Synthesis of (-)-hygromycin a: application of Mitsunobu glycosylation and tethered aminohydroxylation

Full text of DOI:10.1002/anie.200902840

Angewandte
Chemie
DOI: 10.1002/anie.200902840
Total Synthesis
Synthesis of (À)-Hygromycin A: Application of Mitsunobu
Glycosylation and Tethered Aminohydroxylation**
Timothy J. Donohoe,* Aida Flores, Carole J. R. Bataille, and Fꢀtima Churruca
The natural product (À)-hygromycin A, (À)-1, was isolated
from the fermentation of Streptomyces hygroscopicus in 1953
and shown to be a broad-spectrum antibiotic which also
exhibits immunosuppressant activity.[1] The mode of antibac-
terial action of the natural product has been shown to involve
inhibition of the ribosomal peptidyl transferase activity, and
hygromycin A binds to the ribosome in a manner that is
closely related, but not identical, to that of chloroampheni-
col.[2] Recent interest from the pharmaceutical industry has
centered around the possibility of using this compound, or
analogues thereof, in the control of swine dysentery and other
infectious diseases of economic importance in animal
health.[3]
that the sequence to the natural product itself could be
shortened and that the key glycosylation reaction could be
made selective for the desired b anomer.
Our retrosynthesis of this molecule consisted of discon-
nection into three portions, namely the furanose sugar A, the
aromatic linker B, and the aminocyclitol C. The key chal-
lenges were thought to be the formation of the sterically
crowded b-configured sugar linkage to the phenol and the
avoidance of epimerization at C4 of the sugar adjacent to the
sensitive ketone unit.[7] While our earlier studies had culmi-
nated in a short synthesis of cyclitol C, we anticipated that we
could improve this further and generate material suitable for
completion of the synthesis of (À)-1. One other noteworthy
point about our proposed route is the importance of the
protecting groups at C2 and C3 of the furanose sugar A. We
anticipated that installation of a bulky group here (we chose
to use a triisopropylsilyl (TIPS) group) might have two
beneficial effects on the synthesis. Firstly, the large group at
C2, adjacent to the anomeric position, might exhibit a bias on
the diastereoisomeric ratio of the anomeric hydroxy group
that is a precursor to sugar coupling (favoring the a anomer,
see A; Scheme 1). According to the innovative work of Smith
Despite its biological activity and interesting structure,
there has been only one successful synthesis of this product,
by Ogawa et al. in 1991,[4] although note should be made of
the elegant synthesis of C-2-epihygromycin A by Trost et al.
in 2002.[5] Although Ogawaꢀs synthesis was inspiring, several
areas could be improved upon. Most notably the formation of
the cis sugar linkage between the phenol and furanose was
essentially nonselective, and the sequence was lengthy with
low overall yield. Our interest in the molecule was originally
sparked by the aminocyclitol moiety which was a challenging
target on which to test our recently developed tethered
aminohydroxylation (TA) methodology.[6] In addition, we felt
Scheme 1. Retrosynthetic analysis of (À)-hygromycin A. PG=protect-
[*] Prof. T. J. Donohoe, Dr. A. Flores, Dr. C. J. R. Bataille, Dr. F. Churruca
Department of Chemistry, University of Oxford
Chemistry Research Laboratory
ing group.
Mansfield Road, Oxford, OX13TA (UK)
Fax: (+44)1865-275-708
E-mail: timothy.donohoe@chem.ox.ac.uk
et al. and then Roush et al.,[8] and assuming that SN2 displace-
ment is faster than anomerization, then such a bias may well
manifest itself in a b-selective glycosylation reaction with a
phenol acceptor under Mitsunobu-type coupling conditions.
Secondly, it has been noted in previous studies that the
natural product and its precursors are sensitive towards
epimerization at C4 (in Ogawaꢀs synthesis this sensitivity
[**] We thank the Basque Government, the Fundaciꢀn Ramꢀn Areces,
the Leverhulme Trust, and the European Union for supporting this
project.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200902840.
Angew. Chem. Int. Ed. 2009, 48, 6507 –6510
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6507

Communications
required the introduction of a ketal protecting group for the
ketone).[7] We hoped that a bulky protecting group at C3
would hinder the vulnerable proton and thus enable us to
carry the kinetically stable ketone through the synthesis
without protection.
Our synthesis began with commercially available and
inexpensive d-arabinose (drawn in the furanose form for
clarity), which was substituted with an allyl group at the
anomeric position before ensuing tritylation of the primary
hydroxy group and TIPS protection of the two secondary
alcohols (see 2, Scheme 2). Then, selective cleavage of the
Scheme 3. Selective Mitsunobu glycosylation.
Table 1: Mitsunobu glycosidation of 6 leading to 9.
Entry Conditions[a]
6 (a/b ratio)[b]
Yield of 9 (a/b)[c]
1
2
3
4
5
6
DIAD, PPh3, THF, RT
DIAD, PPh3, CH2Cl2, RT
DIAD, PPh3, toluene, RT
DIAD, PPh3, toluene, 608C C6D6 (86:14)
DIAD, PPh3, toluene, 608C C6D6 (86:14)
ADDP, PBu3, toluene, RT
[D8]THF (83:17) 47% (40:60)
CD2Cl2 (86:14) 68% (29:71)
C6D6 (86:14)
80% (25:75)
75% (17:83)
83% (10:90)[c]
69% (100:0)
[a] The ratio of diazo compound/phosphine/6 was 1.6:1.6:1 throughout.
[b] The relative configurations were determined from NMR spectra
recorded in the solvent given. [c] The relative configurations were
assigned by examination of NOE data and JH1,H2 coupling constants for
a-9 and b-9.[4b,10] [c] Slow addition of DIAD. ADDP=1,1’(azodicarbo-
nyl)dipiperidine.
Scheme 2. Preparation of the furanose segment. DMSO=dimethyl
sulfoxide, NBS=N-bromosuccinimide, py=pyridine, Tf=trifluorome-
thanesulfonyl, Tr=triphenylmethyl (trityl).
trityl ether gave alcohol 3, which was oxidized to the
aldehyde, reacted with MeMgBr, and then oxidized with the
Dess–Martin reagent to yield ketone 4 in 68% yield from 2.
Finally the allyl protecting group at C1 was removed by a two-
step procedure that began with alkene isomerization under
the action of RuH generated in situ by the decomposition of
the Grubbs II catalyst.[9] The enol ether (5) thus formed was
cleaved by reaction with aqueous NBS to generate the free
alcohol 6 in 71% yield from allyl ether 4. This sequence
concluded a nine-step preparation of the sugar unit of
hygromycin A that proceeded in 24% overall yield.
The pivotal glycosylation reaction was then investigated
using phenol 8 (made in two steps from 7) as the nucleophile.
Given the acidic nature of the pronucleophile, we decided to
examine the Mitsunobu glycosylation reaction, searching for
conditions that favored the more hindered b isomer, which we
required (Scheme 3). Table 1 shows that there was a signifi-
cant variation in the a/b ratio depending upon the solvent that
was used for the coupling. Toluene was clearly the solvent of
choice, and further optimization of the reaction conditions led
us to perform the reaction at 608C with slow addition of
diisopropylazodicarboxylate (DIAD) (Table 1, entry 6).
Pleasingly, this procedure furnished the product in 83%
yield and with 90:10 selectivity for b-9 (note that the
diastereoisomers were not separable). In general terms
there was only a broad correlation between the a/b ratio of
the starting material 6 and the a/b ratio of the coupled
products, which shows that other processes (such as SN1 or
anomerization) can compete with SN2 attack; these results are
not as clear-cut as those reported with pyranose sugar
donors.[8] Interestingly, complete a-selectivity was achieved
with Bu3P and ADDP, and this was interpreted in terms of
prior ionization of the anomeric hydroxy group to an
oxocarbenium ion, which was then intercepted by the
nucleophile in a stereoselective manner.
The inositol unit of the natural product was then prepared
using a route developed in our laboratories, starting from
commercially available dienone 10 (Scheme 4). The key step
in the sequence was the tethered aminohydroxylation reac-
tion of the primary carbamate derived from alcohol 11.
However, we have recently shown that significant improve-
ments can be made to the tethered aminohydroxylation
protocol by embedding the re-oxidant into the substrate and
preparing N,O-acylated derivatives (ROCONHOCOAr),
rather than the parent carbamates (ROCONH2- which must
be chlorinated in situ).[6b] Therefore, alcohol 11 was activated
(CDI) and then quenched with hydroxylamine to provide a N-
OH carbamate, which was subsequently esterified on the free
hydroxy group by reaction with mesitoyl chloride (Scheme 4).
6508 www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 6507 –6510

Angewandte
Chemie
Scheme 4. Synthesis of the aminocyclitol moiety of hygromycin A.
CDI=1,1’-carbonyldiimadazole.
Ensuing aminohydroxylation of 12 was accomplished by
reaction with potassium osmate (1 mol%) in aqueous butanol
at room temperature; no halogenating agent was required and
the yields and catalyst loadings are far improved over those of
the original procedure. The original protocol transformed 11
into 13 in 61% yield using 4 mol% of the osmium catalyst;
now we could accomplish the same transformation in 74%
using 1 mol% of the catalyst. The oxazolidinone thus
produced was converted into aminocyclitol 14 in a further
seven steps to constitute a 15-step route with 20% overall
yield; this compares well with the yields of the two routes
reported in the literature (Ogawa group: 20 steps, 2.7%
overall yield; Trost group: 13 steps, 10% overall yield from
benzoquinone or 10 steps, 23% yield from conduritol B
tetraacetate).
Scheme 5. Completion of the synthesis. BOP=1-benzotriazolyloxytris-
(dimethylamino)phosphonium hexafluorophosphate.
addition, recent improvements to the reagents used for the
tethered aminohydroxylation mean that the synthesis of the
aminocyclitol ring is improved considerably.
Received: May 27, 2009
Published online: July 31, 2009
The final sequence required deprotection of the aromatic
linker. The two protecting groups on the arene unit were
removed by reaction with aluminum tribromide and dime-
thylsulfide,[11] yielding free acid 15 in 61% yield (Scheme 5).
(Note that at this point the minor diastereoisomer from the
sugar coupling could be separated by chromatography.) This
free acid was then coupled with aminocyclitol 14 (see
Scheme 4), furnishing amide 16 in 70% yield after chroma-
tography. Pleasingly, no epimerization was detected at C4
during either of these two steps, suggesting that the OTIPS
group at C3 was acting as planned. Finally, the two TIPS
groups were removed using a mild source of fluoride
Keywords: aminohydroxylation · Mitsunobu glycosylation ·
natural products · total synthesis
.
[1] a) R. C. Pittenberger, R. N. Wolfe, M. M. Hohen, P. N. Marks,
W. A. Daily, M. McGuire, Antibiot. Chemother. 1953, 3, 1268;
b) R. L. Mann, R. M. Gale, F. R. VanAbeele, Antibiot. Chemo-
ther. 1953, 3, 1279; c) Y. Sumiki, G. Nakamura, M. Kawasaki, S.
Yamashita, K. Anazi, K. Isono, Y. Serizawa, Y. Tomiyama, S.
Suzuki, J. Antibiot. Ser. A 1955, 8, 170; d) K. Isono, S. Yamashita,
Y. Tomiyama, S. Suzuki, J. Antibiot. 1957, 10, 21; e) Y. Wakisaka,
K. Koizumi, Y. Nishimoto, M. Kobayashi, N. Tsuji, J. Antibiot.
1980, 33, 695; f) M. Uyeda, M. Mizukami, K. Yokomizo, K.
Suzuki, Biosci. Biotechnol. Biochem. 2001, 65, 1252.
[2] M. C. Guerrer, J. Modolell, Eur. J. Biochem. 1980, 107, 409.
[3] a) S. J. Hecker, M. L. Minich, K. M. Werner, Bioorg. Med. Chem.
Lett. 1992, 2, 533; b) S. J. Hecker, S. C. Lilley, M. L. Minich,
K. M. Werner, Bioorg. Med. Chem. Lett. 1992, 2, 1015; c) S. J.
Hecker, S. C. Lilley, K. M. Werner, Bioorg. Med. Chem. Lett.
1992, 2, 1043; d) S. J. Hecker, C. B. Cooper, K. T. Blair, S. C.
Lilley, M. L. Minich, K. M. Werner, Bioorg. Med. Chem. Lett.
1993, 3, 289; e) C. B. Cooper, K. T. Blair, C. S. Jones, M. L.
Minich, Bioorg. Med. Chem. Lett. 1997, 7, 1747; f) S. Omura, A.
Hakagawa, T. Fujimoto, K. Saito, K. Otoguro, J. Antibiot. 1987,
40, 1619; g) A. Nakagawa, T. Fujimoto, S. Omura, J. C. Walsh,
R. L. Stoish, B. George, J. Antibiot. 1987, 40, 1627.
(tris(dimethylamino)sulfonium
difluorotrimethylsilicate
(TSAF))[12] at 08C. (At this point care must be taken to
ensure that there is no epimerization at C4.) This sequence
gave synthetic (À)-hygromycin A (À)-1; its spectroscopic
data closely matched that reported in the literature (1H/
13C NMR, m.p., [a]D).
In conclusion, we have synthesized (À)-hygromycin A
(À)-(1) by a sequence that is noteworthy for its brevity and
efficiency with just 17 linear steps and 10% overall yield. This
route compares well with the only previous synthesis of this
compound by Ogawa (24 linear steps and < 1% yield).
Key points are the successful Mitsunobu glycosylation
strategy which can be tuned to produce good selectivity for
the desired anomer (and the unnatural anomer if desired). In
Angew. Chem. Int. Ed. 2009, 48, 6507 –6510
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 6509

Communications
[4] a) N. Chida, M. Ohtsuka, K. Nakazawa, S. Ogawa, J. Chem. Soc.
[8] a) A. B. Smith III, K. J. Hale, R. A. Rivero, Tetrahedron Lett.
1986, 27, 5813; b) A. B. Smith III, R. A. Rivero, K. J. Hale, H. A.
Vaccaro, J. Am. Chem. Soc. 1991, 113, 2092; c) W. R. Roush,
X. F. Lin, J. Org. Chem. 1991, 56, 5740; d) for early work see, G.
Grynkiewicz, Carbohydr. Res. 1977, 53, C11.
[9] For a review see: T. J. Donohoe, T. J. C. OꢀRiordan, C. P. Rosa,
Angew. Chem. 2009, 121, 1032 – 1035; Angew. Chem. Int. Ed.
2009, 48, 1014 – 1017.
[10] J. D. Stevens, H. G. Fletcher, J. Org. Chem. 1968, 33, 1799.
[11] M. Node, K. Nishide, M. Sai, K. Fuji, E. Fujita, J. Org. Chem.
1981, 46, 1991.
[12] K. A. Scheidt, H. Chen, B. C. Follows, S. R. Chemler, D. S.
Coffey, W. R. Roush, J. Org. Chem. 1998, 63, 6436.
Chem. Commun. 1989, 436; b) N. Chida, M. Ohtsuka, K.
Nakazawa, S. Ogawa, J. Org. Chem. 1991, 56, 2976.
[5] a) B. M. Trost, J. Dudash, O. Dirat, Chem. Eur. J. 2002, 8, 259;
b) B. M. Trost, J. Dudash, E. J. Hembre, Chem. Eur. J. 2001, 7,
1619.
[6] a) T. J. Donohoe, P. D. Johnson, R. J. Pye, M. Keenan, Org. Lett.
2005, 7, 1275 – 1277; b) T. J. Donohoe, C. J. R. Bataille, W.
Gattrell, J. Kloesges, E. Rossignol, Org. Lett. 2007, 9, 1725 –
1728.
[7] Hygromycin A itself is isolated together with its epimer at this
position, see Ref. [1e].
6510
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 6507 –6510