Mild aryl ether formation in the semisynthesis of the novel macrolide immunosuppressant L-732,531

Full text of DOI:10.1021/jo980451q

J . Org. Chem. 1998, 63, 6721-6726
6721
Sch em e 1
Mild Ar yl Eth er F or m a tion in th e
Sem isyn th esis of th e Novel Ma cr olid e
Im m u n osu p p r essa n t L-732,531
Karel M. J os Brands,* Ulf-H. Dolling, Ronald B. J obson,
George Marchesini, Robert A. Reamer, and
J . Michael Williams
Merck Research Laboratories, Department of Process
Research, P.O. Box 2000, Rahway, New J ersey 07065
Received March 10, 1998
The macrolide FK-506 (1) which was first isolated and
characterized at Fujisawa in 19871 has elicited great
interest among medicinal chemists.2 In an effort to
reduce undesired side effects associated with the clinical
use of FK-506,3 many derivatives of FK-506 have been
prepared and evaluated.4 Recently, we needed a large
quantity of arylated macrolide 3, which has demonstrated
a particularly promising biological profile.
paper describes a practical synthesis of bismuthane 8 as
well as its coupling with the macrolide alcohol 2 to yield
the interesting immunosuppressant 3 after deprotection.
A tert-butyldimethylsilyl protecting group was found
to provide the optimal balance between ease of introduc-
tion in bismuthane 7 and ease of removal from the
sensitive macrolide 14. A two-step preparation of 7
Few methods have been identified that allow the
preparation of highly functionalized aryl alkyl ethers
under mild conditions. A copper-mediated arylation of
alcohols using pentavalent organobismuth reagents has
been described for relatively simple substrates.5 These
reactions showed an unexpected and synthetically useful
regioselectivity in the arylation of 1,2-diols.6 Recently,
this chemistry was used at Merck to prepare various O-32
arylated derivatives of ascomycin (FK-520; 2).7 This
starting from 5-bromoindole (5) was targeted to proceed
via 6 (Scheme 1), a compound known to be crystalline.8
Temporary protection of 2-bromoethanol (9) was achieved
by adding an equimolar amount of 2-methoxypropene
(10). Presumably, residual HBr in 9 catalyzed this
exothermic reaction (Scheme 2). According to 13C NMR
a statistical mixture of 11/12/13 (2/1/1 ratio, respectively)
was obtained in quantitative yield. The use of a mixture
is inconsequential since the corresponding alkylated
products will be hydrolyzed to a single product following
(1) (a) Kino, T.; Hatanaka, H.; Hashimoto, M.; Nishiyama, M.; Goto,
T.; Okuhara, M.; Koshaka, M.; Aoki, H.; Iminaka, H. J . Antibiot. 1987,
40, 1249-1255. (b) Kino, T.; Hatanaka, H.; Miyata, S.; Inamura, N.;
Nishiyama, M.; Yajima, T.; Goto, T.; Okuhara, M.; Koshaka, M.; Aoki,
H.; Ochiai, T. J . Antibiot. 1987, 40, 1256-1265. (c) Tanaka, H.; Kuroda,
A.; Murasawa, H.; Hatanaka, H.; Kino, T.; Goto, T.; Hashimoto, M.;
Taga, T. J . Am. Chem. Soc. 1987, 109, 5031-5033.
(2) Parsons, W. H.; Sigal, N. H.; Wyvratt, M. J . Ann. N. Y. Acad.
Sci. 1993, 685, 22-36.
(3) Parsons, W. H.; Sigal, N. H. In ImmunoPharmaceuticals; Kim-
ball, E. S., Ed.; CRC: Boca Raton, 1996; p 1-30.
(4) Goulet, M. T.; Rupprecht, K. M.; Sinclair, P. J .; Wyvratt, M. J .;
Parsons, W. H. Perspect. Drug Discovery Des. 1994, 2, 145-162.
(5) Reviews: (a) Barton, D. H. R.; Finet, J .-P. Pure Appl. Chem.
1987, 59, 937-946. (b) Abramovitch, R. A.; Barton, D. H. R.; Finet,
J .-P. Tetrahedron 1988, 44, 3039-3071. (c) Finet, J .-P. Chem. Rev.
1989, 89, 1487-1501. (d) Barton, D. H. R. In Heteroatom Chem. (Int.
Conf.); Block, E., Ed.; VCH: New York, 1990; pp 95-104. (e) Dodonov,
V. A.; Gushin, A. V. Russian Chem. Bull. 1993, 42, 1955-1959.
(6) (a) David, S.; Thieffry, A. Tetrahedron Lett. 1981, 22, 5063-5066.
(b) David, S.; Thieffry, A. J . Org. Chem. 1983, 48, 441-447. (c) Barton,
D. H. R.; Finet, J .-P.; Pichon, C. J . Chem. Soc. Chem. Commun. 1986,
65-66.
(7) (a) Sinclair, P. J .; Wong, F.; Wyvratt, M.; Staruch, M. J .; Dumont,
F. Bioorg. Med. Chem. Lett. 1995, 5, 1035-1038. (b) Sinclair, P. J .;
Wong, F.; Staruch, M. J .; Wiederrecht, G.; Parsons, W. H.; Dumont,
F.; Wyvratt, M. Bioorg. Med. Chem. Lett. 1996, 6, 2193-2196.
(8) We thank Dr. P. J . Sinclair of these laboratories for this
observation. The corresponding TBDMS ether is not a crystalline
compound.
S0022-3263(98)00451-4 CCC: $15.00 © 1998 American Chemical Society
Published on Web 08/22/1998

6722 J . Org. Chem., Vol. 63, No. 19, 1998
Notes
Ta ble 1. Solven t Effects in th e Rea ction of 2 w ith 8a
Sch em e 2
entry
solventb
equiv of 8
2 in %
14 in %
15 in %
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
MeOH
NMP
IPA
MeCN
EtOAc
MTBE
DMF
MeNO2
DCM
1.25
1.26
1.22
1.14
1.06
1.23
1.24
1.29
1.18
1.24
1.25
1.21
1.11
1.26
1.47
1.80
86
95
87
79
45
37
62
22
16
36
20
12
17
9
<1
ND
8
ND
ND
ND
<1
2
4
8
6
5
2
5
7
5
8
12
21
17
53
57
57
62
70
72
78
80
78
83
86
80
THF
acetone
toluene
MEK
MEK
MEK
indole alkylation without isolation. Often sodium hy-
dride in DMF is used for deprotonation. However, these
conditions are considered a safety hazard on large scale.9
Therefore, the more practical powdered potassium hy-
droxide10 was used as the base for the alkylation in a
solution of DMSO in THF. Only 10 vol % of DMSO was
found to suffice to ensure regioselective N vs C-3 alky-
lation of 5. Upon completion of the alkylation, water was
added to the reaction mixture. The salt load effected a
phase separation between a THF layer containing the
product and a solution of DMSO in water. Treatment of
the THF solution of crude alkylated products with
aqueous acid followed by workup and crystallization
yielded the crystalline 6 (84% overall isolated yield from
5). A one-pot silylation, halogen-metal exchange, and
quench with bismuth(III) chloride was developed for the
conversion of 6 to 7. Treatment of 6 under standard
silylating conditions in THF (TBDMSCl, Et3N, DMAP)
led to complete conversion to the corresponding silyl
ether. The crystalline triethylamine hydrochloride was
filtered under nitrogen, and the resulting clear THF
solution was directly used for halogen-metal exchange
with n-butyllithium between -75 and -70 °C.11 The
resulting indolyllithium derivative was then directly
quenched with a solution of BiCl3 in THF yielding 7 in
80% yield after crystallization.
3
1
MEK
a
All reactions were run using isolated 8 at rt under ambient
atmosphere for 16 h in the presence of 0.2-0.3 equiv of Cu(OAc)2
using dodecahydrotriphenylene as an internal standard. Yields
were determined by HPLC (ND ) not detected). b All solvents were
used as commercially obtained without further purification and/
or drying. The following less common abbreviations are used:
NMP, N-methyl-2-pyrrolidinone; IPA, 2-propanol; MTBE, methyl
tert-butyl ether; DCM, dichloromethane; MEK, 2-butanone.
Analysis of ascomycin (2) and its reaction products was
quite complicated due to the fact that (in analogy to FK-
506) this compound has been demonstrated to form a
mixture of a major (with structure 2) and two minor
tautomers (one of which has been identified as 4) in
Various protocols were evaluated for the conversion of
7 to a pentavalent bismuthane required for coupling with
2. Most literature methods were deemed inefficient since
they require the isolation of intermediates.12 After
evaluating various more direct protocols, oxidation of 7
with an equimolar quantity of benzoyl peroxide (a stable
and safe crystalline peroxide) was selected as the best
method.13 The procedure yielded 8 in excellent yields.
Although 8 could be isolated as a crystalline solid, direct
use for the arylation reaction in the same pot was
preferred. In such a direct protocol benzoyl peroxide was
slightly undercharged with respect to 7 (0.95 mol equiv)
in order to minimize the formation of an analogue of 8
in which one of the indole ligands had been benzoyloxy-
genated, presumably at the 3′ position.14
solution. Each of these isomers can exist in cis and trans
amide rotamers due to restricted rotation around the
N7-C8 bond (rotamers not only obscure NMR spectra
but also necessitate HPLC analysis at 55 °C at which
temperature they convert to a single species).15,16 The
ascomycin available to us typically contained upon dis-
solution 2, 4, and the third unidentified tautomer in a
ratio of 95/4/1. An additional complicating factor is the
fact that 2, obtained via fermentation from Streptomyces
hygroscopicus va. ascomyceticus, is contaminated with a
lower and a higher methylene homologue (2; R21 ) methyl
and 1-propyl, respectively) which are typically present
at <1% each. As expected, any reaction of this mixture
of starting materials will give a corresponding mixture
of products. To simplify the following discussion our
starting material is presented as consisting solely of 2.
The key coupling step between 2 and pentavalent
organobismuth 8 catalyzed by copper salts gave numer-
(9) DeWall, G. Chem. Eng. News 1982, 60(37), 5, 43.
(10) Milling of potassium hydroxide pellets to a fine powder can
easily be achieved on small scale using a Waring blender.
(11) The intermediate indolyllithium derivative was found to be
rather unstable under the reaction conditions mandating that the
halogen metal exchange be performed as fast as possible at the lowest
possible temperature. Efficient heat transfer during this exothermic
addition posed particular challenges on the larger scale (up to 12 kg
of 6 have been prepared using this procedure) which lowered the yield
for this step by 10-15%.
(12) Freedman, L. D.; Doak, G. O. Chem. Rev. 1982, 82, 15-57.
(13) (a) Challenger, F.; Wilson, V. K.; J . Chem. Soc. 1927, 209-
213. (b) After completion of this work catalysis of the oxidative addition
of benzoyl peroxide was reported: Arnauld, T.; Barton, D. H. R.; Doris,
E. Tetrahedron Lett. 1997, 38, 365-366.
(15) Namiki, Y.; Kihara, N.; Koda, S.; Hane, K.; Yasuda, T. J .
Antibiot. 1993, 46, 1149-55.
(16) Gailliot, F. P.; Natishan, T. K.; Ballard, J . M.; Reamer, R. A.;
Kuczynski, D.; McManemin, G. J .; Egan, R. S.; Buckland, B. C. J .
Antibiot. 1994, 47, 806-11.
(14) Kanaoka, Y.; Aiura, M.; Hariya, S. J . Org. Chem. 1971, 36, 458-
460.

Notes
J . Org. Chem., Vol. 63, No. 19, 1998 6723
ous reaction products according to TLC and HPLC
analysis. The major products of the catalyzed reaction
are the desired 14, accompanied by “bis-indolyzed” 15,17
benzoic acid, and indole derivatives 17 and 18.18 Several
marginally effective while FeCl3 was completely ineffec-
tive. On the basis of all data, 2-butanone and cupric
acetate were selected as the best solvent and catalyst,
respectively. It was found that the amount of the catalyst
did not significantly effect the yield although at <10 mol
% the reaction became markedly slower. Typically, 15-
20 mol % relative to 2 are used in the reaction. The
presence of limited amounts of water or an ambient
atmosphere does not adversely effect the reaction, thus
obviating scrupulous drying of solvent, substrate 2, and
catalyst. On the basis of these considerations, an opti-
mum protocol was developed for large scale synthesis of
3. Under the optimized conditions, the reaction is
completed in about 8 h at 20-25 °C and affords the
desired 14 in 80-90% assay yield. The product can be
isolated from the crude reaction mixture by silica gel
flash chromatography.20 The separation of 14 from the
slightly less polar 15 is difficult to achieve in this manner
in a variety of solvent systems, thus substantially lower-
ing the isolated yield for pure 14 (as compared to the
crude assay yields). However, it was found that upon
desilylation, the resulting 3 and the more polar 16 were
more readily separated. Therefore, chromatography of
the crude coupling reaction mixture over a short silica
gel column was only used to effect separation of the 14/
15 mixture from the bulk of less polar impurities (mainly
17 and 18) and the more polar unconverted 2.21 The
resulting partially purified mixture of 14 and 15 was then
directly used in the deprotection step.
In an effort to gain better understanding of this
coupling reaction the structures of several side products
were elucidated, and attempts were made to close the
various mass balances. A balance for the indole ligands
of 7 in the optimum protocol indicated that only 25 mol
% of these ligands are coupled with the macrocycle
(yielding 14 and 15); approximately 70 and 5 mol % end
up in 17 and 18, respectively (a multitude of other
products could be detected at <2 area% by HPLC). When
substrate 2 is omitted from the reaction, complete disap-
pearance of 8 occurs on a similar time scale! Under these
conditions 50-55 mol % and approximately 10 mol % of
the indole ligands are incorporated in 17 and 18, respec-
tively. In this case there is a more substantial discrep-
ancy in the ligand balance albeit that, again, a multitude
of minor reaction products are detected by HPLC. The
structure of some of these were tentatively assigned on
the basis of HPLC-MS data (19-22). In more mecha-
nistically oriented experiments it was determined that
the 5-H of 17 originated from water present during the
reaction (solvent, catalyst, and substrate all contain
substantial amounts of water). Performing the coupling
reaction in the presence of a small amount of D2O in dry
toluene (2.4 mol equivalents relative to indole ligands)
led to 80% 5-D incorporation in 17. When the coupling
reaction was carried out in the presence of a large amount
of 1,1-diphenylethylene (40 equiv) as a radical trap, the
yield for 14 was not significantly lowered. Therefore, it
appears that free radicals are not involved in this
reaction. Unfortunately, our experiments did not shed
reaction parameters were optimized to achieve the best
yield with respect to the macrocycle. Table 1 shows some
solvent effects on the coupling reaction in the presence
of a catalytic amount of Cu(OAc)2. The conversion of 2
with 1.1-1.3 equiv of the pentavalent organobismuth was
incomplete in all solvents examined. Not surprisingly,
in methanol none of the desired product was formed and
2 was largely recovered (entry 1). However, in the more
hindered 2-propanol (present in 300-fold molar excess
relative to 2) 8% of 14 was formed (entry 3)! This result
is a reflection of the activating role in the arylation by
the neighboring methoxy group.2 The data in Table 1
show little correlation between the nature of the solvent
and the reaction yield. The best solvent examined was
2-butanone (methyl ethyl ketone, MEK). Attempts to
drive the reaction to completion in this solvent by adding
a larger excess of 8 resulted in formation of more 15 with
an optimum yield for 14 at 1.4-1.5 equiv.
The effect of the catalyst was also investigated. As was
observed by the pioneers in this field19 virtually all
copper(I) and copper(II) salts effect the reaction albeit
with a marginally different efficiency. The less soluble
CuO and Cu2O as well as Cu powder itself show an
initiation period before the reaction starts. Some other
salts known to engage in what is presumably a one-
electron-transfer reaction were tested with CoCl2 being
(17) A pure reference sample of this compound was obtained via
silylation of 16 using TBSOTf and 2,6-di-tert-butylpyridine in CH2Cl2
followed by SiO2 chromatography in 65% yield.
(18) Structure assignments for 17 and 18 are based on HPLC-MS
analysis of the crude reaction mixture and was corroborated by NMR
analysis of the chromatographically pure compounds.
(19) (a) Dodonov, V. A.; Gushin, A. V.; Brilkina, T. G. Zh. Obshch.
Khim. 1985, 55, 2514-2519. (b) Barton, D. H. R.; Finet, J .-P.; Pichon,
C. J . Chem. Soc. Chem. Commun. 1986, 65-66.
(20) Still, C. W.; Kahn, M.; Mitra, A. J . Org. Chem. 1978, 50, 2923.
(21) The best chromatographic separation between 14 and 15 was
obtained using acetone/hexanes for elution. However, acetone readily
condenses with the indole moiety of these products under the desily-
lation conditions, thus requiring its rigorous removal. Therefore, ethyl
acetate/hexanes is preferred for elution although the separation
between 14 and 15 is not as good.

6724 J . Org. Chem., Vol. 63, No. 19, 1998
Notes
has been reported for FK-506 by Fujisawa workers in the
patent literature.24
In conclusion, an efficient and practical synthesis of
triindolylbismuthane 7 was developed. This reagent was
then used for the two-step preparation of 3 from asco-
mycin (2) in 55-60% overall yield using an optimized
copper catalyzed chemo- and regioselective arylation
reaction as the key step.
Exp er im en ta l Section
All commercially available materials and solvents were used
as received. The quality of the benzoyl peroxide was checked
by iodometric titration before use. Reaction temperatures were
measured internally, unless indicated otherwise (rt ) 18-23 °C).
Melting points are not corrected. Merck F254 plates were used
for thin-layer chromatography (UV and ceric ammonium mo-
lybdate stain were used for visualization). Preparative chro-
matography was performed using silica gel-60 (230-400 mesh).
The macrolides exist as a mixture of amide rotamers. Therefore,
selected 1H and complete 13C NMR data are reported for the
major rotamer only.25 HRMS data were obtained using liquid
SIMS ionization.
more light on the mechanism of these copper-catalyzed
alcohol arylation reactions.22
In the final step 14 was deprotected under mildly acidic
conditions. The crude product was purified by silica gel
chromatography in which 3 was relatively easily sepa-
rated from the more polar “bis-indolyzed” 16. The overall
isolated yield for the conversion of 2 into 3 via the
through process was typically 55-60%. The amorphous
3 was obtained with a purity of 94 area% according to
HPLC. Two equilibrium species were detected at 0.9 and
3.6 area% (tentatively assigned as the indolyzed analogue
of 4), respectively. In addition, lower and higher meth-
ylene homologues of 3 were detected at 0.5 and 0.4 area%,
respectively (structure based on LC-MS). An impurity
in which the indole moiety of 3 was presumably benzoyl-
oxygenated at C-3′ (structure tentatively assigned based
on LC-MS; vide supra) is typically present at 0.4 area%.
Finally, it is of interest to note that 3 is not photochemi-
cally stable. Upon storing purified 3 in wet acetonitrile
and exposing to light, a new product was formed accord-
ing to HPLC analysis. Some of this product could be
isolated via chromatography and identified on the basis
of detailed NMR analysis as 23 (the stereochemistry of
the aminal proton could not unambiguously be deter-
1-(2-Hydr oxyeth yl)-5-br om oin dole (6). A solution of 2-bro-
moethanol (35 mL; 95%, 493 mmol) in THF (230 mL) was cooled
to 0-5 °C, and 2-methoxypropene (48.8 mL; 97%, 494 mmol)
was added dropwise. The reaction was judged complete when
a 35 ppm signal in the 13C NMR spectrum of a reaction sample
(0.4 mL sample diluted with 0.1 mL of pyridine-d5) had disap-
peared (30 min). The resulting solution was added slowly under
mechanical stirring to a mixture of 5-bromoindole (80.0 g, 408
mmol) and powdered KOH (40.0 g) in a mixture of DMSO (80
mL) and THF (480 mL) at rt. After stirring overnight at rt,
water (180 mL) was added, dissolving the solids in approxi-
mately 10 min. The lower aqueous layer was separated. To the
remaining orange-red layer was added a solution of 1 vol % of
phosphoric acid in water (400 mL). After 4 h toluene (400 mL)
was added, and the mixture was agitated for 10 min. The
organic layer was washed with water (2 × 400 mL) and then
partially concentrated in vacuo (110 mmHg; 35-55 °C internal
temp) to a total volume of 400 mL. This solution was flushed
with 400 mL of fresh toluene. Upon cooling to rt, the crystal-
lization started. After aging for 2 h, the crystals were filtered,
washed with 100 mL of 20% toluene in hexanes, and dried,
yielding 80.0 g of crystalline 6 (80% yield). This material was
99.6 area% pure according to HPLC analysis (Dupont Zorbax
Phenyl, 25 cm × 4.6 mm; 50/50 Water (0.1% H3PO4)/CH3CN;
1.0 mL/min; 220 nm). Mp: 88 °C. 1H NMR (250.1 MHz, CDCl3)
δ 7.75 (d, J ) 1.8, 1H), 7.29 (dd, J ) 8.7, 1.8, 1H), 7.22 (d, J )
8.8, 1H), 7.14 (d, J ) 3.2, 1H), 6.44 (d, J ) 3.3, 1H), 4.21 (t, J )
5.2, 2H), 3.88 (t, J ) 5.2, 2H), 1.62 (s, 1H); 13C NMR (62.9 MHz,
CDCl3) δ 134.7, 130.2, 129.4, 124.4, 123.4, 112.8, 110.8, 101.0,
61.3, 48.7. Anal. Calcd for C10H10BrNO: C, 50.03; H, 4.20; N,
5.83. Found: C, 49.96; H, 4.05; N, 5.69.
Tr is[1-(2-ter t-Bu tyld im eth ylsilyloxyeth yl)in d ol-5-yl]bis-
m u t h a n e (7). A solution of 6 (40.2 g, 167 mmol), tert-
butyldimethylsilyl chloride (97%; 27.8 g, 184 mmol), 4-(dimeth-
ylamino)pyridine (200 mg, 1.6 mmol), and triethylamine (26.0
mL, 186 mmol) in THF (400 mL) was stirred at ambient
temperature for 72 h. After cooling to 0-5 °C, the precipitate
was filtered under N2. The filter cake was washed with THF
(80 mL). The combined filtrates were cooled to -72 °C, and
n-BuLi (120 mL, 1.4 M solution in hexane, 168 mmol) was added
over 5 min, maintaining the temperature below -65 °C. After
the mixture was aged for an additonal 30 min at -72 °C, a
solution of BiCl3 (18.6 g, 59.0 mmol) in THF (100 mL) was added
while maintaining the temperature below -60 °C (10 min).
After warming to rt over 1 h, a slurry of 20 g of cellulose in 100
mined). This general type of photochemical reaction has
been reported in the literature,23 and a similar reaction
(23) Aoyama, H.; Hasegawa, T.; Watabe, M.; Shiraishi, H.; Omote,
Y. J . Org. Chem. 1978, 43, 419-422.
(24) Namiki, Y.; Kihara, N. (Fujisawa Pharmaceutical Co., Ltd.,
J apan). PCT Int. Appl. WO 92/13862, 1992.
(22) Several mechanistic proposals have been made in the literature
(ref 5) but none of these has been tested experimentally. Our experi-
ments were not designed to achieve this goal.
(25) (a) Karuso, P.; Kessler, H.; Mierke, D. F. J . Am. Chem. Soc.
1990, 112, 9434-9436. (b) Rosen, M. K.; Standaert, R. F.; Galat, A.;
Nakatsuka, M.; Schreiber, S. L. Science 1990, 248, 863-866.

Notes
J . Org. Chem., Vol. 63, No. 19, 1998 6725
mL of THF and 13 mL of water was added. The resulting
suspension was stirred for 1 h and then filtered. The filter cake
was washed with 200 mL of THF. The filtrates were concen-
trated to a total volume of 200 mL (160 mmHg, internal T < 24
°C). The residual solvent was displaced in vacuo with 95%
ethanol (600 mL), maintaining the total volume at 200 mL. The
resulting slurry was aged for 1.5 h at 20 °C and filtered. The
solids were washed with 100 mL of 95% ethanol and dried under
a stream of nitrogen for 15 h to give 46.6 g of 7 (80% yield).
This material was 99 area% pure according to HPLC analysis
(Vydac Protein C4, 25 cm × 4.6 mm; 30/70 water (0.005 M Na2-
HPO4 and 0.005 M NaH2PO4)/CH3CN; 1.5 mL/min; 276 nm).
Mp: 135-136 °C. 1H NMR (250.1 MHz, CDCl3) δ 8.08 (s, 1H),
7.58 (dd, J ) 8.3, 0.7, 1H), 7.33 (d, J ) 8.2, 1H), 7.09 (d, J ) 3.2,
1H), 6.37 (d, J ) 2.7, 1H), 4.21 (t, J ) 5.7, 2H), 3.90 (t, J ) 5.7,
2H), 0.84 (s, 9H), -0.11 (s, 6H); 13C NMR (62.9 MHz, CDCl3) δ
144.2, 135.6, 131.3, 130.6, 128.2, 111.4, 100.9, 62.3, 48.5, 25.8,
18.2, -5.7. Anal. Calcd for C48H72N3O3Si3Bi: C, 55.85; H, 7.03;
N, 4.07. Found: C, 55.63; H, 7.05; N, 4.10.
J ) 6.8, 3H), 0.82 (t, J ) 7.2, 3H), 0.79 (s, 9H), -0.16 (s, 6H);
13C NMR (100.6 MHz, CD3CN) δ 212.8, 198.6, 170.4, 166.5,
153.3, 139.4, 133.0, 132.9, 132.3, 130.6, 130.1, 124.7, 114.2, 111.3,
107.4, 100.95, 98.2, 82.9, 82.8, 79.7, 76.2, 74.5, 73.9, 70.4, 63.4,
57.9, 57.5, 56.7, 55.8, 49.6, 49.6, 46.1, 41.0, 39.9, 37.4, 35.5(2C),
34.1, 33.4, 31.2, 30.6, 28.5, 27.2, 26.2, 25.5, 25.1, 22.0, 20.4, 18.8,
16.5, 16.1, 13.7, 12.0, 11.9, 10.1, -5.5. Anal. Calcd for
C59H92N2O13Si: C, 66.51; H, 8.70; N, 2.63. Found: C, 66.32; H,
8.61; N, 2.77.
[3S -[3R *-[E (1S *,3S *,4S *)],4S *,5R *,8S *,9E ,12R *,14R *,
15S*,16R*,18S*,19S*,26a R*]]-8-E t h yl-5,6,8,11,12,13,14,15,
16,17,18,19,24,25,26,26a -h exa d eca h yd r o-5,19-d ih yd r oxy-3-
[2-[4-[[1-(2-h yd r oxyeth yl)-1H-in d ol-5-yl]oxy]-3-m eth oxycy-
cloh exyl]-1-m eth yleth en yl]-14,16-d im eth oxy-4,10,12,18-tet-
r a m e t h yl-15,19-e p oxy-3H -p yr id o[2,1-c][1,4]oxa za cyclo-
tr icosin e-1,7,20,21(4H,23H)-tetr on e (3). To a solution of 14
and 15 (14.11 and 2.36 assay g, respectively; 13.2 and 1.8 mmol,
respectively) in methanol (400 mL) was added 1.5 mL of 1.0 N
hydrochloric acid. After 2 h the solution was partitioned
between a solution of NaH2PO4‚H2O (1.75 g) and Na2HPO4 (1.80
g) in water (500 mL) and isopropyl acetate (400 mL). The
organic layer was washed with brine solution (120 g of NaCl in
400 mL of water) and then concentrated in vacuo (bath tem-
perature 45 °C). The crude product was purified by chroma-
tography using 150 g of SiO2-60 eluting with a gradient of
acetone in hexanes (from 25 vol % to 40 vol %). Fractions were
Bis(ben zoyloxy)tr is[1-(2-ter t-Bu tyld im eth ylsilyloxyeth -
yl)in d ol-5-yl]bism u th (8). Bismuthane 7 (10.5 g, 10.2 mmol)
was solubilized in toluene (100 mL) with heating. The mixture
was filtered through cellulose (10 g). The filtrate was heated
with benzoyl peroxide (3.1 g, 12.5 mmol) to 75 °C for 90 min.
Hexane (200 mL) was added, and the mixture was allowed to
cool to rt overnight. The crystals were filtered and dried in vacuo
1
at 50 °C, yielding 9.0 g of 8 (69% yield). Mp: 174-176 °C; H
combined based on HPLC analyses (Phenomenex CustomSil C18,
NMR (250.1 MHz, CDCl3) δ 8.62 (d, J ) 1.2, 3H), 8.28 (dd, J )
8.7, 1.3, 3H), 7.96 (dd, J ) 8.2, 1.4, 4H), 7.53 (d, J ) 8.8, 3H),
7.29 (m, 6H), 7.18 (d, J ) 3.2, 3H), 6.51 (d, J ) 3.1, 3H), 4.22 (t,
J ) 5.5, 6H), 3.88 (t, J ) 5.5, 6H), 0.78 (s, 27H), -0.18 (s, 18H);
13C NMR (62.9 MHz, CDCl3) δ 171.7, 151.2, 136.7, 134.4, 130.9,
130.2, 130.0, 129.6, 127.6, 127.5, 126.9, 111.6, 102.4, 62.4, 48.8,
25.8, 18.2, -5.7. Anal. Calcd for C62H82N3O7Si3Bi: C, 58.43;
H, 6.48; N, 3.30. Found: C, 58.28; H, 6.26; N, 3.14.
250 × 4.6 mm; water (0.1% H3PO4)/CH3CN gradient from 30/70
at 0 min to 5/95 at 40 min; 1.0 mL/min; 60 °C; 220 nm) and
concentrated in vacuo (bath temperature <40 °C). Bis-adduct
16 can be obtained from the trailing fractions. A solution of the
chromatographed 3 in acetonitrile (200 mL) was washed with
hexanes (2 × 100 mL) and then evaporated to dryness to yield
3 as a slightly yellow foam (9.0 g; 72% from assayed 14; 60%
overall from 2) after drying. The product can be obtained as an
easier to handle amorphous white solid by dissolution in 6 parts
of 2-propanol followed by slow addition into 60 parts of water
containing 0.5 wt % of NaCl.
[3S -[3R *-[E (1S *,3S *,4S *)],4S *,5R *,8S *,9E ,12R *,14R *,
15S*,16R*,18S*,19S*,26a R*]]-3-[2-[4-[[1-[2-[[1,1-Dim et h yl-
eth yl)d im eth ylsilyl]oxy]eth yl]-1H-in d ol-5-yl]oxy]-3-m eth -
oxycycloh e xyl]-1-m e t h yle t h e n yl]-8-e t h yl-5,6,8,11,12,13,
14,15,16,17,18,19,24,25,26,26a -h exa d eca h yd r o-5,19-d ih y-
dr oxy-14,16-dim eth oxy-4,10,12,18-tetr am eth yl-15,19-epoxy-
3H-pyr ido[2,1-c][1,4]oxazacyclotr icosin e-1,7,20,21(4H,23H)-
tetr on e (14). A mixture of 7 (25.0 g, 24.2 mmol) and benzoyl
peroxide (98 wt %; 5.5 g, 22.3 mmol) in 2-butanone (200 mL)
was stirred at ambient temperature for 24 h. Then 2 (98 wt %;
12.9 g, 16.0 mmol) and copper(II) acetate (0.5 g, 2.7 mmol) were
added. The mixture turned green after a few minutes and was
stirred at rt for 24 h. The suspension was diluted with 600 mL
of hexanes and then filtered through cellulose. The filterbed
was thoroughly washed with 2-butanone (150 mL). The com-
bined filtrates were concentrated in vacuo to yield a green syrup.
The crude product was dissolved in 100 mL of a 2-butanone/
hexanes 15/85 mixture and loaded onto 200 g of SiO2-60. The
column was eluted with 4 L of 10% ethyl acetate in hexanes
followed by 2 L of 50% ethyl acetate in hexanes. Fractions were
combined based on HPLC analyses (Dupont Zorbax Rx C8, 250
× 4.6 mm; water (0.1% H3PO4)/CH3CN gradient from 40/60 and
1.0 mL/min at 0 min to 10/90 and 1.5 mL/min at 30 min followed
by isocratic at 10/90 and 1.5 mL/min until 45 min; 60 °C; 220
nm). The combined rich cut containing 14 (14.11 assay g by
HPLC corresponding to 83% assay yield) and 15 (2.36 assay g
by HPLC corresponding to 11% assay yield) was concentrated
in vacuo to a dark foam. This mixture was used for the next
step. An analytically pure sample of 14 could be obtained by
careful chromatography using acetone/hexanes gradient to elute
(Rf for 14 and 15 in hexanes/acetone 3/1 are 0.26 and 0.33,
respectively).
3: Rf ) 0.40 (hexanes/acetone 1/1). [R]405 ) -358° (c ) 2;
acetonitrile). IR (KBr) ν 3840, 3355, 1744, 1706, 1649, 1483
cm-1
.
1H NMR (400.1 MHz, CD3CN) δ 7.28 (d, J ) 8.7, 1H),
7.18 (d, J ) 3.2, 1H), 7.11 (d, J ) 2.4, 1H), 6.82 (dd, J ) 8.7, 2.4,
1H), 6.32 (dd, J ) 3.2, 0.8, 1H), 5.18, (m, 2H), 4.95 (br d, J )
9.9, 1H), 4.52 (br m, J ) 5.2, 1H), 4.41 (d, J ) 1.2, 1H), 4.30 (br
d, J ) 13.5, 1H), 4.17 (t, J ) 5.6, 2H), 4.00 (m, 1H), 3.89 (m,
1H), 3.78 (q, J ) 5.6, 2H), 3.60 (dd, J ) 9.5, 1.2, 1H), 3.56 (ddd,
J ) 11.1, 3.6, 1.2, 1H), 3.40 (s, 3H), 3.40-3.25 (m, 3H), 3.33 (s,
3H), 3.28 (s, 3H), 2.99 (d, J ) 4.4, 1H), 2.89 (m, 2H), 2.74 (dd, J
) 14.7, 4.8, 1H), 2.40 (m, 1H), 2.33-1.02 (m, 24H), 1.61 (d, J )
0.8, 3H), 1.59 (d, J ) 1.2, 3H), 0.90 (d, J ) 6.3, 3H), 0.88 (d, J
) 6.7, 3H), 0.86 (d, J ) 7.1, 3H), 0.82 (t, J ) 7.5, 3H); 13C NMR
(100.6 MHz, CD3CN) δ 212.8,198.6, 170.4, 166.4, 153.4, 139.4,
132.9, 132.9, 132.3, 130.3, 130.1, 124.7, 114.3, 111.1, 107.4, 101.1,
98.2, 83.0, 82.8, 79.7, 76.2, 74.5, 73.9, 70.4, 61.9, 57.9, 57.5, 56.7,
55.8, 49.6, 49.6, 46.1, 41.0, 39.9, 37.4, 35.5, 34.1, 33.4, 31.2, 30.7,
28.5, 27.2, 25.5, 25.1, 22.0, 20.3, 16.7, 16.1, 13.7, 12.0, 10.1. (+)-
LSI-HRMS calcd for C53H78N2O13 m/z 950.5504, obsd m/z
950.5530. Anal. Calcd for C53H78N2O13: C, 66.92; H, 8.27; N,
2.95. Found: C, 66.47; H, 8.15; N, 3.02.
16: Rf ) 0.27 (hexanes/acetone 1/1). [R]405 ) -445° (c ) 2;
acetonitrile). 1H NMR (400.1 MHz, CD3CN) δ 7.29 (d, J ) 8.7,
1H), 7.28 (d, J ) 8.7, 1H), 7.19 (d, J ) 3.2, 1H), 7.17 (d, J ) 3.2,
1H), 7.12 (d, J ) 2.4, 1H), 7.06 (d, J ) 2.4, 1H), 6.81 (dd, J )
8.7, 2.4, 1H), 6.78 (dd, J ) 8.7, 2.4, 1H), 6.34 (dd, J ) 3.2, 0.8,
1H), 6.33 (dd, J ) 3.2, 0.8, 1H), 5.20 (d, J ) 7.5, 1H), 4.99 (d, J
) 8.7, 1H), 4.81, dq, J ) 10.3, 1.2, 1H), 4.62 (d, J ) 5.2, 1H),
4.55 (d, J ) 0.8, 1H), 4.49 (m, 1H), 4.32 (br d, J ∼ 13, 1H), 4.17
(t, J ) 5.6, 2H), 4.14 (t, J ) 5.6, 2H), 3.79 (m, 2H), 3.74 (m, 2H),
3.70 (dd, J ) 9.5, 1.6, 1H), 3.59 (m, 1H), 3.341 (s, 3H), 3.339 (s,
3H), 3.31 (s, 3H), 2.54 (dd, J ) 14.3, 4.8, 1H), 1.72 (d, J ) 0.8,
3H), 1.45 (d, J ) 1.2, 3H), 1.03 (d, J ) 6.7, 3H), 0.91 (d, J ) 6.3,
3H), 0.85 (d, J ) 6.3, 3H), 0.77 (t, J ) 7.5, 3H); 13C NMR (100.6
MHz, CD3CN) δ 211.2, 198.8, 170.4, 166.5, 153.3, 152.5, 139.7,
135.5, 133.1, 132.9, 132.1, 130.6, 130.4, 130.11, 130.09, 124.8,
114.2, 114.1, 111.4, 111.2, 107.3, 107.2, 101.11, 101.07, 98.5,
82.72, 82.68(2C), 77.8, 76.3, 74.4, 73.9, 61.9(2C), 57.9, 57.7, 57.2,
56.6, 56.3, 50.0, 49.58, 49.55, 45.1, 39.8, 39.6, 37.2, 35.7, 35.5,
14: [R]405 ) -322° (c ) 2; acetonitrile). 1H NMR (399.9 MHz,
CD3CN) δ 7.27 (br d, J ) 8.8, 1H), 7.16 (d, J ) 3.2, 1H), 7.10 (d,
J ) 2.4, 1H), 6.81 (dd, J ) 8.8, 2.4, 1H), 6.30 (dd, J ) 3.2, 0.8,
1H), 5.19 (d, J ) 4.4, 1H), 5.17 (d, J ) 9, 1H), 4.96 (d, J ) 10.0,
1H), 4.59 (br d, J ) 4.8, 1H), 4.40 (d, J ) 1.2, 1H), 4.30 (br d, J
) 13, 1H), 4.18 (t, J ) 5.2, 2H), 4.00 (m, 1H), 3.88 (t, J ) 5.2,
2H), 3.60 (dd, J ) 9.6, 1.2, 1H), 3.56 (m, 1H), 3.40 (s, 3H), 3.34
(s, 3H), 3.29 (s, 3H), 2.99 (d, J ) 4.4, 1H), 2.89 (td, J ) 13.3, 3.2,
1H), 2.74 (dd, J ) 14.5, 4.8, 1H), 1.61 (d, J ) 0.8, 3H), 1.59 (d,
J ) 1.2, 3H), 0.90 (d, J ) 6.4, 3H), 0.88 (d, J ) 6.4, 3H), 0.86 (d,

6726 J . Org. Chem., Vol. 63, No. 19, 1998
Notes
35.3, 33.3, 30.6, 30.5, 28.8, 26.7, 25.2, 25.0, 21.6, 19.7, 16.4, 16.1,
12.4, 11.9, 11.1. Anal. Calcd for C63H87N3O14: C, 68.15; H, 7.90;
N, 3.78. Found: C, 67.31; H, 7.86; N, 3.66.
0.81 (d, J ) 7.1, 3H); 13C NMR (100.6 MHz, CD3CN) δ 213.5,
171.7, 169.4, 153.4, 140.2, 132.9, 132.4, 130.4, 130.1, 129.6, 123.7,
114.3, 111.2, 107.5, 101.0, 98.7, 88.0, 83.1, 82.9, 77.1, 77.0, 76.8,
74.8, 72.5, 70.0, 58.0, 57.3, 56.6, 55.5, 53.7, 49.6, 49.2, 44.4, 40.5,
37.5, 36.0, 35.4, 33.4, 32.6, 31.8, 31.3, 30.6, 29.7, 26.6, 26.0, 25.6,
19.7, 19.2, 16.2, 15.9, 14.7, 12.1, 10.4.
P h otop r od u ct 23: 1H NMR (400.1 MHz, CD3CN) δ 7.28 (d,
J ) 8.7, 1H), 7.18 (d, J ) 3.2, 1H), 7.11 (d, J ) 2.4, 1H), 6.82
(dd, J ) 8.7, 2.4, 1H), 6.32 (dd, J ) 3.2, 0.8, 1H), 5.18 (br s, 1H),
5.15 (ddd, J ) 10.3, 4.4, 1.2, 1H), 4.94 (d, J ) 9.1, 1H), 4.77 (d,
J ) 2.0, 1H), 4.61 (d, J ) 6.7, 1H), 4.33 (d, J ) 1.2, 1H), 4.17 (t,
J ) 5.6, 2H), 3.99 (m, 1H), 3.78 (q, J ) 5.6, 2H), 3.60 (dd, J )
9.5, 2.2, 1H), 3.48 (m, 1H), 3.40 (s, 3H), 3.31 (s, 3H), 3.27 (s,
3H), 2.87 (t, J ) 5.6, 1H), 2.65 (dd, J ) 17.4, 1.6, 1H), 2.31 (dd,
J ) 17.4, 9.5, 1H), 1.65, (d, J ) 0.8, 3H), 1.63 (d, J ) 0.8, 3H),
0.92 (d, J ) 6.7, 3H), 0.84 (d, J ∼ 7, 3H), 0.83, (t, J ) 7.1, 3H),
Ack n ow led gm en t. We thank Dr. Larry Colwell, J r.,
for collecting the HRMS data. We thank Pat Gailliot
and J ames Corry for providing us with ascomycin.
J O980451Q