Use of polystyrene-supported DBU in the synthesis and α-selective glycosylation study of the unstable schmidt donor of L-kedarosamine

Article abstract of DOI:10.1021/ol036353v

In the α-glycosylation study of the unusual, 2-deoxy amino sugar of kedarcidin, polystyrene-supported DBU (PDBU) was found to be invaluable to the clean preparation of the highly labile Schmidt donor of L-kedarosamine. By further recognition that the C4-a

Full text of DOI:10.1021/ol036353v

ORGANIC
LETTERS
2004
Vol. 6, No. 5
719-722
Use of Polystyrene-Supported DBU in
the Synthesis and r-Selective
Glycosylation Study of the Unstable
Schmidt Donor of L-Kedarosamine
Isao Ohashi, Martin J. Lear,* Fumihiko Yoshimura, and Masahiro Hirama*
Department of Chemistry, Graduate School of Science, Tohoku UniVersity, COE,
Sendai 980-8578, Japan
martin@ykbsc.chem.tohoku.ac.jp; hirama@ykbsc.chem.tohoku.ac.jp
Received December 3, 2003
ABSTRACT
In the r-glycosylation study of the unusual, 2-deoxy amino sugar of kedarcidin, polystyrene-supported DBU (PDBU) was found to be invaluable
to the clean preparation of the highly labile Schmidt donor of L-kedarosamine. By further recognition that the C4-alcohol should be left free
for favorable acceptor reactivity, we could for the first time successfully assemble the C13-O-r-glycoside of the ansamacrolide substructure
of the kedarcidin chromophore.
A central consideration in the total synthesis of the kedarcidin
chromophore 1 resides in timing the construction of the
fragile, nine-membered enediyne core with the attachment
of the labile 2,6-dideoxy sugars, L-kedarosamine and L-
mycarose (Figure 1).1,2 Most conveniently, glycosylation
events when performed at a late stage on unstable aglycon
acceptors will ensure that a minimal number of reaction steps
will be required in a total synthesis. This consideration
demands for the glycosyl donor (cf. 2) to be attached in a
direct and exceptionally mild manner and thus demands that
(1) Isolation of kedarcidin and chromophore structure (1): (a) Lam, K.
S.; Hesler, G. A.; Gustavson, D. R.; Crosswell, A. R.; Veitch, J. M.;
Forenza, S.; Tomita, K. J. Antibiot. 1991, 44, 472-478. Hofstead, S. J.;
Matson, J. A.; Malacko, A. R.; Marquardt, H. J. Antibiot. 1992, 45, 1250-
1254. (b) Leet, J. E.; Schroeder, D. R.; Langley, D. R.; Colson, K. L.; Huang,
S.; Klohr, S. E.; Lee, M. S.; Golik, J.; Hofstead, S. J.; Doyle, T. W.; Matson,
J. A. J. Am. Chem. Soc. 1993, 115, 8432-8443. (c) Structure revision:
Kawata, S.; Ashizawa, S.; Hirama, M. J. Am. Chem. Soc. 1997, 119, 12012-
12013.
Figure 1. Structure of kedarcidin chromophore 1 featuring the
unusual R-linked, 2,6-dideoxy sugar, L-kedarosamine.
(2) For a very inspired and elegant synthesis of the aglycon of 1, see:
Myers, A. G.; Hogan, P. C.; Hurd, A. R.; Goldberg, S. D. Angew. Chem.,
the amino group be left unprotected in its NMe2 form.3,4 As
Int. Ed. 2002, 41, 1062-1067.
a consequence, synthetic difficulties are anticipated in the
10.1021/ol036353v CCC: $27.50 © 2004 American Chemical Society
Published on Web 02/10/2004

formation and handling of not only the activated sugar donor
but also of the resulting 2,6-dideoxy-4-dimethylaminopyra-
noside. Not surprisingly, there are few reports describing
direct glycosylation protocols for unprotected amino sugar
donors, especially for 2,6-dideoxy systems.5,6 Further chal-
lenges involve controlling the stereochemical outcome of the
glycosylation event to be R-selective in a sugar that lacks a
stereodirecting 2-substituent. Nevertheless, by extrapolating
the stereochemistry of kedarosamine to fuco-like 2-deoxy
systems that bear axially orientated 4-substituents, we
anticipated that high R-selectivity would be inherently
realized through substrate control.3-5,7
remained totally unreactive to these conditions or decom-
posed. It was therefore necessary to determine alternative
glycosylation conditions. Yet, despite a multitude of re-
agent systems that were explored for donors derived from
thioglycoside 2, which worked well on simple alcohols such
as i-Pr2CHOH to give 3, all efforts failed for the pro-
pargylic alcohol acceptor 4.9 Further efforts were plagued
with a combination of the instability of the glycosyl donors
derived from 2 (X ) SPh) and reagent incompatibility to
the free NMe2 group. Specifically, under standard Schmidt
conditions using DBU or K2CO3,10 the isolation of the
glycosyl trichloroacetimidate in pure form from 2 (X ) OH)
was unsuccessful and attempts to generate glycoside products
of 4 could not be achieved or investigated in a reliable and
systematic manner.
Our study began with the 2-deoxythioglycoside (2) of
L-kedarosamine.8 Encouraged by the successful R-selective
glycosylation of 2 (X ) SPh) with simple alcohols to give
glycosides such as 3 using AgPF6/2,6-di-tert-butyl-4-meth-
ylpyridine (DTBMP),5 we investigated the glycosylation of
the diyne alcohol 4 (Scheme 1). To our dismay, regardless
It became increasingly apparent that specific reagents and
isolation conditions had to be judiciously chosen for the case
at hand, especially since we favored the amino sugar (5) to
bear a TES-ether as an easily removable protective group
for the concluding stages to 1 (Scheme 2). First, we deter-
Scheme 1. Low Acceptor Reactivity of Diyne Fragment 4
Scheme 2. Reliable Route to Schmidt Donor 8a
of the protective groups employed, the secondary alcohol in
4 (e.g., R1 ) Piv, R2 ) H, TES, MOM; or R1 ) R2 ) CMe2)
(3) Reviews on 2-deoxyglycoside formation: (a) Marzabadi, C. H.;
Franck, R. W. Tetrahedron 2000, 56, 8385-8417. (b) Toshima, K.; Tatsuta,
K. Chem. ReV. 1993, 93, 1503-1531.
mined the best method to form the free sugar 6 from thio-
glycoside 5 and found AgNO3/2,6-lutidine in wet THF to
be aptly suited to the task, which allowed for a nonaqueous
isolation procedure.11 Next, we pursued the clean formation
and isolation of the Schmidt donor 8 from 6. Assessment of
the most desirable preparative conditions led to the notion
of using a polymer-supported base, which would necessitate
a simple filtration and evaporation step at workup. Indeed,
(4) For R-selective glycosylations using NH-protected kedarosamine
donors, see: Vuljanic, T.; Kihlberg, J.; Somfai, P. J. Org. Chem. 1998, 63,
279-286.
(5) For a general R-selective method for both sugars of kedarcidin using
AgPF6 activation of 2-deoxythioglycosides, see: Lear, M. J.; Yoshimura,
F.; Hirama, M. Angew. Chem., Int. Ed. 2001, 40, 946-949.
(6) Glycosylation examples of sugar donors bearing unprotected amino-
groups in total synthesis settings: Tatsuta, K.; Kobayashi, Y.; Gunji, H.;
Masuda, H. Tetrahedron Lett. 1988, 29, 3975-3978. Paquette, L. A.;
Collado, I.; Purdie, M. J. Am. Chem. Soc. 1998, 120, 2553-2562. Myers,
A. G.; Liang, J.; Hammond, M.; Harrington, P. M.; Wu, Y.; Kuo, E. Y. J.
Am. Chem. Soc. 1998, 120, 5319-5320. Toshima, K.; Maeda, Y.; Ouchi,
H.; Asai, A.; Matsumura, S. Bioorg. Med. Chem. Lett. 2000, 10, 2163-
2165. Kolar, C.; Dehmel, K.; Knoedler, U.; Paal, M.; Hermentin, P.; Gerken,
M. J. Carbohydr. Chem. 1989, 8, 295-305.
(7) Instances of high R-selectivity with fuco-like 2-deoxypyranosyl
donors: (a) Roush, W. R.; Bennett, C. E. J. Am. Chem. Soc. 2000, 122,
6124-6125. (b) Kolar, C.; Kneissl, G. Angew. Chem., Int. Ed. Engl. 1990,
29, 809-811. (c) Schene, H.; Waldmann, H. Synthesis 1999, 1411-1422.
(d) Florent, J.-C.; Gaudel, G.; Monneret, C.; Hoffmann, D.; Kraemer, H.-
P. J. Med. Chem. 1993, 36, 1364-1368.
(9) Reagent systems explored include activation of 2 (X ) SPh) with
NIS, Selectfluor/BF3‚OEt2 (via glycosyl fluoride), PhSeCl/Mg(ClO4)2
(unpublished findings), AgOAc/DTBMP and TBSOTf activation of re-
sulting glycosyl acetate (unpublished work), Tf2O activation of derived
sulfoxide, and dehydrative glycosylations with 2 (X ) OH): Burkart, M.
D.; Zhang, Z.; Hung, S.-C.; Wong, C.-H. J. Am. Chem. Soc. 1997, 119,
11743-11746. Kim, S.-H.; Augeri, D.; Yang, D.; Kahne, D. J. Am. Chem.
Soc. 1994, 116, 1766-1775. Crich, D.; Sun, S. J. Am. Chem. Soc. 1998,
120, 435-436. Garcia, B. A.; Gin, D. Y. J. Am. Chem. Soc. 2000, 122,
4269-4279.
(8) 2 and 5 were prepared by thiolysis of protected, methyl L-
kedarosaminides: Lear, M. J.; Hirama, M. Tetrahedron Lett. 1999, 40,
4897-4900. Hanessian, S.; Guindon, Y. Carbohydr. Res. 1980, 86,
C3-C6.
(10) Schmidt, R. R.; Michel, J. Angew. Chem., Int. Ed. Engl. 1980, 19,
731-732. Schmidt, R. R. Pure Appl. Chem. 1989, 61, 1257-1270.
(11) Method adapted from conditions to deprotect methylthiomethyl
ethers: Corey, E. J.; Bock, M. G. Tetrahedron Lett. 1975, 3269-3270.
720
Org. Lett., Vol. 6, No. 5, 2004

use of polystyrene-supported DBU (PDBU, 7) developed by
Tomoi and co-workers furnished the Schmidt donor 8 in a
reproducible and clean manner.12 After testing with i-Pr2-
CHOH, which gave 3 in a 53% yield under TESOTf acti-
vation (Scheme 2), reaction of the diyne fragment 9 with
the imidate 8 generated the glycosylated adducts 10/11 for
the first time, exclusively as R-anomers (Table 1).
MgBr2‚OEt2, Yb(OTf)3, CCl3CHO, Sc(OTf)3, and PPTS were
observed (even at room temperature over 15 h), the use of
BF3‚OEt2 was found most successful and equally effective
in either the presence of 4 Å molecular sieves (mildly basic)
or Lewis acidic, AW-300 MS (entries 2-4).3,7c In these cases,
the starting temperature was essential to optimizing the
glycosylation yield, and 85% of 10/11 was attained at 0 °C
by adding catalytic amounts of BF3‚OEt2 to 9 in the presence
of excess imidate 8 (entry 4). In comparison, at -78 °C,
moderate yields and similar ratios were obtained regardless
of reaction time, even by using excess BF3‚OEt2 or by
warming to -30 or to 0 °C (cf. entry 2).
Table 1. R-Selective Glycosylation of Diyne 9 with Imidate 8a
As anticipated empirically,3-5,7 we were pleased to dis-
cover that all reactions produced the 2-deoxy-R-glycosides
of 10/11 exclusively. It was nevertheless surprising that
acetonitrile had no effect on the stereochemical outcome and
solvent effects in general were found to be rather negligible
(cf. entry 3), which was confirmed more thoroughly for the
C4-epimer of 9 (entry 4). These results are consistent with
high substrate control and an unusually high preference for
kinetically controlled, pseudoaxial attack onto an oxocarbe-
nium intermediate like 12 (Figure 2).13
entry reagent (equiv)
temp (time)
yield (10:11)
9
1b
2
3c
4d
5e
6
TBSOTf (2.2)
BF3‚OEt2 (3.0)
BF3‚OEt2 (2.0)
BF3‚OEt2 (0.4)
Cp2HfX2 (0.2)
TiCl4 (0.2)
-78 °C (5.0 h) 38% (3.5:1)
-78 °C (3.0 h) 33% (2.0:1)
-30 °C (3.0 h) 43% (1.8:1)
0 °C (2.5 h) 85% (1.8:1)
0 °C (2.5 h) 74% (1.0:1)
0 °C (4.5 h) 72% (1.3:1)
0 °C (3.0 h) 36% (3.9:1)
0 °C (4.5 h) 20% (3.2:1)
31%
52%
56%
3%
26%
27%
49%
79%
Figure 2. Proposed origin of high R-anomeric stereocontrol for
L-kedarosamine.
7
8
Hf(OTf)4 (0.6)
HOTf (2.4)
a See text and Scheme 2 for details. b Silylated products of 9 also
recovered in 15% yield. c CH3CN was used as solvent instead of
CH2Cl2. d Similar results were obtained for C4-epimer of 9 in CH2Cl2, THF,
CH3CN, and PhCH3. e Cp2HfX2 ) Cp2HfCl2 (0.2 equiv)/AgClO4 (0.4
equiv).
In attempts to improve the reactivity and prolong the life-
time of sugar donors, various additives were added in com-
bination with BF3‚OEt2. However, all additives (for example,
TMSOTf,14 DTBMP, Bu3SnCl, and TBAC) did not improve
the combined yield of 10/11 but tended to destroy the starting
material 9. Although Hf(OTf)4 and HOTf gave moderate to
low conversions favoring the desired C13-O-glycoside 10
(entries 7 and 8), catalytic amounts of TiCl4 or Cp2HfCl2/
AgClO4 were recognized as high-yielding reagent systems
and were exceptionally clean (entries 5 and 6). Indeed, it
should be noted that unreacted 9 was largely recovered for
all the entries given in Table 1, and full recovery of precious
starting material is clearly an important factor when deter-
mining viable glycosylation conditions in an advanced total
synthesis setting of 1.
In the prospect that each acceptor would behave differently
for any given glycosylation reaction with 8, we focused upon
optimizing the combined chemical yield of 10/11 (10 being
major) from the diyne alcohol 9 (Table 1). Besides minor
amounts of 10/11, TMSOTf and TESOTf mostly gave
O-silylation-derived products of 9 due to the amino sugar
acting as a base. As a more bulky and less reactive promoter,
TBSOTf minimized, but never suppressed totally, this
competing side-reaction and gave a 38% combined yield of
10/11 more cleanly (entry 1). Although no glycosides with
1.1-4.0 equiv of activators such as LiB(C6F6)4, LiPF6, ZnBr2,
Armed with the experience described herein, we next
focused on screening glycosylation conditions for various
kedarcidin aglycon fragments (Scheme 3).15 In short, leaving
(12) (a) Polystyrene-supported DBU (PDBU, 7) can be prepared simply
by adding Merrifield resin to n-BuLi/DBU in THF at -78 °C: Tomoi, M.;
Kato, Y.; Kakiuchi, H. Makromol. Chem. 1984, 185, 2117-2124. Iijima,
K.; Fukuda, W.; Tomoi, M. J. Macromol. Sci., Pure Appl. Chem. 1992,
A29 (3), 249-261. (b) Cs2CO3 can be used to give imidate 8 (R/â ) 1:2),
which can also be removed by centrifugation or filtration (Urban, F. J.;
Moore, B. S.; Breitenbach, R. Tetrahedron Lett. 1990, 31, 4421-4424),
but the larger particle size of PDBU merited the most convenient,
contaminant-free isolation procedure.
(13) Intramolecular participation of the 4-NMe2 lone-pair, which tran-
siently stabilizes the oxocarbenium species and directs pseudoaxial attack,
is also conceivable; cf.: Jiao, H.; Hindsgaul, O. Angew. Chem., Int. Ed.
1999, 38, 346-348. Romero, J. A. C.; Tabacco, S. A.; Woerpel, K. A. J.
Am. Chem. Soc. 2000, 122, 168-169.
(14) Nicolaou, K. C.; Mitchell, H. J.; van Delft, F. L.; Ru¨bsam, F.;
Rodr´ıguez, R. M. Angew. Chem., Int. Ed. 1998, 37, 1871-1874.
Org. Lett., Vol. 6, No. 5, 2004
721

be virtually inert (see Supporting Information). Ultimately,
BF3‚OEt2 with a large excess of the imidate 8 [15-25 equiv]
proved to be most successful for the diols 13 and 15. In the
case of 13, the use of 1:1 CH3CN/CH2Cl2 at 0 °C moderated
the reactivity of BF3‚OEt2, so that unreacted starting material
could be completely recovered in 18% yield, giving a
marginal preference for desired 14 in 42% yield and 40%
of the C4-O-R-glycoside. For the ansamacrolide 15, higher
reaction temperatures were necessary, and employment of
ClCH2CH2Cl as the solvent at 40 °C gave a 38% yield of
the desired C13-O-R-glycoside 16, together with 27%
recovery of 15 and about 10% impure C4-O-glycoside.
Notwithstanding the chemical challenges and inherent
instabilities of nine-membered, epoxyenediyne aglycon units
of 1, this latter result holds promise toward finalizing current
efforts to a most formidable target of contemporary organic
synthesis.2,5,15
Scheme 3. Successful R-Glycosylation of the Kedarcidin
Subunits 15 and 17 with L-Kedarosaminea
Acknowledgment. This work was supported under the
CREST program from JST and a Grant-in-Aid for Scientific
Research, Wakate B, by MEXT (to M.J.L.).
Supporting Information Available: Full experimental
details and a general glycosylation procedure for the prepara-
tion of 6, 8, 10, and 11 (and their C4-epimers) and 14 and
16, including a representative scheme of glycosylation results
with C4-O-protected and C4/C5-olefin counterparts of 13
and 15. This material is available free of charge via the
Internet at http://pubs.acs.org.
a Reagents and conditions: (a) 8 (15.0 equiv), BF3‚OEt2 (3.0
equiv), CH2Cl2/MeCN (1:1), 4 Å molecular sieves, rt, 12 h; (b) 8
(25.0 equiv), BF3‚OEt2 (13.0 equiv), ClCH2CH2Cl, 4 Å molecular
sieves, 40 °C, 0.5 h.
OL036353V
(15) 13 and 15 were prepared by modification of our previous strategy:
Yoshimura, F.; Kawata, S.; Hirama, M. Tetrahedron Lett. 1999, 40, 8281-
8285. Full details, including advanced studies of 1, will be disclosed
elsewhere.
the C4-OH free and intact was found to be vital to acceptor
reactivity, since acceptors akin to 13 and 15, which feature
C4-O-protective groups or a C4-C5 olefin, were found to
722
Org. Lett., Vol. 6, No. 5, 2004