Synthetic and computational studies on the tricarboxylate core of 6,7-dideoxysqualestatin H5 involving a carbonyl ylide cycloaddition- rearrangement

Article abstract of DOI:10.1039/c004496b

Reaction of diazodiketoesters 17 and 28 with methyl glyoxylate in the presence of catalytic rhodium(ii) acetate generates predominantly the 6,8-dioxabicyclo[3.2.1]octanes 29 and 30, respectively. Acid-catalysed rearrangement of the corresponding alcohol 3

Full text of DOI:10.1039/c004496b

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Synthetic and computational studies on the tricarboxylate core of
6,7-dideoxysqualestatin H5 involving a carbonyl ylide
cycloaddition–rearrangement†
David M. Hodgson,*a Carolina Villalonga-Barber,b Jonathan M. Goodmanc and Silvina C. Pellegrinetd
Received 19th March 2010, Accepted 8th June 2010
First published as an Advance Article on the web 2nd July 2010
DOI: 10.1039/c004496b
Reaction of diazodiketoesters 17 and 28 with methyl glyoxylate in the presence of catalytic rhodium(II)
acetate generates predominantly the 6,8-dioxabicyclo[3.2.1]octanes 29 and 30, respectively.
Acid-catalysed rearrangement of the corresponding alcohol 31 favours, at equilibrium, the
2,8-dioxabicyclo[3.2.1]octane skeleton 33 of the squalestatins–zaragozic acids. Force field calculations
on the position of the equilibrium gave misleading results. DFT calculations were correct in suggesting
that the energy difference between 31 and 33 should be small, but did not always suggest the right
major product. Calculation of the NMR spectra of the similar structures could be used to assign the
isomers with a high level of confidence.
Introduction
More people die annually from cardiovascular diseases (CVDs),
a group of disorders of the heart and blood vessels, than from
any other cause.1 High levels of cholesterol are strongly associated
with CVDs and statins are effective in preventing heart diseases
by inhibiting HMG-CoA reductase, the rate-controlling enzyme of
the cholesterol biosynthetic pathway. Squalene synthase inhibitors
have shown promise as an alternative to statins in reducing levels
of cholesterol.2 As part of the search for more effective medicines,
Fig. 1
fungal extracts were screened for inhibition of squalene synthase,
which lead to the isolation of the zaragozic acids (squalestatins)
in the early 1990s [e.g. zaragozic acid A (squalestatin S1) (1)
(Fig. 1)].3–5 Squalene synthase catalyses the first committed step
of cholesterol biosynthesis, and zaragozic acids are picomolar
inhibitors of this enzyme. The structurally novel and synthetically
challenging 2,8-dioxabicyclo[3.2.1]octane-3,4,5-tricarboxylic acid
core, together with the biological activity of the zaragozic acids,
has resulted in this family of molecules becoming irresistible
targets for synthetic chemists. Total syntheses of several zaragozic
acids and many synthetic studies towards the highly oxygenated
core have been reported.6
As part of a research programme exploring the scope of
diazocarbonyl-derived carbonyl ylide [3+2] dipolar cycloaddi-
tions as a powerful complexity-generating strategy in target and
asymmetric synthesis,7 we detail here our successful synthetic
approach (together with associated computational studies) to
the triester of an anhydrofuranose core 38 (Scheme 1) of 6,7-
dideoxysqualestatin H5 (2),9 the least oxygenated member of the
squalestatin family. Our approach to this core 3 envisages as
keys steps an acid-catalysed intramolecular transketalisation of
a 6,8-dioxabicyclo[3.2.1]octane system 4 forged by a carbonyl
ylide cycloaddition (6→4). The latter involves a glyoxylate as
dipolarophile to directly produce the triacid functionality at the
correct oxidation level, and requires preferential generation of 1
out of 8 possible cycloadducts.
aDepartment of Chemistry, Chemistry Research Laboratory, University of
Oxford, Mansfield Road, Oxford, UK OX1 3TA. E-mail: david.hodgson@
chem.ox.ac.uk
In earlier cycloaddition studies, we demonstrated the viability
of glyoxylates as dipolarophiles with more commonly used 2-
diazo-3,6-diketoesters substrates.10,11 While these cycloadditions
occurred with the regioselectivity desired in Scheme 1 unfortu-
nately, and under all conditions examined, the glyoxylate ester
group was preferentially incorporated endo with respect to the
ylide-containing ring (e.g., 7→8 in Scheme 2).
bInstitute of Organic and Pharmaceutical Chemistry, National Hellenic
Research Foundation, Vas. Constantinou 48, Athens, 11635, Greece
cUnilever Centre for Molecular Science Informatics, Department of Chem-
istry, Lensfield Road, Cambridge, CB2 1EW
dInstituto de Qu´ımica Rosario (CONICET), Facultad de Ciencias
Bioqu´ımicas y Farmace´uticas, Universidad Nacional de Rosario, Suipacha
531, Rosario (2000), Argentina
By incorporating the a-hydroxy ester functionality at C-3 shown
in the cycloaddition substrate 6 (Scheme 1), we aimed to reverse
the earlier endo preference exemplified in Scheme 2—which was
ascribed to secondary orbital overlap between ketone(dipole) and
ester(dipolarophile) functionality. Also, by appropriate choice
of hydroxyl protection, it was hoped to provide a strong
† Electronic supplementary information available: Experimental details
for compounds 11–19, 20 (via 19), 21 (via 24), 22–24 and 29, assignment
of spirolactones 34, structures arising from opposite facial selectivity in
cycloaddition, comparison of NMR spectra of alcohols 31 and 33 with
natural product 2, 1H and 13C NMR spectra of new compounds, Cartesian
coordinates and energies of the DFT structures, and CP3 calculations. See
DOI: 10.1039/c004496b
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3975–3984 | 3975
©

View Article Online
Scheme 4 Reagents and conditions: i, methyl vinyl ether, t-BuLi, THF,
-78 ◦C (92%); ii, TMSCl, pyridine, THF, 0 ◦C, (92%); iii, O3, CH2Cl2,
-78 ◦C, then Ph3P (12 : 33% and 13: 30%).
the generated amide by a masked carboxylate nucleophile was
not considered necessary in this unsubstituted lactone. Indeed,
addition of 1-methoxyvinyllithium was found to be straightfor-
ward, generating the unstable ring-opened hydroxyketone 10, in
92% yield. However, if the reaction was allowed to warm to room
temperature before being quenched, a mixture of hydroxyketone
10 and the corresponding lactol was obtained; only for highly
substituted lactones is the lactol generated quantitatively on
addition of a 1-alkoxyvinyllithium13 and therefore requires an
initial transformation into an amide, as seen in Carreira and
Dubois’ total synthesis of zaragozic acid C.14 Moreover, in the
current studies, a strategy proceeding through an intermediate
amide did not provide significant improvement (vide infra).
Protection of the free hydroxyl group of hydroxyketone 10 using
TMSCl-pyridine provided the TMS ether 11, in 92% yield [this
silyl group was found to be quite labile (vide infra) and would
later be (preferably) replaced by the less labile triethylsilyl group].
Ozonolysis of TMS ether 11 provided a-ketoester 12 in low yield.
Cleavage of the TMS ether was found to occur in solution, and
ozonolysis of a solution of TMS ether 11 which had been standing
at room temperature for a couple of hours gave lactol 13 as the
major product. Repetition of the reaction using freshly prepared
11 gave the same components, indicating that desilylation was
also occurring to some extent at low temperatures, though now
favouring the desired a-ketoester 12 in the crude mixture (1.6 : 1)
albeit isolated in only 33% yield (lactol 13 was also isolated, in
30% yield).
Scheme 1
Scheme 2 Reagents and conditions: methyl glyoxylate, Rh2(OAc)4 (cat.),
PhMe, 110 ◦C (60%).10,11
diastereofacial bias in the ylide 5 to the incoming dipolarophile.
In this regard, we were encouraged by a related carbonyl
ylide formation–cycloaddition from a triester in Hashimoto
and co-workers’ approach to zaragozic acid C (Scheme 3,
note however, that the diastereomeric triester failed to undergo
cycloaddition).6h,12
Addition of lithiated methyl diazoacetate15 to a-ketoester 12
gave a-diazoester 14 in 44% yield (53% based on recovered 12,
Scheme 5). TMS protection of the free tertiary hydroxyl group was
achieved using TMS-imidazole in THF; however, deprotection (or,
possibly, silyl migration to the tertiary alcohol) occurred at the
labile secondary TMS ether to some extent during this step and
the crude reaction mixture contained disilyl and monosilyl ethers
15 and 16 in 1 : 0.4 ratio, respectively. The latter was not incon-
venient since selective cleavage of the secondary TMS ether was
next required, although the lability of the secondary TMS ether
may account for some of the low yields throughout this synthetic
sequence. Reaction of the crude mixture of silyl ethers 15 and
16 with PCC effected concomitant deprotection of the secondary
TMS ether in 15 and oxidation, giving the desired cycloaddition
Scheme 3 Reagents and conditions: (E)-3-hexene-2,5-dione, Rh2(OAc)4
(cat.), benzene, 80 ◦C (47%).12
Results and discussion
So as to examine the chemistry outlined in Scheme 1, a synthesis
of cycloaddition substrate 6 (R = Me, [Si] = SiMe3 was developed.
The synthesis of TMS-protected a-diazoester 17 commenced with
g-valerolactone (9) and proceeded as shown in Scheme 4 and
Scheme 5. Ring-opening of lactone 9 by an amine followed by
suitable protection of the free alcohol and nucleophilic attack on
3976 | Org. Biomol. Chem., 2010, 8, 3975–3984
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Scheme 5 Reagents and conditions: i, LDA, N2CHCO2Me, THF, -78 ◦
C
(44%); ii, TMS-Im, THF, 25 ◦C, 15 : 16 1 : 0.4; iii, PCC, CH2Cl2, 25 ◦C
(17 : 58% and 16: 7%); iv, PCC, NaOAc, CH2Cl2, 25 ◦C (62%).
Scheme 6 Reagents and conditions: i, pyrrolidine, PhMe, 25 ◦C (86%);
ii, TESOTf, 2,6-lutidine, CH2Cl2, -78 ◦C (65% from 10 and 92% from
18); iii, methyl vinyl ether, t-BuLi, THF, -78 ◦C (92%); iv, O3, pyridine,
CH2Cl2, -78 ◦C, then Ph3P (57%); v, NaOH, MeOH (88%); vi, TESCl,
Imidazole, DMF, 100 ◦C (42%); vii, K2CO3, MeOH, THF, H2O, 25 ◦C
substrate a-diazoester 17 in 58% yield from a-diazoester 14. A
small amount of alcohol 16 (7%) was also isolated from this last
reaction, which could be oxidised with PCC to provide further
a-diazoester 17 (62%).
(71%); viii, Ph3P CHCN, EDCI, DMAP, CH2Cl2, 0 ◦C to 25 ◦C (50%);
=
ix, O3, pyridine, CH2Cl2, MeOH, -78 ◦C, then Ph3P (34%).
A cycloaddition substrate possessing an alternative silyl ether
to that in TMS-protected a-diazoester 17 (such as TBDMS
protection) was also sought, to enable comparison of the effect of
a more sterically demanding ether in the subsequent cycloaddition
chemistry. This also provided an opportunity to improve on the
above synthetic sequence by moving away from the rather labile
secondary TMS ether. However, the protecting group for the
secondary alcohol would need to be sufficiently stable to survive
the previously developed chemistry (Scheme 4 and Scheme 5), and
at the same time be labile enough to be cleaved in the presence
of a TBDMS-protected tertiary alcohol under conditions which
would not destroy the diazo group, whose stability is lowered due
to the presence of the a-hydroxy group. Triethylsilyl ethers are
typically 10–100 times more stable than TMS ethers, but are labile
to mild acidic conditions.16 With the above considerations in mind,
TES-protected ketone 20 was obtained in 65% yield by silylation
of hydroxyketone 10 with TESOTf–2,6-lutidine (Scheme 6). An
alternative route to triethylsilyl ether 20 was also achieved, via
amide 18.17 Addition of excess a-lithio methyl vinyl ether to TES-
protected amide19 gave ketone20 (92%yield), withoutany signs of
overaddition according to 1H NMR analysis of the crude reaction
mixture. Optimal yields of a-ketoester 21 (57%) were obtained by
submitting freshly prepared ketone 20 to ozonolysis in the presence
of pyridine18 [to minimise (acid-catalysed) hydrolysis of the vinyl
ether]; no lactol was observed, which is indicative of the robustness
of the TES protecting group.
(9) by hydroxide-induced ring-opening (88%),20 TES-capping of
both acid and hydroxyl groups (42%), then selective cleavage of the
resulting silyl ester 2 using K2CO3 to reveal acid 23 (71%), which
was finally coupled with (cyanomethylene)triphenylphosphorane
in moderate yield (50%). Ozonolysis of b-ketocyanophosphorane
24 was first attempted in a solution of MeOH and CH2Cl2;21
however, these reaction conditions led to the formation of lactol
13, presumably due to the in situ formation of HCN which cleaved
the silyl protecting group and promoted ring-closure. Addition
of pyridine to the reaction mixture circumvented this problem
and a-ketoester 21 was obtained in 34% yield. The moderate
yields from the ozonolyses reactions are attributed to hydration of
the a-ketoesters on silica;6g however, in the present case addition
of Me2S instead of Ph3P following ozonolysis (in order to avoid
chromatographic purification), did not lead to improvement.
Addition of lithiated methyl diazoacetate15 to a-ketoester 21
gave a-diazoester 25, in 62% yield (Scheme 7). Pleasingly, the
derived TBDMS ether 26 (72% from 25 using TBDMSOTf
and 2,6-lutidine) underwent selective deprotection (76%) of the
secondary TES ether using AcOH in THF–water after 1 h, without
affecting either the tertiary silyl ether or the diazo group. Oxidation
of the resulting secondary alcohol 27 using PCC generated the
desired cycloaddition substrate 28, in 90% yield.
a-Diazoesters 17 and 28 underwent Rh2(OAc)4-catalysed tan-
dem carbonyl ylide formation–cycloaddition with methyl glyoxy-
late to each give a mixture of three cycloadducts in ratios 8 : 1 : 1
and 12 : 2 : 1, respectively (Scheme 8). The isomer proportions
were assigned from the clear singlets of the bridge methines in
An alternative strategy to a-ketoester 21 utilised chemistry
of Wasserman and Ho, wherein a b-ketocyanophosphorane is
ozonolysed to yield the corresponding a-keto ester.19 The required
b-ketocyanophosphorane 22 was synthesised from g-valerolactone
1
the crude H NMR spectra. The isomers could not be totally
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3975–3984 | 3977
©

View Article Online
to the silyloxy group) to avoid steric interactions, and with the
methyl glyoxylate orienting itself exo to the ylide-containing
ring.
With the major cycloadducts possesing the desired stereochem-
istry, their propensity to undergo the transketalisation process
was next examined. Firstly, desilylation of the cycloadducts 29
and 30 was performed using TBAF in THF. From the reaction of
cycloadduct 29 the resulting isomeric alcohols were now separable,
however, the desired alcohol 31 was isolated in only 15% yield.
More promising results were obtained from the desilylation of
cycloadduct 30 using TBAF, with alcohol 31 being isolated in 72%
yield as a single isomer (Scheme 9). From a larger scale desilylation
using cycloadduct 30, containing some of a minor isomer, a
fraction consisting of mainly the corresponding desilylated minor
isomer (tentatively assigned as alcohol 32), could be isolated (5.5%
yield) and differences between the NOE spectra of the two different
alcohol isomers could then be compared in order to further
strengthen the assigned structure of the major cycloadduct. An
NOE for the major desilylated cycloadduct 31 of 6.8% betweeen
C7–H and one H of C3–H2 and the absence of a corresponding
NOE for the minor cycloadduct 32, together with the NOE data
on the rearranged material (vide infra) further suggests the major
isomer 31 to have the glyoxylate-derived ester group exo-disposed
relative to the six-membered ring.
Scheme 7 Reagents and conditions: i, LDA, N2CHCO2Me, THF, -78 ◦
C
(62%); ii, TBDMSOTf, 2,6-lutidine, CH2Cl2, 0 ◦C to 25 ◦C (72%); iii,
AcOH–THF–H2O, 25 ◦C (76%); iv, PCC, NaOAc, CH2Cl2, 25 ◦C (90%).
Scheme 8 Reagents and conditions: i, methyl glyoxylate, PhMe, 110 ◦C,
then Rh2(OAc)4 (cat.) (29 : 54%; 30 : 63.5%).
separated after column chromatography. From a-diazoester 17 a
54% combined yield of cycloadducts was obtained, with a major
fraction (47%) predominantly consisting of cycloadduct 29 (Fig. 2)
along with traces (10 : 0 : 1) of a minor cycloadduct. Cycloadducts
30 (63.5%) was similarly isolated along with traces (10 : 0 : 1) of a
minor cycloadduct.
Scheme 9 Reagents and conditions: i, TBAF, THF, 25 ◦C (31 : 15% from
29; 72% from 30).
Treatment of cycloadduct alcohol 31 under the rearrangement
conditions previously described by Nicolaou and co-workers (2%
HCl in MeOH at reflux)22 gave, after 15 h, a mixture of four bicyclic
compounds in a ratio of 69 : 21 : 5 : 5, as suggested by integration
Fig. 2
1
of the methine singlets in the d 5.60 to 4.80 region of the H
The stereochemistry of the major cycloadduct isomer in both
cases was assigned as that required for 6,7-dideoxysqualestatin
H5 synthesis. Configurational assignments were made following
NOE studies on the major isomers (Fig. 2). An NOE betweeen
C7–H and one H of C3–H2 of 6% for cycloadduct 29 and 4.6%
for cycloadduct 30 suggest the major isomers to be exo- with
respect to the six-membered ring. Facial selectivity was assigned
on the basis of Hashimoto’s precedent,12 together with the fact
that the cycloaddition on the more sterically demanding a-diazo
ester 28 provided a higher proportion of the major cycloadduct.
Data supporting the cycloaddition regioselectivity in cycloadducts
29 and 30 are ketal carbon C5 [dC 110.2 for both cycloadducts, and
dC 111.0 for C5 in cycloadduct 8 (Scheme 2), where the structure
was previously established by X-ray crystallographic analysis11]
and HMBC conectivities with C4–H2 and C5–Me. It is apparent
that modification of the a-keto group of the ylide to bulkier
a-silyloxy ester functionality led to 1,3-dipolar cycloaddition
occurring preferentially on the less-hindered face (the one opposite
NMR spectrum of the crude mixture; the major component of
the mixture being starting material. Separation of the two major
components could be achieved after column chromatography in
40% and 13% yield respectively, while the two minor less polar
compounds were isolated as a 1 : 1.5 mixture in 5% combined
yield. The latter mixture was tentatively assigned as a mixture of
diastereomeric spirolactones 34 (Scheme 10, further information
is given in the ESI†).
Assignment of the most polar component as the desired 2,8-
1
dioxabicyclic core 33 was made on the basis of H NMR and
NOE studies. Small vicinal coupling constants around the ring
suggested that the dimethylene group was confined to a five-
membered ring,23 while NOE enhancements were obtained on
two protons, one of each methylene, when H-3 was irradiated. The
observed NOE enhancements further corroborate the assigned
facial selectivity in the cycloaddition step; opposite facial selectiv-
ity would have resulted in a rearranged bicyclic compound where
NOE enhancements between the H-3 and the 5-membered ring
3978 | Org. Biomol. Chem., 2010, 8, 3975–3984
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Table 1 Effect of experimental conditions on the transketalisation
process
Crude ratio
31 : 33 : 34
Entry Conditions
Substrate
1
2
3
4
5
6
7
8
9
2% HCl, MeOH, D, 15 h
CSA, MeOH, D, 24 h
69 : 21 : 10
83 : 13 : 14
100 : 0 : 0
100 : 0 : 0
75 : 25 : 0
31
31
31
31
31
31
33
31
33
2% HCl, CHCl3, D, 15 h
Triflic acid, DMSO-d6, 25 ◦C, 4 d
Triflic acid, CDCl3, 25 ◦C, 4 d
CH2Cl2/TFA/H2O 20 : 10 : 1, D, 15 h 40 : 60 : 0
CH2Cl2/TFA/H2O 20 : 10 : 1, D, 15 h 28 : 78 : 0
CH2Cl2/TFA/H2O 20 : 10 : 1, D, 68 h 34 : 66 : 0
CH2Cl2/TFA/H2O 20 : 10 : 1, D, 68 h 34 : 66 : 0
Scheme 10 Reagents and conditions: H+ (see Table 1).
protons would have been absent (see ESI†). Moreover, the 1H and
13C NMR spectra of 2,8-dioxabicyclo[3.2.1]octanetrioate 33 were
very similar with regard to the corresponding signals reported for
the natural product 29 (particularly ring methine chemical shifts,
for a detailed comparison see the ESI†).
it dominated the reaction mixture (31:33, 34 : 66), thus validating
our rearrangement strategy to the 2,8-dioxabicyclo[3.2.1]octane
core of the zaragozic acids/squalestatins. The desired core 33 was
then isolated in 54% yield (83% based on recovered 31). That true
equilibrium had been reached was established by subjecting core
33 to these latter reaction conditions, which resulted in the same
ratio of alcohols 31 : 33 (entry 9).
Molecular mechanics calculations for cycloadduct alcohol 31
and rearranged alcohol 33 using the MM2* force field26 and the
OPLSAA force field27 as implemented in MacroModel28 lead to
the prediction that rearranged alcohol 33 should not be observed
at equilibrium, as the energy difference between the two isomers
is greater than 25 kJ mol-1 in both cases (Table 2, entries 1–
3). However, such force fields underestimate the energies of 1,3-
dioxolanes compared to 1,3-dioxanes, because key atoms are now
in a 1,4 relationship instead of a 1,3 relationship, and so the intra-
ring electrostatic interactions are treated inconsistently.24 This
anomaly can be treated by scaling the electrostatic interactions,
or, in the case of reasonably small molecules like 31 and 33, using
a higher level of theory. The DFT methods, using Jaguar,29 all
suggest that the energy differences between 31 and 33 should
A range of other acidic conditions were then screened for the
rearrangement of alcohol 31 (Table 1). CSA in MeOH gave some
signs of isomerisation to the core 33, although a 1 : 1 mixture
of spirolactones 34 was again observed and starting 31 was still
the major component present (Table 1, entry 2). To avoid the
formation of spirolactones 34 other solvents were examined. No
reaction occurred with 2% HCl in CHCl3 at reflux after 15 h
(entry 3) or with triflic acid in DMSO-d6 after 4 days at 25 ◦C
(entry 4).24 Triflic acid in CDCl3 generated 33 but only in a
poor ratio with respect to starting 31 (entry 5). Encouraging
results were found under Evans’ conditions (CH2Cl2/TFA/H2O
20 : 10 : 1)25 (40 : 60 31 : 33, entry 6). However, submitting the
isolated rearranged material 33 to the same conditions did not
provide the same ratio (entry 7), suggesting that true equilibrium
had not been reached. Prolonged exposure to Evans’ conditions
(68 h) slightly increased the proportion of 33 (entry 8), so that
Table 2 Calculated equilibrium ratios 31 : 33a
Entry
Theory
E(33) - E(31)/kJ mol-1
Calcd. 31 : 33
1
2
3
4
5
6
7
8
9
MM2*
OPLSAA
26.6
25.9
25.1
0.0
-1.3
-2.0
-0.4
3.5
100 : 0
100 : 0
100 : 0
50 : 50
38 : 62
31 : 69
46 : 54
80 : 20
83 : 17
OPLSAA with continuum water
B3LYP/6-31G**
B3LYP/6-31G** with CH2Cl2 modelbc
B3LYP/6-31G** with CH2Cl2 modelbd
B3LYP/6-31++G**
LMP2/6-31G**
ce
LMP2/6-31G**
4.0
de
E(33) - E(31)/kJ mol-1
Calcd. 31 : 33
6-31G**
6-311+G**
cc-pVTZ
6-31G**
6-311+G**
cc-pVTZ
B3LYP
BHandHLYP
M05
M06
MPW1PW91
0.0
3.3
2.8
4.6
-2.5
2.2
3.5
1.5
6.0
-0.2
-4.0
1.2
-1.4
2.1
50 : 50
79 : 21
76 : 24
86 : 14
27 : 73
71 : 29
80 : 20
64 : 36
92 : 8
17 : 83
62 : 38
36 : 64
70 : 30
29 : 71
-2.2
48 : 52
a Conformation searches were carried out with MM2* and OPLSAA. The lowest energy OPLSAA geometries were reminimised using the DFT
methods. b Continuum model using Jaguar’s Poisson–Boltzmann continuum solvation model and default parameters for dichloromethane.36a c Single
point. d Optimisation. e Local MP2, all pairs.36b
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3975–3984 | 3979
©

View Article Online
be small (entries 4–7), and LMP2 calculations reinforce this
(entries 8–9). Only B3LYP30,31 used with a solvent model and
MPW1PW9132 suggested that rearranged alcohol 33 should be
thermodynamically preferred to cycloadduct alcohol 31 by a small
amount, as is experimentally observed. BHandHLYP,33 M0534 and
M0635 all show, correctly, that the energy differences between the
structures are small, but suggest a favoured structure opposite to
that found from experiment. Increasing the size of the basis set
(Table 2, section 2) did not give consistent results. 6-311+G** give
similar results to 6-31G**, slightly increasing the preference for 31
in most cases. The largest basis set investigated, cc-pVTZ, slightly
overestimates the preference for 33.
We conclude that calculations on equilibria with these structures
must be treated with caution, even when quite a high level
of theory is being used. The force fields give very misleading
estimates of the energy differences, as expected. The DFT methods,
however, do not give a reliable guide to the major isomer in
most cases, and in the seven where they do (B3LYP/6-31G**
with a solvent model, MPW1PW91/6- 31G**, B3LYP/6-31++G**,
MPW1PW91/6-311+G**, B3LYP/cc-pVTZ, M05/cc-pVTZ, and
MPW1PW91/cc-pVTZ) it may be that the agreement is fortuitous.
LMP2 (entries 8 and 9) gives similar results to the DFT methods,
intermediate between M05 and M06 with the same basis set. This
is a particularly difficult comparison, because the experimental
results suggest that the energy difference should be small, and so a
high degree of precision is required to ensure that the right sense of
selectivity is obtained. The calculations could have been improved
by performing DFT calculations on all fifty of the OPLSAA
calculated conformations, and further improved by doing a further
conformation search on the DFT potential energy surface. Both
of these procedures would take a prohibitive amount of computer
power. Calculations on these systems appear to be useful at the
DFT level, provided that they are not expected to be accurate to
more than a few kJ mol-1 in comparing the isomers.
5,7-Dioxabicyclo[2.2.1]heptanetrioate 35 could not be unam-
biguously ruled out as the major product from the acid-catalysed
rearrangement of cycloadduct alcohol 31, as alcohols 33 and
35 would be anticipated to have very similar NMR spectra (see
Table 3). In order to check that bicyclo[2.2.1]heptane 35 is higher
in energy, we carried out an OPLSAA conformation search on 35
and re-calculated the energy using B3LYP/6-31G** in the same
way as we had done for alcohols 31 and 33 in Table 2, entry 4.
These calculations show that 35 is much higher in energy (about
20 kJ mol-1) than alcohol 33. The comparison of 31 and 33
suggested that these calculations were probably accurate to within
a few kJ mol-1. This level of precision makes a large difference
when comparing two structures with similar energies, such as 31
and 33. For the comparison of 33 and 35 this uncertainty is less
than 10% of the difference, and so we can be confident that under
equilibrium conditions the ratio between bicyclic alcohols 31, 33
and 35 would contain almost no 35.
theory, using the same structures reported above and the GIAO
method.37 Our earlier studies suggest that this is an appropriate
level of theory for structures of this type.38 The results are
given in Table 3. We have developed a comparison parameter,
CP3, to help to pair computational and experimental NMR
data and to give a measure of the confidence that we can have
in the results.39 CP3 is positive for the assignment of alcohols
31 and 33 illustrated in Scheme 10, suggesting that this is the
correct assignment. Swapping the assignments of alcohols 31 and
33 gives a negative CP3 value, suggesting this is the incorrect
assignment. A Bayesian analysis of these values indicates that we
can be confident (>99.9%) that the spectra have been assigned
the correct way around, consistent with the result expected from
the cycloaddition reaction. If we assume that one of the two
compounds is bicyclo[2.2.1]heptane 35, then the CP3 value is
negative for all pairings, indicating that the experimental data do
not fit the calculated values, further reaffirming the assignment of
31 and 33.
Table 3 Experimental and calculated 13C data for bicyclic alcohols 31, 33
and 35
C atoma
31 expt
31 calc
33 expt
33 calc
35 calc
B-CO2
173.8
169.2
166.7
111.0
89.4
77.6
73.7
53.7
53.0
52.6
30.8
29.6
23.6
170.4
165.1
163.2
111.7
94.6
80.3
76.1
52.0
52.6
51.7
32.7
32.8
24.5
168.7
169.9
167.4
107.6
74.9
74.9
88.3
53.3
52.9
52.7
32.8
29.4
23.6
163.9
163.8
162.3
107.6
79.5
78.7
92.8
52.1
51.7
51.6
34.0
31.3
24.0
160.9
165.8
167.8
112.3
92.0
71.4
94.7
52.0
51.6
52.4
38.0
33.2
19.3
A-CO2
G-CO2
E
A
G
B
A-CO2Me
B-CO2Me
G-CO2Me
D
C
E-Me
a The assignments of the ester and methoxy C atoms are based on matching
their peaks as closely as possible to the calculated values.
Conclusions
The carbonyl ylide cycloaddition precursors a-diazoesters 17
and 28 [each 7 steps from g-valerolactone (9)] were designed to
remove the potential secondary orbital effect between previously
studied carbonyl ylides and a glyoxylate dipolarophile. This
effect was considered a significant influence on the cycloadduct
stereochemistry, which was undesired for projected squalestatin
synthesis. In the current work, the putative carbonyl ylides
generated in situ from the rhodium-catalysed decomposition
of a-diazoesters 17 and 28 underwent highly diastereoselective
cycloadditions using methyl glyoxylate as the dipolarophile, with
the major cycloadduct diastereomers 29 and 30 now possessing
the required stereochemistry for 6,7-dideoxysqualestatin H5 (2)
synthesis. Use of a-diazoesters 29 and 30 not only reversed the
endo-selectivity previously observed simple a-diazo-b-ketoesters,
As a further check, we calculated the 13C NMR spectra for
bicyclic alcohols 31, 33, and 35, at the B3LYP/6-31G** level of
3980 | Org. Biomol. Chem., 2010, 8, 3975–3984
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
but also delivered strongly biased face selectivity arising from the
a-silyloxyester stereocentre. Transketalisation of cycloadduct 31
under thermodynamic conditions resulted in favourable isomeri-
sation to the bicyclic core 33 contained in 6,7-dideoxysqualestatin
H5 (2). All of the DFT calculations calculated, correctly, that
the energy difference between the isomers should be small.
However, the methods disagreed about the degree and sense of the
preference. Application of the above chemistry to the synthesis of
6,7-dideoxysqualestatin H5 is currently under investigation, and
the results will be reported in due course.
H of =CH2), 3.90–3.70 (1 H, m, CH), 3.64 (3 H, s, OMe), 2.82
(2 H, t, J 7.5, CH2), 1.90–1.60 (2 H, m, CH2) and 1.21 (3 H, d, J
6, Me); dC(100 MHz) 197.9 (C(3), quat.), 158.4 (C(2), quat.), 90.7
(CH2=), 67.3 (CH), 55.2 (OMe), 34.2 (CH2), 32.6 (CH2) and 23.5
(Me).
6-(Triethylsilyloxy)-2-methoxyhept-1-en-3-one 20. From 10:
TESOTf (1.0 cm3, 4.7 mmol) was added to a stirred solution of
◦
2,6-lutidine (1.0 cm3, 4.7 mmol) in CH2Cl2 (8 cm3) at 0 C. The
◦
resulting solution was stirred at 0 C for 30 min, then cooled to
◦
-78 C. A solution of crude alcohol 10 (500 mg, 3.16 mmol) in
CH2Cl2 (5 cm◦3) was then added, and the reaction mixture was
stirred at -78 C for 5 min. Water (10 cm3) was then added, the
layers were separated and the aqueous layer extracted with CH2Cl2
(3 ¥ 10 cm3). The combined organic layers were washed with brine
(10 cm3), water (10 cm3), dried (MgSO4) and evaporated under
reduced pressure to give a colourless oil, silyl ether 20 (0.56 g,
65%) which was used directly in the next step; Rf 0.56 (20% Et2O
in light petroleum); nmax/cm-1 2956 m, 2913 s, 2877 s, 1709 m,
1613 s, 1458 m, 1376 m, 1139 m, 1055 s, 1005 s, 848 w and 743 s;
dH(400 MHz) 5.20 (1 H, d, J 3, =CH2), 4.46 (1 H, d, J 3, =CH2),
3.87–3.81 (1 H, m, CH), 3.63 (3 H, s, Me), 2.81–2.67 (2 H, m,
CH2), 1.80–1.63 (2 H, m, CH2), 1.15 (3 H, d, J 6, Me), 0.95 (9 H, t,
J 8, Si(CH2Me)3) and 0.58 (6 H, q, J 8, Si(CH2Me)3); dC(100 MHz)
197.7 (C(3), quat.), 158.5 (C(2), quat.), 90.6 (CH2=), 67.4 (CH),
55.2 (OMe), 34.0 (CH2), 33.2 (CH2), 23.7 (Me), 6.8 (Si(CH2Me)3)
and 4.9 (Si(CH2Me)3); m/z (APCI) 141 (M–OTES, 100%).
Experimental
General details
All reactions requiring dry or inert conditions were conducted in
flame-dried equipment under an atmosphere of argon. Syringes
and needles were oven-dried and allowed to cool in a desic-
cator over P2O5 before use. Ethers were distilled from sodium-
benzophenone; (chlorinated) hydrocarbons, amines, MeOH and
DMF from CaH2 and TMSCl from quinoline; acetone was dried
˚
over 4 A molecular sieves for 30 min. Reactions were monitored
by TLC using commercially available Merck silica gel 60 F254
glass-backed plates. Visualisation of reaction components was
achieved with a UV lamp (254 nm) and with vanillin or KMnO4
stains. Column chromatography was carried out on silica gel
60 (particle size 40–63 mm) as supplied by Merck. Preparative
TLC was performed using 2 mm pre-coated glass-backed plates
(Merck). Light petroleum refers to the fraction with bp 40–
60 ◦C. IR spectra were recorded as thin films unless stated
otherwise using a Perkin-Elmer 1733 instrument. Peak intensities
Methyl-5-(triethylsilyloxy)-2-oxohexanoate 21. From 20: A
mixture of O3/O2 was bubbled through a stirred solution of silyl
ether 20 (138 mg, 0.51 mmol) and pyridine (0.5 cm3) in CH2Cl2
(50 cm3) at -78 ◦C. After 40 min, the stream of O3/O2 was stopped
and argon was bubbled through for 10 min. Ph3P (146 mg, 0.56
mmol) was then added, the reaction mixture allowed to warm to
are specified as strong (s), medium (m) or weak (w). H and 13C
1
NMR spectra were recorded in CDCl3 unless stated otherwise on
Varian Gemini 200, Bru¨ker DPX400, Bru¨ker DPX330 or Bru¨ker
AMX330 instruments. Chemical shifts are reported relative to
the internal solvent [e.g. dH 7.27, dC (central line of t) 77.0 for
CDCl3]. Coupling constants (J) are given in Hz to the nearest
0.5 Hz. Mass spectra were obtained from the EPSRC National
Mass Spectrometry Service Centre, Swansea with a VG Micromass
Zab-E instrument under EI or CI (NH3) conditions, or at Oxford
with a Micromass Platform APCI spectrometer. Organolithiums
were titrated before use.40
◦
25 C and then evaporated under reduced pressure. Purification
of the residue by column chromatography (gradient elution: 5%
Et2O in light petroleum to 10% Et2O in light petroleum) gave a
colourless oil, a-ketoester 21 (80 mg, 57%); Rf 0.44 (20% Et2O in
light petroleum); nmax/cm-1 2956 s, 2913 m, 2878 s, 1732 s, 1458 w,
1378 m, 1062 m, 1005 s and 743s; dH(400 MHz) 3.95–3.80 (1 H,
m, CH), 3.85 (3 H, s, OMe), 2.98–2.71 (2 H, m, CH2), 1.88–1.65
(2 H, m, CH2), 1.14 (3 H, d, J 6, Me), 0.93 (9 H, t, J 8, Si(CH2Me)3)
and 0.56 (6 H, q, J 8, Si(CH2Me)3); dC(100 MHz) 194.1 (C(2),
quat.), 161.4 (C(1), quat.), 67.0 (CH), 52.8 (OMe), 35.3 (CH2),
32.6 (CH2), 23.5 (Me), 6.8 (Si(CH2Me)3) and 4.8 (Si(CH2Me)3);
6-Hydroxy-2-methoxyhept-1-en-3-one 10. t-BuLi (1.7 mol
dm-3 in pentane; 100 cm3, 170 mmol) was added to a stirred
solution of methyl vinyl ether (22.0 g, 380 mmol) in THF (100 cm3)
at -78 ◦C. The resulting yellow solution was allowed to warm to
+
m/z (CI, NH3) 292 (M + NH4 , 20%), 275 (100), 245 (10), 217
(15), 160 (20) and 143 (30) (Found M + H+: 275.1678. C13H27O5Si
0
◦C and stirred at this temperature for 10 min. The solution
requires M, 275.1681).
◦
was then cooled to -78 C and a solution of g-valerolactone (9)
(13 cm3, 137 mmol) in THF (20 cm3) was added. After 16 h at
-78 ◦C, saturated aq. NH4Cl (10 cm3) was added and the reaction
mixture was allowed to warm to 25 ◦C, then diluted with EtOAc
(15 cm3) and the organic layer separated. The aqueous layer was
extracted with EtOAc (3 ¥ 15 cm3) and the combined organic
layers were washed with brine (10 cm3), water (10 cm3), dried
(MgSO4) and evaporated under reduced pressure to give a yellow
oil, crude alcohol 10 (20.0 g, 92%) which was used directly in the
next step; Rf 0.23 (60% Et2O in light petroleum); nmax/cm-1 3400 s,
2968 s, 1698 s, 1613 s, 1454 m, 1377 m, 1299 m, 1058 s and 852 m;
dH(200 MHz) 5.23 (1 H, d, J 3, H of =CH2), 4.48 (1 H, d, J 3,
Methyl 2-diazo-3-hydroxy-3-(methoxycarbonyl)-6-triethylsily-
loxy heptanoate 25. A solution of LDA (3.88 mmol) [prepared
from the addition of n-BuLi (1.9 mol dm-3 in hexanes; 2.0 cm3, 3.88
mmol) to a solution of◦diisopropylamine (0.55 cm3, 3.88 mmol) in
THF (30 cm3) at -78 C] was slowly added to a stirred solution
of a-ketoester 21 (626 mg, 2.28 mmol) and methyl diazoacetate15
(92% w/w in CH2Cl2, 421 mg, 3.88 mmol) in THF (10 cm3) at
-78 ◦C. After 30 min at -78 ◦C, saturated aq. NH4Cl (10 cm3) was
added, the mixture allowed to warm to 25 ◦C and then diluted with
Et2O (5 cm3). The layers were separated and the aqueous layer was
extracted with Et2O (3 ¥ 10 cm3). The combined organic layers
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3975–3984 | 3981
©

View Article Online
◦
were washed with brine (15 cm3), water (15 cm3), dried (MgSO4)
and evaporated under reduced pressure. The residue was purified
by column chromatography (20% Et2O in light petroleum). First
to elute was recovered ketone 21 (40 mg, 6%). Second to elute
was a yellow oil, a-diazoester 25 [530 mg, 62% (69% based on
recovered 21); 1 : 1 mixture of diastereomers, by 1H NMR analysis
of the hydroxyl signals)]; Rf 0.50 (50% Et2O in light petroleum);
THF (1 cm3) was stirred at 25 C for 1 h. The reaction mixture
was then evaporated, the residue redissolved in CH2Cl2 (5 cm3)
and the solution washed with saturated aq. NaHCO3 (2 cm3),
water (2 cm3), dried (MgSO4) and evaporated under reduced
pressure. Purification of the residue by column chromatography
(gradient elution: 40% Et2O in light petroleum to 50% Et2O in
light petroleum) gave a yellow oil, alcohol 27 (210 mg, 76%; 1 : 1
1
n
max/cm-1 3490 m, 2956 s, 2913 m, 2877 s, 2098 s, 1752 s, 1703 s,
mixture of diastereomers, by H NMR of the SiMe groups); Rf
1439 m, 1321 s, 1244 m, 1141 m, 1058 m and 744 s; dH(400 MHz)
4.31 (1 H, br, OH), 4.27 (1 H, br, OH), 3.90–3.70 (2 H, m, 2 ¥
CH), 3.80 (6 H, s, 2 ¥ OMe), 3.76 (6 H, s, 2 ¥ OMe), 2.10–1.60
(6 H, m, 6 ¥ H of CH2), 1.40–1.20 (2 H, m, H of CH2), 1.14
(6 H, d, J 6, 2 ¥ Me), 0.95 (9 H, t, J 8, Si(CH2Me)3), 0.94 (9
H, t, J 8 Si(CH2Me)3), 0.59 (6 H, q, J 8, Si(CH2Me)3) and 0.57
(6 H, q, J 8, Si(CH2Me)3); dC(100 MHz) 173.5 (C, quat.), 173.4
(C, quat.), 166.1 (C, quat.), 166.1 (C, quat.), 73.8 (C, quat.), 73.7
(C, quat.), 68.1 (CH), 67.5 (CH), 53.4 (OMe), 53.4 (OMe), 52.0
(2 ¥ OMe), 32.8 (CH2), 32.7 (CH2), 32.6 (CH2), 31.8 (CH2), 23.9
(Me), 23.5 (Me), 6.8 (2 ¥ Si(CH2Me)3), 4.9 (Si(CH2Me)3), and 4.8
(Si(CH2Me)3); m/z (APCI) 397 (M + Na+, 10%), 315 (5), 287 (5),
215 (55), 183 (100) and 155 (50).
0.31 (50% Et2O in light petroleum); nmax/cm-1 3417 w, 2956 m,
2931 m, 2858 m, 2098 s, 1747 m, 1704 s, 1438 m, 1317 s, 1142 s
and 838 m; dH(400 MHz) 3.86–3.76 (2 H, m, 2 ¥ CH), 3.75 (6
H, s, 2 ¥ OMe), 3.73 (6 H, s, 2 ¥ OMe), 2.18–2.06 (2 H, m, CH2),
2.01–1.88 (2 H, m, CH2), 1.75 (1 H, br, OH), 1.63–1.35 (5 H, m,
2 ¥ CH2 and OH), 1.19 (6 H, d, J 6, 2 ¥ Me), 0.87 (18 H, s, 2 ¥
SiCMe3), 0.12 (3 H, s, SiMe), 0.11 (3 H, s, SiMe), 0.06 (3 H, s,
SiMe) and 0.05 (3 H, s, SiMe); dC(100 MHz) 171.6 (C, quat.),
171.5 (C, quat.), 165.1 (2 ¥ C, quat.), 76.0 (C, quat.), 76.0 (C,
quat.), 67.8 (CH), 67.7 (CH), 52.7 (2 ¥ OMe), 51.8 (2 ¥ OMe),
34.5 (CH2), 34.4 (CH2), 32.9 (2 ¥ CH2), 25.7 (2 ¥ SiCMe3), 23.6
(2 ¥ Me), 18.4 (2 ¥ C, quat.), -3.6 (2 ¥ SiMe) and -4.1 (2 ¥ SiMe);
m/z (ES) 397 (M + Na+, 10%), 369 (15), 347 (5), 287 (10) and 215
(55) (Found M + Na+: 397.1770. C16H30N2NaO6Si requires M,
397.1771).
Methyl
3-tert-butyldimethylsilyloxy-2-diazo-6-(triethylsily-
loxy)-3-(methoxycarbonyl)heptanoate 26. TBSOTf (5.2 cm3,
22.6 mmol) was added to a stirred solution of 2,6-lutidine (3.7 cm3,
33.5 mmol) in CH2Cl2 (10 cm3) at 0 ◦C. After 30 min, a solution of
a-diazoester 25 (4.22 g, 11.3 mmol) in CH2Cl2 (5 cm3) was added
Methyl 3-tert-butyldimethylsilyloxy-2-diazo-3-methoxycarbo-
nyl-6-oxoheptanoate 28. PCC (136 mg, 0.63 mmol) was added
to a stirred solution of alcohol 27 (215 mg, 0.57 mmol) and
NaOAc (24 mg, 0.28 mmol) in CH2Cl2 (10 cm3). After 16 h at
25 ◦C, a slurry of silica gel in Et2O (5 cm3) was added, the mixture
was filtered, evaporated under reduced pressure and purified by
column chromatography (40% Et2O in light petroleum) to give a
yellow oil, ketone 28 (191 mg, 90%); Rf 0.50 (50% Et2O in light
petroleum); nmax/cm-1 2955 w, 2858 m, 2099 s, 1749 m, 1713 s,
1438 m, 1317 s, 1253 s, 1137 s, 1032 m and 838 s; dH(400 MHz)
3.74 (3 H, s, OMe), 3.73 (3 H, s, OMe), 2.64–2.56 (1 H, m, H of
CH2), 2.49–2.40 (1 H, m, H of CH2), 2.25–2.21 (2 H, m, 2 ¥ H
of CH2), 2.14 (3 H, s, Me) 0.87 (9 H, s, SiCMe3), 0.10 (3 H, s,
SiMe) and 0.05 (3 H, s, SiMe); dC(100 MHz) 207.0 (C(6), quat.),
171.1 (C, quat.), 165.0 (C, quat.), 75.4 (C, quat.), 52.8 (OMe),
51.9 (OMe), 37.8 (CH2), 32.0 (CH2), 30.0 (Me), 25.7 (SiCMe3),
18.4 (C, quat.), -3.7 (SiMe) and -4.1 (SiMe); m/z (ES) 395 (M +
Na+, 100%), 367 (30), 313 (20), 285 (10) and 213 (20) (Found: M
+ Na+, 395.1614. C16H28NaO6Si requires M, 395.1614).
◦
and the reaction mixture stirred at 25 C. After 54 h, water was
added (10 cm3), the layers separated and the aqueous solution
extracted with CH2Cl2 (3 ¥ 10 cm3). The combined organic layers
were washed with brine (10 cm3), water (10 cm3), dried (MgSO4)
and evaporated under reduced pressure. Purification of the residue
by column chromatography (10% Et2O in light petroleum) gave a
yellow oil, bis(silyl ether) 26 [4.0 g, 72% (77% based on recovered
1
25); 1 : 1 mixture of diastereomers, by H NMR analysis of the
SiMe group)]; Rf 0.64 (20% Et2O in light petroleum); nmax/cm-1
2955 s, 2878 s, 2859 s, 2096 s, 1765, 1711 s, 1437 m, 1317 m, 1250 m,
1141 s, 1059 m and 839 m; dH(400 MHz) 3.80–3.70 (2 H, m, 2 ¥
CH), 3.74 (3 H, s, OMe), 3.74 (3 H, s, OMe), 3.73 (6 H, s, 2 ¥
OMe), 2.09 (1 H, ddd, J 12.5, 12.5 and 4, H of CH2), 1.99 (1 H,
ddd, J 12.5, 12.5 and 4.5, H of CH2), 1.90 (1 H, ddd, J 12.5, 12.5
and 4.5, H of CH2), 1.85 (1 H, ddd, J 12.5, 12.5 and 4.0, H of
CH2), 1.63–1.49 (2 H, m, CH2), 1.43–1.29 (2 H, m, CH2), 1.13
(6 H, d, J 6, 2 ¥ Me), 0.95 (9 H, t, J 8, Si(CH2Me)3), 0.94 (9 H, t,
J 8, Si(CH2Me)3), 0.88 (9 H, s, SiCMe3), 0.88 (9 H, s, SiCMe3), 0.58
(6 H, q, J 8, Si(CH2Me)3), 0.57 (6 H, q, J 8, Si(CH2Me)3), 0.13
(3 H, s, SiMe), 0.11 (3 H, s, SiMe), 0.06 (3 H, s, SiMe) and 0.05 (3
H, s, SiMe); dC(100 MHz) 171.6 (C, quat.), 171.5 (C, quat.), 76.2
(C, quat.), 76.0 (C, quat.), 68.1 (CH), 67.9 (CH), 52.6 (OMe),
52.6 (OMe), 51.8 (OMe), 51.8 (OMe), 34.8 (CH2), 34.7 (CH2),
33.3 (2 ¥ CH2), 25.7 (2 ¥ SiCMe3), 23.9 (Me), 23.8 (Me), 18.4 (C,
quat.), 18.4 (C, quat.), 6.8 (2 ¥ Si(CH2Me)3), 4.9 (Si(CH2Me)3),
4.9 (Si(CH2Me)3), -3.6 (SiMe), -3.6 (SiMe), -4.0 (SiMe) and -4.0
(SiMe); m/z (APCI) 511 (M + Na+, 20%), 461 (20), 429 (100),
329 (90) and 183 (50). Second to elute was recovered alcohol 25
(250 mg, 6%); data as above.
Trimethyl (1R*,2S*,5R*,7R*) 2-tert-butyldimethylsilyloxy-5-
methyl-6,8-dioxabicyclo[3.2.1]octane-1,2,7-tricarboxylate
30.
Rh2(OAc)4 (cat.) was added to a stirred solution of diazoester
28 (189 mg, 0.50 mmol) in toluene (1 cm3) and freshly distilled
methyl glyoxylate41 (187 mg, 2.13 mmol) at 110 ◦C. After 1 h,
the reaction mixture was cooled, diluted with Et2O (3 cm3),
R
filtered through Celiteꢀ and evaporated under reduced pressure.
1H NMR analysis of the residue suggested 3 cycloadduct isomers
in a ratio 12 : 2 : 1, as indicated by singlets assigned to the bridge
methines at d 5.57, 4.84 and 4.60, respectively. Purification of the
residue by column chromatography (20% Et2O in light petroleum)
gave a colourless oil, cycloadduct 30 (140 mg, 63.5%, 10 : 0 : 1
with a minor isomeric cycloadduct); Rf 0.38 (50% Et2O in light
petroleum); nmax/cm-1 2954 s, 2894 m, 2857 s, 1753 s, 1437 s, 1387
m, 1261 s, 1164 s, 1117 s, 1097 s 1004 m and 830 s; dH(400 MHz)
Methyl
3-tert-butyldimethylsilyloxy-2-diazo-6-hydroxy-3-
(methoxycarbonyl) heptanoate 27. A solution of silyl ether 26
(361 mg, 0.74 mmol) in acetic acid (2 cm3), water (1 cm3) and
3982 | Org. Biomol. Chem., 2010, 8, 3975–3984
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
5.57 (1 H, s, CH), 3.80 (3 H, s, OMe), 3.70 (3 H, s, OMe), 3.67
(3 H, s, OMe), 2.38 (1 H, ddd, J 14, 13 and 6, Hendo of C(3)H2),
1.93 (1 H, ddd, J 13, 13 and 5.5, H of C(4)H2), 1.77–1.69 (2 H,
m, Hexo of C(3)H2 and H of C(4)H2), 1.67 (3 H, s, Me), 0.86
(9 H, s, SiCMe3), 0.12 (3 H, s, SiMe) and 0.10 (3 H, s, SiMe)
[discernible data for minor isomer, 4.60 (1 H, s, CH) and 3.79 (3
H, s, OMe)]; 1H NMR NOE experiment: irradiation at d 5.57 saw
enhancement at 2.38 (4.6%); dC(100 MHz) 173.1 (C(2)CO, quat.),
169.5 (C, quat.), 166.6 (C, quat.), 110.2 (C(5), quat.), 90.4 (C(1),
quat.), 77.4 (CH), 77.1 (C(2), quat.), 52.6 (OMe), 52.4 (OMe),
52.4 (OMe), 30.7 (C(4)H2), 29.9 (C(3)H2), 25.7 (SiCMe3), 23.9
(Me), 18.9 (C, quat.), -2.8 (SiMe) and -3.1 (SiMe) [discernible
data for minor isomer, 79.6 (CH), 31.3 (CH2) and 29.5 (CH2)];
ddd, J 14, 13 and 6, H of CH2), 2.04 (1 H, ddd, J 14, 13 and 5.5, H
of CH2), 1.90 (1 H, ddd, J 14, 6 and 1, H of CH2), 1.78 (1 H, ddd,
J 14, 5.5 and 1, H of CH2), and 1.61 (3 H, s, Me); 1H NMR NOE
experiment: irradiation at d 4.83 saw no enhancements in 2.7–
1.7 region; dC(100 MHz) 172.3 (C, quat.), 167.4 (C, quat.), 167.2
(C, quat.), 110.7 (C(5), quat.), 86.8 (C(1), quat.), 79.2 (CH), 73.9
(C(2), quat.), 53.4 (OMe), 53.1 (OMe), 52.4 (OMe), 31.4 (CH2),
28.0 (CH2) and 23.8 (Me); m/z (CI, NH3) 319 (M + H+, 100%).
Trimethyl (1R*,3R*,4R*,5S*) 4-hydroxy-1-methyl-2,8-dioxa-
bicyclo[3.2.1]octane-3,4,5-tricarboxylate 33 and (5R*,8R*,9S*)
9-Hydroxy-2-methoxy-2-methyl-8,9-bis(methoxycarbonyl)-1,7-
+
m/z (CI, NH3) 450 (M + NH4 , 100%) and 433 (70) (Found: M +
dioxaspiro[4,4]nonan-6-one
34. Method
A:
6,8-Dioxa-
H+, 433.1892. C19H33O9Si requires M, 433.1894).
bicyclo[3.2.1]octan-2-ol 31 (20 mg, 0.06 mmol) was stirred
in 2% HCl in MeOH (1 cm3) at reflux. After 15 h, the reaction
mixture was concentrated in vacuum and the residue redissolved
in CH2Cl2 (5 cm3). NaHCO3 was then added, the mixture was
filtered and evaporated under reduced pressure. The residue
was purified by column chromatography (50% Et2O in light
petroleum). First to elute was spirolactones 34 (1 mg, 5%, 1 : 1.5
Trimethyl (1R*,2S*,5R*,7R*) 2-hydroxy-5-methyl-6,8-dioxa-
bicyclo[3.2.1]octane-1,2,7-tricarboxylate 31 and trimethyl (1R*,
2S*,5R*,7S*) 2-hydroxy-5-methyl-6,8-dioxabicyclo[3.2.1]octane-
1,2,7-tricarboxylate 32. From 29: TBAF (1.0 mol dm-3 in THF;
1.3 cm3, 1.3 mmol) was added to a solution of cycloadduct 29
(170 mg, 0.43 mmol) in THF (5 cm3). After 1 h at 25 ◦C, the
mixture was diluted with Et2O (3 cm3), water (2 cm3) added
and the layers separated. The aqueous layer was extracted with
Et2O (3 ¥ 5 cm3) and the combined organic layers were washed
with brine (2 cm3), water (2 cm3), dried (MgSO4) and evaporated
under reduced pressure. Purification of the residue by column
chromatography (gradient elution: 50% Et2O in petroleum ether
to 100% Et2O) gave a colourless oil, alcohol 31 (20 mg, 15%); Rf
0.24 (100% Et2O); nmax/cm-1 3436 br, 2957 w, 1747 s, 1640 m, 1298
m, 1207 m and 1059 m; dH(500 MHz) 5.47 (1 H, s, CH), 3.92 (3
H, s, OMe), 3.75 (3 H, s, OMe), 3.73 (1 H, s, OH), 3.71 (3 H, s,
OMe), 2.32 (1 H, ddd, J 14, 12.5 and 6, Hendo of C(3)H2), 1.99 (1
H, ddd, J 14, 13 and 5.5, H of C(4)CH2), 1.80–1.74 (2 H, m, Hexo
1
mixture of diastereomers, by H NMR analysis of the methoxy
signals); Rf 0.50 (100% Et2O in light petroleum); dH(500 MHz)
major isomer: 5.08 (1 H, s, CH), 3.92 (3 H, s, OMe), 3.83 (3 H, s,
OMe), 3.82 (1 H, s, OH), 3.16 (3 H, s, OMe), 2.66 (1 H, ddd, J
13, 11 and 8, H of CH2), 2.52–2.05 (3 H, m, 3 ¥ H of CH2 and
1.38 (3 H, s, Me); dH(500 MHz) minor isomer: 5.13 (1 H, s, CH),
3.91 (3 H, s, OMe), 3.84 (3 H, s, OMe), 3.84 (1 H, s, OH), 3.32
(3 H, s, OMe), 2.28–1.76 (4 H, m, 4 ¥ H of CH2) and 1.44 (3
H, s, Me); dC(125 MHz) 170.5 (C, quat.), 170.1 (C, quat.), 166.2
(C, quat.), 166.0 (C, quat.), 163.5 (2 ¥ C, quat.), 110.9 (C(7),
quat.) 110.6 (C(7), quat.), 53.9 (2 ¥ OMe), 52.7 (2 ¥ OMe), 49.7
(2 ¥ OMe), 37.7 (CH2), 37.5 (CH2), 29.8 (CH2), 29.7 (CH2), 21.6
+
(Me) and 21.4 (Me); m/z (CI, NH3) 336 (M + NH4 , 15%), 304
1
+
of C(3)H2 and H of C(4)CH2) and 1.70 (3 H, s, Me); H NMR
(100) and 287 (85) (Found: M + NH4 , 336.1289. C13H22NO9
NOE experiment: irradiation at d 5.47 saw enhancement at 2.32
(6.8%); dC(125 MHz) 173.8 (C(2)CO, quat.), 169.2 (C, quat.),
166.7 (C, quat.), 111.0 (C(5), quat.), 89.4 (C(1), quat.), 77.6 (CH),
73.7 (C(2), quat.), 53.7 (OMe), 53.0 (OMe), 52.6 (OMe), 30.8
(C(4)H2), 29.6 (C(3)H2) and 23.6 (Me); m/z (CI, NH3) 336 (M
+ NH4 , 35%), 319 (5), 176 (100), 164 (5), 147 (30), 132 (55) and
77 (40) (Found M + H+: 319.1019. C13H19O9 requires M,
319.1029).
From 30: TBAF (1.0 mol dm-3 in THF; 110 ml, 0.11 mmol)
was added to a solution of cycloadduct 30 (45 mg, 0.10 mmol,
10 : 1 with a minor isomeric cycloadduct) in THF (1 cm3). After
15 min at 25 ◦C, the mixture was diluted with Et2O (1 cm3), water
(1 cm3) added and the layers separated. The aqueous solution was
extracted with Et2O (3 ¥ 1 cm3) and the combined organic layers
were washed with brine (1 cm3), water (1 cm3), dried (MgSO4) and
evaporated under reduced pressure. Purification of the residue by
column chromatography (gradient elution: 50% Et2O in petroleum
ether to 100% Et2O) gave a colourless oil, alcohol 31 (23 mg, 72%);
data as above. From a larger scale desilylation [using cycloadduct
30 (638 mg, 1.47 mmol) in THF (10 cm3) and TBAF (1.0 mol dm-3
in THF; 1.95 cm3, 1.95 mmol)], was obtained alcohol 31 (304 mg,
65%) and endo-ester-alcohol 32 (26 mg, 5.5%); Rf 0.32 (100%
Et2O); dH(400 MHz) 4.83 (1 H, s, CH), 4.15 (1 H, br, OH), 3.88
(3 H, s, OMe), 3.81 (3 H, s, OMe), 3.78 (3 H, s, OMe), 2.63 (1 H,
requires M, 336.1295). Second to elute was recovered alcohol,
6,8-dioxabicyclo[3.2.1]octan-2-ol 31 (8 mg, 40%); data as above.
Third to elute was a white solid, 2,8-dioxabicyclo[3.2.1]octan-4-ol
33 (2.6 mg, 13%); (Found: C, 49.00; H, 5.78. C13H18O9 requires
C, 49.00; H, 5.70%); Rf 0.14 (100% Et2O in light petroleum); mp
118–123 ◦C; nmax/cm-1 3425w, 2958w, 1732 s, 1436 m, 1221 s, 1123
m, 1063 m and 867m; dH(500 MHz) 4.92 (1 H, s, CH), 3.90 (3 H, s,
OMe), 3.79 (3 H, s, OMe), 3.77 (3 H, s, OMe), 3.74 (1 H, s, OH),
3.15 (1 H, ddd, J 9.5, 9.5 and 1.5, Hendo of C(6)H2), 2.25 (1 H,
ddd, J 9.5, 9.5 and 2, Hendo of C(7)H2), 2.18–2.05 (2 H, m, Hexo
of C(6)H2 and Hexo of C(7)H2) and 1.75 (3 H, s, Me); 1H NMR
NOE experiment: irradiation at d 4.92 saw enhancements at 3.15
(5.8%) and at 2.25 (6.9%); dC(125 MHz) 169.9 (C(4)CO, quat.),
168.7 (C(5)CO, quat.), 167.4 (C(3)CO, quat.), 107.6 (C(1), quat.),
88.3 (C(5), quat.), 74.9 (CH), 74.9 (C(4), quat), 53.3 (OMe), 52.9
(OMe), 52.7 (OMe), 32.8 (C(7)H2), 29.4 (C(6)H2) and 23.6 (Me);
+
+
m/z (CI, NH3) 336 (M + NH4 , 100%) and 319 (10) (Found: M +
H+, 319.1032. C13H19O8 requires M, 319.1029).
Method B: A solution 6,8-dioxabicyclo[3.2.1]octan-2-ol 31
(22 mg, 0.07 mmol) in a mixture of CH2Cl2 (500 ml), TFA (250 ml)
and H2O (25 ml) was stirred at 40 ◦C for 68 h. The reaction
mixture was then evaporated under reduced pressure and the
residue purified by preparative TLC (100% Et2O). First to elute
was recovered alcohol 31 (7.5 mg, 34%); data as above. Second to
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3975–3984 | 3983
©

View Article Online
elute was a white solid, 2,8-dioxabicyclo[3.2.1]octane 33 (12 mg,
54%; 83% based on recovered 31); data as above.
10 D. M. Hodgson, J. M. Bailey and T. Harrison, Tetrahedron Lett., 1996,
37, 4623–4626.
11 D. M. Hodgson, J. M. Bailey, C. Villalonga-Barber, M. G. B. Drew and
T. Harrison, J. Chem. Soc., Perkin Trans. 1, 2000, 3432–3443.
12 O. Kataoka, S. Kitagaki, N. Watanabe, J. Kobayashi, S. Nakamura, M.
Shiro and S. Hashimoto, Tetrahedron Lett., 1998, 39, 2371–2374.
13 S. Jiang, A. D. Rycroft, X.-Z. Wang and Y.-L. Wu, Tetrahedron Lett.,
1998, 39, 3809–3812.
Acknowledgements
We thank the EPSRC National Mass Spectrometry Service Centre
for mass spectra.
14 E. M. Carreira and J. Dubois, J. Am. Chem. Soc., 1994, 116, 10825–
10826.
15 Methyl diazoacetate was obtained according to a procedure for ethyl
diazoacetate: N. E. Searle, Org. Synth., 1963, Coll. Vol. IV, 424–426.
16 T. G. Greene and P. G. M. Wuts, Protecting Groups in Organic Synthesis,
4th edn, Wiley, New York, 2006.
References
1 World Health Organization Fact Sheet No 317, February 2007,
http://www.who.int/mediacentre/factsheets/fs2317/en/index.html.
2 R. Do, R. S. Kiss, D. Gaudet and J. C. Engert, Clin. Genet., 2009, 75,
19–29.
3 M. J. Dawson, J. E. Farthing, P. S. Marshall, R. F. Middleton, M. J.
O’Neil, A. Shuttleworth, C. Stylli, R. M. Tait, P. M. Taylor, H. G.
Wildman, A. D. Buss, D. Langley and M. V. Hayes, J. Antibiot., 1992,
45, 639–647.
17 A. Azzouzi, M. Dufour, J.-C. Gramain and R. Remuson, Heterocycles,
1988, 27, 133–148.
18 M. R. Angelastro, N. P. Peet and P. Bey, J. Org. Chem., 1989, 54,
3913–3916.
19 H. H. Wasserman and W. B. Ho, J. Org. Chem., 1994, 59, 4364–4366.
20 J. W. Labadie, D. Tueting and J. K. Stille, J. Org. Chem., 1983, 48,
4634–4642.
4 P. J. Sidebottom, R. M. Highcock, S. J. Lane, P. A. Procopiou and N. S.
Watson, J. Antibiot., 1992, 45, 648–658.
5 K. Hasumi, K. Tachikawa, K. Sakai, S. Murakawa, N. Yoshikawa, S.
Kumazawa and A. Endo, J. Antibiot., 1993, 46, 689–691.
21 G. Bhalay, S. Clough, L. McLaren, A. Sutherland and C. L. Willis,
J. Chem. Soc., Perkin Trans. 1, 2000, 901–910.
22 K. C. Nicolaou, E. W. Yue, S. LaGreca, A. Nadin, Z. Yang, J. E.
Leresche, T. Tsuri, Y. Naniwa and F. Dericcardis, Chem.–Eur. J., 1995,
1, 467–494.
23 V. K. Aggarwal, M. F. Wang and A. Zaparucha, J. Chem. Soc., Chem.
Commun., 1994, 87–88.
6 For a review, see: (a) A. Armstrong and T. J. Blench, Tetrahedron, 2002,
58, 9321–9349. For subsequent studies, see: (b) S. Naito, M. Escobar,
P. R. Kym, S. Liras and S. F. Martin, J. Org. Chem., 2002, 67, 4200–4208;
(c) M. A. Calter, C. Zhu and R. J. Lachicotte, Org. Lett., 2002, 4, 209–
212; (d) A. Bierstedt, J. Roels, J. L. Zhang, Y. Z. Wang, R. Frohlich and
P. Metz, Tetrahedron Lett., 2003, 44, 7867–7870; (e) A. Armstrong, P. A.
Barsanti, T. J. Blench and R. Ogilvie, Tetrahedron, 2003, 59, 367–375;
(f) P. Perlmutter, W. Selajerern and F. Vounatsos, Org. Biomol. Chem.,
2004, 2, 2220–2228; (g) S. Nakamura, H. Sato, Y. Hirata, N. Watanabe
and S. Hashimoto, Tetrahedron, 2005, 61, 11078–11106; (h) Y. Hirata,
S. Nakamura, N. Watanabe, O. Kataoka, T. Kurosaki, M. Anada,
S. Kitagaki, M. Shiro and S. Hashimoto, Chem.–Eur. J., 2006, 12,
8898–8925; (i) J. O. Bunte, A. N. Cuzzupe, A. M. Daly and M. A.
Rizzacasa, Angew. Chem., Int. Ed., 2006, 45, 6376–6380; (j) M. Tsubuki,
H. Okita and T. Honda, Heterocycles, 2006, 67, 731–748; (k) Y. Wang,
S. Gang, A. Bierstedt, M. Gruner, R. Frohlich and P. Metz, Adv. Synth.
Catal., 2007, 349, 2361–2367; (l) D. A. Nicewicz, A. D. Satterfield, D. C.
Schmitt and J. S. Johnson, J. Am. Chem. Soc., 2008, 130, 17281–17283;
(m) N. Shimada, S. Nakamura, M. Anada, M. Shiro and S. Hashimoto,
Chem. Lett., 2009, 38, 488–499; (n) Y. Wang, O. Kataeva and P. Metz,
Adv. Synth. Catal., 2009, 351, 2075–2080.
7 (a) A. Padwa and M. D. Weingarten, Chem. Rev., 1996, 96, 223–
269; (b) G. Mehta and S. Muthusamy, Tetrahedron, 2002, 58, 9477–
9504; (c) D. A. Selden and D. M. Hodgson, in Comprehensive Organic
Functional Group Transformations II, ed. K. Jones, Elsevier, Oxford,
2004, vol. 3, pp 309–353; (d) Z. Zhang and J. Wang, Tetrahedron, 2008,
64, 6577–6605.
8 Preliminary communication: D. M. Hodgson and C. Villalonga-Barber,
Tetrahedron Lett., 2000, 41, 5597–5600.
9 W. M. Blows, G. Foster, S. J. Lanes, D. Noble, J. E. Piercey, P. J.
Sidebottom and G. Webb, J. Antibiot., 1994, 47, 740–754.
24 A. P. Dominey and J. M. Goodman, Org. Lett., 1999, 1, 473–475.
25 D. A. Evans, J. C. Barrow, J. L. Leighton, A. J. Robichaud and M.
Sefkow, J. Am. Chem. Soc., 1994, 116, 12111–12112.
26 N. L. Allinger, J. Am. Chem. Soc., 1977, 99, 8127–8134.
27 W. L. Jorgensen, D. S. Maxwell and J. TiradoRives, J. Am. Chem. Soc.,
1996, 118, 11225–11236.
28 BatchMin V9.6, Schrodinger LLC, New York, 2007.
29 Jaguar 7.0, Schrodinger LLC, New York, 2007.
30 A. D. Becke, J. Chem. Phys., 1993, 98, 5648–5652.
31 C. T. Lee, W. T. Yang and R. G. Parr, Phys. Rev. B: Condens. Matter,
1988, 37, 785–789.
32 C. Adamo and V. Barone, J. Chem. Phys., 1998, 108, 664–675.
33 A. Frisch and M. J. Frisch, Gaussian 98 Users Reference, Gaussian Inc.,
Pittsburgh, PA, 1998.
34 Y. Zhao, N. E. Schultz and D. G. Truhlar, J. Chem. Phys., 2005, 123.
35 Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2008, 120, 215–241.
36 (a) B. Marten, K. Kim, C. Cortis, R. A. Friesner, R. B. Murphy, M. N.
Ringnalda, D. Sitkoff and B. Honig, J. Phys. Chem., 1996, 100, 11775–
11788; (b) J. Pipek and P. G. Mezey, J. Chem. Phys., 1989, 90, 4916–
4926.
37 (a) F. London, J. Phys. Radium, 1937, 8, 397–409; (b) R. Ditchfield,
J. Chem. Phys., 1972, 56, 5688–5691; (c) K. Wolinski, J. F. Hinton and
P. Pulay, J. Am. Chem. Soc., 1990, 112, 8251–8260.
38 S. G. Smith, R. S. Paton, J. W. Burton and J. M. Goodman, J. Org.
Chem., 2008, 73, 4053–4062.
39 S. G. Smith and J. M. Goodman, J. Org. Chem., 2009, 74, 4597–4607.
40 J. Suffert, J. Org. Chem., 1989, 54, 509–510.
41 J. M. Hook, Synth. Commun., 1984, 14, 83–87.
3984 | Org. Biomol. Chem., 2010, 8, 3975–3984
This journal is
The Royal Society of Chemistry 2010
©