A versatile approach toward the ansamycin antibiotics

Article abstract of DOI:10.1021/ol060022b

The ansamycin antibiotics contain metacyclophanic macrolactams, many of which possess potent antitumor activity. Only a few total syntheses of this family of natural products have been reported, and modifications to increase potency have not been describe

Full text of DOI:10.1021/ol060022b

ORGANIC
LETTERS
2006
Vol. 8, No. 5
975-978
A Versatile Approach toward the
Ansamycin Antibiotics
Weimin Peng and Brian S. J. Blagg*
Department of Medicinal Chemistry and The Center for Chemical Methodology and
Library DeVelopment, The UniVersity of Kansas, 1251 Wescoe Hall DriVe,
Malott 4070, Lawrence, Kansas 66045-7562
bblagg@ku.edu
Received January 4, 2006
ABSTRACT
The ansamycin antibiotics contain metacyclophanic macrolactams, many of which possess potent antitumor activity. Only a few total syntheses
of this family of natural products have been reported, and modifications to increase potency have not been described. Therefore, a method
was developed to prepare the trienomycin A core via resin-bound triphenylphosphonium salts, which serve as both a reagent and a traceless
linker to afford olefinic products that undergo ring-closing metathesis (RCM) to give macrocyclic scaffolds of varying ring sizes.
The ansamycin family of antibiotics consists of more than
20 members, many of which exhibit potent biological
activities such as geldanamycin, from which derivatives are
currently in clinical trials for the treatment of cancer (Figure
1).1,2 Another important member of this family is trienomycin
A, which like geldanamycin is also a macrolactam that
contains an oxygenated aromatic ring.3 Similar to geldana-
mycin, trienomycin A has exceptional activity against several
cancer cell lines4,5 and has been shown to cause a decrease
in nitric oxide synthase,6 suggesting that these molecules may
share a common biological target.7,8 Because the quinone
ring of GDA has demonstrated redox-active behavior,9 it has
been proposed that more stable derivatives are likely to serve
as important leads for clinical development.10 Trienomycin
A lacks the quinone moiety and instead contains a phenol,
which is not subject to the same redox-active behavior. The
ancillary ring of trienomycin A contains several function-
alities that differ from those of geldanamycin and perhaps
can be optimized for increased biological activity. Such
attributes make trienomycin A an excellent target for the
development of analogues and elucidation of its biological
target.
In an effort to provide synthetic access to members of the
ansamycin family, we sought to develop a versatile method
(1) Banerji, U. Proc. Am. Assoc. Cancer Res. 2003, 44, 677.
(2) Sausville, E. A. Curr. Cancer Drug Targets 2003, 3, 377-383.
(3) Funayama, S.; Okada, K.; Komiyama, K.; Umezawa, I. J. Antibiot.
1985, 38, 1107-1109.
(7) Garcia-Cardena, G.; Fan, R.; Shah, V.; Sorrentino, R.; Cirino, G.;
Papapetropoulos, A.; Sessa, W. C. Nature 1998, 392, 821-824.
(8) Pritchard, K. A.; Ackerman, A. W.; Gross, E. R.; Stepp, D. W.; Shi,
Y.; Fontana, J. T.; Baker, J. E.; Sessa, W. C. J. Biol. Chem. 2001, 276,
17621-17624.
(4) Komiyama, K.; Hirokawa, Y.; Yamaguchi, H.; Funayama, S.;
Masuda, K.; Anraku, Y.; Umezawa, I.; Omura, S. J. Antibiot. 1987, 40,
1768-1772.
(9) Dikalov, S.; Landmesser, U.; Harrison, D. G. J. Biol. Chem. 2002,
277, 25480-25485.
(5) Umezawa, I.; Funayama, S.; Okada, K.; Iwasaki, K.; Satoh, J.;
Masuda, K.; Komiyama, K. J. Antibiot. 1985, 38, 699-705.
(6) Kim, W.-G.; Song, N.-K.; Yoo, I.-D. J. Antibiot. 2002, 55, 204-
207.
(10) Patel, K.; Piagentini, M.; Rascher, A.; Tian, Z.-Q.; Buchanan, G.
O.; Regentin, R.; Hu, Z.; Hutchinson, C. R.; McDaniel, R. Chem. Biol.
2004, 11, 1625-1633.
10.1021/ol060022b CCC: $33.50
© 2006 American Chemical Society
Published on Web 02/04/2006

that could enable access to related analogues for elucidation
of structure-activity relationships. As a starting point for
the development of such a library, the trienomycin A
macrocycle was disconnected into three main parts (Figure
1): (1) the A-sublibrary contained the aromatic ring and two
additional appendages for connection to other sublibraries;
(2) the B-sublibrary contained olefinic acids, which could
be coupled directly with the amino group from the A-
sublibrary and undergo ring-closing metathesis with a
terminal olefin from the C-sublibrary; (3) members of the
C-sublibrary contained carbonyls to serve as substrates for
Wittig olefination as well as an olefin for ring-closing
metathesis (RCM).11
reduce the R,â-unsaturated esters without affecting the nitro
substituent were problematic and eventually led to a two-
step process that required reduction of the olefin with
potassium diazodicarboxylate13 to give intermediates 6 and
7, the esters of which were subsequently reduced with
diisobutylaluminum hydride to afford alcohols 8 and 9. The
hydroxyls were converted to the corresponding iodides (10
and 11) and then reacted with polymer-bound triphenyl-
phoshine14 to afford the resin-bound phosphonium salts 12
and 13. Reduction of the nitro group on solid support was
surprisingly difficult as only partial reduction occurred or
entrapment of metal ions resulted. After numerous methods
were explored, it was found that treatment of the nitrated
resin with stannous chloride15 completely dissolved in N,N-
dimethylformamide gave aniline salts 14 and 15, which
represent members of the A-sublibrary.
Scheme 1. Synthesis of Resin-Bound Phosphonium Salts
Figure 1. Geldanamycin and trienomycin A are representative
members of the ansamycin family of antibiotics.
Attempts to couple the requisite acids proved burdensome
on solid support, as incomplete coupling was commonly
observed following product release by Wittig olefination and/
or IR spectroscopy. Consequently, acid chlorides were
coupled with anilines 14 and 15, as shown in Scheme 2.
With this three-component approach toward the trieno-
mycin A core in mind, we began preparation of the
A-sublibrary by the synthesis of two members containing a
deoxygenated and a benzyloxy-protected aromatic ring.
Aldehyde 2 was prepared from 3-hydroxy-5-nitrobenzyl
alcohol (1)12 as shown in eq 1, whereas aldehyde 3 was
readily available from commercial sources.
Scheme 2. Synthesis of Resin-Bound Amides
Homologation of the aldehydes with the anion of trieth-
ylphosphonoacetate provided the Horner-Wadsworth-Em-
mons olefination products, 4 and 5 (Scheme 1). Attempts to
(11) Grubbs, R. H.; Miller, S. J.; Fu, G. C. Acc. Chem. Res. 1995, 28,
446-452.
(12) Smith, A. B.; Barbosa, J.; Wong, W.; Wood, J. L. J. Am. Chem.
Soc. 1996, 118, 8316-8328.
976
Org. Lett., Vol. 8, No. 5, 2006

Scheme 4. Synthesis of Macrocyclic Products
Figure 2. IR spectra for 13, 15, and 16.
Addition of 4-N,N-(dimethylamino)pyridine to the reaction
mixture gave the amide products in good yield as determined
by IR spectroscopy.
Figure 2 shows the IR spectra between 1200 and 1700
cm-1 for the resin-bound intermediates. The nitrated resin
(Figure 2, 13) produced strong absorption at 1526 and 1348
cm-1, whereas the corresponding region for the aniline-
derived resin (Figure 2, 15) was devoid of these absorptions
but produced a broad absorption at ∼1600 cm-1. Likewise,
the amide resin (Figure 2, 16) produced a sharpened 1600
cm-1 absorption and a strong signal at 1678 cm-1, corre-
sponding to the stretching frequency of the amide carbonyl.
With the amide-containing resin in hand, we next pursued
the Wittig reaction with the solid-phase phosphonium salts.
The advantage of this method over the solution-phase
approach is that the undesired phosphine oxide product
produced in this reaction remains attached to the resin and
can be easily removed from hydrophobic products by
filtration.
Scheme 3. Synthesis of RCM Precursors
Four carbonyl substrates were chosen to represent mem-
bers of the C-sublibrary and consisted of 5- or 7-carbons as
well as methyl ketones or aldehydes. After a large number
of bases were surveyed, it was found that treatment of a
mixture of both the polymer-supported phosphonium salt and
the carbonyl substrate with 2.2 equiv of sodium bis-
(trimethylsilyl)amide in THF at 0 °C gave the desired
(13) Beruben, D.; Marek, I.; Normant, J. F.; Platzer, N. J. Org. Chem.
1995, 60, 2488-2501.
(14) Bernard, M.; Ford, W. T. J. Org. Chem. 1983, 48, 326-331.
(15) Bellamy, F. D.; Ou, K. Tetrahedron Lett. 1984, 25, 839-842.
Org. Lett., Vol. 8, No. 5, 2006
977

products. When aldehyde substrates were employed, only
cis isomers were obtained. However, ketone substrates gave
a 1:1 mixture of cis and trans products. In total, eight Wittig
products were prepared as outlined in Scheme 3.
via this approach. In addition, access to smaller metacyclo-
phanic rings through relay metathesis is likely to prove useful
in the pursuit of other macrocyclic natural products.
Although Ford19 and Hughes20 have shown that resin-
bound phosphonium salts provide Wittig products, its ap-
plication to natural products and library development has
not been pursued. In this letter, we have shown that the
polymer-supported triphenylphosphonium resin is capable of
providing natural product-like analogues of an important
class of biologically active natural products. Further elabora-
tion of individual members of each sublibrary will provide
structures for biological investigation and may help to explain
the antitumor activity manifested by trienomycin A.
As expected, most of the terminal olefins underwent ring-
closing metathesis; however, it was found that Grubbs’ first
generation catalyst16 gave the most reproducible RCM results,
whereas the second generation catalyst17 afforded a greater
percentage of products resulting from relay metathesis.18
When these eight molecules were treated with Grubbs’ I,
the products obtained represented a mixture of cis/trans
isomers of the metathesis-derived alkenes as illustrated in
Scheme 4. Both cis and trans isomers obtained from the
Wittig reaction were substrates for the RCM reaction.
However, olefins containing a four-carbon linker devoid of
the trisubstituted olefin underwent relay metathesis18 to give
33 or the ring-contracted macrocylcic product 34 in a trans/
cis ratio of 1:1. In contrast, 1,2-substituted alkenes with less
than four methylenes between olefins remained substrates
for RCM.
Acknowledgment. The authors gratefully acknowledge
support of this project by the NIH Center of Excellence in
Combinatorial Methods and Library Development (5 P50
GM069663).
Supporting Information Available: Experimental pro-
cedures and characterization for representative compounds
in this letter. This material is available free of charge via
the Internet at http://pubs.acs.org.
These RCM results suggest that incorporation of trisub-
stituted olefins is sufficient for reducing the amount of relay
metathesis products and that various ring sizes are accessible
OL060022B
(16) Schwab, P.; Grubbs, R. H.; Ziller, J. W. J. Am. Chem. Soc. 1996,
118, 100-110.
(17) Chatterjee, A. K.; Grubbs, R. H. Org. Lett. 1999, 1, 1751-1754.
(18) Hoye, T. R.; Jeffrey, C. S.; Tennakoon, M. A.; Wang, J.; Zhao, H.
J. Am. Chem. Soc. 2004, 126, 10210-10211.
(19) Bernard, M.; Ford, W. T.; Nelson, E. C. J. Org. Chem. 1983, 48,
3164-3168.
(20) Hughes, I. Tetrahedron Lett. 1996, 37, 7595-7598.
978
Org. Lett., Vol. 8, No. 5, 2006