Dihydroxylation of the triene subunit of rapamycin

Full text of DOI:10.1021/jo9816820

J . Org. Chem. 1998, 63, 10069-10073
10069
Dih yd r oxyla tion of th e Tr ien e Su bu n it of
Ra p a m ycin
Richard Sedrani,* Binh Thai, J ulien France, and
Sylvain Cottens
Novartis Pharma AG, Preclinical Research,
CH-4002, Basel, Switzerland
Received August 18, 1998
Rapamycin 1 was discovered in 1975 at Ayerst in a
screen for novel antifungal agents.1 It has attracted
interest since the beginning of the nineties because of
its immunosuppressive properties2 and is presently un-
dergoing clinical trials for the prevention of allograft
rejection.3
The C17-C22 triene subunit represents one of the
characteristic structural subunits of rapamycin. Only a
few modifications of this part of the macrolide have been
reported in the literature. A hetero Diels-Alder reaction
with 4-phenyl-1,2,4-triazoline-3,5-dione involving C19-
C22 as the diene partner has been described.4 Two groups
have published results concerning the acid-catalyzed
displacement or elimination of the C16-methoxyl which
in some cases occurs with concomitant rearrangement
of the triene.5 Finally, a degradation of rapamycin
through exhaustive ozonolysis has been reported.6 We
were interested in exploring the selective functionaliza-
tion of the triene subunit and, with this goal in mind,
turned our attention to dihydroxylation. Further impetus
for exploring this reaction was provided by a recent report
describing a 19,20-dihydroxyrapamycin as one of the
metabolites obtained by incubation of rapamycin with
dexamethasone-induced rat liver microsomes,7 though
the absolute configurations of the newly introduced
hydroxyls were not reported. The problems we expected
at the outset of our investigation concerned the regio-
and chemoselectivity. It was for instance not clear
whether dihydroxylation of one of the double bonds of
the triene would be preferred over oxidation of the
C29,C30 olefin. Also, we would have to avoid polyhy-
droxylation of the substrate.
FKBP12/rapamycin complex.9 Its conformation in solu-
tion in DMSO has also been reported.10 The conforma-
tions of the macrocycle are virtually identical in the three
cases. These structures clearly show that the “bottom”
face of the C19-C22 triene is exposed and accessible to
peripheral attack, while the approach of any reagent on
the internal, “top” face is highly unlikely (Figure 1). On
the basis of this structural information we anticipated
that facial selectivity would not be an issue for the
dihydroxylation if the reaction were to occur in the triene
subunit.
We initially resorted to modifications of the Upjohn
osmylation procedure11 using catalytic osmium tetraoxide
and various amounts of NMO. These attempts met with
little success. When 1 equiv of NMO was used, a mixture
of rapamycin, together with dihydroxylated and tetrahy-
droxylated derivatives was obtained. Each of the latter
consisted of a mixture of isomeric compounds which were
not further characterized. The use of 2 equiv of NMO
afforded an inseparable mixture consisting predomi-
nantly of tetrahydroxyrapamycins according to MS. The
problems encountered with this procedure led us to
contemplate the use of the asymmetric dihydroxylation
(AD) conditions. Indeed, reports by several groups have
documented that the AD process allows for the selective
dihydroxylation of polyenes.12-14 Most relevant for our
purpose were publications from the Sharpless group
demonstrating that “mono”-dihydroxylation can also be
achieved in the case of conjugated polyenes.12a,c This
The stereoselectivity of the reaction, on the other hand,
was considered to be predictable. The three-dimensional
structure of rapamycin has been determined by single-
crystal X-ray analyses of the macrolide alone8 and of the
* Corresponding author: Dr. Richard Sedrani, Novartis Pharma AG,
S-507.1.01, CH-4002 Basel, Switzerland, Tel: (+) 41 61 324 2609,
Fax: (+) 41 61 324 6735, E-mail: richard.sedrani@pharma.novartis.com.
(1) (a) Ve´zina, C.; Kudelski, A.; Sehgal, S. N. J . Antibiot. 1975, 28,
721-726. (b) Sehgal, S. N.; Baker, H.; Ve´zina, C. J . Antibiot. 1975,
28, 727-732.
(2) Dumont, F. J .; Su, Q. Life Sci. 1996, 58, 375-395.
(3) (a) Shoker, A. S. Drugs Today 1997, 33, 221-236. (b) Kahan, B.
D. Transplant. Proc. 1997, 29, 48-50.
(4) Ocain, T. D.; Longhi, D.; Stefan, R. J .; Caccese, R. G.; Sehgal, S.
N. Biochem. Biophys. Res. Commun. 1993, 192, 1340-1346.
(5) (a) Grinfeld, A. A.; Caufield, C. E.; Schiksnis, R. A.; Mattes, J .
F.; Chan, K. W. Tetrahedron Lett. 1994, 35, 6835-6838. (b) Luengo,
J . I.; Konialian-Beck, A.; Rozamus, L. W.; Holt, D. A. J . Org. Chem.
1994, 59, 6512-6513.
(6) Goulet, M. T.; Boger, J . Tetrahedron Lett. 1990, 31, 4845-4848.
(7) Nickmilder, M. J . M.; Latinne, D.; Verbeeck, R. K.; J anssens,
W.; Svoboda, D.; Lhoe¨st, G. J . J . Xenobiotica 1997, 27, 869-883.
(8) Swindells, D. C. N.; White, P. S.; Findlay, J . A. Can. J . Chem.
1978, 56, 2491-2492.
(9) (a) Van Duyne, G. D.; Standaert, R. F.; Schreiber, S. L.; Clardy,
J . J . Am. Chem. Soc. 1991, 113, 7433-7434. (b) Van Duyne, G. D.;
Standaert, R. F.; Karplus, P. A.; Schreiber, S. L.; Clardy, J . J . Mol.
Biol. 1993, 229, 105-124.
(10) Kessler, H.; Haesner, R.; Schu¨ler, W. Helv. Chim. Acta 1993,
76, 117-130.
(11) Van Rheenen, V.; Kelly, R. C.; Cha, D. Y. Tetrahedron Lett.
1976, 1973-1976.
(12) (a) Xu, D.; Crispino, G. A.; Sharpless, K. B. J . Am. Chem. Soc.
1992, 114, 7570-7571. (b) Xu, D.; Park, C. Y.; Sharpless, K. B.
Tetrahedron Lett. 1994, 35, 2495-2498. (c) Becker, H.; Soler, M. A.;
Sharpless, K. B. Tetrahedron 1995, 51, 1345-1376.
(13) Corey, E. J .; Noe, M. C.; Shieh, W. C. Tetrahedron Lett. 1993,
34, 5995-5998.
(14) (a) Vidari, G.; Dapiaggi, A.; Zanoni, G.; Garlaschelli, L. Tetra-
hedron Lett. 1993, 34, 6485-6488. (b) Vidari, G.; Giori, A.; Dapiaggi,
A.; Lanfranchi, G. Tetrahedron Lett. 1993, 34, 6925-6928.
10.1021/jo9816820 CCC: $15.00 © 1998 American Chemical Society
Published on Web 12/01/1998

10070 J . Org. Chem., Vol. 63, No. 26, 1998
Notes
trienes, the internal double bond remained unaffected,
possibly because its dihydroxylation would result in
complete disruption of conjugation.12c In the present case,
the product resulting from dihydroxylation of the internal
olefin predominated. The trisubstituted C17,C18 double
bond remained untouched, although this type of olefin
is generally an excellent substrate for the AD. This is
probably due in large part to the electron-withdrawing
inductive effect of the C16-methoxy group and also, to
some degree, to its steric demands. Electronegative allylic
atoms, i.e., oxygen and nitrogen, have been reported to
decrease the hydroxylation rate in some cases.23 This
observation is confirmed by AD reactions performed on
the diene geraniol and some of its derivatives. In those
cases, dihydroxylation occurs with high preference on the
trisubstituted olefin which is remote from the allylic
O-substituent, again indicating a deactivating effect of
the latter.12b,13,14a As far as the slight selectivity for the
formation of 3 over 4 is concerned, it can be explained
by the fact that the C21,C22 double bond is to some
degree shielded by the C23-methyl group. The isolated
trisubstituted C29,C30 olefin remained unchanged. This
could be due in part to a deactivating effect of the allylic
hydroxyl, as discussed above. But in that particular case,
it is more likely that the inaccessibility of this olefin due
to steric factors is the main reason for the lack of
reactivity24 (Figure 1). These structural features of the
macrolide can explain the unusual regiochemical outcome
of the AD.
When 2 was submitted to the reaction conditions
described in eq 1, but using the pseudoenantiomeric
(DHQ)2PHAL instead of (DHQD)2PHAL, practically no
conversion was observed. Forcing the reaction conditions
by increasing the temperature to 23 °C and by prolonging
the reaction time also was to no avail, but rather led to
some decomposition of starting material. On the basis of
the Sharpless mnemonic device, (DHQ)2PHAL was ex-
pected to induce dihydroxylation of the “top” face of the
triene subunit of either 1 or 2 (vide supra), which is
oriented toward the inside of the macrocycle and is
therefore not readily accessible (Figure 1). This was
predicted to be a very disfavored process. The virtual
absence of reaction in the presence of (DHQ)2PHAL
clearly showed that dihydroxylation of the “top” face of
the triene is not feasible. This result confirmed the
stereochemical outcome of the reaction depicted in eq 1.
Deprotection of 3, using HF-pyridine in acetonitrile,
proceeded uneventfully to afford 19(R),20(R)-dihydroxy-
rapamycin 5 in 61% yield (Scheme 2). When 4 was
treated in the same conditions, desilylation occurred with
concomitant condensation of the 21,22-diol with the C26-
ketone to provide ketal 6 as the sole product (Scheme 3).
This latter compound provided, in addition to the
reactivity pattern discussed above, another opportunity
for verifying the stereochemical outcome of the dihy-
droxylation. Inspection of Dreiding models revealed that
intramolecular ketalization of the 21(R),22(R)-diol with
the C26-ketone can only result in the ketal 6 having the
S configuration at C-26. The 21(S),22(S)-diol, on the other
F igu r e 1. X-ray crystal structure of rapamycin showing that
only one face of the triene subunit is accessible.
selectivity is ascribed to the preclusion of the so-called
“second cycle”.12a,15 The Sharpless mnemonic device16
indicated that the dihydroquinidine-derived ligand
(DHQD)2PHAL, corresponding to AD-mix-â, would pro-
mote dihydroxylation of any double bond within the
triene from its accessible, i.e., peripheral (“bottom”) face
(vide supra). Exposure of 28,40-O-bis(triethylsilyl)rapa-
mycin 217 to modified AD conditions18 using (DHQD)2-
PHAL resulted after 12 h at 0 °C in complete consump-
tion of the starting material and the clean formation of
1
the diols 3 and 4 in a 2:1 ratio as judged by H NMR
analysis of the crude reaction mixture (Scheme 1). After
careful silica gel chromatography, 3 and 4 were isolated
in 39% and 23% yield, respectively (Scheme 1). 1H NMR
analysis in CDCl3 showed the 19,20-dihydroxylated
derivative 3 to consist of a 1:1 mixture of conformers,
whereas the 21,22-dihydroxylated compound 4 exists as
a ca. 4:1 ratio of conformers.19 The fact that the mixture
1
observed for 3 by H NMR corresponds to conformers,
and not to diastereomers, was established by a combina-
tion of DQF COSY20 and ROESY21,22 NMR experiments.
In the 2D ROESY experiment performed with 3 an
exchange cross-peak was observed between the two
signals corresponding to the C2 proton (δ 4.48 and 5.22
ppm), indicating the presence of two slowly equilibrating
forms of the same compound. The fact that both 3 and 4
are single diastereomers was subsequently confirmed by
other means (vide infra). The regiochemical outcome of
this reaction was to some degree unexpected. In previous
reports concerning the AD of less complex conjugated
(15) Kwong, H.-L.; Sorato, C.; Ogino, Y.; Chen, H.; Sharpless, K. B.
Tetrahedron Lett. 1990, 31, 2999-3002.
(16) Kolb, H. C.; Andersson, P. G.; Sharpless, K. B. J . Am. Chem.
Soc. 1994, 116, 1278-1291.
(17) The use of bissilylated rapamycin 2 was made necessary by the
fact that unprotected rapamycin is insoluble in the biphasic tert-butyl
alcohol:water mixture recommended for these reactions. Preclusion of
the “second cycle” and of polyhydroxylation calls for such a biphasic
solvent system. Attempts to use nonwater miscible organic solvents
capable of solubilizing rapamycin, such as methylene chloride, toluene,
or diethyl ether, led to no reaction.
(18) For a standard procedure, see: Sharpless, K. B.; Amberg, W.;
Bennani, Y. L.; Crispino, G. A.; Hartung, J .; J eong, K.-S.; Kwong, H.-
L.; Morikawa, K.; Wang, Z.-M.; Xu, D.; Zhang, X.-L. J . Org. Chem.
1992, 57, 2768-2771. In this study we added the individual compo-
nents separately instead of using the premixed powder. Instead of
potassium osmate, osmium tetraoxide solution was used. More impor-
tantly, the loads of osmium catalyst and phthalazine ligand were 5-fold
higher than those recommended in the published standard procedure
(see Experimental Section for more details).
(19) Rapamycin itself exists in CDCl3 as a mixture of conformers in
a ratio of about 3:1: McAlpine, J . B.; Swanson, S. J .; J ackson, M.;
Whittern, D. N. J . Antibiot. 1991, 44, 688-690.
(23) Wade, P. A.; D′Ambrosio, S. G.; Rao, J . A.; Shah-Patel, S.; Cole,
D.T; Murray, J . K., J r.; Carroll, P. J . J . Org. Chem. 1997, 62, 3671-
3677 and references therein.
(24) Hydrogenation of rapamycin over Pd/C results in complete
saturation of the triene, but the C29,C30 double bond remains
unchanged. This result supports the argument that the latter olefin
experiences severe steric hindrance. Dr. T. Fehr from our laboratories,
unpublished results.
(20) Rance, M.; Sorensen, O. W.; Bodenhausen, G.; Wagner, G.;
Ernst, R. R.; Wu¨thrich, K. Biochem. Biophys. Res. Commun. 1983, 117,
479-485.
(21) Bothner-By, A. A. J . Am. Chem. Soc. 1984, 106, 811-813.
(22) Bax, A.; Davis, D. G. J . Magn. Reson. 1985, 63, 207-213.

Notes
J . Org. Chem., Vol. 63, No. 26, 1998 10071
Sch em e 1
Sch em e 2
Sch em e 3
chiral ligand, leads to a separable mixture of two regio-
isomeric diols 3 and 4. The stereochemical course of the
reaction, predicted on the basis of both the conformation
of rapamycin and the choice of the chiral ligand, was
confirmed by the virtual absence of reaction when the
pseudoenantiomeric ligand (DHQ)2PHAL was used, as
well as by NMR analysis of the ketal 6. The AD reaction
thus proves to be a very useful tool for selectively
functionalizing the triene region of the complex natural
product rapamycin. Moreover, the chemistry described
herein allows to prepare a rapamycin metabolite, or at
least one of its stereoisomers, in sufficient amounts to
evaluate the effect of triene-dihydroxylation on the
immunosuppressive activity. The biological activities of
the compounds described in this paper and of derivatives
are going to be described elsewhere.
hand, would have resulted in the 26(R)-ketal 6′, of which
the most likely conformation is depicted below. Strong
NOEs found in the ROESY spectrum are indicated by
arrows. These are in complete agreement with the 21-
(R),22(R),26(S) configurations and, thus, with structure
6. Most notably, a NOE from H-21 to the axial H-24 is
observed. This effect would not be expected for the
conformation shown for 6′. Inversely, the latter confor-
mation should result in a NOE between H-21 and H-23,
which is not observed. These NMR data confirm the
structure of ketal 6 and, thereby, the stereochemical
outcome of the dihydroxylation of the C21,C22 olefin.
Exp er im en ta l Section
All reagents and solvents were used as received from the
suppliers without further purification. Rapamycin was obtained
from Novartis Pharma Research, Core Technology Area, Bio-
molecules Production, Basel, Switzerland. Osmium tetraoxide
(2.5 wt % solution in tert-butyl alcohol), (DHQ)2PHAL, and
(DHQD)2PHAL, were purchased from Aldrich. Analytical thin-
layer chromatography was performed with precoated glass-
backed plates (Merck Kieselgel 60 F254). Compounds were
purified by flash chromatography25 using Merck Kieselgel 60
(40-63 µm). 1H and 13C NMR spectra were recorded in CDCl3
at 400 and 100 MHz, respectively.
28,40-O-Bis(tr ieth ylsilyl)r apam ycin (2). To a stirred, cooled
(0 °C) solution of rapamycin (50.0 g, 54.68 mmol) and imidazole
(18.61 g, 273.43 mmol) in 280 mL of DMF was added triethyl-
chlorosilane (20.2 mL, 120.30 mmol) over 5 min. The resulting
yellow solution was stirred at 0 °C for 1.5 h, and the reaction
mixture was quenched by slow addition of 300 mL of saturated
aqueous sodium bicarbonate, which was accompanied by sig-
In summary, we have shown that the AD of 28,40-O-
bis(triethylsilyl)rapamycin 2, using (DHQD)2PHAL as a
(25) Still, W. C.; Kahn, M.; Mitra, A. J . Org. Chem. 1978, 43, 2923-
2925.

10072 J . Org. Chem., Vol. 63, No. 26, 1998
Notes
nificant gas evolution. Then 400 mL of ethyl acetate were added,
and the layers were separated. The organic layer was washed
successively with 200 mL of saturated aqueous sodium bicarbon-
ate, twice with 200 mL of water, and finally twice with 200 mL
of saturated brine. The aqueous solutions were combined and
extracted twice with 500 mL of ethyl acetate. The combined
organic extract was dried over anhydrous sodium sulfate,
filtered, and concentrated under reduced pressure. The residue
was purified by column chromatography on silica gel (60:40
hexanes/methyl tert-butyl ether) to afford bis-O-silylated rapa-
mycin 2 (56.6 g, 90%) as an off-white amorphous solid: 1H NMR
(CDCl3) (∼4:1 mixture of conformers, only signals of major
conformer listed) δ 0.54 (m, 6H), 0.64 (m, 6H), 0.73 (dd, J ) 12.1,
23.9 Hz, 1H), 0.84-1.03 (m, 24H), 1.04-1.13 (m, 9H), 1.13-1.30
(m, 4H), 1.32-1.47 (m, 4H), 1.48-1.72 (m, 8H), 1.68 (s, 3H), 1.78
(s, 3H), 1.73-1.97 (m, 5H), 2.02 (m, 2H), 2.30 (m, 2H), 2.37 (dd,
J ) 2.5, 15.6 Hz, 1H), 2.63 (dd, J ) 8.0, 15.6 Hz, 1H), 2.73 (m,
1H), 2.93 (m, 1H), 3.16 (s, 3H), 3.29 (s, 3H), 3.32-3.43 (m, 3H),
3.45 (s, 3H), 3.72 (m, 2H), 3.81 (m, 1H), 3.88 (d, J ) 6.3 Hz,
1H), 4.16 (d, J ) 5.9 Hz, 1H), 4.73 (s, 1H), 5.06 (m, 1H), 5.26 (d,
J ) 10.2 Hz, 1H), 5.34 (d, J ) 4.7 Hz, 1H), 5.59 (dd, J ) 8.0,
15.0 Hz, 1H), 6.08 (d, J ) 10.8 Hz, 1H), 6.19 (dd, J ) 10.3, 15.0
Hz, 1H), 6.34 (dd, J ) 10.3, 14.6 Hz, 1H), 6.44 (dd, J ) 10.8,
14.6 Hz, 1H); 13C NMR (CDCl3) (∼4:1 mixture of conformers,
only signals of major conformer listed) δ 4.6, 4.7, 5.0, 6.6, 6.7,
6.8, 10.1, 12.3, 13.7, 14.9, 15.4, 16.0, 20.4, 21.4, 25.1, 26.9, 27.3,
31.3, 31.8, 33.0, 33.8, 33.9, 34.1, 35.2, 36.1, 38.5, 38.6, 39.8, 41.7,
42.4, 44.0, 47.0, 51.2, 55.8, 58.0, 58.1, 66.9, 75.6, 76.8, 77.2, 79.2,
84.0, 84.1, 84.7, 98.6, 126.8, 127.1, 129.4, 130.7, 132.8, 135.8,
138.1, 139.1, 166.3, 169.6, 193.4, 208.4, 211.5; IR (KBr) 1109,
1648, 1724 cm-1; MS (FAB, nitrobenzyl alcohol + lithium iodide
matrix) calcd for C63H107NO13Si2Li 1148, found 1148; Anal. Calcd
for C63H107NO13Si2 C, 66.22%; H, 9.44%; N, 1.23%. Found C,
65.80%; H, 9.29%; N, 1.12%; Rf 0.32 (60:40 hexanes/methyl tert-
28.3, 31.7, 31.9, 32.0, 32.9, 33.0, 33.2, 33.4, 33.7, 34.0, 34.4, 34.9,
35.0, 36.0, 38.1, 38.3, 38.7, 39.6, 40.9, 41.0, 42.4, 42.7, 44.1, 46.9,
47.2, 51.5, 55.8, 56.1, 58.1, 58.2, 59.0, 66.5, 70.2, 70.4, 74.7, 75.1,
75.6, 75.7, 75.8, 77.5, 78.2, 83.5, 84.1, 84.3, 85.5, 98.1, 99.1, 126.9,
127.6, 127.9, 128.3, 130.0, 131.5, 133.5, 135.3, 137.1, 137.2, 137.9,
138.2, 166.0, 166.6, 169.6, 169.8, 194.5, 197.1, 208.6, 208.8, 210.9,
211.3; IR (KBr) 1108, 1651, 1722, 3439 cm-1; MS (FAB, ni-
trobenzyl alcohol + lithium iodide matrix) calcd for C63H109NO15
-
Si2Li 1182, found 1182. Anal. Calcd for C63H109NO15Si2 C,
64.30%; H, 9.34%; N, 1.19%. Found C, 63.91%; H, 9.20%; N,
1.12%; Rf 0.35 (70:30 hexanes/acetone); [R]25 ) -54.6 (c 0.25,
D
CHCl3).
4: 1H NMR (CDCl3) (∼4:1 mixture of rotamers, only signals
corresponding to major rotamer listed) δ 0.49-0.72 (m, 13H),
0.84-1.11 (m, 30H), 1.12-2.08 (m, 33H), 2.32 (bd, J ) 13.7 Hz,
1H), 2.65 (dd, J ) 4.5, 13.7 Hz, 1H), 2.76 (m, 2H), 3.00 (bs, 1H),
3.11 (s, 3H), 3.24-3.36 (m, 5H), 3.37-3.62 (m, 6H), 3.63-3.80
(m, 3H), 4.12 (t, J ) 8.23 Hz, 1H), 4.30 (d, J ) 7.9 Hz, 1H), 5.05
(s, 1H), 5.23 (bd, J ) 5.1 Hz, 1H), 5.30 (m, 1H), 5.42 (d, J ) 8.6
Hz, 1H), 5.63 (dd, J ) 7.7, 15.0 Hz, 1H), 5.96 (d, J ) 11.0 Hz,
1H), 6.59 (dd, J ) 11.0, 15.0 Hz, 1H); 13C NMR (CDCl3) (∼4:1
mixture of rotamers, only signals corresponding to major rotamer
listed) δ 4.6, 5.0, 6.7, 6.8, 9.9, 11.7, 13.9, 16.4, 16.7, 17.1, 17.5,
20.6, 25.4, 26.9, 27.3, 27.6, 31.4, 32.2, 32.8, 33.2, 34.0, 34.1, 35.7,
36.6, 37.3, 38.1, 40.7, 44.1, 45.2, 46.8, 51.0, 55.8, 57.9, 58.1, 60.0,
66.7, 73.3, 74.7, 75.5, 75.7, 83.9, 84.2, 84.5, 98.1, 127.8, 129.4,
131.8, 135.9, 138.4, 167.1, 168.9, 191.5, 208.7, 212.8; IR (KBr)
1106, 1652, 1719 3442 cm-1; MS (FAB, nitrobenzyl alcohol +
lithium iodide matrix) calcd for C63H109NO15Si2Li 1182, found
1182. Anal. Calcd for C63H109NO15Si2 C, 64.30%; H, 9.34%; N,
1.19%. Found C, 64.13%; H, 9.20%; N, 1.12%; Rf 0.30 (70:30
hexanes/acetone); [R]25 ) -33.9 (c 0.25, CHCl3).
D
19(R),20(R)-Dih yd r oxyr a p a m ycin (5). To a stirred, cooled
(0 °C) solution of 3 (353 mg, 0.30 mmol) in 3 mL of acetonitrile
was added 0.3 mL of HF-pyridine complex. Stirring was
continued for 2 h at 0 °C, and the reaction was quenched by the
addition of 1 N aqueous sodium bicarbonate. The resulting
mixture was extracted twice with methyl tert-butyl ether. The
combined organic solution was washed with brine, dried over
anhydrous sodium sulfate, filtered, and concentrated under
reduced pressure. The residue was purified by column chroma-
tography on silica gel (50:50-40:60 hexanes/acetone) to give 5
(180 mg, 63%) as a white amorphous solid: 1H NMR (CDCl3)
(∼2:1 mixture of rotamers) δ 0.66 (m, 1H), 0.80 (d, J ) 6.6 Hz,
3H), 0.84-1.87 (m, 37H), 1.96 (bm, 2H), 2.13 (bm, 2H), 2.26 (bm,
2H), 2.50-2.80 (m, 2H), 2.81-3.04 (m, 2H), 3.07-3.23 (m, 6H),
3.30-3.86 (m, 12H), 4.06 (bm, 1H), 4.25 (d, J ) 6.2 Hz, 1H),
4.40 (m, 3H), 5.00 (bm, 1H), 5.16 (bm, 1H), 5.40-5.51 (m, 1.5H),
5.54-5.70 (m, 1.5H), 5.75 (m, 1H), 6.08 (dd, J ) 6.6, 15.8 Hz,
1H); 13C NMR (CDCl3) (∼2:1 mixture of rotamers) δ 10.0, 10.8,
13.6, 14.1, 15.1, 15.6, 15.8, 16.1, 16.2, 16.6, 17.0, 17.8, 20.6, 20.8,
21.0, 21.4, 24.1, 26.6, 26.7, 26.9, 27.3, 28.0, 29.2, 31.3, 31.6, 31.7,
31.8, 32.4, 33.1, 33.2, 33.4, 33.7, 33.9, 34.1, 35.0, 38.1, 38.2, 38.8,
38.9, 39.0, 39.7, 40.0, 40.5, 40.6, 40.8, 44.0, 45.9, 46.4, 49.4, 51.5,
53.7, 55.9, 56.0, 56.6, 58.8, 66.4, 66.7, 70.2, 73.9, 74.5, 75.4, 75.7,
83.5, 84.4, 86.1, 87.9, 98.1, 99.3, 125.2, 125.8, 127.0, 128.6, 129.8,
131.9, 133.1, 135.2, 136.2, 136.5, 138.9, 166.1, 166.9, 169.3, 170.1,
194.6, 197.7, 209.2, 210.4, 213.7, 215.1; IR (KBr) 1093, 1648,
1718, 3447 cm-1; MS (ESI) calcd for C51H81NNaO15 970, found
970. Anal. Calcd for C51H81NO15 C, 64.60%; H, 8.61%; N, 1.48%.
Found C, 64.36%; H, 8.46%; N, 1.21%; Rf 0.37 (40:60 hexanes/
butylether); [R]25 ) -112.1 (c 0.26, CHCl3).
D
28,40-O-Bis(tr ieth ylsilyl)-19(R),20(R)-d ih yd r oxyr a p a m y-
cin (3) an d 28,40-O-Bis(tr ieth ylsilyl)-21(R),22(R)-dih ydr oxy-
r a p a m ycin (4). In a mixture of 10 mL tert-butyl alcohol and
20 mL of water was dissolved, under stirring, potassium ferri-
cyanide (1.98 g, 6.00 mmol), potassium carbonate (0.84 g, 6.00
mmol), methanesulfonamide (0.38 g, 4.00 mmol), (DHQD)2PHAL
(78 mg, 0.10 mmol), and osmium tetraoxide (0.26 mL of a 2.5
wt % solution in tert-butyl alcohol, 0.02 mmol). The resulting
yellow emulsion was cooled to 0 °C, and a solution of 2 (2.28 g,
2.00 mmol) in 10 mL of tert-butyl alcohol was added. Stirring
was continued at 0 °C for 12 h. After this time TLC analysis
indicated complete consumption of the starting material and the
appearance of two new polar products. To the reaction mixture
was added 40 mL of 1 N aqueous sodium bicarbonate solution,
and the resulting emulsion was poured into an Erlenmeyer flask
containing 100 mL of methyl tert-butyl ether. After stirring for
15 min, the organic layer was transferred to another Erlenmeyer
flask, charged with solid sodium sulfite (3.00 g, 23.00 mmol).
Water was added until the salt had dissolved, and the mixture
was stirred for 0.5 h. The layers were separated, and the aqueous
solution was extracted once with methyl tert-butyl ether. The
combined organic solution was washed with saturated aqueous
sodium bicarbonate and saturated brine, then dried over anhy-
drous sodium sulfate, filtered, and concentrated under reduced
pressure. The residue was carefully purified by column chro-
matography on silica gel (80:20-70:30 hexanes/acetone) to afford
the 19,20-dihydroxy derivative 3 (910 mg, 39%) and the 21,22-
dihydroxy derivative 4 (544 mg, 23%) as white amorphous solids.
3: 1H NMR (CDCl3) (∼1:1 mixture of rotamers) δ 0.46-0.65
(m, 12H), 0.67 (m, 1H), 0.82-1.00 (m, 24H), 1.01-1.21 (m, 12H),
1.22-1.50 (m, 6H), 1.51-1.71 (m, 9H), 1.72-1.86 (m, 7H), 1.91-
2.06 (m, 3H), 2.11-2.29 (m, 3H), 2.50 (dd, J ) 4.8, 17.1 Hz, 0.5
H), 2.66-2.93 (m, 3.5H), 3.10-3.22 (m, 5H), 3.31-3.56 (m, 9.5H),
3.62 (dd, J ) 5.3, 10.5 Hz, 0.5H), 3.72-3.82 (m, 2.5H), 4.04 (bm,
1H), 4.26 (2d, J ) 5.8 and 7.2 Hz, 1H), 4.40 (m, 1H), 4.45 (broad,
0.5H), 4.48 (d, J ) 3.8 Hz, 0.5H), 5.15 (m, 1H), 5.22 (d, J ) 4.8
Hz, 0.5H), 5.33 (d, J ) 9.6 Hz, 0.5H), 5.40 (d, J ) 7.9 Hz, 0.5H),
5.43 (d, J ) 9.9 Hz, 0.5H), 5.52 (d, J ) 4.8 Hz, 1H), 5.56 (d, J )
4.8 Hz, 0.5H), 5.63 (d, J ) 8.2 Hz, 0.5H), 5.79 (m, 1.5H); 13C
NMR (CDCl3) (∼1:1 mixture of rotamers, some signals overlap)
δ 4.6, 4.7, 5.0, 6.7, 6.8, 10.1, 10.5, 11.9, 12.2; 15.2, 15.6, 15.8,
16.0, 16.2, 16.3, 16.4; 20.7, 21.4; 21.5, 24.3, 25.0, 26.8, 26.9, 27.3,
acetone); [R]25 ) -28.5 (c 0.26, CHCl3).
D
Keta l (6). To a stirred, cooled (0 °C) solution of 4 (206 mg,
0.17 mmol) in 2 mL of acetonitrile was added 0.2 mL of HF-
pyridine complex. Stirring was continued for 1.5 h at 0 °C, and
the reaction was quenched by the addition of 1 N aqueous
sodium bicarbonate. The resulting mixture was extracted twice
with methyl tert-butyl ether. The combined organic solution was
washed with brine, dried over anhydrous sodium sulfate, filtered,
and concentrated under reduced pressure. The residue was
purified by column chromatography on silica gel (70:30-60:40
hexanes/acetone) to give 6 (118 mg, 75%) as a white amorphous
solid:1H NMR (CDCl3) δ 0.75 (dd, J ) 11.8, 23.8 Hz, 1H), 0.93
(m, 9H), 1.01 (d, J ) 6.4 Hz, 3H), 1.03-1.17 (m, 2H), 1.24 (d, J
) 6.8 Hz, 3H), 1.27-1.81 (m, 24H), 1.88-2.22 (m, 6H), 2.33 (bd,
J ) 13.7 Hz, 1H), 2.44 (dd, J ) 2.3, 16.4 Hz, 1H), 2.66 (s, 1H),
2.85 (dd, J ) 8.1, 16.4 Hz, 1H), 2.99 (m, 1H), 3.17 (s, 3H), 3.26

Notes
J . Org. Chem., Vol. 63, No. 26, 1998 10073
(d, J ) 8.0 Hz, 1H), 3.38-3.48 (m, 8H), 3.51-3.65 (m, 3H), 3.78
(bt, J ) 11.2 Hz, 1H), 3.87 (dd, J ) 4.7, 11.1 Hz, 1H), 4.10 (d, J
) 2.1 Hz, 1H), 4.52 (d, J ) 9.6 Hz, 1H), 4.62 (d, J ) 7.8 Hz, 1H),
5.41 (m, 1H), 5.47 (d, J ) 4.5 Hz, 1H), 5.57 (d, J ) 1.4 Hz, 1H),
5.64 (d, J ) 8.8 Hz, 1H), 6.03 (d, J ) 10.8 Hz, 1H), 6.10 (dd, J
) 9.8, 15.1 Hz, 1H), 6.39 (dd, J ) 10.8, 15.1 Hz, 1H); 13C NMR
(CDCl3) δ 9.6, 15.3, 16.5, 16.7, 17.0, 17.7, 21.0, 25.8, 27.2, 29.2,
31.3, 31.4, 31.6, 32.5, 33.1, 33.3, 33.6, 34.1, 34.4, 34.8, 38.3, 39.4,
39.7, 44.0, 46.3, 51.1, 55.8, 56.5, 60.9, 66.8, 71.2, 71.4, 73.9, 74.8,
77.8, 83.8, 83.9, 84.4, 85.2, 98.6, 111.7, 122.6, 127.5, 128.6, 132.7,
136.3, 142.0, 167.1, 168.5, 190.0, 208.0; IR (KBr) 1099, 1624,
1744, 3503; MS (FAB, nitrobenzyl alcohol + lithium iodide
matrix) calcd for C51H79LiO14 936, found 936. Anal. Calcd for
Ack n ow led gm en t. We thank Dr. J u¨rgen Wagner
for carefully reading the manuscript and Mr. Charles
Quiquerez for performing the mass spectroscopic analy-
ses.
Su p p or tin g In for m a tion Ava ila ble: Copies of 1H and 13C
NMR spectra for compounds 2-6 (10 pages). This material is
contained in libraries on microfiche, immediately follows this
article in the microfilm version of the journal, and can be
ordered from the ACS; see any current masthead page for
ordering information.
C
51H79NO14 C, 65.85%; H, 8.56%; N, 1.51%. Found C, 65.43%;
H, 8.38%; N, 1.32%; Rf 0.21 (60:40 hexanes/acetone); [R]25
-57.0 (c 0.26, CHCl3).
)
D
J O9816820