Furanose Synthesis via Regioselective Dihydroxylation of 1-Silyloxy-1,3-dienes: Application to the Furanose Unit of 4-epi-Hygromycin A

Article abstract of DOI:10.1055/s-2003-44979

Asymmetric dihydroxylation of a 1-silyloxy-1,3-diene is shown to occur regioselectively on the non-oxygenated alkene. Second dihydroxylation is complicated by over-oxidation to the dione and epimerisation of reaction products. However, subsequent manipula

Full text of DOI:10.1055/s-2003-44979

350
LETTER
Furanose Synthesis via Regioselective Dihydroxylation of 1-Silyloxy-1,3-
dienes: Application to the Furanose Unit of 4-epi-Hygromycin A
F
uranose Synthesi
l
s
via Re
a
gioselectiv
n
e
D
ihydroxylation
A
of 1-Silyloxy-1,3-
r
dienes mstrong,*a David M. Gethin,b Christopher J. Wheelhousea
a
Department of Chemistry, Imperial College London, South Kensington, London SW7 2AZ, UK
E-mail: A.Armstrong@imperial.ac.uk
b
Veterinary Medicine Research and Development, Pfizer Animal Health, Sandwich, Kent CT13 9NJ, UK
Received 7 October 2003
mnemonic1 to require opposite chiral ligand series in the
two dihydroxylation steps – as well as selective oxidation
of the terminal, primary hydroxyl group in 4. In this paper
we report successful regioselective dihydroxylation of a
Abstract: Asymmetric dihydroxylation of a 1-silyloxy-1,3-diene is
shown to occur regioselectively on the non-oxygenated alkene. Sec-
ond dihydroxylation is complicated by over-oxidation to the dione
and epimerisation of reaction products. However, subsequent ma-
nipulations including selective oxidation of a primary alcohol allow 1-silyloxy-1,3-diene and synthetic manipulations leading
synthesis of a protected furanose corresponding to that present in
C4-epi-hygromycin.
to 16, corresponding to the C4-epimer of 3.
Key words: alkenes, dihydroxylation, osmium, regioselective, oxi-
dations
HO
HO
OH
OH
a, b
BnO
BnO
OBn
OBn
OBn
OBn
HO
HO
The use of chiral-pool materials is a well established
approach to the synthesis of complex polyhydroxylated
natural products. Whilst this strategy conveniently pro-
vides enantiomerically pure compounds, the excision of
unwanted functionality can lead to lengthy synthetic
routes. Additionally, the unnatural enantiomeric series
can be difficult to access. In our total synthesis of (+)-
zaragozic acid C, we utilised an alternative approach
which involved preparation of a stereodefined 1,3-diene
unit 1 containing the required carbon framework, fol-
lowed by double asymmetric dihydroxylation (AD)1 to in-
troduce the hydroxyl functionality in 2 with a high degree
of stereocontrol (Scheme 1).2 The first dihydroxylation
step gave a mixture of regioisomers, which was inconse-
quential in the context of this total synthesis but would
preclude the use of a different functionalisation reaction
(such as the enantiomeric chiral ligand series) in the sec-
ond step. In order to make this synthetic strategy more
general and flexible, we sought to discover dienes where
initial dihydroxylation could be achieved regioselective-
ly. 1,3-Dienes with an ester substituent in the 1-position
have been reported by Sharpless to undergo AD reaction
at the more electron rich 3,4-alkene.3 We wondered
whether 1-oxygenated-1,3-dienes would also undergo
regioselective AD, allowing a flexible route to substituted
polyol derivatives. A particular target was the furanose
unit 3 of the important antibiotic hygromycin,4 previous
syntheses of which have required multi-step routes from
carbohydrate precursors.5 Our retrosynthetic approach to
this furanose 3 is outlined in Scheme 2, where the key
requirements are the regioselective dihydroxylation of
diene 5 – predicted on the basis of the Sharpless AD
1
2
Scheme 1 Reagents and conditions: (a) AD-mix b, tert-BuOH/
H2O, 78%; (b) (DHQD)2PHAL, OsO4, acetone/H2O, 58%, >80% de,
76% ee.
OR2
AD-mix α
O
O
O
OH
OR1
1
OR1
4
OR1
HO
OH
HO
OH
AD-mix β
3
4
5
Scheme 2
The first requirement was a synthetic route to the substi-
tuted silyloxy diene 8. This was accomplished as shown in
Scheme 3. Deprotonation of propargyl alcohol benzyl
ether 6 and reaction with formaldehyde was followed by
stereoselective reduction to the E-allylic alcohol with RE-
DAL,6 conversion to the bromide, displacement with 2-
lithio-2-methyl-1,3-dithiane and subsequent removal of
the 1,3-dithiane unit to afford 7. Thermodynamic silyl
enol ether formation was then readily accomplished, pro-
viding the desired 1,3-diene 8 as a 2.5:1 Z:E mixture. We
were now able to test the dihydroxylation of diene 8. In-
terestingly, when 8 was subjected to AD conditions,7 di-
hydroxylation proceeded with apparently complete
regiocontrol, with reaction taking place on the non-sily-
loxy-substituted olefin (Scheme 3). The two silyl enol
ether isomers (Z)-9 and (E)-9 proved separable by flash
chromatography, and the enantiomeric excess of (Z)-9
was shown to be 84% by chiral HPLC.7 The regiochemi-
cal outcome of this reaction was perhaps surprising, since
it is the same as that observed for 1,3-dienes with an elec-
tron withdrawing ester group in the 1-position.3 In order
SYNLETT 2004, No. 2, pp 0350–0352
0
2.
0
2.
2
0
0
4
Advanced online publication: 08.12.2003
DOI: 10.1055/s-2003-44979; Art ID: D24703ST
© Georg Thieme Verlag Stuttgart · New York

LETTER
Furanose Synthesis via Regioselective Dihydroxylation of 1-Silyloxy-1,3-dienes
351
to rule out the influence of alkene substitution pattern, we
also tested 1-tert-butyldimethylsilyloxy-1,3-pentadiene,
where both alkenes are disubstituted. Interestingly, analy-
sis of the reaction mixture at low conversion indicated that
this diene also reacted regioselectively on the 3,4-alkene,
suggesting that the differing substitution pattern between
the two alkenes in 8 is not responsible for the AD-regio-
selectivity.
O
O
a, b
(Z/E)-9
2.5:1
TBSO
BnO
R2
R1
TBSO
BnO
+
O
OTBS
OTBS
10a R1=H, R2=OH
10b R1=OH, R2=H
11
Scheme 4 Reagents and conditions: (a) TBSCl, imidazole, DMF,
94%; (b) (i) super AD mix-a, tert-BuOH/H2O, r.t., 120 h,: 10a: 34%,
10b: 6%; 11: 13%.
O
TBSO
(a)-(e)
(f)
BnO
BnO
BnO
make 10b the major product. Use of AD-mix b again af-
forded predominantly 10a.
6
7
8
With 10a in hand, we investigated the manipulations
needed for furanose formation (Scheme 5). Protection of
the ketone as a dioxolane and removal of the benzyl group
led to 12. Mosher’s ester analysis of this material indicat-
ed that it had >95% ee. We now required selective oxida-
tion of the primary alcohol in the presence of the
secondary one. We were aware of recent progress in the
selective oxidation of primary alcohols using TEMPO-
based oxidation systems.11a Use of TEMPO with NCS as
co-oxidant11b afforded a mixture of the lactone 14 (45%)
TBSO
HO
OTBS
(g)
HO
+
BnO
BnO
OH
(Z)-9
OH
(E)-9
Scheme 3 Reagents and conditions: (a) n-BuLi, THF, –78 °C, then
(CH2O)n, 69%; (b) REDAL, –20 °C, THF, 79%; (c) NBS, PPh3,
–40 °C, THF, 79%; (d) 2-lithio-2-methyl-1,3-dithiane, DMPU,
–78 °C, THF, 75%; (e) AgNO3, NCS, MeCN/H2O, 78%: (f) KH,
TBSCl, THF, 76%; (g) super AD-mix b, tert-BuOH/H2O, 4 d, 0 °C, and the lactol 13 (36%), which could be oxidised to 14 us-
41% (Z)-9, 21% (E)-9.
ing PCC. Interestingly, the recently reported TEMPO/iso-
cyanuric acid system11c led directly to the lactone 14 in
good yield. This selective primary oxidation therefore al-
lowed realisation of our aim of using 1,3-dienes as precur-
sors to furanoses.
The second dihydroxylation step was now examined.
Sharpless and co-workers have previously reported the ef-
ficient AD of silyl enol ethers.8 Importantly, their studies
indicated that E- and Z-isomers of trisubstituted enol
ethers afforded the same product enantiomer, indicating
that an E/Z-mixture could be employed. Accordingly, the
(E/Z)-9 mixture was taken on through the synthesis. These
diols were protected as their bis-TBS ethers in order to fa-
cilitate product analysis and to allow efficient hydroxyl
differentiation later in the synthesis. AD reaction using
the AD-mix-a ligand series, which was predicted to lead
ultimately to the required C4-stereochemistry, afforded
recovered starting material (10%) along with three prod-
ucts: two diastereomeric a-hydroxy ketones 10a and 10b,
along with the dione 11.9 Dione 11 presumably arises
from conversion of 10 to the ene-diol tautomer and subse-
quent dihydroxylation. Indeed, resubmission of 10a/10b
to the AD conditions led to formation of 11. This side-re-
action has been reported once before, in the AD of silyl
ketene acetals,10 but to the best of our knowledge has not
been reported previously in the AD of silyl enol ethers.
O
O
O
OH
O
TBSO
OTBS
13
c or d
a, b
O
e
10a
+
TBSO
HO
OH
OTBS
O
O
O
O
12
TBSO
OTBS
14
Scheme 5 Reagents and conditions: (a) Ethylene glycol, triethyl
orthoformate, PPTS, 30 h, 79%; (b) H2, Pd/C, MeOH; (c) 5 mol%
TEMPO, 2 equiv isocyanuric acid, CH2Cl2, 79% 14; (d) TEMPO,
NCS, Bu4NCl, CH2Cl2/H2O, pH 8.6, 45% 14, 36% 13; (e) PCC,
Subsequent manipulations suggest that the major a-hy- CH2Cl2, 4Å ms, 4 h, 78%.
droxyketone 10a has the unexpected and undesired con-
figuration a- to the ketone (vide infra). However, control
Since NOE studies on 14 proved inconclusive, we opted
experiments suggest that 10a and 10b undergo epimerisa-
tion and over-oxidation under the AD conditions at differ-
ent rates, which means that the observed ratio of 10a/b
may not reflect the outcome of the initial AD step. Despite
experimentation with the AD-reaction conditions, we
were not able to prevent the formation of dione 11 or to
to assign relative configuration by correlation of lactol 13
with known literature compounds. Thus, coupling of 13
with phenol afforded a 3:1 mixture of anomers with the
major assigned structure 15.12 Deketalisation of the major
anomer gave 16 (Scheme 6), silyl deprotection of which
1
gave a product whose H NMR spectrum did not match
Synlett 2004, No. 2, 350–352 © Thieme Stuttgart · New York

352
A. Armstrong et al.
LETTER
Cox, L. R.; DeBoos, G. A.; Fullbrook, J. J.; Percy, J. M.;
Spencer, N. S.; Tolley, M. Org. Lett. 2003, 5, 337.
(4) Kakinuma, K.; Sakagami, Y. Agric. Biol. Chem. 1978, 42,
279.
(5) (a) Chida, N.; Ohtsuka, M.; Nakazawa, K.; Ogawa, S. J.
Org. Chem. 1991, 56, 2976. (b) Buchanan, J. G.; Hill, D. G.;
Wightman, R. H. Tetrahedron 1995, 51, 6033. (c) For a
recent synthesis of C2-epi-hygromycin A, see: Trost, B. M.;
Dudash, J. J.; Dirat, O. Chem.–Eur. J. 2002, 8, 259.
(6) Sola, L.; Castro, J.; Moyano, A.; Pericas, M.; Riera, A.
Tetrahedron Lett. 1992, 33, 2836.
that of the correct hygromycin C4-epimer reported by
Buchanan.5b NOE measurements on 16 gave firm evi-
dence that it was the C4-epimer. A 14% enhancement of
H4 upon irradiation of H3 was particularly diagnostic.
Assuming that complete epimerisation did not occur in
any of the transformations employed,13 this suggests that
the major product 10a in the second AD step has the
configuration depicted in Scheme 4.
O
(7) ‘Super AD-mix’ refers to commercial AD-mix
supplemented with 1 mol% K2OsO2(OH)4, 5 mol% ligand
and 1 equiv MeSO2NH2. The absolute configuration of 9 is
assigned based upon the Sharpless mnemonic.1 The ee of
(Z)-9 was determined using HPLC (Chiralcel OD, 1 mL per
min, 5% IPA/hexane, detection at 210 nm, retention times:
11.8 min (minor enantiomer), 27.8 min (major enantiomer).
We have been unable to determine the ee of (E)-9.
(8) Hashiyama, T.; Morikawa, K.; Sharpless, K. B. J. Org.
Chem. 1992, 57, 5067.
O
O
a
O
b
O
OPh
OPh
1
4
13
TBSO
OTBS
TBSO
OTBS
15
16
Scheme 6 Reagents and conditions: (a) Phenol (1.1 equiv), PPh3
(1.2 equiv), THF, then DEAD (1.2 equiv), 2.5 h, 98% (3:1 15:C1-epi-
15); (b) TFA, THF, H2O, 79%.
(9) Rf values (20% Et2O-petroleum Et2O): 11, 0.85; 10a, 0.21;
10b, 0.14. Characteristic data for dione 11 include carbonyl
In conclusion, we have demonstrated that 1-silyloxy-1,3-
dienes can be regioselectively dihydroxylated on the 3,4-
alkene. In the second dihydroxylation step, we have noted
an unusual side-reaction, namely over-oxidation to the di-
one. We have shown that the double dihydroxylation
products may be manipulated via selective oxidation of a
primary alcohol to allow the synthesis of furanose deriva-
tives, including compounds related to the C4-epimer of
the furanose contained in the antibiotic hygromycin.
Efforts to refine this strategy by improving the yield and
selectivity of the second alkene functionalisation step are
currently under investigation.
stretches in the IR spectrum at 1716 and 1675 cm–1, and 13
NMR resonances at d = 201 and 200 ppm.
C
(10) Monenschein, H.; Drager, G.; Jung, A.; Kirschning, A.
Chem.–Eur. J. 1999, 5, 2270.
(11) (a) Review: De Nooy, A. E. J.; Besemer, A. C.; van Bekkum,
H. Synthesis 1996, 1153. (b) Einhorn, J.; Einhorn, C.;
Ratajczak, F.; Pierre, J. P. J. Org. Chem. 1996, 61, 7452.
(c) De Luca, L.; Giacomelli, G.; Porcheddu, A. Org. Lett.
2001, 3, 3041. (d) de Luca, L.; Giacomelli, G.; Masala, S.;
Porcheddu, A. J. Org. Chem. 2003, 68, 4999.
(12) H1 in the major anomer appeared as a singlet, while that in
the minor was d, J = 3.6 Hz. Comparison to ref.5a,b and to
various pentafuranoses then suggests the anomeric
configuration indicated. Stevens, J. D.; Fletcher, H. G. J. J.
Org. Chem. 1968, 33, 1799.
Acknowledgement
(13) We have demonstrated that deketalisation of the benzyl ether
of 12 regenerates 10a, suggesting that epimerisation has not
occurred here. We cannot rule out epimerisation at C4 in the
deprotection of 15. However, molecular mechanics (MMFF)
indicate that the 16 is 7.7 kcal/mol less stable than its C4-
epimer, suggesting that epimerisation is unlikely.
We thank the EPSRC and Pfizer Central Research (CASE
studentship to CJW) for their support of this work. We also thank
Pfizer, Merck Sharp and Dohme, and Bristol-Myers Squibb for
generous unrestricted support of our research programmes.
20
Data for 16: white amorphous solid; mp 79–81 °C; [a]D
–37.5 (c 0.1, CH2Cl2). IR (film): nmax = 2949, 2928, 2857,
1718, 1587, 1490, 1254, 1254 cm–1. 1H NMR (250 MHz,
CDCl3) d = 7.35–7.01 (5 H, m, Ph), 5.82 (1 H, d, J = 3.1 Hz,
2-H), 4.72 (1 H, d, J = 4.4 Hz, 5-H), 4.42 (1 H, dd, J = 2.6
and 4.4 Hz, 4-H), 4.16 (1 H, app. t, J = 3.0 Hz, 3-H), 2.28
(3 H, s, 6-CH3), 0.95 [9 H, s, SiC(CH3)3], 0.86 [9 H, s,
SiC(CH3)3], 0.15 [3 H, s, Si(CH3)2], 0.13 [3 H, s,
ROSi(CH3)2], 0.10 [3 H, s, Si(CH3)2], 0.05 [3 H, s,
Si(CH3)2]. 13C NMR (62.5 MHz, CDCl3): d = 209.0 (C),
157.3 (C), 129.5 (CH), 122.2 (CH), 116.3 (CH), 102.2 (CH),
85.7 (CH), 79.2 (CH), 77.9 (CH), 28.6 (CH3), 25.7 (CH3),
25.6 (CH3), 18.2 (C), 17.8 (C), –4.4 (C), –4.7 (C), –5.1 (C),
–5.2 (C). MS (CI): m/z = 484 (100) [M + NH4]+, 390 (5), 201
(31), 132 (24). Anal. Calcd for C24H42Si2O5 [M+NH4]+
requires 484.2915. Found: 484.2919.
References
(1) For an excellent review, see: Kolb, H. C.; VanNieuwenhze,
M. S.; Sharpless, K. B. Chem. Rev. 1994, 94, 2483.
(2) Armstrong, A.; Barsanti, P. A.; Jones, L. H.; Ahmed, G. J.
Org. Chem. 2000, 65, 7020.
(3) (a) Becker, H.; Soler, M. A.; Sharpless, K. B. Tetrahedron
1995, 51, 1345; Other examples of regioselective AD of
ester-substituted dienes include:. (b) Armstrong, R. W.;
Tellew, J. E.; Moran, E. J. Tetrahedron Lett. 1996, 37, 447.
(c) Nicolaou, K. C.; Yue, E. W.; La Greca, S.; Nadin, A.;
Yang, Z.; Leresche, J. E.; Tsuri, T.; Naniwa, Y.; De
Riccardis, F. Chem.–Eur. J. 1995, 1, 467. (d) Bruckner, R.;
Harcken, C. Tetrahedron Lett. 2001, 42, 3967. (e) For a
recent example of synthesis of a xylulose derivative using
regioselective AD of a 2-alkoxy-1,1-difluoro-1,3-diene, see:
Synlett 2004, No. 2, 350–352 © Thieme Stuttgart · New York