An abiotic strategy for the enantioselective synthesis of erythromycin B

Article abstract of DOI:10.1002/anie.200351136

Sweet success: The classical approach to macrolide antibiotics involves a macrocyclization followed by the introduction of the carbohydrate residue(s). Now for the first time, a glycosylated seco-acid (see scheme, left) was cyclized to give an intermediat

Full text of DOI:10.1002/anie.200351136

Communications
Natural Product Synthesis
An Abiotic Strategy for the Enantioselective
Synthesis of Erythromycin B**
Paul J. Hergenrother, Anne Hodgson, Andrew S. Judd,
Wen-Cherng Lee, and Stephen F. Martin*
Erythromycin A (1) and erythromycin B (2), which owe their
antibiotic activity to their ability to inhibit ribosomal-depend-
ent protein biosynthesis, are arguably the best-known mem-
bers of the macrolide family of antibiotics.[1] Despite the
advent of a myriad of newer antibiotics, 1 continues to be the
drug of choice for the clinical treatment of numerous
pathogenic bacteria. Consequent to their important antibiotic
activity and their complex structures, these natural products
have been the objects of synthetic investigations for more
than two decades,[2] and a substantial amount of novel and
useful chemistry has emerged from these efforts. Indeed,
these antibiotics have served as a veritable testing ground for
evaluating methods for effecting acyclic stereochemical con-
trol. All these advances notwithstanding, it is noteworthy that
there has only been a single report of the total synthesis of 1
and of 2. In an elegant series of studies that culminated in
1981 with the total synthesis of 1, the Woodward group not
only solved problems of stereochemical control and glyco-
sylation of the macrocyclic lactone, but also identified many
of the control elements associated with the macrolactoniza-
tion of seco-acid derivatives.[3] Oishi and co-workers subse-
quently recorded a formal synthesis of 1,[4] and the Tatsuta
group developed an alternative strategy for glycosylating the
macrocyclic aglycone.[5] More recently we reported the total
synthesis of 2 through a plan in which the carbohydrate
residues were also introduced following the macrolactoniza-
tion step.[6] Indeed, the synthetic strategy of installing the
carbohydrate groups after cyclization of a suitable seco-acid
derivative follows the general lines of the putative biosyn-
thesis of the macrolide antibiotics.[7,8]
We have long been interested in exploring a different end
game for the synthesis of the erythromycin antibiotics in
which either one or both of the carbohydrate substituents
[*] Prof. Dr. S. F. Martin, P. J. Hergenrother, Dr. A. Hodgson,
Dr. A. S. Judd, W.-C. Lee
Department of Chemistry and Biochemistry, The University of Texas
Austin, TX 78712 (USA)
Fax: (+1)512-471-4180
E-mail: sfmartin@mail.utexas.edu
[**] Taken in part fromthe PhD Dissertations of W.-C.L, The University of
Texas, Austin, 1992, and P.J.H, The University of Texas, Austin, 1999
(present address, Department of Chemistry, University of Illinois,
Urbana, IL 61801 (USA)). We thank the NIH (GM 31077), the
Robert A. Welch Foundation, Pfizer, Inc, and Merck Research
Laboratories for their generous support of this research. A.S.J. is
also grateful for an NRSA Postdoctoral Fellowship fromthe NIH
(5F32GM19992). We thank Dr. Paul A. Lartey (Abbott Laboratories)
for a generous gift of erythromycin A for use as a source of
d-desosamine and l-cladinose, and we are grateful to Dr. Erica
Kraynack and Dr. Philippe Breton for conducting exploratory
glycosylation studies.
3278
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/anie.200351136
Angew. Chem. Int. Ed. 2003, 42, 3278 – 3281

Angewandte
Chemie
would be appended prior to the macrolactonization step
(Scheme 1). In principle, the carbohydrate residues them-
selves might serve as protecting groups for the hydroxy
functions at C3 and C5, thereby potentially resulting in a
more concise route to 2. This strategy was not without its risks,
however, as the Woodward group had nicely defined many of
11 with 12 under conditions that were defined only after
considerable experimentation gave a mixture of 13 and 14
(1.2:1) in a combined yield of 70%, together with 20–25% of
recovered 11 (Scheme 3). Glycosylation of the tertiary
hydroxy group at C6 of 11 was thus remarkably facile, and
14 was formed in significant quantities, irrespective of
numerous variations in the coupling procedure. We
also explored the feasibility of protecting the hydroxy
group at C6 but were unable to introduce the desos-
amine subunit on such protected substrates. Ulti-
mately, we had to be content with separating the
isomeric C5- and C6-glycosylated products and con-
tinuing the synthesis with 13.
To introduce the remaining three-carbon unit of
the erythromycin backbone, it was first necessary to
oxidize the alcohol function at C3 to an aldehyde.
However, we found that if the hydroxy group at C6
was not protected prior to generating the aldehyde at
C3, a five-membered lactol that could not be advanced
in the synthesis was unavoidably formed. The most
expeditious solution to this problem involved depro-
tection of the primary alcohol group at C3 of 13
followed by reaction of the resultant diol with
TESOTf to give an intermediate 3,5-diprotected
Scheme 1. Retrosynthetic analysis of erythromycin A (1) and erythromycin B (2).
Pg=protecting group, BOM=benzyloxymethyl.
diol, which underwent desilylation and Swern oxida-
tion of the hydroxy function at C3 to give 15
the structural parameters required for the successful cycliza-
tion of erythromycin seco-acid derivatives.[3] These studies
had clearly revealed the critical importance of rigidifying the
seco-acid backbone in the regions spanning the C3–C5 and
C8–C12 segments, typically by using cyclic derivatives that
incorporated the oxygen substituents at C3, C5, C9, and C11.
Hence, to establish the underlying viability of this abiotic
approach to the erythromycins, we first established in a key
experiment that a derivative of 3 did indeed undergo macro-
lactonization.[9] We now report the successful implementation
of the strategy outlined in Scheme 1for the synthesis of
2
from subunits 4–6.
The synthesis commences with selective protection of the
primary hydroxy group of the triol 5, which we had previously
prepared in seven steps from ethyl furan,[10] to give 7
(Scheme 2). Conversion of the diol moiety in 7 into a cyclic
carbonate followed by removal of the dithiolane protecting
group afforded the ketone 8.[11] The subsequent aldol reaction
of the enolate of 8 with the known aldehyde 4[10] proceeded
with high diastereoselectivity (d.r. > 95:5) to give 9. The C9
ketone function was stereoselectively (d.r. = 98:2) reduced
with Me4NBH(OAc)3 according the Evans protocol to give
10,[12] thereby setting the S configuration at C9 that would be
required for the macrolactonization. The C9–C11 anti diol
moiety was then protected as its cyclic mesitylene (Mes)
acetal, and the cyclic carbonate protecting group was
removed to furnish the glycosylation substrate 11, thus
completing the first phase of the synthesis.
Prior art in the area of introducing desosamine residues
onto macrocyclic lactones suggested that 11 should react
selectively with a suitable desosamine donor to give 13.[3,5,6]
Much to our surprise, however, we found that the reaction of
Scheme 2. Synthesis of C3–C15 backbone subunit 11. TBS=tert-butyl-
dimethylsilyl, ImH=imidazole, DMF=N,N-dimethylformamide,
CDI=N,N-carbonyl diimidazole, HMDS=1,1,1,3,3,3-hexamethyldisila-
zane, CSA=(À)-camphorsulfonic acid.
Angew. Chem. Int. Ed. 2003, 42, 3278 – 3281
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3279

Communications
(Scheme 4).[13] The reaction of 15 with crotylstannane in the
presence of BF3·OEt2 then furnished 16[14] in 84% yield,
together with about 12% of a mixture of other diastereomeric
adducts. We briefly explored the possibility of introducing the
remaining carbon atoms into 15 through a number of different
aldol reactions,[15] but these efforts did not yield significant
quantities of the desired adduct.
À
It then remained to oxidize the terminal C C double bond
to generate the C1carboxylic acid. In previous work, we had
developed two- and three-step protocols for effecting this
transformation,[6,10] but we found that the oxidation of 16 to
give 17 by using a recently developed procedure for organo-
metallic ozonolysis provided an efficient one-step alterna-
tive.[16] Removal of the protecting group from the C13
hydroxy function by hydrogenolysis under carefully defined
conditions gave an unstable hydroxy acid, which was then
cyclized according to the Yamaguchi protocol to furnish 18.[17]
Deprotection of the hydroxy group at C6 then delivered 19,
which we had previously converted in five steps into
erythromycin B (2), thereby completing a synthesis of eryth-
romycin B (2).[6]
Scheme 3. Synthesis of glycosylated C3–C15 backbone subunit 13.
pyrm=2-pyrimidinyl, Tf=trifluoromethanesulfonyl, MS=molecular
sieves, Py=pyridine, Mc=methoxycarbonyl.
In conclusion, we have completed a novel synthesis of
erythromycin B by using an abiotic strategy in which a key
step is the cyclization of a glycosylated seco-acid derivative.
As such, this synthesis represents the first time that a naturally
occuring macrolide antibiotic has been prepared through an
approach in which a sugar residue is appended prior to the
macrolactonization step. Moreover, the ability to cyclize the
hydroxy acid derived from 17 clearly illustrates that more
structural flexibility in the backbone can be tolerated in the
cyclization step than was previously recognized.[3]
Received: February 7, 2003 [Z51136]
Keywords: asymmetric synthesis · cyclization · glycosylation ·
.
macrolactonization · natural products · total synthesis
[1] Macrolide Antibiotics (Ed.: S. Omura), Academic Press,
Orlando, FL, 1984.
[2] For leading references and representative synthetic approaches,
see: a) E. J. Corey, P. B. Hopkins, S. Kim, S.-E. Yoo, K. P.
Nambiar, J. R. Falck, J. Am. Chem. Soc. 1979, 101, 71 31 ; b) S.
Masamune, M. Hirama, S. Mori, S. A. Ali, D. S. Garvey, J. Am.
Chem. Soc. 1981, 103, 1568; c) B. Bernet, P. M. Bishop, M.
Caron, T. Kawamata, B. L. Roy, L. Ruest, G. Sauve, P. Soucy, P.
Deslongchamps, Can. J. Chem. 1985, 63, 2810, 2814, 2818; d) G.
Stork, S. D. Rychnovsky, J. Am. Chem. Soc. 1987, 109, 1564,
1565; e) A. R. Chamberlin, M. Dezube, S. H. Reich, D. J. Sall, J.
Am. Chem. Soc. 1989, 111, 6247; f) I. Paterson, D. J. Rawson,
Tetrahedron Lett. 1989, 30, 7463; g) M. Nakata, M. Arai, K.
Tomooka, N. Ohsawa, M. Kinoshita, Bull. Chem. Soc. Jpn. 1989,
62, 2618; h) M. Hikota, H. Tone, K. Horita, O. Yonemitsu, J. Org.
Chem. 1990, 55, 7; i) D. C. Myles, S. J. Danishefsky, G. Schulte, J.
Org. Chem. 1990, 55, 1636; j) M. Hikota, H. Tone, K. Horita, O.
Yonemitsu, Tetrahedron 1990, 46, 4613; k) A. F. Sviridov, V. S.
Borodkin, M. S. Ermolenko, D. V. Yashunsky, N. K. Kochetkov,
Tetrahedron 1991, 47, 2291, 2317; l) J. Mulzer, P. A. Mareski, J.
Buschmann, P. Luger, Synthesis 1992, 215; m) R. Sturmer, K.
Ritter, R. W. Hoffmann, Angew. Chem. 1993, 105, 1 1 2A; ngew.
Chem. Int. Ed. Engl. 1993, 32, 101; n) D. A. Evans, A. S. Kim,
Tetrahedron Lett. 1997, 38, 53.
Scheme 4. Synthesis of erythromycin B (2). TBAF=tetrabutylammonium fluoride,
TES=triethylsilyl, DMSO=dimethyl sulfoxide, NMM=N-methylmorpholine,
DMAP=4-dimethylaminopyridine.
3280
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 3278 – 3281

Angewandte
Chemie
[3] R. B. Woodward, E. Logusch, K. P. Nambiar, K. Sakan, D. E.
Ward, B.-W. Au-Yeung, P. Balaram, L. J. Browne, P. J. Card,
C. H. Chen, R. B. Chenevert, A. Fliri, K. Frobel, H. J. Gais, D. G.
Garratt, K. Hayakawa, W. Heggie, D. P. Hesson, D. Hoppe, I.
Hoppe, J. A. Hyatt, D. Ikeda, P. A. Jacobi, K. S. Kim, Y. Kojima,
K. Krowicki, V. J. Lee, T. Leutert, S. Malchenko, J. Martens, R. S.
Matthews, B. S. Ong, J. B. Press, T. V. Rajan Babu, G. Rousseau,
H. M. Sauter, M. Suzuki, K. Tatsuta, L. M. Tolbert, E. A.
Truesdale, I. Uchida, Y. Ueda, T. Uyehara, A. T. Vasella, W. C.
Vladuchick, P. A. Wade, R. M. Williams, H. N.-C. Wong, J. Am.
Chem. Soc. 1981, 103, 3210; R. B. Woodward, E. Logusch, K. P.
Nambiar, K. Sakan, D. E. Ward, B.-W. Au-Yeung, P. Balaram,
L. J. Browne, P. J. Card, C. H. Chen, R. B. Chenevert, A. Fliri, K.
Frobel, H. J. Gais, D. G. Garratt, K. Hayakawa, W. Heggie, D. P.
Hesson, D. Hoppe, I. Hoppe, J. A. Hyatt, D. Ikeda, P. A. Jacobi,
K. S. Kim, Y. Kojima, K. Krowicki, V. J. Lee, T. Leutert, S.
Malchenko, J. Martens, R. S. Matthews, B. S. Ong, J. B. Press,
T. V. Rajan Babu, G. Rousseau, H. M. Sauter, M. Suzuki, K.
Tatsuta, L. M. Tolbert, E. A. Truesdale, I. Uchida, Y. Ueda, T.
Uyehara, A. T. Vasella, W. C. Vladuchick, P. A. Wade, R. M.
Williams, H. N.-C. Wong, J. Am. Chem. Soc. 1981, 103, 3213;
R. B. Woodward, E. Logusch, K. P. Nambiar, K. Sakan, D. E.
Ward, B.-W. Au-Yeung, P. Balaram, L. J. Browne, P. J. Card,
C. H. Chen, R. B. Chenevert, A. Fliri, K. Frobel, H. J. Gais, D. G.
Garratt, K. Hayakawa, W. Heggie, D. P. Hesson, D. Hoppe, I.
Hoppe, J. A. Hyatt, D. Ikeda, P. A. Jacobi, K. S. Kim, Y. Kojima,
K. Krowicki, V. J. Lee, T. Leutert, S. Malchenko, J. Martens, R. S.
Matthews, B. S. Ong, J. B. Press, T. V. Rajan Babu, G. Rousseau,
H. M. Sauter, M. Suzuki, K. Tatsuta, L. M. Tolbert, E. A.
Truesdale, I. Uchida, Y. Ueda, T. Uyehara, A. T. Vasella, W. C.
Vladuchick, P. A. Wade, R. M. Williams, H. N.-C. Wong, J. Am.
Chem. Soc. 1981, 103, 3215
[4] T. Nakata, M. Fukui, T. Oishi, Tetrahedron Lett. 1988, 29, 2219;
T. Nakata, M. Fukui, T. Oishi, Tetrahedron Lett. 1988, 29, 2223.
[5] K. Toshima, Y. Nozaki, S. Mukaiyama, T. Tamai, M. Nakata, K.
Tatsuta, M. Kinoshita, J. Am. Chem. Soc. 1995, 117, 3717.
[6] S. F. Martin, T. Hida, P. R. Kym, M. Loft, A. Hodgson, J. Am.
Chem. Soc. 1997, 119, 3193.
[7] J. Staunton, B. Wilkinson, Chem. Rev. 1997, 97, 2611.
[8] For a critical overview of strategic approaches, see: J. Mulzer,
Angew. Chem. 1991, 103, 1484; Angew. Chem. Int. Ed. Engl.
1991, 30, 1452.
[9] S. F. Martin, M. Yamashita, J. Am. Chem. Soc. 1991, 113, 5478.
[10] S. F. Martin W.-C. Lee, G. J. Pacofsky, R. P. Gist, T. A. Mulhern,
J. Am. Chem. Soc. 1994, 116, 4674.
[11] The structure assigned to each compound was in accord with its
spectral (1H and 13C NMR, IR, MS) characteristics. Analytical
samples of new compounds were obtained by flash chromatog-
raphy and gave satisfactory identification by high-resolution
mass spectrometry. All yields are based on isolated, purified
materials that were ꢀ 95% pure.
[12] D. A. Evans, K. T. Chapman, E. M. Carreira, J. Am. Chem. Soc.
1988, 110, 3560.
[13] G. A. Tolstikov, M. S. Miftakhov, M. E. Adler, N. G. Komis-
sarova, O. M. Kuznetsov, N. S. Vostrikov, Synthesis 1989, 940.
[14] Preliminary attempts to introduce the cladinose residue to the
masked seco-acid 16 were unsuccessful.
[15] a) D. A. Evans, J. V. Nelson, E. Vogel, T. R. Taber, J. Am. Chem.
Soc. 1981, 103, 3099; b) M. T. Crimmins, B. W. King, E. A. Tabet,
K. Chaudhary, J. Org. Chem. 2001, 66, 894; c) M. T. Crimmins, K.
Chaudhary, Org. Lett. 2000, 2, 775.
[16] B. R. Travis, R. S. Narayan, B. Borhan, J. Am. Chem. Soc. 2002,
124, 3824.
[17] J. Inanaga, K. Hirata, H. Saeki, T. Katsuki, M. Yamaguchi, Bull.
Chem. Soc. Jpn. 1979, 52, 1989.
Angew. Chem. Int. Ed. 2003, 42, 3278 – 3281
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3281