The Total Synthesis of (-)-Lemonomycin

Article abstract of DOI:10.1021/ja039223q

The first total synthesis of the novel glycosylated tetrahydroisoquinoline antitumor antibiotic (?)-lemonomycin has been accomplished (15 steps from 9). The highly convergent synthesis relies on a key asymmetric dipolar cycloaddition to set the stereochem

Full text of DOI:10.1021/ja039223q

Published on Web 11/14/2003
The Total Synthesis of (-)-Lemonomycin
Eric R. Ashley, Ernest G. Cruz, and Brian M. Stoltz*
DiVision of Chemistry and Chemical Engineering, California Institute of Technology, Pasadena, California 91125
Received October 24, 2003; E-mail: stoltz@caltech.edu
Scheme 1
Lemonomycin (1), a member of the tetrahydroisoquinoline family
of antitumor antibiotics that includes the saframycins, ecteinascidins,
and quinocarcins was originally isolated in 1964 from a fermentation
broth of Streptomyces candidus and was found to have potent
antibiotic activity against Staphylococcus aureus and Bacillus
subtilis.1 The structure of lemonomycin, however, was not eluci-
dated until 2000,2 when researchers at Wyeth-Ayerst discovered
its activity against a variety of antibiotic-resistant bacterial strains.
In addition to the connectivity and relative stereochemistry of
lemonomycin, the authors reported antibiotic activity against
methicillin-resistant S. aureus and vancomycin-resistant Entero-
coccus faecium, as well as cytotoxicity against a human colon tumor
cell line (i.e., HCT 116). Lemonomycin is unique among the nearly
60 natural products and hundreds of synthetic analogues in this
family in that it bears a glycoside at C(18).3 This is especially
striking considering the vast numbers of analogues that have been
prepared by derivatization at that position, particularly in the
saframycin, quinocarcin, and ecteinascidin families. Moreover, the
2,6-dideoxy-4-amino sugar present in lemonomycin has never been
prepared and is rare in nature.4 The combination of potent biological
activity and the challenging novel structure makes lemonomycin
Scheme 2
an attractive molecule for target-directed synthesis. Herein, we
describe the first total synthesis of (-)-lemonomycin by use of a
stereoselective dipolar cycloaddition and a novel, diastereoselective
Pictet-Spengler cyclization.
Our retrosynthetic plan for the preparation of lemonomycin is
outlined in Scheme 1 and involves the rapid dissection of the
Scheme 3
alkaloid into three subunits: aryl boronic ester 4, diaza-bicyclo-
[3.2.1]octane 5, and aminoglycosyloxy aldehyde 3. To maximize
convergency in our synthetic design, we pursued a strategy that
would directly incorporate the aminopyranose as part of a Pictet-
Spengler reaction to establish the tetrahydroisoquinoline core of
the molecule (e.g., 2 + 3), rather than append the aglycone via the
glycosidic linkage. The diazabicycle was expected to arise via a
These undesired adducts could be minimized to trace amounts by
Joule dipolar cycloaddition (e.g., 7 + 8) with the absolute stereo-
performing the reaction at -20 °C (Scheme 2).8
chemistry deriving from an auxiliary controller (i.e., Xc in 8).5
With a suitable variant of the Joule reaction in hand, we turned
We initiated our efforts toward the synthesis of lemonomycin
our attention to the development of an asymmetric route to the
by investigating the dipolar cycloadditions of an oxidopyrazinium
diazabicycle. Implementation of the Oppolzer sultam-derived
salt (e.g., 7) with various dipolarophiles. Although these cycload-
ditions could be viewed as a powerful method to access the desired
diazabicylic framework of lemonomycin, such reactions (first
reported by Joule) have suffered from modest yields and regio-
selectivity. Moreover, dipole 7 is typically generated and purified
using ion exchange chromatography. We sought improvements to
produce the quantities of material necessary for advancement
through the synthesis and for the eventual development of asym-
metric variants.6,7 To our delight, in situ generation of 7 by
deprotonation of bromide salt 9 with Et3N followed by treatment
with acrolein, a dipolarophile not studied by Joule, resulted in a
facile cycloaddition to provide bicycle 10 along with significant
quantities of inseparable diastereo- and regioisomeric byproducts.
acrylamide 117 under slightly modified cycloaddition conditions,
employing N-methyl morpholine as base and CH3CN as solvent,
produced a cycloadduct, which upon treatment with NaBH4 in EtOH
provided alcohol 12 in 72% yield and 94% ee (Scheme 3).9
Elaboration of enamide 12 by protection of the free hydroxyl
functionality as a silyl ether and iodination produced Z-iodoenamide
5 as a single isomer (Scheme 4). Suzuki fragment coupling of iodide
5 with boronic ester 49b afforded aryl enamide 13. Catalytic
hydrogenation of the hindered trisubstituted olefin was sluggish
but proceeded at 1000 psi with concomitant cleavage of the
benzylamine unit to produce 14 as a single diastereomer.10 To
convert 14 into a suitable substrate for the key Pictet-Spengler
cyclization, the 2° amine was converted to a urethane (CBZCl,
9
15000
J. AM. CHEM. SOC. 2003, 125, 15000-15001
10.1021/ja039223q CCC: $25.00 © 2003 American Chemical Society

C O M M U N I C A T I O N S
Scheme 4
Scheme 6
In summary, we have developed the first total synthesis of the
novel glycosylated tetrahydroisoquinoline antitumor antibiotic (-)-
lemonomycin (15 steps from 9). The highly convergent synthesis
features an asymmetric dipolar cycloaddition that sets the stereo-
chemistry of the aglycone core, a Suzuki coupling to connect the
diazabicycle to the aryl subunit, and a stereoselective Pictet-
Spengler reaction that incorporates the aminoglycoside directly
without the need for late-stage glycosylation or protecting group
manipulations. The novel aminopyranose was prepared using a
highly diastereoselective Felkin-controlled acetate aldol reaction
to a threonine-derived ketone. The utilization of this strategy for
the synthesis of nonnatural glycosylated tetrahydroisoquinoline
derivatives for biological evaluation is currently under investigation.
Scheme 5
Acknowledgment. This work is dedicated to Professor E. J.
Corey on the occasion of his 75th birthday. The authors are grateful
to Caltech, the University of California TRDRP (predoctoral
fellowship to E.R.A.), and the J. Irvine Foundation (predoctoral
fellowship to E.G.C.) for financial support. We also thank T. Y.
Lam for experimental assistance and Dr. H. He (Wyeth-Ayerst)
for an authentic sample of (-)-1.
DMAP), and the phenol was liberated from its sulfonate ester
(KOTMS) to provide 2. Unfortunately, this amide was highly
unreactive, and all attempts to couple 2 to aldehydes via Pictet-
Spengler cyclization failed. Thus, we turned to the more reactive
aminotriol 15, which was accessed by activation of the amide with
BOC2O (with concomitant protection of the phenol), reduction with
NaBH4,11 and cleavage of the two BOC moieties and TIPS ether
by methanolic HCl. This compound (employed as a TFA salt)
proved much more active in Pictet-Spengler cyclizations, and in
model studies with a variety of R-hydroxy acetaldehyde derivatives
it typically produced diastereomerically pure tetrahydroisoquinoline
products.
Supporting Information Available: Experimental details (PDF)
and crystallographic details (CIF). This material is available free of
charge via the Internet at http://pubs.acs.org.
References
(1) Whaley, H. A.; Patterson, E. L.; Dann, M.; Shay, A. J.; Porter, J. N.
Antimicrob. Agents Chemother. 1964, 8, 83-86.
(2) He, H.; Shen, B.; Carter, G. T. Tetrahedron Lett. 2000, 41, 2067-2071.
(3) For a comprehensive review of the chemistry and biology related to the
tetrahydroisoquinoline alkaloids, see: Scott, J. D.; Williams, R. M. Chem.
ReV. 2002, 102, 1669-1730.
The synthesis of the aminopyranose began following conversion
of D-threonine to ketone 16.12,13 Felkin-controlled addition of the
lithium enolate of ethyl acetate to ketone 16 produced aldol adduct
17 as a single observable diastereomer (Scheme 5).10,14 Hydroxy
ester 17 was further processed by acid-mediated acetonide cleavage
with subsequent lactonization, conversion to an oxazolidine, dia-
stereoselective reduction to the corresponding lactol and incorpora-
tion of allyl alcohol to produce bicycle 18. Reductive cleavage of
the oxazolidine with concomitant removal of the benzenesulfonyl
group provided the 2° amine 19, which was methylated by reductive
amination to afford allyl glycoside 20. Oxidative cleavage of the
allyl group by catalytic dihydroxylation of 20 in the presence of
NaIO4 furnished R-glycosyloxy acetaldehyde derivative 3.
The completion of the total synthesis now relied on the success
of the unprecedented Pictet-Spengler cyclization of the aminogly-
cosyloxy aldehyde 3 and the TFA salt of aminotriol 15 (Scheme
6). In the event, simply mixing the two compounds in EtOH at
room temperature produced the desired adduct 21 as a single
diastereomer at C(1) in 95% yield.15 Elaboration of tetrahydroiso-
quinoline 21 to the natural product was straightforward and involved
hydrogenolytic cleavage of the CBZ group, bis Swern oxidation,
and treatment with CAN to provide (-)-lemonomycin (1). The fully
synthetic material obtained by this sequence proved identical in
all respects to a sample obtained from natural sources.
(4) The aminopyranose of 1 has been found in only two unrelated natural
product families. For a list of references, see the Supporting Information.
(5) (a) Kiss, M.; Russell-Maynard, J.; Joule, J. A. Tetrahedron Lett. 1987,
28, 2187-2190. (b) Allway, P. A.; Sutherland, J. K.; Joule, J. A.
Tetrahedron Lett. 1990, 31, 4781-4782. (c) Yates, N. D.; Peters, D. A.;
Allway, P. A.; Beddeoes, R. L.; Scopes, D. I. C.; Joule, J. A. Heterocycles
1995, 40, 331-347.
(6) Scott and Williams have developed an oxidative dipolar cycloaddition
for the synthesis of related natural products. See: Scott, J. D.; Williams,
R. M. J. Am. Chem. Soc. 2002, 124, 2951-2956 and references therein.
(7) Garner et al. have utilized a related dipolar cycloaddition that resulted in
the elegant enantioselective total synthesis of quinocarcin. See: Garner,
P.; Ho, W. B.; Shin, H. J. Am. Chem. Soc. 1993, 115, 10742-10753.
(8) To facilitate isolation of the desired isomer, we have adopted an in situ
procedure that involves reduction of the aldehyde to a 1° alcohol and
protection as a TIPS ether. The overall yield for this three-step transforma-
tion is 72% and has been carried out on a 10 g scale.
(9) (a) Unless otherwise noted, all reactions were performed at ambient
temperature (20 °C) for 1 h or less. (b) See Supporting Information for
full experimental details.
(10) The relative configuration has been established by X-ray crystallographic
analysis of a derivative; see the Supporting Information.
(11) Fukuyama, T.; Yang, L.; Ajeck, K. L.; Sachleben, R. A. J. Am. Chem.
Soc. 1990, 112, 3712-3713.
(12) Maurer, P. J.; Knudsen, C. G.; Palkowitz, A. D.; Rapoport, H. J. Org.
Chem. 1985, 50, 325-332.
(13) Evans, D. A.; Hu, E.; Tedrow, J. S. Org. Lett. 2001, 3, 3133-3136.
(14) (a) Reetz, M. T.; Schmitz, A. Tetrahedron Lett. 1999, 40, 2737-2740.
(b) Reetz, M. T. Chem. ReV. 1999, 99, 1121-1162.
(15) Although we have not fully characterized any other products of this PS
cyclization, purification by HPLC must remove any trace diastereomers
(<3%) resulting from the initial cycloaddition (94% ee).
JA039223Q
9
J. AM. CHEM. SOC. VOL. 125, NO. 49, 2003 15001