the
solvent by rotary evaporation and column chromatography (1:2
chromatography (1:7 EtOAc/PE) on silica gel gave 11 as a colorless
þ6.9 (c 1.15, CHCl3); 1H
23
oil (97 mg, 0.425 mmol, 83% from 10). [
NMR (400 MHz, CDCl3)
a
]
D
d
3.80–3.87 (m,1H), 3.62 (dd, J¼10.0, 3.9 Hz,
1H), 3.54 (dd, J¼10.0, 7.0 Hz, 1H), 3.12–3.05 (m, 1H), 2.77 (dd, J¼4.6,
4.4 Hz, 1H), 2.55 (d, J¼3.6 Hz, 1H), 2.52 (dd, J¼5.0, 2.7 Hz, 1H), 1.75
(dt, J¼14.4, 4.7 Hz, 1H), 1.70–1.62 (m, 1H), 0.90 (s, 9H), 0.06 (s, 6H);
13C NMR (100 MHz, CDCl3)
d 69.9, 66.8, 49.7, 46.7, 35.7, 25.9, 18.3,
ꢀ5.4 (2C’s); FTIR (film) 3462, 2929, 2858, 1472, 1254, 1107, 837, 778,
670 cmꢀ1
C
;
ESIMS m/z 255.1 ([MþNa]þ). ESIHRMS calcd for
11H24O3SiNa ([MþNa]þ): 255.1387; found: 255.1388.
4.9. Synthesis of (2S,4S)-21
MeLi (3.0 M, in Et2O, 0.5 mL,1.5 mmol) was added (via a syringe)
dropwsie to a mixture of CuCN (134 mg, 1.496 mmol) in dry Et2O
(6.0 mL) stirred at ꢀ78 ꢁC under argon (balloon). After completion
of the addition, the mixture was stirred at the same temperature for
30 min (the pale yellow suspension gradually became a solution).
The lithium species of 20, prepared by dropwise addition (via
a syringe) of t-BuLi (1.5 M, in pentane, 2.5 mL, 3.75 mmol) to a so-
lution of iodide 2011 (450 mg, 1.529 mmol) in dry Et2O (7.0 mL)
stirred at ꢀ78 ꢁC under argon (balloon) followed by stirring at the
same temperature for 30 min, was then introduced via a cannula to
the aforementioned MeLi/CuCN solution stirred at ꢀ78 ꢁC under
argon (balloon). After completion of the addition, the mixture was
stirred at the same temperature for 30 min. A solution of epoxide 11
(105 mg, 0.452 mmol) in dry Et2O (2.0 mL) was added via a syringe.
The mixture was stirred at ꢀ78 ꢁC for another 3 h, when TLC
showed completion of the reaction. Aq satd NH4Cl/aq NH3 (9:1 v/v,
10 mL) was added. The mixture was stirred at ambient temperature
until a clear biphasic solution was formed before being extracted
with Et2O (3ꢂ20 mL). The combined organic layers were washed
with water (20 mL) and brine (20 mL), and dried over anhydrous
Na2SO4. Removal of the solvent by rotary evaporation and column
EtOAc/PE) on silica gel gave epoxide 9 as a colorless oil (1.975 g,
23
8.5 mmol, 85% from 6a). [
(500 MHz, CDCl3)
a]
ꢀ13.8 (c 1.1, CHCl3); 1H NMR
D
d
3.93–3.83 (m, 1H), 3.66 (dd, J¼10.0, 3.8 Hz,
1H), 3.44 (dd, J¼11.0, 7.1 Hz, 1H), 3.15–3.08 (m, 1H), 2.81 (t,
J¼4.5 Hz, 1H), 2.61 (d, J¼3.7 Hz, 1H), 2.53 (dd, J¼4.9, 2.7 Hz, 1H),
1.81 (ddd, J¼14.2, 8.5, 4.1 Hz, 1H), 1.46 (ddd, J¼14.3, 7.1, 4.3 Hz, 1H),
0.88 (s, 9H), 0.07 (s, 6H); 13C NMR (125 MHz, CDCl3)
d 69.8, 67.1,
49.8, 47.2, 36.0, 25.9, 18.3, ꢀ5.4 (2C’s); FTIR (film) 3469, 2929,
2958, 1472, 1255, 1090, 837, 778, 669 cmꢀ1; ESIMS m/z 255.2
([MþNa]þ). ESIHRMS calcd for C11H24O3SiNa ([MþNa]þ):
255.1387; found: 255.1388.
4.7. Tosylation of 6a leading to 10
A
solution of 6a (246 mg, 1.0 mmol), dry Et3N (0.28 mL,
2.0 mmol), and p-TsCl (286 mg, 1.50 mmol) in dry CH2Cl2 (5.0 mL)
was stirred at ambient temperature overnight. The mixture was
diluted with CH2Cl2 (50 mL), washed with water (5 mL) and brine
(5 mL), and dried over anhydrous Na2SO4. Removal of the solvent
by rotary evaporation and column chromatography (1:6 EtOAc/PE)
chromatography (1:2 EtOAc/PE) on silica gel gave epoxide (2S,4S)-
25
21 as a colorless oil (466 mg, 1.163 mmol, 76%). [
a
]
þ3.3 (c 0.1,
D
CHCl3); 1H NMR (300 MHz, CDCl3)
d
5.83 (ddt, J¼17.1, 10.3, 6.5 Hz,
on silica gel afforded 10 as
a
colorless sticky oil (365 mg,
1H), 5.00 (d, J¼17.5 Hz, 1H), 4.92 (d, J¼10.4 Hz, 1H), 4.01–3.80 (m,
2H), 3.60 (dd, J¼9.7, 3.7 Hz, 1H), 3.45 (dd, J¼9.6, 7.2 Hz, 1H), 3.34 (br
s, OH), 2.95 (br s, OH), 2.04 (q, J¼6.8 Hz, 2H), 1.80–1.15 (m, 22H),
0.91 (s, 9H), 0.09 (s, 6H); FTIR (film) 3364, 3077, 2926, 2854, 1641,
1463, 1259, 1093, 909, 837, 778 cmꢀ1; ESIMS m/z 423.2 ([MþNa]þ).
ESIHRMS calcd for C23H48O3SiNa ([MþNa]þ): 423.3265; found:
432.3271.
23
0.911 mmol, 91%). [
CDCl3)
J¼10.1. 9.0 Hz, 1H), 4.50–4.41 (m, 1H), 3.87 (dd, J¼11.6, 3.3 Hz, 1H),
3.68 (dd, J¼11.7, 3.9 Hz,1H), 2.66 (ddd, J¼15.2, 9.1, 6.0 Hz,1H), 2.50–
2.35 (m, 4H); ESIMS m/z 423.1 ([MþNa]þ). ESIHRMS calcd for
a
]
þ19.4 (c 1.45, CHCl3); 1H NMR (300 MHz,
D
d
7.88 (d, J¼8.6 Hz, 2H), 7.38 (d, J¼8.9 Hz, 2H), 5.35 (dd,
C
18H28O6SSiNa ([MþNa]þ): 423.1268; found: 423.1280.

9382
J. Gao et al. / Tetrahedron 64 (2008) 9377–9383
4.10. Synthesis of (2S,4S)-22
water (10 mL) and brine (10 mL), and dried over anhydrous Na2SO4.
Removal of the solvent by rotary evaporation and column chro-
matography (2:1 EtOAc/PE) on silica gel gave the known14 triol
(2S,4R)-22 as a white solid (86 mg, 0.215 mmol, 71% from 9). Mp
A solution of (2S,4S)-21 (137 mg, 0.342 mmol) and n-Bu4NF
(1.0 M solution in THF, 2.0 mL, 2.0 mmol) in THF (5.0 mL) was
stirred at ambient temperature until TLC showed completion of the
reaction. Aq satd NH4Cl (5.0 mL) was added. The mixture was
extracted with EtOAc (3ꢂ20 mL). The combined organic layers
were washed with water (10 mL) and brine (10 mL), and dried over
anhydrous Na2SO4. Removal of the solvent by rotary evaporation
and column chromatography (2:1 EtOAc/PE) on silica gel gave the
27
20
83–84 ꢁC (lit.14 mp 82.5 ꢁC); [
a
]
ꢀ6.7 (c 1.1, CHCl3) (lit.14
[
a
]
ꢀ7.3
D
D
(c 1.0, CHCl3)); 1H NMR (300 MHz, CDCl3)
d
5.81 (ddt, J¼16.9, 10.1,
7.0 Hz,1H), 4.99 (d, J¼17.5 Hz,1H), 4.92 (d, J¼10.6 Hz,1H), 4.10–3.85
(m, 2H), 3.65 (dd, J¼11.4, 3.4 Hz, 1H), 3.51 (dd, J¼11.0, 7.2 Hz, 1H),
2.95 (br s, OH), 2.17 (br s, OH), 2.02 (q, J¼7.0 Hz, 2H), 1.80–1.20 (m,
22H).
known14 triol (2S,4S)-22 as a white solid (93 mg, 0.325 mmol, 95%).
22
Mp 66–67 ꢁC (lit.14 mp 65.5–66 ꢁC); [
a
]
þ6.4 (c 1.2, CHCl3) (lit.14
D
4.13. Synthesis of (2S,4R)-15
20
[
a]
þ6.0 (c 1.0, CHCl3)); 1H NMR (400 MHz, CDCl3)
d 5.81 (ddt,
D
J¼17.0, 10.3, 6.7 Hz, 1H), 4.99 (dd, J¼17.1, 1.8 Hz, 1H), 4.92 (dd,
J¼10.1, 0.9 Hz, 1H), 4.00–3.85 (m, 3H), 3.62 (dd, J¼11.1, 2.5 Hz, 1H),
3.45 (dd, J¼11.0, 6.5 Hz, 1H), 3.15 (br s, OH), 2.90 (br s, OH), 2.03 (q,
J¼7.0 Hz, 2H), 1.60–1.20 (m, 22H).
A solution of (2S,4R)-22 (16 mg, 0.0558 mmol) and Novozyme
435 (2 mg) in CH2]CHOAc (2.0 mL) was stirred at ambient tem-
perature for ca. 20 min, when TLC showed completion of the re-
action. The solids were filtered off. The filtrate was concentrated on
a rotary evaporator. The residue was chromatographed (1:2 EtOAc/
4.11. Synthesis of (2S,4S)-15
PE) on silica gel to give (2S,4R)-15 as a white solid (17 mg,
28
0.0508 mmol, 91%). Mp 44–45 ꢁC; [
a
]
D
ꢀ11.4 (c 0.15, CHCl3); 1H
A solution of (2S,4S)-21 (20 mg, 0.0698 mmol) and Novozyme
435 (2 mg) in CH2]CHOAc (2.0 mL) was stirred at ambient
temperature for ca. 20 min, when TLC showed completion of the
reaction. The solids were filtered off. The filtrate was concen-
trated on a rotary evaporator. The residue was chromatographed
NMR (400 MHz, CDCl3)
d
5.81 (ddt, J¼18.0, 10.4, 6.8 Hz, 1H), 4.99
(dd, J¼17.2, 1.6 Hz, 1H), 4.92 (dd, J¼10.4, 1.2 Hz, 1H), 4.22–4.12 (m,
2H), 4.03 (dd, J¼11.2, 7.6 Hz, 1H), 3.97–3.89 (m, 1H), 2.77 (br s, OH),
2.10 (s, 3H), 2.03 (q, J¼7.2 Hz, 2H), 1.72–1.22 (m, 22H); 13C NMR
(100 MHz, CDCl3)
d 171.2, 139.3, 114.1, 72.5, 70.8, 68.6, 39.1, 38.2,
(1:2 EtOAc/PE) on silica gel to give (2S,4S)-15 as a white solid
33.8, 29.6, 29.5, 29.2, 28.9, 25.3, 20.9; FTIR (KBr) 3389, 3077, 2921,
2850, 1741, 1641, 1464, 1370, 1260, 1089, 1035, 909, 799 cmꢀ1. ESIMS
m/z 351.1 ([MþNa]þ). ESIHRMS calcd for C19H36O4Na ([MþNa]þ):
351.2506; found: 351.2505.
21
(21 mg, 0.0649 mmol, 93%). Mp 52–53 ꢁC; [
a
]
þ6.7 (c 0.35,
D
CHCl3); 1H NMR (400 MHz, CDCl3)
d
5.81 (ddt, J¼17.0, 10.2,
6.6 Hz, 1H), 4.99 (dd, J¼17.0, 1.6 Hz, 1H), 4.93 (d, J¼10.4 Hz, 1H),
4.15–4.06 (m, 2H), 4.04–3.96 (m, 1H), 3.93–3.84 (m, 1H), 3.38 (br
s, OH), 2.65 (br s, OH), 2.10 (s, 3H), 2.03 (q, J¼6.8 Hz, 2H), 1.65–
Acknowledgements
1.18 (m, 22H); 13C NMR (100 MHz, CDCl3)
d 171.2, 139.3, 114.1,
72.5, 70.8, 68.6, 39.1, 38.2, 33.8, 29.6, 29.5, 29.1, 28.9, 25.3, 20.9;
FTIR (KBr) 3439, 3075, 2919, 2851, 1710, 1641, 1468, 1262, 1139,
This work was supported by the National Natural Science
Foundation of China (20372075, 20321202, 20672129, 20621062,
20772143) and the Chinese Academy of Sciences (‘Knowledge In-
novation’, KJCX2.YW.H08).
1093, 1043, 909, 800 cmꢀ1
;
ESIMS m/z 351.3 ([MþNa]þ).
ESIHRMS calcd for
351.2499.
C
19H36O4Na ([MþNa]þ): 351.2506; found:
References and notes
4.12. Synthesis of (2S,4R)-22
1. (a) Okabe, M.; Sun, R.-C.; Zenchoff, G. B. J. Org. Chem. 1991, 56, 4392–4397; (b)
Sun, R. C.; Okabe, M. Org. Synth. 1995, 72, 48–52.
2. Compound 3 is known as a natural hunger substance, see: Enders, D.; Sun,
H.-B.; Leusink, F. R. Tetrahedron 1999, 55, 6129–6138.
MeLi (3.0 M, in Et2O, 0.34 mL, 0.990 mmol) was added (via
a syringe) dropwsie to a mixture of CuCN (89 mg, 0.994 mmol) in
dry Et2O (4.0 mL) stirred at ꢀ78 ꢁC under argon (balloon). After
completion of the addition, the mixture was stirred at the same
temperature for 30 min (the pale yellow suspension gradually be-
came a solution). The lithium species of 20, prepared by dropwise
addition (via a syringe) of t-BuLi (1.5 M, in pentane, 1.7 mL,
2.55 mmol) to a solution of iodide 20 (300 mg, 1.020 mmol) in dry
Et2O (7.0 mL) stirred at ꢀ78 ꢁC under argon (balloon) followed by
stirring at the same temperature for 30 min, was then introduced
via a cannula to the aforementioned MeLi/CuCN solution stirred at
ꢀ78 ꢁC under argon (balloon). After completion of the addition, the
mixture was stirred at the same temperature for 30 min. A solution
of epoxide 9 (70 mg, 0.301 mmol) in dry Et2O (2.0 mL) was added
via a syringe. The mixture was stirred at ꢀ78 ꢁC for another 3 h,
when TLC showed completion of the reaction. Aq satd NH4Cl/aq
NH3 (9:1 v/v, 10 mL) was added. The mixture was stirred at ambient
temperature until a clear biphasic solution was formed before being
extracted with Et2O (3ꢂ20 mL). The combined organic layers were
washed with water (20 mL) and brine (20 mL), and dried over an-
hydrous Na2SO4. The solvent was removed by rotary evaporation.
The residue was dissolved in THF (5.0 mL). n-Bu4NF (1.0 M solution
in THF, 1.3 mL, 1.3 mmol) was added. The mixture was stirred at
ambient temperature until TLC showed completion of the reaction.
Aq satd NH4Cl (5.0 mL) was added. The mixture was extracted with
EtOAc (3ꢂ20 mL). The combined organic layers were washed with
3. (a) Wu, Y.; Guan, M.; Cheng, L.; Wu, X. C. Chin. Chem. Lett. 2001, 12, 765–766; (b)
Wu, Y.; Guan, M.; Ji, T.; Cheng, L. Ind. J. Chem. B 2001, 40, 136–138; (c) Wu, Y.;
Guan, M. Ind. J. Chem. B 1998, 37, 844–846; (d) Our 4 (tR¼13.025 min) was >99.29%
pure by HPLC analysis, performed on a Kromasil C18 column (150ꢂ4.6 mm,
5 mM) eluting with MeCN/H2O (starting from 90:10 linearly reaching 100:0 over
20 min and kept 100:0 afterwards) at a flow rate of 1.0 mL/min with the UV
detector set to 220 nm. (e) This solid (tosylate 4) ‘partially melt’ at ca. 67 ꢁC and
soon fully resolidified. The newly formed solid started to melt again at 101 ꢁ
C
and fully melt at 102 ꢁC. The crystal formed after full melting at 102 ꢁC still
showed mp 101–102 ꢁC, but without the phase change at 67 ꢁC.
4. Schurrle, K.; Piepersberg, W. J. Carbohydr. Chem. 1996, 15, 435–447 and refer-
ences cited therein.
5. (a) Vargeese, C.; Abushanab, E. Nucleosides Nucleotides 1992, 11, 1549–1559;
(b) The location of the Ts or TBS group on the primary OH in 4a and 6a,
respectively, was confirmed by the absence of any aldehyde signal in IR and
1H NMR spectra of the crude product of Dess–Martin oxidation of each
compound.
6. (a) Martinelli, M. J.; Nayyar, N. K.; Moher, E. D.; Dhokte, U. P.; Pawlak, J. M.;
Vaidyanathan, R. Org. Lett. 1999, 1, 447–450; (b) Martinelli, M. J.; Vaidyana-
than, R.; Pawlak, J. M.; Nayyar, N. K.; Dhokte, U. P.; Doecke, C. W.; Zollars, L.
M. H.; Moher, E. D.; Van Khau, V.; Kosmrlj, B. J. Am. Chem. Soc. 2002, 124,
3578–3585.
7. (a) Hashimura, H.; Ueda, C.; Kawabata, J.; Kasai, T. Biosci. Biotechnol. Biochem.
2001, 65, 1656–1658; (b) Kashman, Y.; Neeman, I.; Lifshitz, A. Tetrahedron 1969,
25, 4617–4631; (c) Kashman, Y.; Neeman, I.; Lifshitz, A. Tetrahedron 1970, 25,
1943–1951; (d) Momergue, F.; Helms, G. L.; Prusky, D.; Browse, J. Phytochemistry
2000, 54, 183–189.
8. Baldwin, J. E.; Chen, D.; Russell, A. T. Chem. Commun. 1997, 2389–2390.
9. Quinkert, G.; Billhardt, U.-M.; Jakob, H.; Fischer, G.; Glenneberg, J.; Nagler, P.;
Autze, V.; Heim, N.; Wacker, M.; Schwalbe, T.; Kurth, Y.; Bats, J. W.; Du¨rner, G.;
Zimmermann, G.; Kessler, H. Helv. Chim. Acta 1987, 70, 771–861.

J. Gao et al. / Tetrahedron 64 (2008) 9377–9383
9383
10. Kato, M.; Mori, K. Agric. Biol. Chem. 1985, 49, 2479–2480.
11. Bach, T.; Lemarchand, A. Synlett 2002, 1302–1304.
12. Negishi, E.-i.; Swanson, D. R.; Rousset, C. J. J. Org. Chem. 1990, 55, 5406–5409.
13. (a) Lipshutz, B. H.; Wilhelm, R. S.; Kozlowski, J. A. J. Org. Chem. 1984, 49, 3938–
3942; (b) Lipshutz, B. H. Synthesis 1987, 325–341.
15. (a) Teng, R.; Ang, C.; McManus, D.; Armstrong, D.; Mau, S.; Bacica, A. Tetrahe-
dron Lett. 2003, 44, 5661–5664; (b) Teng, R.; Ang, C.; McManus, D.; Armstrong,
D.; Mau, S.; Bacica, A. Helv. Chim. Acta 2004, 87, 1860–1872.
16. There exist many such compounds in nature. The oxo polyene macrolide family
antibiotics alone, for instance, already contain more than 200 members. See:
Rychnovsky, S. D. Chem. Rev. 1995, 95, 2021–2040.
14. Sugiyama, T.; Sato, A.; Yamashita, K. Agric. Biol. Chem. 1982, 46, 481–485.