Total synthesis of formamicin

Article abstract of DOI:10.1021/ja048493l

The enantioselective total synthesis of the cytotoxic plecomacrolide natural product formamicin (1) is described. Key aspects of this synthesis include the efficient transacetalation reactions of MOM ethers 28 and 38 to form the seven-membered formyl acet

Full text of DOI:10.1021/ja048493l

Published on Web 07/09/2004
Total Synthesis of Formamicin
Timothy B. Durham, Nicolas Blanchard, Brad M. Savall, Noel A. Powell, and
William R. Roush*
Contribution from the Department of Chemistry, UniVersity of Michigan, 930 North UniVersity,
Ann Arbor, Michigan 48109
Received March 15, 2004; E-mail: roush@umich.edu
Abstract: The enantioselective total synthesis of the cytotoxic plecomacrolide natural product formamicin
(1) is described. Key aspects of this synthesis include the efficient transacetalation reactions of MOM ethers
28 and 38 to form the seven-membered formyl acetals 29 and 39, a late-stage Suzuki cross-coupling reaction
of the highly functionalized vinyl boronic acid 6 and vinyl iodide 7, a highly â-selective glycosidation reaction
of â-hydroxy ketone 4 with 2,6-dideoxy-2-iodoglucopyranosyl fluoride 3, and the global desilylation of
penultimate intermediate 77 mediated by in situ generated Et3N‚2HF.
Introduction
The biological activity of many members of the pleco-
macrolide family originates from their ability to act as selective
The plecomacrolides (formerly known as the hygrolidins)1
are a large family of natural products, some representative
examples of which include the bafilomycins,2 leucanicidin,3,4
hygrolidin,5 the concanamycins,6 and FD-8917-9 (Figure 1).
These macrolides display potent insecticidal,3 antiparasitic,10,11
antifungal,2 antibacterial,2 immunosuppressive,12 cytotoxic,7 and
anthelmintic13 activities. The family name, plecomacrolide, was
inspired by the hemiketal side chain of these molecules and
originates from the Greek word “pleco,” meaning “I fold.”14
Members of this family are typified by a 16- or 18-membered
macrolactone containing four olefin units joined to a side chain
which, in most members of the family, contains a six-membered
hemiketal unit that is separated from the macrocycle by a three
carbon linker (Figure 1). This lactone/linker/hemiketal structural
motif forms a distinctive intramolecular hydrogen-bonding
network.14 The presence of this hydrogen bonding network is
important to the biological activity of these molecules,15
although it is not essential.16
inhibitors of vacuolar H+-ATPases (V-ATPases).17,18 V-ATPases
are ubiquitous within eukaryotic organisms and utilize energy
derived from ATP hydrolysis to maintain a proton gradient for
the acidification of organelles.19 Because of their highly specific
inhibition of V-ATPases, the bafilomycins and concanomycins
have proven to be useful tools for studying cellular processes
involving V-ATPases.19 Further, the inhibition of the V-ATPases
of osteoclasts has been identified as a potential mechanism to
prevent bone resorption, the major indication of postmenopausal
osteoporosis.18,20
Due to the potent and diverse biological activity of the
plecomacrolides, this class of natural products has been
subjected to substantial efforts directed toward their synthesis.
Total syntheses of bafilomycin A1 have been reported by our
laboratory21,22 as well as by those of Evans,23 Toshima,24-26
and Hanessian,27 while Marshall28 has reported the total
synthesis of bafilomycin V1. In addition, total syntheses of
concanamycin F (the aglycon of concanamycin A) have been
reported by Toshima29,30 and Paterson,31 while Yonemitsu32 has
reported a total synthesis of hygrolidin.
(1) Corey, E. J.; Ponder, J. W. Tetrahedron Lett. 1984, 25, 4325.
(2) Werner, G.; Hagenmaier, H.; Drautz, H.; Baumgartner, A.; Za¨hner, H. J.
Antibiot. 1984, 37, 110.
(3) Isogai, A.; Sakuda, S.; Matsumoto, S.; Ogura, M.; Furihata, K.; Seto, H.;
Suzuki, A. Agric. Biol. Chem. 1984, 48, 1379.
(15) Dro¨se, S.; Bindseil, K. U.; Bowman, E. J.; Siebers, A.; Zeeck, A.; Altendorf,
K. Biochemistry 1993, 32, 3902.
(4) O’Shea, M. G.; Rickards, R. W.; Rothschild, J. M. J. Antibiot. 1997, 50,
1073.
(16) Yoshimoto, Y.; Jyojima, T.; Arita, T.; Ueda, M.; Imoto, M.; Matsumura,
S.; Toshima, K. Bioorg. Med. Chem. Lett. 2002, 12, 3525.
(5) Seto, H.; Akao, H.; Furihata, K.; Otake, N. Tetrahedron Lett. 1982, 23,
2667.
(17) Dro¨se, S.; Altendorf, K. J. Exp. Biol. 1997, 200, 1.
(18) Beutler, J. A.; McKee, T. C. Curr. Med. Chem 2003, 10, 787.
(19) Finbow, M. E.; Harrison, M. A. Biochem. J. 1997, 324, 697.
(20) Gagliardi, S.; Nadler, G.; Consolandi, E.; Parini, C.; Morvan, M.; Legave,
M.-N.; Belfiore, P.; Zocchetti, A.; Clarke, G. D.; James, I.; Nambi, P.;
Gowen, M.; Farina, C. J. Med. Chem. 1998, 41, 1568.
(6) Kinashi, H.; Someno, K.; Sakaguchi, K.; Higashijima, T.; Miyazawa, T.
Tetrahedron Lett. 1981, 39, 3861.
(7) Seki-Asano, M.; Okazaki, T.; Yamagishi, M.; Sakai, N.; Hanada, K.;
Mizoue, K. J. Antibiot. 1994, 47, 1226.
(8) Seki-Asano, M.; Tsuchida, Y.; Hanada, K.; Mizoue, K. J. Antibiot. 1994,
47, 1234.
(21) Scheidt, K. A.; Tasaka, A.; Bannister, T. D.; Wendt, M. D.; Roush, W. R.
Angew. Chem., Int. Ed. 1999, 38, 1652.
(9) Eguchi, T.; Kobayashi, K.; Uekusa, H.; Ohashi, Y.; Mizoue, K.; Mat-
sushima, Y.; Kakinuma, K. Org. Lett. 2002, 4, 3383.
(10) Otoguro, K.; Liu, Z.-X.; Fukuda, K.; Li, Y.; Iwai, Y.; Tanaka, H.; Omura,
S. J. Antibiot. 1988, 41, 573.
(22) Scheidt, K. A.; Bannister, T. D.; Tasaka, A.; Wendt, M. D.; Savall, B. M.;
Fegley, G. J.; Roush, W. R. J. Am. Chem. Soc. 2002, 124, 6981.
(23) Evans, D. A.; Calter, M. A. Tetrahedron Lett. 1993, 34, 6871.
(24) Toshima, K.; Jyojima, T.; Yamaguchi, H.; Noguchi, Y.; Yoshida, T.;
Murase, H.; Nakata, M.; Matsumura, S. J. Org. Chem. 1997, 62, 3271.
(25) Toshima, K.; Jyojima, T.; Yamaguchi, H.; Murase, H.; Yoshida, T.;
Matsumura, S.; Nakata, M. Tetrahedron Lett. 1996, 37, 1069.
(26) Toshima, K.; Yamaguchi, H.; Jyojima, T.; Noguchi, Y.; Nakata, M.;
Matsumura, S. Tetrahedron Lett. 1996, 37, 1073.
(11) Lacey, E.; Gill, J. H.; Power, M. L.; Rickards, R. W.; O’Shea, M. G.;
Rothschild, J. M. Int. J. Parasitol. 1995, 25, 349.
(12) Kinashi, H.; Someno, K.; Sakaguchi, K.; Higashijima, T.; Miyazawa, T.
Tetrahedron Lett. 1981, 39, 3857.
(13) Hensens, O. D.; Monaghan, R. L.; Huang, L.; Albers-Scho¨nberg, G. J.
Am. Chem. Soc. 1983, 105, 3672.
(27) Hanessian, S.; Ma, J.; Wang, W. J. Am. Chem. Soc. 2001, 123, 10200.
(28) Marshall, J. A.; Adams, N. D. J. Org. Chem. 2002, 67, 733.
(14) Bindseil, K. U.; Zeek, A. Liebigs Ann. Chem. 1994, 305.
9
10.1021/ja048493l CCC: $27.50 © 2004 American Chemical Society
J. AM. CHEM. SOC. 2004, 126, 9307-9317
9307

A R T I C L E S
Durham et al.
lines, having IC50 values of 0.15-0.13 ng/mL against L1210,
EL4, and P388 leukemia cell lines and 3.45 ng/mL against S180
sarcoma cells.33 Coupled with its impressive cytotoxicity,
formamicin (1) contains a variety of synthetically challenging
architectural elements including the seven-membered formyl
acetal unit imbedded within the 16-membered macrocycle, as
well as a hemiketal side chain that contains a 2,6-dideoxy-â-
glucopyranoside unit. Based on the potent cytotoxicity and
unique architectural challenges of formamicin, we initiated a
program toward the total synthesis of this natural product and
have published two preliminary reports detailing our synthesis
of the aglycon, formamicinone (2).35,36 In this paper, we provide
a full account of these efforts as well as the completion of the
total synthesis of 1 itself.
Synthetic Strategy. In developing a strategy for the synthesis
of formamicin (1), we recognized the need to address the two
most compelling synthetic challenges in the molecule: (i)
generation of the highly functionalized seven-membered formyl
acetal unit and (ii) introduction of the glycoside onto the C(21)
hydroxyl. We hypothesized that the formation of the seven-
membered acetal could be achieved through an intramolecular
transacetalation reaction (see 20 f 29). However, introduction
of the 2,6-dideoxy-â-glucopyranoside unit proved to be a more
challenging problem (as subsequently described).
Previous efforts in our group on the glycosidation of hydroxy
hemiketal acceptors (deriving from â,δ-bishydroxy ketones)
suggested that a synthetic strategy requiring the introduction
of the carbohydrate onto the hemiketal directly would be
unsuccessful due to the instability of these acceptors to Lewis
acidic glycosidation reaction conditions.37 Thus, we envisaged
that the carbohydrate must be introduced onto the acyclic form
of the â-hydroxy ketone side chain (i.e., 4, Scheme 1). This
bond formation represents an especially difficult challenge in
carbohydrate chemistry due to the reduced nucleophilicity of
hydrogen bound â-hydroxy ketones and the high sensitivity of
these systems to Lewis acids.38 In this context, donor 3 was
developed in these laboratories specifically for the â-selective
glycosidation of â-hydroxy ketone acceptors (Scheme 1).38
Therefore, we designed a synthetic strategy in which the fully
protected aglycon 436 could be used as an acceptor in a
diastereoselective glycosidation reaction with 2,6-dideoxy-2-
iodoglucopyranosyl donor 3 (Scheme 1). We anticipated that
aglycon 436 could be accessed from the coupling of three
fragments: aldehyde 5,39 vinyl boronic acid 6 (a key intermedi-
ate in our synthesis of bafilomycin A1),22,40 and vinyl iodide
7.35,36
Figure 1. Selected plecomacrolide natural products.
Those members of the plecomacrolide family which contain
substitution on the hemiketal pyran hydroxyl, either in the form
of acyl or carbohydrate groups, such as occur in hygrolidin and
concanamycin A, represent particularly difficult synthetic chal-
lenges (Figure 1). These substituents are prone to undergo
elimination if the pyran ring opens to reveal the latent â-alkoxy
ketone functionality. While Yonemitsu has successfully ad-
dressed the introduction of acyl groups in these systems,32 to
the best of our knowledge, no glycosidated member of this
family has been successfully prepared.
In the synthetic direction, our plan was to join fragments 6
and 7 via a late-stage Suzuki cross-coupling reaction and then
to append aldehyde 5 via a methyl ketone aldol reaction (Scheme
1). We anticipated that the seven-membered formyl acetal unit
of 7 could be prepared through an intramolecular transacetalation
reaction. A more pressing concern was to devise a strategy for
introducing the four contiguous stereocenters of vinyl iodide 7.
We hypothesized that this could be accomplished through either
Formamicin (1) is a recently reported plecomacrolide (Scheme
1). This natural product was isolated from the culture broth of
Saccharothrix sp. MK27-91F2.33,34 Formamicin (1) displays
impressive cytotoxicity against a variety of murine tumor cell
(29) Toshima, K.; Jyojima, T.; Miyamoto, N.; Katohno, M.; Nakata, M.;
Matsumura, S. J. Org. Chem. 2001, 66, 1708.
(30) Jyojima, T.; Miyamoto, N.; Katohno, M.; Nakata, M.; Matsumura, S.;
Toshima, K. Tetrahedron Lett. 1998, 39, 6007.
(31) Paterson, I.; Doughty, V. A.; McLeod, M. D.; Trieselmann, T. Angew.
Chem., Int. Ed. 2000, 39, 1308.
(35) Powell, N. A.; Roush, W. R. Org. Lett. 2001, 3, 453.
(36) Savall, B. M.; Blanchard, N.; Roush, W. R. Org. Lett. 2003, 5, 377.
(37) Blanchard, N.; Roush, W. R. Unpublished results.
(38) Blanchard, N.; Roush, W. R. Org. Lett. 2003, 5, 81.
(39) Bestmann, H. J.; Liepold, B.; Kress, A.; Hoffmann, A. Chem.sEur. J. 1999,
5, 2984.
(32) Makino, K.; Nakajima, N.; Hashimoto, S.-i.; Yonemitsu, O. Tetrahedron
Lett. 1996, 37, 9077.
(33) Igarashi, M.; Kinoshita, N.; Ikeda, T.; Nakagawa, E.; Hamada, M.; Takeuchi,
T. J. Antibiot. 1997, 50, 926.
(34) Igarashi, M.; Nakamura, H.; Naganawa, H.; Takeuchi, T. J. Antibiot. 1997,
50, 932. Correction: J. Antiobiot. 1998, 51, C1.
(40) Savall, B. M.; Powell, N. A.; Roush, W. R. Org. Lett. 2001, 3, 3057.
9
9308 J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004

Total Synthesis of Formamicin
A R T I C L E S
Scheme 1
Scheme 2
a diastereoselective aldol reaction41-44 of fragments 8 and 9 or
via a chelate-controlled addition of Grignard reagent 10 to
R-alkoxy ketone 11 (Scheme 2).
hydroxy groups were then differentiated by conversion to the
p-methoxybenzylidene acetal and subsequent selective, reductive
opening via treatment with DIBAL-H.49 This provided primary
alcohol 17 in 67% overall yield from aldol product 16.
Oxidation of alcohol 17 under standard Swern conditions50
provided the requisite aldehyde 9 for use in the lactate aldol
reaction (Scheme 3). Treatment of aldehyde 9 with the lithium
enolate of MOM-protected lactate ester 835 at -78 °C in THF
provided an inseparable mixture of aldol products with a
diastereomeric ratio of 7.1:1, favoring the desired product 18.51
Reduction of the diastereomeric mixture with LiAlH4 provided
diols 20 and 21, which were separable by flash chromatography,
in a combined overall yield of 67% from alcohol 17. Further
experimentation showed a striking dependence of the dia-
stereoselectivity of the aldol reaction on the reaction temperature
(Table 1). In the optimized protocol, aldehyde 9 was dissolved
in THF and added slowly to a -95 °C solution of the lactate
enolate, such that the internal temperature of the mixture was
not allowed to rise above -92 °C. By carefully controlling the
reaction conditions in this way, the diastereoselectivity of the
aldol coupling was increased to 11:1 with a concomitant modest
increase in the overall reaction efficiency.51
Results and Discussion
First Generation Synthesis of Fragment 7. Our first
generation synthesis of vinyl iodide 7 focused on the aldol
reaction of O-alkyl lactate ester 8 and aldehyde 9 to generate
the C(6)-C(7)-C(8)-C(25) stereotetrad. Toward this end,
2-methyl-2-propene-1-ol (12) was silylated and subjected to a
retro-Brook rearrangement (Scheme 3).45 The resulting alkoxide
was then quenched with acetic anhydride to give allyl acetate
13.46 Treatment of ester 13 with TBSOTf and Hunig’s base
promoted an Ireland-Claisen rearrangement47 to give the
corresponding silyl ester which was then hydrolyzed to give
carboxylic acid 14 in 70% overall yield from alcohol 12. Acid
14 was then converted to the mixed pivaloyl anhydride and
treated with N-lithio-(R)-4-benzyl-2-oxazolidinone to provide
acyl oxazolidinone 15. Asymmetric aldol reaction of 15 with
hexanal provided the syn-aldol product 16 in 97% yield with
>95:5 diastereoselectivity. The acyl unit of 16 was then reduced
48
to the primary alcohol with NaBH4 in 78% yield. The two
At this point, our attention turned to the construction of the
seven-membered formyl acetal subunit through an intramolecu-
lar transacetalation reaction. During initial efforts directed at
the synthesis of fragment 7, we prepared intermediate 2235
(Scheme 4, eq 1). Upon attempting to silylate the free C(7)
(41) Heathcock, C. H.; Hagen, J. P.; Jarvi, E. T.; Pirrung, M. C.; Young, S. D.
J. Am. Chem. Soc. 1981, 103, 4972.
(42) Heathcock, C. H.; Pirrung, M. C.; Young, S. D.; Hagen, J. P.; Jarvi, E. T.;
Badertscher, U.; Ma¨rki, H.-P.; Montgomery, S. J. Am. Chem. Soc. 1984,
106, 8161.
(43) Heathcock, C. H.; Young, S. D.; Hagen, J. P.; Pilli, R.; Badertscher, U. J.
Org. Chem. 1985, 50, 2095.
(44) Hoagland, S.; Morita, Y.; Bai, D. L.; Ma¨rki, H.-P.; Kees, K.; Brown, L.;
Heathcock, C. H. J. Org. Chem. 1988, 53, 4730.
(48) Prashad, M.; Har, D.; Kim, H.-Y.; Repic, O. Tetrahedron Lett. 1998, 39,
7067.
(49) Takano, S.; Akiyama, M.; Sato, S.; Ogasawara, K. Chem. Lett. 1983, 1593.
(50) Mancuso, A. J.; Swern, D. Synthesis 1981, 165.
(51) Diastereoselectivity was determined by 1H NMR analysis of the crude
reaction mixture.
(45) Danheiser, R. L.; Fink, D. M.; Okano, K.; Tsai, Y.-M.; Szczepanski, S.
W. J. Org. Chem. 1985, 50, 5393.
(46) Calter, P.; Evans, D. A. Ph.D. Thesis, Harvard University, 1993.
(47) Ireland, R. E.; Varney, M. D. J. Am. Chem. Soc. 1984, 106, 3668.
9
J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004 9309

A R T I C L E S
Durham et al.
Scheme 3 a
Scheme 4 a
a Conditions: (a) (i) n-BuLi, THF, -78 °C, 15 min; (ii) TMSCl, 2 h;
(iii) t-BuLi, -78 °C, 2 h; then warm to -40 °C; (iv) Ac2O; -40 f 23 °C;
(b) TBSOTf, i-Pr2NEt, CH2Cl2; (c) LiOH, H2O/MeOH; (d) PivCl, Et3N,
Et2O; (e) LiXc, THF, -78 °C; (f) Bu2BOTf, hexanal, CH2Cl2, -78 to 0
°C; (g) NaBH4, THF, H2O, 23 °C; (h) 4-MeOPhCH(OMe)2, PPTS, CH2Cl2,
23 °C; (i) DIBAL-H, PhCH3, 23 °C; (j) oxalyl chloride, DMSO, Et3N,
CH2Cl2, -78 °C to 0 °C; (k) 8, LDA, THF (see Table 1); (l) LAH, THF,
23 °C.
Table 1. Diasteroselectivity of the Lactate Aldol Reaction of 8 and
9 as a Function of Temperature
temperature
yield (20 + 21)
entry
(°C)
(%)
dr (18:19)
1
2
3
-95 f -78
-78
-60
76
67
79
11:1
7.1:1
5.5:1
alcohol of 22, we observed that treatment of 22 with 4 equiv of
TESOTf directly provided seven-membered formyl acetal 25
in 78% yield!35 Small amounts of the five-membered acetals
26 (5%) and 27 (5%) were also formed under these conditions.
Treatment of 22 with 1 equiv of TESOTf showed that the initial
intermediate formed was C(7) silyl ether 23. Rigorous experi-
mentation showed that the transacetalation reaction was pro-
moted by trace triflic acid impurities in the TESOTf.35 The
reaction presumably proceeds by way of oxonium ion 24, which
suffers loss of the p-methoxybenzyl cation en route to 25.
Hoping to capitalize on this efficient approach to formation
of the seven-membered acetal, we set out to investigate the
generality of this reaction with substrates 28 and 30, either of
which could be converted to fragment 7 (Scheme 4). Treatment
of both vinyl silanes 28 and 30 with TESOTf provided the
a Conditions: (a) TESOTf (4.0 equiv), 2,6-lutidine, CH2Cl2, 23 °C; (b)
PivCl, Py, CH2Cl2; (c) Ac2O, Py, CH2Cl2; (d) TESOTf, 2,6-lutidine, CH2Cl2,
reflux, 24 h.
desired seven-membered acetals 29 and 31, albeit with strikingly
different efficiencies (Scheme 4, eqs 2 and 3). Best results were
obtained with pivaloyl ester 28, which underwent the trans-
acetalation reaction to give 29 in 79% yield. However, we
recognized that achieving a similar cyclization on the lactate
aldol product 18 (Scheme 3) would shorten the synthetic route
to fragment 7 by two steps. However, the reaction of â-hydroxy
ester 18 with TESOTf was both significantly slower and less
efficient than that of 28, producing the desired product 32 in
only 58% yield along with 23% of the five-membered acetal
33 (Scheme 4, eq 4). Therefore, from the perspective of overall
9
9310 J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004

Total Synthesis of Formamicin
A R T I C L E S
Scheme 5 a
Scheme 6 a
a Conditions: (a) NIS, CH3CN, 0 °C; (b) DIBAL-H, CH2Cl2, -78 °C;
(c) oxalyl chloride, DMSO, Et3N, CH2Cl2, -78 °C to 0 °C; (d)
Ph3PdC(Me)CO2Et, PhCH3, reflux; (e) Ph3PdCHCO2Me, PhH, 65 °C, 18
h.
synthetic efficiency, it was more advantageous to utilize pivaloyl
ester 29 as an intermediate in the synthesis of vinyl iodide 7.
Treatment of 29 with NIS effected the conversion of the vinyl
silane to the corresponding vinyl iodide (Scheme 5). Removal
of the pivaloate ester was achieved through treatment of the
ester with DIBAL-H. Oxidation of the resultant alcohol under
standard Swern oxidation conditions provided aldehyde 34.
Homologation of 34 via Horner-Wadsworth-Emmons reac-
tion52 with Ph3PdC(Me)CO2Et provided the corresponding
enoate, which was reduced to the primary alcohol using
DIBAL-H and then oxidized under Swern conditions to give
aldehyde 35. Finally, olefination of aldehyde 35 with Ph3Pd
CHCO2Me completed the synthesis of the C(1)-C(11) fragment
7. Overall, this synthesis of vinyl iodide 7 proceeded in 23 linear
steps and 16% overall yield.
a Conditions: (a) MOMCl, i-Pr2NEt, CH2Cl2, 0 °C; (b) LiBH4, EtOH,
Et2O, 0 °C; (c) oxalyl chloride, DMSO, Et3N, CH2Cl2, -78 °C to 0 °C; (d)
EtO(Li)CdCH2, THF, -118 °C, (e) KHMDS, Et3N, PMBBr, THF then 1
N HCl; (f) CH3CtCsMgBr, THF, -45 °C; (g) Bu3SnH (10 equiv), (Po-
Tol3)2PdCl2 (10 mol%), THF; (h) Me2BBr, 2,6-di-tert-butyl-4-methyl-
pyridine, CH2Cl2, -78 f -50 °C.
Second Generation Synthesis of Fragment 7. In parallel
to our development of the lactate aldol approach to fragment 7,
investigations were also conducted to develop a synthesis based
on the chelate-controlled addition of an alkynyl nucleophile to
aldehyde 11. In our original lactate aldol strategy, the MOM
ether necessary for the transacetalation reaction had been
introduced at the C(6) alcohol. However, we recognized that
placing the MOM group on the C(25) alcohol might provide
the opportunity to shorten the synthetic sequence and minimize
protecting group manipulations. Toward this end, protection of
aldol 16 as a MOM ether followed by reductive removal of the
chiral auxiliary by using LiBH453 provided primary alcohol 36
(Scheme 6). Alcohol 36 was then subjected to standard Swern
oxidation conditions followed by aqueous workup. The crude
aldehyde was then coupled with R-lithio ethyl vinyl ether54 at
-118 °C to give the corresponding carbonyl addition product
with g20:1 diastereoselectivity. Protection of the resulting
alcohol using p-methoxybenzyl bromide (PMBBr) followed by
hydrolysis of the enol ether with 0.1 N HCl provided R-alkoxy
ketone 11 in 79% overall yield for the four steps. Methyl ketone
11 was then coupled with 1-propynylmagnesium bromide under
chelate-controlled conditions,55,56 thereby generating the de-
sired tertiary alcohol 37 in 94% yield with >20:1 diastereo-
selectivity.51
We next explored transition-metal-catalyzed chain extension
reactions of 37. Initial experiments36 showed that hydrostan-
nation of alkyne 37 could be accomplished using Pd2dba3‚CHCl3
(4 mol %) and Bu3SnH (20 equiv).57 Unfortunately, full
consumption of alkyne 37 under these conditions did not occur,
requiring that recovered starting material be recycled twice to
achieve a synthetically useful 79% yield of vinyl stannane 38.
A more efficient approach was eventually found which utilized
(o-Tol3P)2PdCl2 as the hydrostannation catalyst.58 Use of this
more reactive catalyst allowed a reduction in the equivalents
of Bu3SnH (10 equiv) and induced complete consumption of
alkyne 37. In this way, vinyl stannane 38 was obtained in 89%
yield (Scheme 6).
With vinyl stannane 38 in hand, formation of the seven-
membered formyl acetal was investigated. To our delight,
treatment of alcohol 38 with Me2BBr59 at -78 °C promoted
the transacetalation and provided cyclic acetal 39 in 98% yield
(Scheme 6).
We hoped that a more direct approach to fragment 7 could
be achieved by avoiding the use of the PMB protecting group
in 37. To investigate this possibility, substrates 40 and 44
(Scheme 7) were prepared from alcohol 36.60 Surprisingly,
treatment of 40 with 2 equiv of TESOTf did not produce the
(52) Wadsworth, W. S., Jr.; Emmons, W. D. J. Am. Chem. Soc. 1961, 83, 1733.
(53) Penning, T. D.; Djuric, S. W.; Haack, R. A.; Kalish, V. J.; Miyashiro, J.
M.; Rowell, B. W.; Yu, S. S. Synth. Commun. 1990, 20, 307.
(54) Baldwin, J. E.; Ho¨fle, G. A.; Lever, O. W., Jr. J. Am. Chem. Soc. 1974,
96, 7125.
(55) Reetz, M. T. Angew. Chem., Int. Ed. Engl. 1984, 23, 556.
(56) Mengel, A.; Reisser, O. Chem. ReV. 1999, 99, 1191.
(57) Smith, N. D.; Mancuso, J.; Lautens, M. Chem. ReV. 2000, 100, 3257.
(58) Greeves, N.; Torode, J. S. Synlett 1994, 537.
(59) Guindon, Y.; Yoakim, C.; Morton, H. E. J. Org. Chem. 1984, 49, 3912.
9
J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004 9311

A R T I C L E S
Durham et al.
Scheme 7 a
Scheme 8 a
a Conditions: (a) methyl (E)-3-iodopropenoate (1.9 equiv), copper(I)
2-thiophene carboxylate (CuTC, 2 equiv), tetrabutylammonium diphenyl-
phosphinate (1.1 equiv), NMP; (b) DDQ, pH ) 7 buffer/CH2Cl2, 0 °C; (c)
TESOTf, 2,6-lutidene, CH2Cl2, 0 °C, 78% for two steps; (d) NIS, CHCN,
96% to quant..
It was anticipated that the carbon skeleton of fragment 7 could
be completed through a Stille cross-coupling61 reaction between
vinyl stannane 39 and methyl (E)-3-iodopropenoate62 to give
dienoate 47 (Scheme 8). Initial attempts to achieve this coupling
under a variety of Pd(0)-catalyzed conditions proved to be
unsuccessful. However, combined use of copper(I) 2-thiophene
carboxylate (CuTc)63 and the trialkyltin halide scavenger
tetrabutylammonium diphenylphophinate64 eventually proved to
be effective (Scheme 8). Initial experimentation, in which the
vinyl stannane and vinyl iodide were mixed prior to addition
of CuTc, showed that homocoupling of methyl (E)-3-iodo-
propenoate was a competitive background reaction. Facile
homocoupling of vinyl iodides upon exposure to CuTc has been
previously observed by Liebeskind, who has suggested that, in
some systems, oxidative insertion of Cu (I) with alkenyl iodides
is facile and occurs more rapidly than transmetalation with the
vinyl stannane.63 This undesired reaction could be minimized
by syringe pump addition of a solution of the vinyl iodide into
a mixture of vinyl stannane 39 and CuTc.
However, as the reaction scale was increased to >100 mg of
vinyl stannane 39, proteodestannylation also became a competi-
tive side reaction. Careful experimentation showed that this
byproduct arose from proton abstraction from the reaction
solvent at the stage of a putative vinyl copper species produced
by transmetalation of stannane 39.63 This observation suggests
that in the reaction of 39 with CuTc the rate of transmetalation
of the stannane is relatively fast at the beginning of the reaction
time scale but slows as the reaction progresses to a more
moderate rate.65 Suppression of this second nonproductive
reaction pathway, while simultaneously limiting the degree to
which methyl (E)-3-iodopropenoate underwent homocoupling,
could be achieved by adding ∼0.2 equiv of methyl (E)-3-
iodopropenoate to the stannane prior to addition of CuTc.
a Conditions: (a) TESOTf (2 equiv), 2,6-lutidine, CH2Cl2, 0 f 23 °C;
(b) TESOTf (1 equiv), 2,6-lutidine, CH2Cl2, 0 °C; (c) Me2BBr, 2,6-di-tert-
butyl-4-methylpyridine, CH2Cl2, -70 to -60 °C.
expected cyclization product but instead gave a 1:1 mixture of
the mono- and bis-silylated products 41 and 42 (Scheme 7, eq
5). Treatment of C(7) TES ether 41 with Me2BBr gave six-
membered acetal 43 with concomitant silyl migration (Scheme
7, eq 6).
Recalling that the cyclization of C(5) pivaloate ester 28 had
been more efficient than that of C(5) acetate 30 (Scheme 4, eqs
2 and 3), we hypothesized that increasing the steric environment
around the C(5) carbon might improve the chemoselectivity of
the cyclization. Accordingly, we attempted to form the seven-
membered acetal from vinyl stannane 44 (Scheme 7, eq 7). Vinyl
stannane 44 is analogous to our previous PMB ether 38 which
had cyclized with both high efficiency and chemoselectivity
(Scheme 6). However, treatment of alcohol 44 with either
TESOTf or Me2BBr lead only to decomposition (Scheme 7, eq
7). A stepwise approach, in which the C(7) alcohol of 44 was
first converted to TES ether 45 and then subsequently exposed
to Me2BBr, also did not provide the seven-membered acetal
product (Scheme 7, eq 8). Instead, the six-membered acetal 46,
in which the C(7) TES group had undergone a silyl migration
to the tertiary alcohol, was isolated in 90% yield.
(61) Farina, V.; Krishnamurthy, V.; Scott, W. J. Org. React. 1997, 50, 1.
(62) Dixon, D. J.; Ley, S. V.; Longbottom, D. A. J. Chem. Soc., Perkin Trans.
1 1999, 2231.
(63) Allred, G. D.; Liebeskind, L. S. J. Am. Chem. Soc. 1996, 118, 2748.
(64) Srogl, J.; Allred, G. D.; Liebeskind, L. S. J. Am. Chem. Soc. 1997, 119,
12376.
(65) While the exact cause of the purported difference in rate of transmetalation
of 39 with CuTc is not known, it should be noted that this reaction is
heterogeneous. If transmetalation is occurring at the surface of the solid
CuTc, this could explain the variation in reaction rate, as well as the
differences in the efficiency of the coupling depending on the scale of the
reaction.
(60) Compound 40 was prepared from 36 by the following sequence: (a) oxalyl
chloride, DMSO, Et3N, CH2Cl2, -78 °C to 0 °C; (b) EtO(Li)CdCH2, THF,
-118 °C; then 1 N HCl 0 °C, 88%, >20:1 dr; (c)10, THF, -45 °C, >20:1
dr. Compound 44 was generated from 40 via hydrostannation: Bu3SnH
(10 equiv), Pd2dba3‚CHCl3 (4 mol %), THF (77%).
9
9312 J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004

Total Synthesis of Formamicin
A R T I C L E S
Scheme 9 a
Subsequent syringe pump addition of the remaining solution
of the vinyl iodide (1.7 equiv) to the reaction mixture over 15-
20 min then provided 47 in 85% yield. This optimized
experimental protocol could be used to conduct this cross-
coupling on a multigram scale.
Dienoate 47 was then elaborated to fragment 7 by a two-
step modification of the C(7) protecting group and conversion
of the vinyl silane to the iodide via treatment with NIS at 0 °C
(Scheme 8). Using the approach shown in Schemes 6 and 8,
this second synthesis of fragment 7 required 21 linear steps and
proceeded with an overall yield of 20%. The diastereoselectivity
of the least selective step was 20:1.
Synthesis of Formamicinone. With access to synthetically
useful quantities of vinyl iodide 7, elaboration of this fragment
to the formamicin aglycon, formamicinone (2), proceeded
smoothly. Suzuki cross-coupling of vinyl iodide 7 and vinyl
boronic acid 622,40 occurred readily under Pd(0)-catalyzed
conditions in the presence of either TlOEt66 or Tl2CO367 to give
tetraene 48 in 88% yield (Scheme 9). Deprotection of the methyl
68
ester 48 with KOSiMe3 gave the seco acid. The carboxylic
acid was then subjected to Yamaguchi macrolactonization69
which produced lactone 49 in a 50-60% yield from ester 48.
The oxidation state of C(19) was then adjusted by selective
deprotection of the C(19) TES ether under acidic conditions
and subsequent oxidation of the C(19) hydroxyl with Dess-
Martin periodinane.70 This produced methyl ketone 50 in 80-
90% yield over two steps. Mukaiyama aldol coupling of 50 and
aldehyde 539 proceeded with excellent diastereoselectivity to
give alcohol 4 in 86% yield and >95:5 diastereomeric ratio
(dr).51 The stereochemistry of the â-hydroxy ketone was
assigned by NMR methods.71 Deprotection of aldol 4 using tris-
(dimethylamino)sulfur (trimethylsilyl)difluoride (TASF) in DMF/
H2O provided the aglycon of the natural product, 2, in 80%
1
yield.72 Comparison of the H and 13C NMR data of 2 and
formamicin (1) showed excellent agreement, with the only
noticeable differences in chemical shifts being those for the
C(20) and C(21) residues, the site at which the natural product
is glycosylated.
a Conditions: (a) Pd(PPh3)4 (cat.), Tl2CO3, THF/H2O 3:1; (b) KOTMS,
Et2O/THF; (c) 2,4,6-trichlorobenzoyl chloride, DMAP, toluene, reflux, 12
h; (d) AcOH/H2O/THF 6:1:6; (e) Dess-Martin periodinane, pyridine,
CH2Cl2; (f) LHMDS, Et3N/TMSCl, THF, -78 °C; (g) 5, BF3‚OEt2, CH2Cl2,
-78 °C; (h) TASF, H2O, DMF.
Studies of the Glycosidation Reaction. The next issue to
be addressed was the introduction of the 2,6-dideoxygluco-
pyranoside unit, which we anticipated would be best ac-
complished using 4 as the acceptor (vide infra). While methods
for the â-selective glycosidation of â-hydroxy carbonyl com-
pounds had been reported in the literature when we began our
own studies in this area,38 a general method providing both high
efficiency and selectivity had yet to be established. The
weakened reactivity of the â-hydroxy group due to intra-
molecular hydrogen bonding with the carbonyl moiety73 and
the instability of â-hydroxy and â-alkoxy carbonyl derivatives
to Lewis acidic reaction conditions (vide infra) are most likely
the sources of difficulty inhibiting the development of a general
protocol. Among the few reports of such glycosidations in the
chemical literature, the following three instances are illustrative.
The glycosidation of a â-hydroxy ketone using a 2-deoxy-
glucopyranosyl fluoride74,75 donor was demonstrated by Tatsuta
and Kinoshita to proceed with only modest selectivity and low
efficiency (30% yield).76,77 Use of both a 2-deoxyglucopyranosyl
phosphite and a 2-deoxyglucopyranosyl bromide donor in an
attempt to effect the â-selective glycosidation of a similar
â-hydroxy ketone by Paterson also resulted in low isolated yields
(66) Frank, S. A.; Chen, H.; Kunz, R. K.; Schnaderbeck, M. J.; Roush, W. R.
Org. Lett. 2000, 2, 2691.
(67) Marko´, I. E.; Murphy, F.; Dolan, S. Tetrahedron Lett. 1996, 37, 2507.
(68) Lagnais, E. D.; Chenard, B. L. Tetrahedron Lett. 1984, 25, 5831.
(69) Meng, Q.; Hesse, M. Top. Curr. Chem. 1991, 161, 107.
(70) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155.
(71) Roush, W. R.; Bannister, T. D.; Wendt, M. D.; VanNieuwehnze, M. S.;
Gustin, D. J.; Dilley, G. J.; Lane, G. C.; Scheidt, K. A.; Smith, W. J., III.
J. Org. Chem. 2002, 67, 4284.
(72) Scheidt, K. A.; Chen, H.; Follows, B. C.; Chemler, S. R.; Coffey, D. S.;
Roush, W. R. J. Org. Chem. 1998, 63, 6436.
(74) For a recent review on the use of glycosyl fluorides in synthesis, see:
Toshima, K. Carbohydr. Res. 2000, 327, 15.
(73) Results from our laboratories suggest that â-hydroxy ketones such as 58,
which are capable of intramolecular hydrogen-bonding, are far less reactive
than alcohol acceptors such as 60 and 62. However, a similar hydrogen-
bonding pattern has been reported to increase the nucleophilicity of
hydrogen-bound acceptors; Mitchell, S. A.; Pratt, M. R.; Hruby, V. J.; Polt,
R. J. Org. Chem. 2001, 66, 2327.
(75) For a recent review on the synthesis of glycosyl fluorides and their use in
synthesis, see: Shimizu, M.; Togo, H.; Yokoyama, M. Synthesis 1998, 799.
(76) Toshima, K.; Misawa, M.; Ohta, K.; Tatsuta, K.; Kinoshita, M. Tetrahedron
Lett. 1989, 30, 6417.
(77) Paterson, I.; McLeod, M. D. Tetrahedron Lett. 1995, 36, 9065.
9
J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004 9313

A R T I C L E S
Durham et al.
Scheme 10 a
Scheme 11 a
a Conditions: (a) H2NNH2, MeOH-Et2O, 0 °C; (b) DAST, CH2Cl2, 0
°C; (c) SnCl2, AgClO4,Et2O, 4 Å MS, -15 °C, 20-30 min.
a Conditions: (a) TBSOTf (0.3 equiv), CH2Cl2, -78 °C.
(21-39%) and proceeded with only modest â-selectivity
(e2.5:1 â/R).77 Finally, Evans has also reported a â-selective
glycosidation reaction of a â-hydroxy Weinreb amide acceptor
with a 2-deoxyglycosyl acetate donor (70%, 4:1 â/R).78 The
authors noted that equilibration of the R-glycoside to the
â-glycoside could be achieved by resubjection to the reaction
conditions. Further, the selectivity was highly dependent upon
both reaction conditions and substituents on the donor.
Our laboratory has recently reported methodology for the
stereoselective synthesis of 2-deoxy-â-gluco- and galacto-
pyranosides using 2-deoxy-2-iodo- and 2-deoxy-2-bromo-
glucopyranosyl79-82 and galactopyranosyl83,84 acetates and
trichloroacetimidates. After generation of the â-glycosidic
linkage, the C(2)-halogen directing groups can be easily excised
from the products by reduction with Bu3SnH.85 Given the high
selectivity and efficiency of these reactions, we initially hoped
that these 2-deoxy-2-iodo donors could be used to achieve the
â-selective glycosidation of â-hydroxy ketones. Subsequently
we examined the glycosidation of model â-hydroxy ketone 5838
using a variety of 2-deoxy-2-iodoglucopyranosy donors (51-
57)38,79,80 (Scheme 10) under a variety of Lewis acidic conditions
(TMSOTf, BF3‚OEt2, TrClO4,78 K10 clay,86 LiClO4,87 LiOTf88).
However, these initial attempts led only to decomposition of
â-hydroxy ketone 58. Control experiments showed that the
decomposition of ketone 58 in the presence of TMSOTf at
temperatures ranging from -78 to -30 °C occurred within 20
min. However, use of the milder activator TBSOTf and donor
51 was successful, providing â-glycoside 59 with high stereo-
selectivity (>98:2 â/R) but in low efficiency (21%, Scheme
10).51
In view of the sensitivity of both the acceptors and products
to Lewis acidic reaction conditions, we anticipated a more
efficient glycosidation reaction could be realized if a milder
set of activation conditions was employed. Recalling the
successes of Tatsuta and Kinoshita as well as Paterson using
Ag(I) salts to achieve activation of 2-deoxyglucopyranosyl
fluorides76 and bromides77 under mild conditions, we next
prepared 2-deoxy-2-iodoglucopyranosyl fluoride 3.89,90 Selective
deprotection of the anomeric acetate of 51 using aqueous
hydrazine91 followed by reaction of the hemiacetal with diethyl-
aminosulfur trifluoride (DAST) provided â-glucopyranosyl
fluoride 3 in 78% yield (Scheme 11).92,93 Activation of fluoride
donor 3 in the presence of â-hydroxy ketone 58 using
Mukaiyama’s conditions94 (SnCl2/AgClO4) at -15 °C in ether
provided â-glycoside 59 in 65% yield with >98:2 â/R selectiv-
ity.51 Further experiments showed that glycosidation reactions
of fluoride donor 3 with a variety of acceptors were highly
â-selective (Scheme 11).51 Additionally, these reaction condi-
tions were sufficiently mild to accommodate even highly
sensitive acceptors such as glycal 62.95,96
(89) A single example using a 2-deoxy-2-iodoglucopyranosyl fluoride donor in
the glycosidation of cyclohexanol has been reported: Nishimura, S.;
Washitani, K. (Sumitomo Pharmaceuticals Co., Ltd., Japan) Stereoselective
Production of Glycosyl Compound. Japanese Patent 09241288, 1997.
(90) Thiem has studied 2,6-dideoxy-2,6-dibromoglucopyranosyl fluorides and
2,6-dideoxy-2,6-dibromomannopyranosyl fluorides in the synthesis of
oligosaccharides using a mixture of TiF4 and AgClO4 as the promoter.
Reactions of the 2,6-dideoxy-2,6-dibromoglucopyranosyl fluorides gave
yields of 43-75% and selectivities ranging from 1:1 to 7:1 â/R. The reaction
selectivity showed significant dependence on the solvent, the fluoride
substituent stereochemistry (R vs â), and identity of the acceptor. For the
original report, see: Ju¨nneman, J.; Lundt, I.; Thiem, J. Liebigs Ann. Chem.
1991, 759.
(91) Excoffier, G.; Gagnaire, D.; Utille, J.-P. Carbohydr. Res. 1975, 39, 368.
(92) Ju¨nneman, J.; Lundt, I.; Thiem, J. Acta Chem. Scand. 1991, 45, 494.
(93) For a review on the preparation of glycosyl fluorides, see: Yokoyama, M.
Carbohydr. Res. 2000, 327, 5.
(94) Mukaiyama, T.; Murai, Y.; Shoda, S.-i. Chem. Lett. 1981, 431.
(95) Roush, W. R.; Narayan, S. Org. Lett. 1999, 1, 899.
(78) Evans, D. A.; Kaldor, S. W.; Jones, T. K.; Clardy, J.; Stout, T. J. J. Am.
Chem. Soc. 1990, 112, 7001.
(79) Roush, W. R.; Gung, B. W.; Bennett, C. E. Org. Lett. 1999, 1, 891.
(80) Roush, W. R.; Bennett, C. E. J. Am. Chem. Soc. 1999, 121, 3541.
(81) Chong, P. Y.; Roush, W. R. Org. Lett. 2002, 4523.
(82) Roush, W. R.; Bennett, C. E. J. Am. Chem. Soc. 2000, 122, 6124.
(83) Durham, T. B.; Roush, W. R. Org. Lett. 2003, 5, 1871.
(84) Durham, T. B.; Roush, W. R. Org. Lett. 2003, 5, 1875.
(85) Miura, K.; Ichinose, Y.; Nozaki, K.; Fugami, K.; Oshima, K.; Utimoto, K.
Bull. Chem. Soc. Jpn. 1989, 62, 143.
(86) Nagai, H.; Matsumura, S.; Toshima, K. Tetrahedron Lett. 2002, 43, 847.
(87) Waldmann, H.; Bo¨hm, G.; Schmid, U.; Ro¨ttele, H. Angew. Chem., Int. Ed.
Engl. 1994, 33, 1944.
(88) Lubineau, A.; Drouillat, B. J. Carbohydr. Chem. 1997, 16, 1179.
9
9314 J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004

Total Synthesis of Formamicin
A R T I C L E S
Scheme 12 a
a Conditions: (a) BzCl, DMAP, Et3N, CH2Cl2, quant.; (b) AcOH/THF/H2O 6:6:1; (c) Dess-Martin periodinane, pyridine, CH2Cl2, 96% over two steps;
(d) LHMDS, Et3N/TMSCl, THF, -78 °C; (e) 66 or 5, BF3‚OEt2, CH2Cl2; (f) 3, SnCl2, AgClO4,Et2O, 4 Å MS, -10 °C.
Scheme 13 a
To determine the suitability of this glycosidation reaction for
use in our synthesis of formamicin (1), we next examined the
glycosidation reaction of donor 3 using model compounds 67
and 68 (Scheme 12). Preparation of the 67 and 68 commenced
with compound 64, which is an intermediate in the synthesis
of vinyl boronic acid 6.22,40 Benzoylation of the C(15) hydroxyl
followed by deprotection of the C(19) TES protecting group
under acidic conditions and oxidation of the resultant alcohol
with Dess-Martin periodinane provided methyl ketone 65.
Mukaiyama aldol coupling of 65 and aldehydes 66 and 5
proceeded with excellent diastereoselectivity to give aldols 67
and 68 in 59% and 66% yields, respectively. Glycosidation of
these â-hydroxy ketones using 2-iodoglucopyranosyl fluoride
donor 3 proceeded with >98:2 â-selectivity and gave â-glyco-
sides 69 (27-42% yield) and 70 (66-84% yield).51
Studies of the Deprotection Sequence. Satisfied that fluoride
donor 3 could be used to install the 2,6-dideoxy pyranoside of
formamicin, we turned our attention to investigating conditions
a Conditions: (a) AcOH, 1% HF‚KF, THF, 33°C, 93 h; (b) Et3B,O2,
to effect desilylation of the final intermediate. For these studies
we chose to use model compounds 71 and 72. Synthesis of
glycoside 71 was achieved in 72% yield from aldol 69 by
selective deprotection of the DEIPS ether via treatment with
1% HF‚KF/AcOH in THF at 33 °C over 4 days (Scheme 13).
Model glycoside 72 was easily prepared from 70 by way of
reductive removal of the C(2′)-iodo and C(6′)-bromo substituents
using Bu3SnH85 in 90% yield.
Initial investigations to develop a set of global desilylation
conditions applicable to the synthesis of formamicin focused
on hemiketal 71 (Table 2). We hoped that the hemiketal unit
would serve to mask the â-alkoxy ketone and thus suppress
elimination of the 2,6-dideoxyglucopyranoside unit during the
deprotection step. Having found that TASF was highly effective
in the desilylation of protected aglycon 4, it was most surprising
that treatment of 71 with TASF provided only elimination
products 75, even under buffered97 conditions (Table 2, entries
1 and 2)!
Bu3SnH, toluene, rt, 90%.
phenol97 provided only elimination products 75 (Table 2, entries
3-5). Use of (Bu4N)Bu3SnF298 promoted decomposition of 72
(Table 2, entry 6). Based on these observations, we recognized
that basic fluoride reagents would not be effective in promoting
the required deprotections. Consequently, we next investigated
the use of inorganic fluoride sources (e.g., KHF2) in mildly
acidic reaction conditions, but glycoside 72 proved unreactive
under these conditions (Table 2, entry 7). Use of more acidic
conditions such as Dowex 50W-X4200 resin/MeOH or HF‚Py
lead to decomposition of the starting material (Table 2, entries
8 and 9).
Given the data summarized in Table 2, it became clear that
a neutral set of silyl deprotection conditions had the best chance
of being effective. This realization lead to our consideration of
Et3N‚3HF as a potential desilylation reagent.99 While use of
HF‚Py had not been effective, presumably due to acid-catalyzed
decomposition, it has been shown that Et3N‚3HF is a milder
fluoride source, with a pH close to neutral.100 Further, the
addition of Et3N to this reagent leads to formation of Et3N‚
More extensive studies were performed using 72, which is
available with better efficiency than 71 (cf., Schemes 12 and
13). Subjection of 72 to a variety of standard desilylation
conditions including TBAF, TBAF/AcOH, and TBAF/2-nitro-
(98) Gingras, M. Tetrahedron Lett. 1991, 32, 7381.
(99) McClinton, M. A. Aldrichimica Acta 1995, 28, 31.
(100) Saluzzo, C.; Alvernhe, G.; Anker, D. J. Fluorine Chem. 1990, 47.
(96) Bennett, C. E. Ph.D. Thesis, Indiana University, Bloomington, IN, 2000.
(97) Myers, A. G.; Goldberg, S. D. Angew. Chem., Int. Ed. 2000, 39, 2732.
9
J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004 9315

A R T I C L E S
Durham et al.
Scheme 14 a
a Conditions: (a) SnCl2, AgClO4, Et2O, 4 Å MS, -20 to -15 °C; (b) Bu3SnH, Et3B, O2, 23 °C; (c) Et3N‚3HF, Et3N, CH3CN/THF 1:1, 23 °C, 3 d; (d)
Et3N‚3HF, Et3N, CH3CN, 23 °C, 11 d.
Table 2. Desilylation of Glycosidated Model â-Hydroxy Ketones
product 74 in 39% yield (Table 2, entry 10). Interestingly, TBS
ether 73 was also recovered from the reaction as a mixture of
diastereomeric ketals in 47% yield.
Completion of the Total Synthesis of Formamicin. With a
potential set of conditions to allow final desilylation in hand,
our attention turned to the completion of the total synthesis of
formamicin. In the event, glycosidation of aglycon 4 by donor
94
3 with SnCl2 and AgClO4 in Et2O at -20 °C provided the
desired glycoside 76 in 68% yield and with >98:2 â/R
selectivity (Scheme 14).51 Removal of the carbohydrate C(2′)
and C(6′) halogen substituents with Bu3SnH in toluene at 23
°C, via promotion with Et3B/O2,85 proceeded in excellent yield
to provide 2,6-dideoxy glycopyranoside 77, the penultimate
intermediate in the synthesis.
Interestingly, glycoside 77 was insoluble in acetonitrile, which
required the use of THF as a cosolvent for our planned global
desilylation reaction. After exposure of 77 to in situ generated
Et3N‚2HF101 in 1:1 acetonitrile/THF, all but one of the TBS
ether groups were removed (Scheme 11). Unfortunately, further
exposure of this mono-TBS derivative102 to these conditions
failed to remove this lone remaining silyl group. We anticipated
that this could be the result of the lower dielectric constant of
the reaction medium,99 originating from the introduction of the
THF cosolvent. Therefore, the reaction was subjected to an
aqueous workup103 and the mono-TBS ether was treated with
in situ generated Et3N‚2HF in acetonitrile. This reaction
eventually proved successful, albeit slow (11 days), and provided
entry
precursor
reagent
solvent
time
24 h
result
1
2
71
71
TASF
DMF
DMF
75
75
2,6-DCPa
TASF
24 h
3
4
72
72
2-NPb
THF
THF
5 min.
3.5 h
75
75
TBAF
AcOH
TBAF
5
72
THF
24h
75
2-NP
Bu4N+ Bu3SnF2
KHF2(aq)
AcOH
Dowex
50W-X4200
HF‚Py
Et3N‚HF
Et3N
THF
THF
24 h
20 h
decc
NRe
-
6
72
72
7d
8
72
MeOH
24 h
decc
9
10
72
72
THF
CH3N
20 min
5 d
decc
73 (47%)
74 (39%)
(101) Giudicelli, M. B.; Picq, D.; Veyron, B. Tetrahedron Lett. 1990, 31, 6527.
(102) It is clear from the long reaction time (11 days) required to remove the
last TBS ether from intermediate 77 that the environment around this group
is highly sterically congested. Based on the recovery of 73 in the
deprotection of model system 72, we suspect that this intermediate is the
C(17)-OTBS derivative of 1. While we have not been able to unequivo-
cally prove this assignment, support for this hypothesis is found in 1H
NMR analysis of the mono-TBS ether intermediate. This material exists
as a mixture of diastereomeric ketals, just as 73. In the 1H NMR spectrum
of the mono-TBS ether of 1, a broad multiplet at 4.3 ppm is observed,
rather than the doublet of doublets of doublets (ddd) at 4.2 ppm
corresponding to the C(17)-H in 1. Further, the sharp doublet at 4.9 ppm
in 1, originating from the C(17)-OH, is notably absent in this material.
The C(17)-OH is involved in a characteristic hydrogen-bonding pattern
in this natural product, and the absence of a resonance corresponding to
C(17)-OH in the mono-TBS ether strongly points to this compound being
the C(17)-OTBS derivative.
a 2,6-DCP ) 2,6-dichlorophenol. b 2-NP ) 2-nitrophenol. c dec )
decomposition. d Reaction stirred at 23 °C for 12 h and then at 50 °C for
6 h. e NR ) no reaction, starting material recovered.
2HF and Et3N‚HF.101 Of the various forms, Et3N‚2HF is the
most nucleophilic fluoride source and has been used to prepare
alkyl fluorides via displacement reactions.101 Therefore, we were
gratified to find that exposure of glycoside 72 to in situ generated
Et3N‚2HF in acetonitrile over 5 days indeed provided the desired
9
9316 J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004

Total Synthesis of Formamicin
A R T I C L E S
synthetic formamicin 1 in 58% overall yield for the two steps.
The identity of synthetic 1 was confirmed by comparison to an
authentic sample graciously provided by Igarashi and co-
membered formyl acetals 29 and 39, an efficient and highly
diastereoselective glycosidation of the protected aglycon 4 using
fluoride donor 3, and global desilylation of penultimate inter-
mediate 77 using Et3N‚2HF to complete the total synthesis.
1
workers33,34 using H and 13C NMR spectroscopy and TLC
mobility in several solvent systems.
Acknowledgment. We thank the National Institutes of Health
(GM 38436) for support of this research. We also acknowledge
additional financial support in the form of a Lavoisier Fellowship
from the Ministe`re Francais des Affaires Etrange`res to N.B.,
an Abbott Graduate Fellowship to B.M.S., and an NIH post-
doctoral fellowship (GM 20278) to N.A.P..
Summary. We have achieved the first total synthesis of
formamicin (1) by a highly stereoselective sequence. Highlights
of this work include development of both lactate aldol and
chelate controlled carbonyl addition strategies to the C(1)-C(11)
fragment 7 of the natural product, a Lewis acid promoted
intramolecular transacetalation reaction to form the seven-
(103) Following dilution of the reaction with Et2O and washing with pH 7 buffer,
the ethereal solution was dried over KHCO3/Na2SO4 for 1 h prior to
concentration on the rotary evaporator. Failure to use KHCO3 resulted in
decomposition of the material upon concentration. Presumably, this
decomposition is promoted by residual HF. This protocol was first reported
by Shotwell, J. B.; Hu, S.; Medina, E.; Abe, M.; Cole, R.; Crews, C. M.;
Wood, J. L. Tetrahedron Lett. 2000, 41, 9639; see ref 14 therein.
Supporting Information Available: Experimental procedures
for all new compounds. This material is available free of charge
via the Internet at http://pubs.acs.org.
JA048493L
9
J. AM. CHEM. SOC. VOL. 126, NO. 30, 2004 9317