Total synthesis of efomycine M

Article abstract of DOI:10.1002/anie.200701065

(Chemical Equation Presented) Two-directional synthesis: Key steps in the first total synthesis of the anti-inflammatory agent efomycine M (1) are a two-directional C11-C12 chain elongation and an early-stage dimerization reaction. The central stereopentad is synthesized by a highly stereoselective sequence consisting of an onti-aldol reaction and a diastereoselective reduction.

Full text of DOI:10.1002/anie.200701065

Angewandte
Chemie
DOI: 10.1002/anie.200701065
Polyketides
Total Synthesis of Efomycine M**
Roland Barth and Johann Mulzer*
Dedicated to Professor Ekkehard Winterfeldt on the occasion of his 75th birthday
Characteristically, inflammatory skin diseases such as psor-
iasis and atopic dermatitis are accompanied by an expression
of E- and P-selectins by endothelial cells. These selectins
mediate T-cell rolling by their interaction with T-cell-
[1]
expressed sialyl Lewisx (sLex)epitopes.
Specific small-
molecule inhibitors that structurally mimic the binding sites
of selectin ligands reduce the number of skin-infiltrating T
cells. The macrodiolide efomycine M (1)was reported to
exhibit significant anti-inflammatory activity in two different
mouse models of psoriasis by interfering with the binding of
E- and P-selectins.[2] This result was hailed as a new
therapeutic approach in the treatment of human inflamma-
tory disorders,[3] until, quite recently, von Bonin et al., though
confirming the anti-inflammatory profile of 1, strongly
doubted the reported mode of action.[4] In light of this
controversy, detailed studies on the structure–activity rela-
tionships (SAR)appear appropriate for more insight into the
biological profile of the compound.[5] In previous studies, 1
had been obtained semisynthetically from the natural product
elaiophylin (azalomycin B, 2)by base-catalyzed b elimination
of the l-deoxyfucose moiety (Scheme 1).[6–8] Severe efforts to
obtain suitable analogues from a chemical derivatization of 1
failed owing to the lability of the molecule.[5] Thus, we
initiated a total synthesis of 1 which should be flexible enough
to provide suitable derivatives for later SAR experiments.
Structurally, 1 features a 16-membered macrodiolide core,
seven stereogenic centers and a labile a,b-enone moiety.
Prompted by the C2 symmetry of 1 we envisaged a two-
directional approach (Scheme 2).[9] The C11–C12 bond was to
be formed at a late stage by a double nucleophilic attack of an
organometallic species obtained from vinyliodide 4 to dia-
ldehyde 3. The central stereopentad C5–C11 could be
synthesized by an anti-aldol reaction followed by a diaste-
reoselective ketone reduction. The dimerization of 2E,4E-
seco acid 6 was to be achieved by Yamaguchi macrolactoni-
zation. Fragment C12–C16 could be synthesized from methyl
Scheme 1. Base-catalyzed formation of efomycine M (1)from elaiophy-
lin (azalomycin B, 2).[6]
Scheme 2. Retrosynthetic analysis of 1. TBS=tert-butyldimethylsilyl,
TIPS=triisopropylsilyl, Bn=benzyl, PG=protecting group.
[*] R. Barth, Prof. Dr. J. Mulzer
Institut für Organische Chemie
Universität Wien
(R)-3-hydroxybutyric acid (5)by a Frµter–Seebach alkylation
and C1 homologation.
Our synthesis started with the selective formation of the E
enolate of b-ketoimide 8 using Cy2BCl/Me2NEt (Scheme 3).
Subsequent reaction with aldehydes 9a[10] or 9b[11] gave b-
Währingerstrasse 38, 1090 Vienna (Austria)
Fax: (+43)1-4277-9521
E-mail: johann.mulzer@univie.ac.at
[**] R.B. gratefully acknowledges financial support from the Austrian
Academy of Sciences (DOC fellowship). We thank Dr. B. Buchmann,
Schering AG, for helpful discussions and an authentic sample of
efomycine M. We also thank Dr. H. Kählig and S. Felsinger for NMR
spectra and M. Zinke for HPLC analysis.
hydroxyketones 10a (62%, d.r. 91:9)and
10b (80%,
d.r. 95:5), respectively, in high yield. The excellent diastereo-
selectivity of the addition can be explained in terms of
matched double stereodifferentiation, as aldehyde 9 regularly
prefers the Felkin–Anh mode.[12] The use of TBDPS as a
protecting group at C11 gave higher yields and better
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2007, 46, 5791 –5794
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5791

Communications
seco Acid 6 was obtained after cleavage of the PMB
protecting group and base-induced hydrolysis of the methyl
ester.
Dimerization of the seco methyl ester 16 with distannox-
ane catalyst 17 as reported by Panek et al. failed and gave seco
acid 6 only.[18] Therefore, 6 was dimerised by means of the
modified Yamaguchi macrolactonization protocol (Yone-
mitsu conditions)[19] to give the crystalline dimer 18a (59%)
along with minor amounts of the uncyclized dimer 18b. The
primary TBS group was easily removed in the presence of the
secondary one with a dilute solution of HF·pyridine in THF
(Scheme 4), and the resulting dialcohol 19 was oxidized to
give dialdehyde 3.
The synthesis of fragment C12–C16 started with the
diastereoselective alkylation of methyl (R)-3-hydroxybutyric
[20]
acid (5)(Scheme 5).
Protection of the hydroxy group as a
Scheme 3. Preparation of stereopentad 14. Reagents and conditions:
a)Cy 2BCl, Me2NEt, Et2O, ꢀ788C!08C (10a: 62%, d.r. 91:9; 10b: 80%,
d.r. 95:5); b) NaBH(OAc) 3, AcOH, MeCN, ꢀ408C (11a: 72%, d.r. 94:6; 11b:
79%, d.r. 96:4); c) DDQ, MS 4 Š, CH 2Cl2, 08C (33%); d) TBSOTf, 2,6-lutidine,
CH2Cl2, 08C (99%); e) LiBH 4, H2O, Et2O, 08C (69%); f) LiBH 4, H2O, Et2O,
08C (87%); g) PMPCH(OMe) 2, (ꢁ)-CSA, CH2Cl2, RT; h)TBAF, THF, RT (99%
over two steps); i) TBSOTf, 2,6-lutidine, CH 2Cl2, 08C (91%); j) DIBAL, CH 2Cl2,
ꢀ308C (69%). Xp=(4R)-methylphenyloxazolidin-2-one-3-yl, PMB=para-
methoxybenzyl, TBDPS=tert-butyldiphenylsilyl, PMP=para-methoxyphenyl,
Cy =cyclohexyl, DDQ=2,3-dichloro-5,6-dicyanobenzoquinone, MS=molecular
sieves, TBS=tert-butydimethylsilyl, OTf=triflouromethanesulfonate, PMB-
(OMe)2 =para-methoxybenzaldehyde dimethylacetal, CSA=camphorsulfonic
acid, TBAF=tetrabutylammonium fluoride, DIBAL=diisobutylaluminum hy-
dride.
TIPS ether followed by DIBAL reduction and oxidation gave
aldehyde 21 in 97% overall yield. The TIPS protecting group
diastereoselectivity than the PMB ether. A stereoselective
anti reduction with NaBH(OAc)3 completed the synthesis of
the stereopentads 11a (72%, d.r. 94:6)and 11b (79%,
d.r. 96:4, Scheme 3).[13]
Next we had to differentiate between the two secondary
hydroxy groups. In our initial approach we used the primary
PMB group of 11a to protect the neighboring free hydroxy
group as a PMP acetal.[14] However, the protection of the
remaining hydroxy group and reductive removal of the
auxiliary proved to be problematic because of a reductive
opening of the isoxazolidinone ring to give 12 (Scheme 3).
A satisfactory solution was finally achieved by reductive
removal of the auxiliary of 11b followed by regioselective
protection of the terminal 1,3-diol as a PMP acetal by
treatment with PMPCH(OMe)2 to give 13. At a later stage,
the differentiation of the C9 and C11 OH functions was
envisaged. This was deemed easier by introducing two TBS
groups. Regioselective reductive opening of the acetal with
DIBAL[15] furnished stereopentad 14 in gram quantities.
We next planned to construct the diene ester moiety of 1
in a single olefination reaction (Scheme 4). Oxidation of the
primary alcohol was best achieved with Dess–Martin period-
inane; Swern oxidation resulted in partial epimerization at
C6. Wittig reaction using phosphonium salt 7a[16] gave
moderate yields only (46%), and the phosphine oxide was
difficult to remove by chromatography. In contrast, the
Horner–Wadsworth–Emmons olefination with phosphonate
7b and LDA[17] delivered 2E,4E dienoate 15 as a single
stereoisomer in excellent yield (up to 96%, 4E:4Z > 50:1).
Scheme 4. Horner–Wadsworth–Emmons olefination and Yamaguchi
dimerization. Reagents and conditions: a)Dess–Martin periodinane,
CH2Cl2, 08C!RT; b) 7b, LDA, THF, ꢀ788C (89% over two steps,
E:Z>50:1); c) DDQ, phosphate buffer pH 7, CH2Cl2, 08C!RT (98%);
d)LiOH, H 2O, THF, RT (91%); e) 2,4,6-Cl 3C6H2C(O)Cl, NEt3, toluene,
RT, then DMAP (18a: 59%); f) 7% HF·pyridine, THF, RT (86%);
g)Dess–Martin periodinane, CH 2Cl2, 08C (94%). LDA=lithium diiso-
propylamide, DMAP=4-dimethylaminopyridine.
Scheme 5. Synthesis of vinyliodide 4. Reagents and conditions: a) 5,
LDA, THF, ꢀ408C, then EtI, ꢀ788C!RT (76%, d.r. 97:3); b) TIPSOTf,
2,6-lutidine, CH2Cl2, 08C; c)DIBAL, toluene, ꢀ788C (99% over two
steps); d) Dess–Martin periodinane, CH 2Cl2, RT (99%); e) TMSCHN2,
nBuLi, Et2O, ꢀ788C, then 21, 08C (58%); f) [Cp 2Zr(H)Cl], THF, RT,
then I2 (66%, E:Z>50:1); g) CrCl 2, CHI3, THF, 08C (18%, E:Z>40:1).
Cp=cyclopentadienyl.
5792
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 5791 –5794

Angewandte
Chemie
proved to be the most compatible one in this sequence among
a variety of alternatives such as TBS, PMB, and MOM. We
first tried to synthesize vinyliodide 4 by Takai iodoolefina-
tion.[21] However, all attempts gave 4 in low yields only (18%)
although with a high E:Z ratio (> 40:1). In a more convenient
approach, aldehyde 21 was converted into the alkyne 22 by a
Colvin rearrangement using TMSCHN2/nBuLi.[22] The more
popular Corey–Fuchs homologation[23] turned out to be very
sluggish and required repeated chromatographic purification.
Finally, 22 was hydrozirconated with the Schwartz reagent
according to a protocol devised by Negishi et al.[24] The
organozirconium intermediate was quenched with iodine to
give E vinyliodide 4 (66%, E:Z > 50:1).
ing. In this connection, stereochemical variations as well as
modifications of the C12–C16 side chain will be examined.
Received: March 10, 2007
Published online: June 26, 2007
Keywords: aldol reaction · asymmetric synthesis · macrolides ·
.
polyketides · total synthesis
[1] a)T. Springer, Cell 1994, 76, 301 – 314; b)E. C. Butcher, L. J.
Picker, Science 1996, 272, 60 – 66; c)A. Varki, Proc. Natl. Acad.
Sci. USA 1994, 91, 7390 – 7397.
[2] M. P. Schön, T. Krahn, M. Schön, M.-L. Rodriguez, H. Antoni-
cek, J. E. Schultz, R. J. Ludwig, T. M. Zollner, E. Bischoff, K.-D.
Bremm, M. Schramm, K. Henninger, R. Kaufmann, H. P. M.
Gollnick, C. M. Parker, W.-H. Boehncke, Nat. Med. 2002, 8,
366 – 372.
For the coupling of C11 and C12 (Scheme 6)we first
attempted a CrCl2/NiCl2-mediated Nozaki–Hiyama–Kishi
reaction,[25] which gave diallylic alcohol 23 in a disappoint-
[3] G. Todderud, X. Nair, D. Lee, J. Alford, L. Davern, P. Stanley, C.
Bachand, P. Lapointe, A. Marinier, A. Martel, M. Menard, J. J.
Wright, J. Bajorath, D. Hollenbaugh, A. Aruffo, K. M. Tram-
posch, J. Pharmacol. Exp. Ther. 1997, 282, 1298 – 1304.
[4] A. von Bonin, B. Buchmann, B. Bader, A. Rausch, K. Venstrom,
M. Schäfer, S. Gründemann, J. Günther, L. Zorn, R. Nubbe-
meyer, K. Asadullah, T. M Zollner, Nat. Med. 2006, 12, 873 – 874.
[5] B. Buchmann et al., 18. Irseer Naturstofftage der DECHEMA
e.V. 2006, poster session and private communication.
[6] P. Hammann, G. Kretzschmar, Z. Naturforsch. B 1990, 45, 515 –
517.
[7] Isolation of 2: a)F. M. Arcamone, C. Bertazzoli, M. Ghione,
T. G. Scotti, Giorn. Microbiol. 1959, 7, 207 – 216; b)M. Arai, J.
Antibiot. Ser. A 1960, 13, 46 – 50.
[8] Total synthesis of 2: a)K. Toshima, K. Tatsuta, M. Kinoshita,
Bull. Chem. Soc. Jpn. 1988, 61, 2369 – 2381; total syntheses of the
aglycon of 2: b)D. Seebach, H.-F. Chow, R. F. W. Jackson, K.
Lawson, M. A. Sutter, S. Thaisrivongs, J. Zimmermann, J. Am.
Chem. Soc. 1985, 107, 5292 – 5293; c)D. A. Evans, D. M. Fitch, J.
Org. Chem. 1997, 62, 454 – 455; d)I. Paterson, D.-G. Lombart, C.
Allerton, Org. Lett. 1999, 1, 19 – 22.
Scheme 6. The final C11–C12 bond connection. Reagents and condi-
tions: a) 4, CrCl2, NiCl2, THF, DMF, RT (34%); b) 22, [Cp2Zr(H)Cl],
THF, RT, then Et2Zn, ꢀ788C, toluene, then 3, RT (<5%); c) 4, tBuLi,
Et2O, ꢀ788C, then 3, ꢀ788C!08C (89%); d) Dess–Martin period-
inane, CH2Cl2, 08C!RT (82%); e) 70% HF·pyridine, THF, MeCN, RT
(70%). DMF=N,N-dimethylformamide.
[9] For a discussion of symmetry-based synthesis, see C. R. Poss,
S. L. Schreiber, Acc. Chem. Res. 1994, 27, 9 – 17.
[10] Preparation of aldehyde 9a: J. Hassfeld, M. Kalesse, Synlett
2002, 2007 – 2010.
[11] Preparation of aldehyde 9b: PP. R. R. Meira, L. C. Diaz, J. Org.
Chem. 2005, 70, 4762 – 4773.
[12] D. A. Evans, H. P. Ng, S. Clark, D. L. Rieger, Tetrahedron 1992,
48, 2127 – 2142.
[13] a)D. A. Evans, K. T. Chapman, E. M. Carreira, J. Am. Chem.
Soc. 1988, 110, 3560 – 3578. b)NaBH(OAc) 3 gave higher yields
than the corresponding tetramethylammonium salt.
[14] Y. Oikawa, T. Yoshioka, O. Yonemitsu, Tetrahedron Lett. 1982,
23, 889 – 892.
ingly low yield of 34%. Hydrozirconation of alkyne 22
followed by transmetalation with Et2Zn[26] and addition to 3
failed (< 5%). Finally, lithiation of vinyliodide 4 with tBuLi
and subsequent addition of 3 gave 23 in excellent yield (89%)
as a statistical 2:1:1 mixture of stereoisomers. The macro-
lactone carbonyl functions were completely unreactive even
in the presence of excess vinyllithium reagent. Oxidation of
the diastereomeric mixture to give enone 24 followed by
desilylation with 70% HF·pyridine in MeCN/THF gave 1 in
57% yield over two steps. The use of TBAF resulted in
elimination to give the 10E,12E dienone, whereas AcOH-
buffered TBAF gave no reaction at all. The 1H and 13C NMR,
IR, and mass spectra and the optical rotation of our synthetic
sample of 1 were in complete agreement with those obtained
from authentic material.[27]
[15] S. Takano, M. Akiyama, S. Sato, K. Ogasawara, Chem. Lett.
1983, 1593 – 1596.
[16] E. Buchta, F. Andree, Chem. Ber. 1959, 92, 3111 – 3116.
[17] a)K. Sato, S. Mizuno, M. Hirayama, J. Org. Chem. 1967, 32, 177 –
180; b)W. R. Roush, J. Am. Chem. Soc. 1980, 102, 1390 – 1404.
[18] a)T. Yano, Y. Himeno, H. Nozaki, J. Otera, Tetrahedron Lett.
1986, 27, 4501 – 4504; b)Q. Su, A. B. Beeler, E. Lobkovsky, J. A.
Porco, Jr., J. S. Panek, Org. Lett. 2003, 5, 2149 – 2152.
[19] a)J. Inanaga, K. Hirata, H. Saeki, T. Katsuki, M. Yamaguchi,
Bull. Chem. Soc. Jpn. 1979, 52, 1989 – 1993; b)H. Tone, T. Nishi,
Y. Oikawa, M. Hikota, O. Yonemitsu, Tetrahedron Lett. 1987, 28,
4569 – 4572; c)Macrolactonization according to the Yamaguchi
protocol was expected to be more compatible with our system
In conclusion, we have completed the first total synthesis
of efomycine M (1)by a convergent approach in 17 steps over
the longest linear sequence in 7% overall yield. Currently we
are extending our stereo- and regiocontrolled synthesis for
the preparation of simplified analogues for biological screen-
Angew. Chem. Int. Ed. 2007, 46, 5791 –5794
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5793

Communications
than other macrolactonization methods: M. Sutter, D. Seebach,
Liebigs Ann. Chem. 1983, 939 – 949.
[24] a)D. W. Hart, J. Schwartz, J. Am. Chem. Soc. 1974, 96, 8115 –
8116; b)Z. Huang, E.-I. Negishi, Org. Lett. 2006, 8, 3675 – 3678.
[25] a)K. Takai, K. Kimura, T. Kuroda, T. Hiyama, H. Nozaki,
Tetrahedron Lett. 1983, 24, 5281 – 5284; b)Review: A. Fürstner,
Chem. Rev. 1999, 99, 991 – 1045. The Nozaki–Hiyama–Kishi
coupling between 4 and 3 gave moderate yields in DMF (34%)
and low yields in DMSO (10%).
[26] a)W. Xu, P. Wipf, Tetrahedron Lett. 1994, 35, 5197 – 5200;
b)Review: H. Jahn, P. Wipf, Tetrahedron 1996, 52, 12853 –
12910.
[20] a)G. Frµter, Helv. Chim. Acta 1979, 62, 2825 – 2828; b)H.-F.
Chow, R. F. W. Jackson, M. A. Sutter, S. Thaisrivongs, J.
Zimmermann, D. Seebach, Liebigs Ann. Chem. 1986, 1281 –
1308.
[21] K. Nitta, K. Utimoto, K. Takai, J. Am. Chem. Soc. 1986, 108,
7408 – 7410.
[22] K. Miwa, T. Aoyama, T. Shioiri, Synlett 1994, 107 – 108.
[23] P. L. Fuchs, E. J. Corey, Tetrahedron Lett. 1972, 13, 3769 – 3772.
[27] See the acknowledgments.
5794
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 5791 –5794