Enantioselective synthesis of oasomycin A, part I: Synthesis of the C1-C12 and C13-C28 subunits

Article abstract of DOI:10.1002/anie.200603653

(Chemical Equation Presented) Putting the pieces together: The total synthesis of the natural macrolide oasomycin A has been realized. Key fragment couplings include an anti-Felkin selective aldol addition (green), Kociensky-Julia olefinations (red), and competitive Weinreb amide acylation reaction (blue). The utility of the 4,5-diphenyloxazole as a carboxy surrogate and the late-stage macrolactonization affording the 42-membered macrocycle of oasomycin A are also described.

Full text of DOI:10.1002/anie.200603653

Angewandte
Chemie
DOI: 10.1002/anie.200603653
Natural Products Synthesis
Enantioselective Synthesis of Oasomycin A, Part I: Synthesis of the
C1–C12 and C13–C28 Subunits**
David A. Evans,* Pavel Nagorny, Kenneth J. McRae, Dominic J. Reynolds,
Louis-Sebastian Sonntag, Filisaty Vounatsos, and Risheng Xu
Dedicated to Professor Y. Kishi on the occasion of his 70th birthday.
Oasomycin A (I) is a member of the desertomycin family of
macrolide natural products and was isolated in 1993 by
Thiericke and co-workers during the screening of secondary
metabolites of Streptoverticillium baldacii.[1] The overall
structure of oasomycin A was established through extensive
NMR spectroscopic studies and by direct comparison to other
members of the desertomycin family.[1a] In 2001, Kishi and co-
workers issued a proposal for the relative and absolute
stereochemistry of oasomycin A through the use of their
“universal NMR database” (Scheme 1).[2] To validate the
structural assignment of oasomycin A, we decided to pursue
the synthesis of the reported structure, I.
Herein, and in the following Communications,[23] we
describe our efforts which culminated in the synthesis of
oasomycin A (I) and confirm the stereochemical assignment
by Kishi and co-workers. First, we describe the synthesis of
the C1–C12 and C13–C28 subunits II and III, respectively,
which will be united by using the Julia coupling reaction
(Scheme 1).
Our plan for the synthesis of the C13–C28 fragment III is
outlined in Scheme 2. Aldehyde III could be derived from a
chelate-controlled reduction of the a,b-unsaturated ketone
ꢀ
IV, which in turn can be disconnected at the C21 C22 bond to
afford fragments V and VI. We predict successful union of
these two fragments because of the greater reactivity of the
Weinreb amide VI, which is required to avoid self-condensa-
tion of V (M = Li, MgX). This prediction is based on the
supposition that the inductive effect of the a-alkoxy sub-
stituent should elevate the amide reactivity of VI above that
of its reaction partner V.
Scheme 1. Retrosynthetic analysis of oasomycin A (I).
Synthesis of subunit II commenced with the aldol addition
of known b-ketoimide 1[3] to aldehyde 2,[4] which afforded the
desired anti,anti aldol adduct[5] as an 84:16 mixture of
[*] Prof. D. A. Evans, P. Nagorny, Dr. K. J. McRae, Dr. D. J. Reynolds,
Dr. L.-S. Sonntag, Dr. F. Vounatsos, R. Xu
Department of Chemistry & Chemical Biology
Harvard University
Cambridge, MA 02138 (USA)
Fax: (+1)617-495-1460
E-mail: evans@chemistry.harvard.edu
[**] Financial support has been provided by the National Institutes of
Health (GM-33327-19), the Merck Research Laboratories, Amgen,
and Eli Lilly. A postdoctoral fellowship was provided to L.-S.S. by the
Deutscher Akademischer Austauschdienst and the Novartis Foun-
dation, and to D.J.R. by the Glaxo Foundation.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Scheme 2. Retrosynthesis of the C13–C28 subunit III.
Angew. Chem. Int. Ed. 2007, 46, 537 –540
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
537

Communications
tetrazole-5-thiol, and oxidation (H2O2, (NH4)6Mo7O24,
EtOH)[12,13] of the resulting sulfide.
Synthesis of the C13–C21 fragment V (Scheme 2) began
with the aldol addition of the (Z)-dibutylboron enolate of
imide 14 to known aldehyde 9[14] to afford the expected aldol
adduct 10 (Scheme 4). The chiral auxiliary was removed with
sodium methoxide in methanol (63%, 2 steps) to afford the
known ester 11a.[14] This ester 11a was successively silylated
to the known TBS ether 11b[14] (TBSOTf, lutidine) and then
transformed into the unstable a,b-unsaturated aldehyde 13 by
a reduction/Wittig olefination/reduction sequence in 88%
yield over the three steps. Reaction of the (Z)-dibutylboron
enolate of chiral imide 14 with aldehyde 13 proceeded
smoothly to give aldol adduct 15 (84%, > 95:5 d.r.). The
chiral auxiliary was then removed ((MeO)MeNH·HCl,
AlMe3), and the resulting alcohol protected as the TBS
ether (TBSOTf, lutidine, 94%, 2 steps) to complete the
synthesis of the C13–C21 fragment 16.
Scheme 3. Synthesis of the C1–C12 fragment II. Reagents and con-
ditions: a) Cy2BCl, Me2NEt, Et2O, ꢀ788C; then 2, Et2O, ꢀ788C,
(84:16 d.r.); b) TBSOTf, lutidine, CH2Cl2, ꢀ788C, (59%, 2 steps);
c) Zn(BH4)2, CH2Cl2, ꢀ78!ꢀ258C, (10:1 d.r.); d) TESOTf, lutidine,
CH2Cl2, ꢀ788C; e) EtSLi, THF, ꢀ208C; f) DIBALH, CH2Cl2, ꢀ908C,
(69%, 4 steps); g) 5, nBuLi, THF, ꢀ788C; then 4, 96%; h) CSA,
MeOH/CH2Cl2 1:1; i) Crabtree cat. (10 mol%), CH2Cl2, RT; j) TESOTf,
lutidine, CH2Cl2, ꢀ788C, (80%, 3 steps); k) DDQ, CH2Cl2/H2O 12:1;
l) 1-phenyl-1H-tetrazole-5-thiol, DEAD, PPh3, THF; m) (NH4)6Mo7O24,
H2O2, EtOH, (77%, 3 steps). Bn=benzyl, cod=cycloocta-1,5-diene,
CSA=camphorsulfonic acid, Cy=cyclohexyl, DDQ=2,3-dichloro-
5,6-dicyano-1,4-benzoquinone, DEAD=diethyl azodicarboxylate,
DIBALH=diisobutylaluminum hydride, PMB=4-methoxybenzyl,
TBS=tert-butyldimethylsilyl, TES=triethylsilyl, Tf=trifluoromethane-
sulfonyl.
diastereomers (Scheme 3). This mixture was immediately
silylated (TBSOTf, lutidine) to provide 3 as a single diaste-
reomer after flash chromatography. Successive chelate-con-
trolled reduction of 3 with Zn(BH4)2 (10:1 d.r.),[6,7] alcohol
protection (TESOTf, lutidine), removal of the oxazolidinone
auxiliary (EtSH, nBuLi THF, ꢀ208C), and reduction of the
derived thioester with DIBALH afforded aldehyde 4 (69%,
4 steps). Aldehyde 4 was condensed with the known phos-
phonate 5[8] under standard Horner–Wadsworth–Emmons
conditions to afford diene 6 (96%, E/Z 10:1).
Scheme 4. Synthesis of the C13–C21 fragment V. Reagents and
conditions: a) 14, Bu2BOTf, NEt3, CH2Cl2, ꢀ78!08C; then 9,
CH2Cl2, ꢀ788C; b) NaOMe, MeOH/CH2Cl2, ꢀ258C, (63%, 2 steps);
c) TBSOTf, lutidine, CH2Cl2, ꢀ108C, 97%; d) DIBALH, toluene,
ꢀ90!ꢀ808C; e) 12, CH2Cl2; f) DIBALH, CH2Cl2, ꢀ908C, (88%,
3 steps); g) 14, Bu2BOTf, NEt3, CH2Cl2, ꢀ10!08C; then 13, CH2Cl2,
ꢀ788C, 84% h) AlMe3, Me(MeO)NH·HCl, THF, 08C; i) TBSOTf,
lutidine, CH2Cl2, ꢀ108C; (94%, 2 steps).
The selective hydrogenation of the trans D4 olefin in 6 was
complicated by the generation of inseparable side products
that resulted from overreduction.[9] This problem was over-
come by resorting to a hydroxy-directed hydrogenation.
Selective deprotection of the C7 hydroxy group in 6 (CSA,
CH2Cl2/MeOH 1:1) and hydrogenation of the derived homo-
allylic alcohol with the Crabtree catalyst [Ir(cod)py-
The synthesis of the C22–C28 fragment (Scheme 5) began
with LiAlH4 reduction of the d-malic acid derivative 17,[15]
followed by diol protection to form the 3,4-dimethoxybenz-
ylidene acetal 18 (69%, 2 steps). The regioselective reduction
of 18 was accomplished with DIBALH, and the primary
alcohol was then oxidized under Parikh–Doering condi-
tions[16] to provide aldehyde 19 (65%, 2 steps). Allylation of
19 under chelate-controlled conditions (AlMe2Cl, allyltri-
methylsilane) afforded 20 as a 9:1 mixture of diastereomeric
alcohols that were separable by medium pressure liquid
chromatography.[17] Protection of 20 as its PMB ether
(PMBBr, NaHMDS), removal of the TIPS group with
[10]
(PCy3)]PF6 afforded a now-separable 11:1 mixture of the
desired olefin and the fully hydrogenated by-product.[11] The
C7 hydroxy group was then reprotected using TESOTf, thus
affording 7 (80%, 3 steps). Methyl ester 7 was next trans-
formed into sulfone 8 (synthon II, Scheme 1) by removal of
the C12 PMB group (DDQ, CH2Cl2/H2O 12:1), Mitsunobu
reaction of the resulting primary alcohol with 1-phenyl-1H-
538
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 537 –540

Angewandte
Chemie
Scheme 5. Synthesis of the C22–C28 fragment 22. Reagents and
conditions: a) LiAlH4, Et2O, 08C; b) DMPCH(OMe)2, CSA, CH2Cl2,
(69%, 2 steps); c) DIBALH, CH2Cl2, ꢀ78!ꢀ108C; d) SO3·Py, NEt3,
DMSO/CH2Cl2 1:1, ꢀ108C, (65%, 2 steps); e) AlMe2Cl (2 equiv),
=
CH2 CHCH2TMS, CH2Cl2, ꢀ938C, (9:1 d.r., 79% yield); f) PMBBr,
NaHMDS, TBAI, DMF/THF 1:2, ꢀ158C; g) TBAF, THF, 08C;
h) SO3·Py, NEt3, DMSO/CH2Cl2 1:1, ꢀ108C, (63%, 3 steps); i) NaClO2,
NaH2PO4, 2-methylbut-2-ene, tBuOH/H2O 3:2; j) TMSCHN2, MeOH/
PhH 10:1; k) Me(MeO)NH·HCl, iPrMgCl, THF, ꢀ108C, (89%, 3
steps). DMB=3,4-dimethoxybenzyl, DMP=3,4-dimethoxyphenyl,
HMDS=1,1,1,3,3,3-hexamethyldisilazane, Py=pyridine, TBA=tert-
butylammonium, TIPS=triisopropylsilyl, TMS=trimethylsilyl.
Scheme 6. Synthesis of the C13–C28 fragment 26. Reagents and
conditions: a) 1. 16 (1.00 equiv), nBuLi (1.09 equiv), Et2O, ꢀ788C;
2. MgBr2·Et2O (1.31 equiv), Et2O/THF ꢀ78!ꢀ208C; 3. 22
(0.965 equiv), Et2O/THF, ꢀ20!08C, 91%; b) Zn(BH4)2, Et2O/CH2Cl2,
ꢀ308C, 91%; c) TBSOTf, lutidine, THF, 08C, 95%; d) DDQ, H2O/
CH2Cl2 1:10, ꢀ38C; e) TBSOTf, lutidine, CH2Cl2, 08C; f) DIBALH,
toluene, ꢀ788C, (64%, 3 steps).
TBAF, followed by Parikh–Doering oxidation afforded
aldehyde 21 (63%, 3 steps). Aldehyde 21 could then be
converted into the Weinreb amide 22 by a three-step
sequence involving Kraus–Pinnick oxidation,[18] methylation
(TMSCHN2, MeOH/PhH), and amidation by using the Merck
procedure (89%, 3 steps).[19]
the synthesis of the C29–C46 fragment and the assembly of
the fragments to afford oasomycin A.
Received: September 6, 2006
Published online: December 8, 2006
The coupling of the C13–C21 and C22–C28 fragments
gave us the opportunity to test the relative reactivities of
Weinreb amides 16 and 22 (Scheme 6). Metalation of vinyl
iodide 16a afforded vinyllithium 16b. We predicted that this
intermediate would react with 22 rather than with itself based
on the observation that a-alkoxycarbonyl compounds are
usually more reactive electrophiles than the corresponding
a-alkylcarbonyl derivatives.[20] The desired chemoselectivity
was indeed achieved; however, transmetalation of 16b to the
corresponding alkenylmagnesium 16c was required to attenu-
ate the reactivity of the nucleophile. Thus, under the afore-
mentioned conditions, the coupling of 16 and 22 afforded the
coupled C13–C28 product 23[21] in 91% yield. Ketone 23 was
then reduced under chelate-controlled conditions (Zn(BH4)2)
to give alcohol 24 in 91% yield (12:1 d.r.).[7] Protection of 24
as its derived TBS ether (TBSOTf, lutidine) afforded 25 (95%
yield), which was then elaborated to aldehyde 26 by a three-
step sequence: selective deprotection (DDQ, CH2Cl2/H2O) of
the 3,4-dimethoxybenzyl ether at C23,[22] protection of the
resulting alcohol with TBSOTf, and reduction of the Weinreb
amide with DIBALH (64%, 3 steps).
Keywords: aldol reaction · Kocienski–Julia olefination ·
macrolactonization · natural products · total synthesis
.
[1] a) S. Grabley, G. Kretzschmar, M. Mayer, S. Philipps, R.
Thiericke, J. Wink, A. Zeeck, Liebigs Ann. Chem. 1993, 5,
573 – 579; b) M. Mayer, R. Thiericke, J. Chem. Soc. Perkin Trans.
1 1993, 21, 2525 – 2531.
[2] For the stereochemical assignment of oasomycin A, see: Y.
Kobayashi, S.-H. Tan, Y. Kishi, J. Am. Chem. Soc. 2001, 123,
2076 – 2078, and references therein.
[3] For the preparation of 1, see: a) D. A. Evans, M. D. Ennis, T. Le,
N. Mandel, G. Mandel, J. Am. Chem. Soc. 1984, 106, 1154 – 1156;
b) D. A. Evans, J. R. Gage, Org. Synth. 1989, 68, 83 – 91.
[4] T. Ishikawa, S. Ikeda, M. Ibe, S. Saito, Tetrahedron 1998, 54,
5869 – 5882.
[5] A. M. P. Koskien, K. Karisalmi, Chem. Soc. Rev. 2005, 34, 677 –
690.
[6] The relative stereochemistry of the reduction was proven by
converting the product resulting from reduction of 3 into an
acetonide. The 13C NMR signals (d = 97.9, 29.9, and 19.4 ppm)
for the acetonide moiety confirmed that the aforementioned
reduction proceeded to afford the syn diastereomer. a) S. D.
Rychnovsky, D. J. Skalitzky, Tetrahedron Lett. 1990, 31, 945 –
948; b) S. D. Rychnovsky, B. Rogers, G. Yang, J. Org. Chem.
Thus we have described the stereoselective synthesis of
the C1–C12 and C13–C28 fragments of oasomycin A, a
sequence that allowed us to prepare significant quantities of
these materials. The following Communications[23] describe
Angew. Chem. Int. Ed. 2007, 46, 537 –540
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
539

Communications
1997, 62, 3511 – 3515; c) D. A. Evans, D. L. Rieger, J. R. Gage,
Tetrahedron Lett. 1990, 31, 7099 – 7100.
scopic characteristics of 22 to analogous compound 22a, the
stereochemistry of which was confirmed by X-ray crystallogra-
phy.
[7] T. Oishi, T. Nakata, Acc. Chem. Res. 1984, 17, 338 – 344.
[8] For the preparation of phosphonate 5, see: a) G. Pattenden,
B. C. L. Weedon, J. Chem. Soc. C 1968, 1984 – 1997; b) T. Kitara,
A. Horiguchi, K. Mori, Tetrahedron 1988, 44, 4713 – 4720.
[9] Attempts to eliminate the inseparable side products were
unsuccessful under a variety of different conditions. The Lindlar
catalyst in ethyl acetate provided the best selectivity
(5:1:1 favoring reduction of the D4 olefin).
[10] R. H. Crabtree, M. W. Davis, J. Org. Chem. 1986, 51, 2655 – 2661.
[11] The mixture was separated by medium-pressure chromatogra-
phy (MPLC).
[12] The acidic nature of [(NH4)6Mo7O24] resulted in partial cleavage
of the secondary TES group of 3 before the sulfoxide oxidation
was complete (8–10%). The corresponding desilylated product
could be recovered and reprotected (TESOTf, lutidine).
[13] H. S. Schultz, H. B. Freyermuth, S. R. Buc, J. Org. Chem. 1963,
28, 1140 – 1142.
[18] B. S. Bal, W. E. Childers, H. W. Pinnick, Tetrahedron 1981, 37,
2091 – 2096.
[19] J. M. Williams, R. B. Jobson, N. Yasuda, G. Marchesini, U.-H.
Dolling, E. J. Grabowski, Tetrahedron Lett. 1995, 36, 5461 – 5464.
[20] For examples of acetoxysuccinic anhydride reactivity that
illustrate the mentioned effect, see: D. J. Hart, T. K. Yang, J.
Org. Chem. 1985, 50, 235 – 242; S. Gogoi, N. P. Argade,
Tetrahedron 2006, 62, 2999 – 3003, and references therein.
[21] Both the presence of THF and transmetalation of 16b to 16c
were essential for this reaction to run to completion. Thus, the
acylations of 16b and 16c in the absence of THF proceeded in
45% and 66% yield respectively.
[14] R. Baker, J. L. Castro, J. Chem. Soc. Perkin Trans. 1 1990, 47 – 65,
and references therein.
[15] For the synthesis of enantiomer 22, see: a) S. D. Rychnovsky, J.
Org. Chem. 1989, 54, 4982 – 4984; b) J. A. Christopher, P. J.
Kocienski, A. Kuhl, R. Bell, Synlett 2000, 463 – 466; c) M. J.
Remuiæµn, G. Pattenden, Tetrahedron Lett. 2000, 41, 7367 –
7371.
[16] J. R. Parikh, W. von Doering, J. Am. Chem. Soc. 1967, 89, 5505 –
5507.
[17] D. A. Evans, B. D. Allison, M. G. Yang, C. E. Masse, J. Am.
Chem. Soc. 2001, 123, 10840 – 10852. The stereochemistry of the
major allylation product was proven by comparing the spectro-
[22] Y. Oikawa, T. Tanaka, K. Horita, T. Yoshioka, O. Yonemitsu,
Tetrahedron Lett. 1984, 25, 5393 – 5396.
[23] a) D. A. Evans, P. Nagorny, K. J. McRae, D. J. Reynolds, L.-S.
Sonntag, F. Vounatsos, R. Xu, Angew. Chem. 2007, 119, 547 –
550; Angew. Chem. Int. Ed. 2007, 46, 541 – 544; b) D. A. Evans, P.
Nagorny, K. J. McRae, L.-S. Sonntag, D. J. Reynolds, F. Vounat-
sos, Angew. Chem. 2007, 119, 551 – 554; Angew. Chem. Int. Ed.
2007, 46, 545 – 548.
540
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 537 –540