Studies Directed toward the Total Synthesis of Lactonamycin: Control of the Sense of Cycloaddition of a Quinone through Directed Intramolecular Catalysis

Article abstract of DOI:10.1002/anie.200352591

The rapid assembly of the ABCD tetracyclic ring system of lactonamycin involved the strong-base-induced regioselective Diels-Alder reaction of 1 and 2 to form advanced precursor 3. Regiochemical control results from the suitably positioned hydroxy group,

Full text of DOI:10.1002/anie.200352591

Angewandte
Chemie
Natural Product Synthesis
Studies Directed toward the Total Synthesis of
Lactonamycin: Control of the Sense of
Cycloaddition of a Quinone through Directed
Intramolecular Catalysis**
Christopher D. Cox, Tony Siu, and
Samuel J. Danishefsky*
As part of a screening program for the discovery of new
antibiotics, Matsumoto et al. isolated lactonamycin (1) from a
culture broth of Streptomyces rishirienisi MJ773–88K4.[1] Its
[*] Prof. S. J. Danishefsky, Dr. C. D. Cox, T. Siu
Department of Chemistry, Columbia University
Havemeyer Hall, New York, NY 10021 (USA)
E-mail: s-danishefsky@ski.mskcc.org
Prof. S. J. Danishefsky
Laboratory for Bioorganic Chemistry
Sloan-Kettering Institute for Cancer Research
1275 York Ave., New York, N.Y. 10021 (USA)
Fax: (+1)212-772-8691
[**] This work was supported by the National Institutes of Health (Grant
number: HL25848). A Postdoctoral Fellowship is gratefully
acknowledged by C.D.C. (NIH, Grant Number F32-CA84758). The
authors thank Yashuiro Itagaki for high-resolution mass spectral
analyses.
Angew. Chem. Int. Ed. 2003, 42, 5625 –5629
DOI: 10.1002/anie.200352591
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5625

Communications
term cycloaddition in an empirical rather than mechanistic
sense, thereby sidestepping difficultly resolvable issues
related to the concerted nature of the reaction. Cycloadduct
5 would be expected to undergo transformation into 6 upon
the loss of carbon dioxide. Although this type of overall
reaction is, in itself, well-known as a Tamura–Diels–Alder
reaction,[4] in this case it was to be applied to the unsym-
metrical quinone 4. Even if unlikely initial cycloaddition at
the more hindered of the quinone double bonds is neglected,
in principle, two regioisomeric products 6 and 8, arising from
primary cycloadducts 5 and 7, respectively, could be pro-
duced.
In Scheme 1, a general format for solving this type of
problem is suggested by building an internal activating group
into the quinone 4. In the case at hand, we imply that the
formal ketone group (a) can be rendered more active than its
counterpart (b) by means of the biasing element X.[5a–c]
Whether through hydrogen bonding or metal-ion bridging
(see below), the consequences of this intramolecular directed
catalysis would be to favor the eventual formation of 6 over 8.
Indeed, in our recently described total synthesis of
rishirilide B, we demonstrated this type of locally biased
catalysis (Scheme 2, 9 + 10!11).[6] In the total synthesis of
structure and absolute configuration were assigned through
spectroscopic and X-ray crystallographic measurements in
conjunction with degradation protocols. Biological evaluation
of lactonamycin revealed it to have antimicrobial activity
against Gram-positive bacteria, including methicillin- and
vancomycin-resistant strains (IC50 = 0.20–1.56 mgmLꢀ1). In
addition to its antimicrobial activities, lactonamycin is cyto-
toxic against various tumor cell lines (IC50 = 0.06–
3.3 mgmLꢀ1).
The novel, highly functionalized hexacyclic aglycone
domain of lactonamycin, lactonamycinone (2), inherently
poses some significant issues at the planning level of a
chemical synthesis. Furthermore, the feasibility of glycosyla-
tion of a tertiary a-hydroxyketone acceptor site with an
appropriate l-rhodinose donor species under high anomeric
stereoselectivity could hardly be taken for granted. These
concerns notwithstanding, the challenges confronting any
attempt to construct lactonamycin in a laboratory setting and
its promising profile of antibiotic action drew us into a total
synthesis venture. We report herein an important milestone in
that effort, that is, the first total synthesis of the aglycone,
lactonamycinone (2).[2] We expand upon the concept of
intramolecular directed catalysis, which enabled the highly
concise assembly of the key tetracyclic intermediate 35 (see
Scheme 7), and we discuss how we overcame the many
difficulties associated with the synthesis of this intermediate
and its conversion into lactonamycinone.[3]
Scheme 2. Application of locally biased catalysis in the total synthesis
of rishirilide B.
lactonamycin, we hoped to extend the local biasing strategy to
quinones rather than to a cyclohexene-1,4-dione 10. Since
quinones are particularly reactive substrates, the applicability
of the findings in the case of rishirilide was open to question.
Moreover, we would be seeking to extend the generalized
idea of regiocontrol by intramolecular catalysis from a clear
cycloaddition case (9 + 10) conducted under neutral condi-
tions to a less-well-defined, anion-mediated setting (3 + 4).
Our strategy hinged on the condensation of a homo-
phthalic anhydride (see generalized structure 3) with a
quinone 4 in a process that is initiated by an anionically
mediated “cycloaddition” reaction (Scheme 1). We use the
Scheme 1. Tamura–Diels–Alder reaction with an unsymmetrical quinone.
5626
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 5625 –5629

Angewandte
Chemie
Viewed from the perspective of lactonamycinone, we
envisioned the cycloaddition of 15 and 16 (Scheme 3). The
orienting function would be the strategically placed hydroxy
group at C3’, which would differentiate the two ketone-like
albeit in 15–25% yield. Hydrogenolysis of the two benzyl
esters gave rise to a diacid. However, dehydration with
trimethylsilylethoxyacetylene[11] only produced trace amounts
of the desired homophthalic anhydride 24, presumably as a
result of solubility issues. It is possible
that the difficulties in the cycloaddi-
tion reaction of 21 and 22 reflect the
usual complexities associated with
Diels–Alder reactions of 1,1-disubsti-
tuted dienes, in which the required 2,3-
s-cis coplanar conformer may be of
high energy. In the case at hand, the
situation is perhaps further aggravated
by the presence of substituents at C2
and C3 in 21. Given the difficulties of
the Diels–Alder reaction and anhy-
dride formation, we also explored an
alternative route, which, though some-
what less concise, lends itself to larger-
scale development.
The second path started with the
known bis(silyl enol ether) 25.[12]
Cycloaddition of 25 with the 1,3-
Scheme 3. Proposed synthesis of lactonamycinone
dicarbomethoxyallene (26)[10] gave
rise to 27 in 75% yield (Scheme 5).
centers of quinone 16. We hoped to reach the homophthalic
anhydride 15 by a Diels–Alder reaction of 12 and 13. In this
way, we would ultimately have connected the diverse
functionalities of the ABCD ensemble through two defining
cycloaddition reactions (see product 17). Some of the
necessary connections to proceed from 17 to lactonamycinone
(2) are suggested in Scheme 3.
Implementation of the ideas outlined above commenced
with tetramic acid derivative 18[7] (Scheme 4). Iodination[8] of
this compound gave rise to 19, which served as a substrate for
a Stille reaction with 20,[9] thereby giving rise to diene 21 in
42% yield (two steps). For our allene component, we took
recourse to the 1,3-dicarbobenzyloxyallene (22).[10] In the
event, cycloaddition of 21 and 22 was possible and gave 23,
To overcome issues of solubility later in the synthesis, the
phenolic function of 27 was protected as an unconventional
octyloxymethyl ether[13] derivative 28. The aryl methyl group
of 28 was brominated with N-bromosuccinimide and benzoyl
peroxide under irradiation from a UV lamp. Revisiting a type
of chemistry that we had studied in 1969, we treated the
resulting compound with methylamine, which gave rise
primarily to lactam 29[14] (30% yield based on recovered
starting material) along with the isomeric isoindolinone
(3.5:1). Hydrolysis of the ester linkages gave rise to a
diacid, which was dehydrated to produce the protected
pyrrolo homophthalic anhydride 30 (Scheme 5).
The building of a suitable quinone system started with the
known 31[15] (Scheme 6). Protection of the primary alcohol as
a benzyl ether followed by deprotection of
the dithiane[16] afforded 32 in 74% yield.
The smooth chain extension of the ben-
zaldehyde derivative by the Rathke meth-
odology[17] gave rise to 33 in excellent
yield. Oxidative demethylation of this
substrate afforded quinone 34 in 96%
yield. It was this compound that would
serve as the coupling partner with the
previously described homophthalic anhy-
dride 30.
With the appropriately functionalized
quinone 34 and homophthalic anhydride
30 in hand, the crucial Tamura–Diels–
Alder reaction was examined. In the
event, treatment of 30 with 2 equivalents
of NaH at ꢀ 788C, followed by the
addition of 2 equivalents of quinone 34
Scheme 4. Reagents and conditions. a) I2, PIFA, pyridine, 51%; b) 20, [PdCl2(PPh3)2], PhMe,
reflux, 83%; c) 22, neat, 2008C, 15–25%; d) H2, Pd/C, 50%; e) trimethylsilylethoxyacetylene,
CH3CN/CH2Cl2, trace amounts of product. PIFA=bis(trifluoroacetoxy)iodobenzene.
and warming to 08C, led to a rapid cyclo-
addition reaction followed by extrusion of
Angew. Chem. Int. Ed. 2003, 42, 5625 –5629
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5627

Communications
Scheme 5. Reagents and conditions. a) 25, 26, neat, 1058C; b) NH4F, MeOH, 75%; c) CH3(CH2)7OCH2Cl, DIPEA, DMF, 95%; d) NBS, benzoyl
peroxide (cat.), UV, benzene; e) K2CO3, NH2Me, MeOH/CH3CN (1:2), 30% yield based on recovered starting material, 3.5:1; f) KOH, MeOH,
98%; g) trimethysilylethoxyacetylene, CH3CN/CH2Cl2, quantitative. DIPEA=N,N-diisopropylethylamine; DMF=N,N-dimethylformamide.
carbon dioxide to provide quinone 35 in
40% yield (Scheme 7). As we had
hoped, only a single regioisomer was
formed, as determined by NMR spec-
troscopic analysis at 500 MHz. The
structure of the product was first infer-
red from precedent and later confirmed
by X-ray crystallography to correspond
to that depicted in 35. Notably, 2 equiv-
alents of quinone were needed for
optimum yield—presumably the extra
equivalent is required to oxidize the
resulting hexacyclic intermediate to
quinone 35. Efforts to use 1 equivalent
of quinone and an external source of
oxidant gave lower yields.
Scheme 6. Reagents and conditions. a) NaH, BnBr, TBAI, THF, 86%; b) PIFA, CH3CN/H2O (1:1),
86%; c) LDA, tBuOAc, ꢀ788C, THF, 99%; d) CAN, 08C, CH3CN/H2O (9:1), 96%. TBAI=tetra-
butylammonium iodide; LDA=lithium diisopropylamide; CAN=ceric ammonium nitrate.
Although we had indeed predicted
the preferred formation of 35 under
mediation by
a strategically placed
activating group, it was appropriate to
test this rationale in a closely related
setting, thereby enabling further under-
standing of the issues involved. Accord-
ingly, the hydroxy group of 34 was
protected as its tert-butyldimethylsilyl
ether 36. The latter compound reacted
smoothly under the same Tamura–
Scheme 7. Regioselective Tamura–Diels–Alder reaction.
Diels–Alder
reaction
conditions
employed for 35 (Scheme 8). However,
in this case a 1:1 mixture of 37 and 38
was produced. Certainly, this experi-
ment is fully in keeping with the gov-
erning paradigm sketched out in
Scheme 1 (see structure 4).
In presenting this model for directed
activation we were appropriately vague
as to the precise nature of the activating
Scheme 8. Disruption of the locally biased catalysis.
5628
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
Angew. Chem. Int. Ed. 2003, 42, 5625 –5629

Angewandte
Chemie
[6] J. G. Allen, S. J. Danishefsky, J. Am. Chem. Soc. 2001, 123, 351 –
effect. One possibility would be that of a local intramolecular
hydrogen bond. Alternatively, carbonyl a (34a) could be
differentially activated by a chelation effect (sodium in the
case of our application of the Tamura reaction). Clearly, a
metal chelate effect might, in principle, have been possible
even if the oxygen atom of the side chain were protected as
the TBS ether. Since TBS ethers projecting from carbons a or
b to a carbonyl function tend to be poor donors to support
metal-mediated chelates,[18] the 1:1 ratio of products resulting
from the reaction of 30 and 36 is not surprising.
Finally, we probed the possibility of exerting control
through an alkoxide-based metal bridge. Toward this end,
compound 34 was pretreated with NaH. Treatment of 34b,
thus produced, with 30 under the same conditions used above
again led to 35 as the only isolated regioisomer (Scheme 7).
Hence, metal-induced positionally directed selective activa-
tion is clearly possible with a hydroxy-based anchor (see 34b).
It remains to be shown whether a hydrogen-bonded counter-
part can intervene under these basic conditions.
352.
[7] H. W. Pinnick, K. S. Kochhar, J. Org. Chem. 1989, 54, 3222 –
3224.
[8] R. Benhida, P. Blanchard, T. L. Fourrey, Tetrahedron Lett. 1998,
39, 6849 – 6852.
[9] J. A. Soderquist, J.-H. Hsu, Organometallics 1982, 1, 830 – 833.
[10] M. Node, T. Fujiwara, S. Ichihashi, K. Nishide, Tetrahedron 1986,
42, 6645 – 6656.
[11] P. Mꢀller, N. Pautex, Helv. Chim. Acta 1991, 74, 55 – 64.
[12] P. Allevi, M. Anastasia, S. Bingham, P. Ciuffreda, A. Fiecchi, G.
Cighetti, A. Scala, J. Tyman, J. Chem. Soc. Perkin Trans. 1 1998,
575 – 582.
[13] A. Warshawsky, A. Deshe, J. Polym. Sci. Part. A 1985, 23, 1839.
[14] S. J. Danishefsky, T. A. Bryson, J. Puthenpurayil, J. Org. Chem.
1975, 40, 1846 – 1848.
[15] G. A. Kraus, L. Chen, R. A. Jacobson, Synth. Commun. 1993, 23,
2041 – 2049.
[16] G. Stork, K. Zhao, Tetrahedron Lett. 1989, 30, 287 – 290.
[17] M. J. Rathke, D. F. Sullivan, J. Am. Chem. Soc. 1973, 95, 3050 –
3051.
[18] G. E. Keck, S. Castellino, Tetrahedron Lett. 1987, 28, 281 – 284.
In summary, we have shown how the otherwise compli-
cated ABCD domain of lactonamycin can be assembled with
high regiocontrol through two cycloaddition reactions from
readily synthesized components. This enabling strategy was
crafted around the notion of gaining regiochemical control
through mediation of a suitably positioned hydroxy group.
The concept was reduced to practice but does not extend to
the corresponding TBS ether of the hydroxy group. In the
following Communication in this issue, we describe the
completion of the total synthesis of lactonamycinone utilizing
the key ideas and intermediates illustrated herein.
Received: August 7, 2003 [Z52591]
Keywords: antibiotics · Diels–Alder reaction · natural products ·
.
quinones · total synthesis
[1] a) N. Matsumoto, T. Tsuchida, M. Maruyama, R. Sawa, N.
Kinoshita, Y. Homma, Y. Takahashi, H. Iinuma, H. Naganawa, J.
Antibiot. 1996, 49, 953 – 954; b) N. Matsumoto, T. Tsuchida, M.
Maruyama, N. Kinoshita, Y. Homma, H. Iinuma, T. Sawa, M.; T.
Hamada, T. Takeuchi, N. Heida, T. Yoshioka, J. Antibiot. 1999,
52, 269 – 275; c) N. Matsumoto, T. Tsuchida, H. Nakamura, R.
Sawa, Y. Takahashi, H. Naganawa, H. Iinuma, T. Sawa, T.
Takeuchi, M. Shiro, J. Antibiot. 1999, 52, 276 – 280.
[2] For synthetic and model studies toward the synthesis of
lactonamycin, see: a) C. Cox, S. J. Danishefsky, Org. Lett. 2000,
2, 3493 – 3496; b) C. Cox, S. J. Danishefsky, Org. Lett. 2001, 3,
2899 – 2902; c) J. P. Deville, V. Behar, Org. Lett. 2002, 4, 1403 –
1405d) T. R. Kelly, D. Xu, G. Martinez, H. Wang, Org. Lett. 2002,
4, 1527 – 1529.
[3] T. Siu, C. D. Cox, S. J. Danishefsky, Angew. Chem. 2003, 115,
5787 – 5792; Angew. Chem. Int. Ed. 2003, 42, 5629 – 5634.
[4] Y. Tamura, F. Fukata, M. Sasho, T. Tsugoshi, Y. Kita, J. Org.
Chem. 1985, 50, 2273 – 2277, and references therein.
[5] For a related use of this concept in a Diels–Alder reaction, see:
a) H. Fujioka, H. Yamamoto, H. Annoura, H. Maeda, Y. Kita,
Chem. Pharm. Bull. 1992, 40, 32 – 35; for a related use of this
concept in juglone and naphthazarin systems, see: b) V. H.
Powell, A. J. Birch, Tetrahedron Lett. 1970, 11, 3467 – 3469;
c) T. R. Kelly, J. W. Gillard, R. N. Goerner, Jr. J. M. Lyding, J.
Am. Chem. Soc. 1977, 99, 5513 – 5514.
Angew. Chem. Int. Ed. 2003, 42, 5625 –5629
www.angewandte.org
ꢀ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5629