Macrocyclic triarylethylenes via heck endocyclization: A system relevant to diazonamide synthesis

Full text of DOI:10.1021/jo981791e

8640
J . Org. Chem. 1998, 63, 8640-8641
Ma cr ocyclic Tr ia r yleth ylen es via Heck
En d ocycliza tion : A System Releva n t to
Dia zon a m id e Syn th esis
Susan J eong, Xin Chen, and Patrick G. Harran*
Department of Biochemistry, The University of Texas
Southwestern Medical Center at Dallas,
Dallas, Texas 75235-9038
Received September 3, 1998
Diazonamide A (1, Figure 1) and its congeners1 are
impressive products of invertebrate secondary metabolism
wherein at least three proteinogenic amino acids have been
incorporated, perhaps by posttranslational modification of
a short oligopeptide, into a heterocyclic network of consider-
able complexity. Notably, 1 has shown potent cytotoxicity
in vitro toward a transformed human cell line (HCT-116),1a
but a subsequent lack of natural material1c has prevented
further study. To enable future mode-of-action studies and
to access a range of related polycyclic structures, we outline
here our approach to diazonamide synthesis and describe
initial results demonstrating that Heck endocyclization can
form highly functionalized lactam rings containing an
imbedded 1,2-diaryl-1-(5′-oxazoyl)ethylene.
When considering possible routes to assemble the diaz-
onamide skeleton,2 it was evident that potentially complicat-
ing issues of atropisomerism could be avoided if the D-E
biaryl linkage was designated as the final carbon-carbon
bond to be formed (Figure 1). In the presence of a confor-
mationally rigid, correctly configured A-F macrolactam,
intramolecular biaryl formation at C16 in 2 necessitates that
the relative rotational orientation of rings A-E adopt a
desired atropisomeric relationship (as depicted in 2) to bring
C18 within bonding distance of C16. We currently envision
that halogenation at C25 and C27 can occur late in the
synthesis and that the D-E linkage can be established by
intramolecular orthophenolic coupling. This implicates
either serotonin or a derivative of 5-HT as the original source
of carbons 18-27. Lactam 3 therefore contains all the
stereochemical information needed to complete natural
diazonamides. A protected derivative of this molecule, with
the C11 hemiacetal in open tautomeric form, is seen as the
ring-contracting pinacol rearrangement product of vicinal
diol 4 or its functional equivalent. To the extent that C10
stereochemistry in 4 correlates to that sought in 3 through
inversion of configuration at the migrating terminus, mac-
rocyclic olefin 5 is a reasonable link to 3 through a sequence
beginning with face-selective oxidation.
F igu r e 1.
were related to natural amino acids. The A ring (C29-C33)
was synthesized in one step from commercial materials by
treating N-Boc-L-Val-OH (6) with aminomalononitrile p-
toluenesulfonate and EDC4 in pyridine (Scheme 1). Follow-
ing trituration of a concentrated residue with water, ami-
nooxazole 7 was isolated by filtration (>95% purity) and
converted directly to bromide 8 through in situ bromination
of a derived nitrosamine.5 This succinct preparation of a
fully functionalized A-ring, combined with the availability
of several halogenated tyrosine derivatives, led us to consider
a third building block (E-ring) which would permit consecu-
tive formation of the C8-C11 and C10-C30 bonds in 5 via
metal-catalyzed cross-couplings. A trifluoroacetate salt de-
rived from 8 was condensed independently (TBTU, DIPEA,
CH3CN) with 3,5-dibromo-N-Boc-L-Tyr-OH (9) and 3-iodo-
N-Boc-L-Tyr-OH (10) to afford dipeptides 11a and 11b,
respectively (Figure 2). Interestingly, when 11b was treated
with bis-stannyl styrene 13c in the presence of 5 mol %
PdCl2(CH3CN)2, a single adduct assigned as R-styryl stan-
nane 12b was isolated in 30% yield. Intractables and
starting material accounted for the remaining mass. Re-
subjecting 12b to the reaction conditions did not induce a
second, ring-closing Stille coupling. The same net results
were observed in reactions between tribromide 11a and 13c.
While synthetically unproductive, it was during these
experiments that we observed a marked susceptibility to
Functionalized lactams related to 5 were without prece-
dent prior to this study,3 although it was fairly obvious how
two (A and F) of the three-component rings in this structure
(1) (a) Lindquist, N.; Fenical, W.; Van Duyne, G. D.; Clardy, J . J . Am.
Chem. Soc. 1991, 113, 2303-2304. (b) Lindquist, N. L. Ph.D. Thesis,
University of California, San Diego, CA, 1989. (c) Petit, C. San Francisco
Chronicle, Friday J an 31, 1997, A4.
(2) For other approaches to diazonamide synthesis, see: (a) Wipf, P.;
Yokokawa, F. Tetrahedron Lett. 1998, 39, 2223-2226. (b) J amison, T. F.
Ph.D. Thesis, Harvard University, Cambridge, MA, 1997. (c) Moody, C. J .;
Doyle, K. J .; Elliott, M. C.; Mowlem, T. J . J . Chem. Soc., Perkin Trans. 1
1997, 16, 2413-2419. (d) Vedejs, E.; Wang, J . B. Abstracts of Papers, 212th
National Meeting of the American Chemical Society, Orlando, FL; American
Chemical Society: Washington, DC, 1996; ORGN 93. (e) Konopelski, J . P.;
Hottenroth, J . M.; Oltra, H. M.; Ve´liz, E. A.; Yang, Z. C. Synlett 1996, 609-
611. (f) Moody, C. J .; Doyle, K. J .; Elliott, M. C.; Mowlem, T. J . Pure Appl.
Chem. 1994, 66, 2107-2110.
(3) Shortly after we began our studies J amison and Schreiber reported
the preparation of cyclic 3-aryl-2-(5′-oxazoyl)benzofurans related to lactams
5. In this approach 3-iodo-N-Cbz-L-Tyr-OMe was acylated with 3-iodo-7-
methoxybenzofuran-2-carboxylic acid and the C8-C11 bond was formed by
internal coupling of the incipient diiodide with excess Ni(PPh3)4. The product
coumarin was reduced to provide a C30 alcohol from which the A-ring
oxazole was assembled. Lactamization subsequently closed the 13-
membered ring. See ref 2b.
(4) This modified version of Freeman’s oxazole synthesis simplifies
product isolation and eliminates the need for a purification step. Confer:
Freeman, F.; Chen, T.; van der Linden, J . B. Synthesis 1997, 861-862.
(5) Doyle, M. P.; Siegfried, B.; Dellaria, J . F., J r. J . Org. Chem. 1977,
42, 2426-2430.
(6) (a) Ashimori, A.; Bachand, B.; Calter, M. A.; Govek, S. P.; Overman,
L. E.; Poon, D. J . J . Am. Chem. Soc. 1998, 120, 6488-6499. (b) Gibson, S.
E.; Middleton, R. J . Contemp. Org. Synth. 1996, 3, 447-471.
(7) Energy profiles for a simulated torsion about the C10-C30 bond in
low-energy conformers of 16d and 16f were generated within the dihedral
driver subroutine of Macromodel V6.0.
10.1021/jo981791e CCC: $15.00 © 1998 American Chemical Society
Published on Web 11/05/1998

Communications
J . Org. Chem., Vol. 63, No. 24, 1998 8641
Sch em e 1a
selective protiodestannylation at the distal carbon-tin bond
(a) in bis-stannane 13c. Treatment of 13c with silica gel
impregnated with oxalic acid provided mono-stannane 14
(Scheme 2) as an air stable, amorphous solid in 87% yield.
This result allowed us to exploit the initially troublesome
preference for Stille coupling at the oxazoyl bromide in
polyhalides 11a /b (an undesired regiochemical result when
metathesizing at bond a with 13c). A mixture of 11a or 11b
and 14 was stirred in the presence of 2 mol % PdCl2(CH3-
CN)2 (DMF, rt), and the resultant mono adducts were
etherified (DEAD, PPh3, BnOH) to afford seco F-A-E systems
16f and 16d in 65% and 57% overall yields, respectively. To
access ring system 5, we attempted to initiate intramolecular
carbometalation of the vinylogous acrylonitrile in 16d /16f
using a variety of Heck cyclization protocols.6 These experi-
ments uniformly met with failure, leading to recovered
starting material, slow degradation or, in the case of iodide
16d with phosphine additives, polar residue presumed to
be phosphonium salt. It was unclear whether this was due
to the substrate poisoning the catalyst or adopting prefer-
entially a conformation which precluded coordination of the
tethered C10-C11 olefin to an intermediate palladium(II)
halide putatively formed at C8. Another explanation was
that the region around C10 was simply too hindered.
Modeling suggested that the E-ring in 16d /16f would rotate
out of a plane defined by the vinyl oxazole to minimize
destabilizing nonbonded interactions.7 As a result, the
E-ring alkoxy group shields one face of the terminal olefin.
To vary functionality at this position while continuing to
differentiate the E- and F-ring phenols, the sequence of
component assembly was reorganized.
F igu r e 2.
Sch em e 2a
Cross-coupling of 14 with bromide 8 and treatment of the
resulting adduct with excess BBr3 gave amino phenol 15 in
high yield (Scheme 2). Peptide coupling between 15 and the
O-benzylated derivatives (9b/10b) of tyrosine halides 9/108
then afforded free phenols 16e and 16b, respectively. The
same sequence, employing vinyl stannanes 13a /13b, was
used to prepare unsubstituted analog 16a and anisole 16c.
Heck cyclization attempts resumed, and, after much experi-
mentation, exposure of 16c to 10 mol % Pd2(dba)3 and Ag3-
PO4 (0.025 M in THF at 75 °C) was found to give an
endocyclization product in only 2% yield. Likewise, 16a gave
6% of a ring-closed material under identical conditions.9 It
was thus intriguing to find that free phenol 16b gave
macrocycle 17a in 66% yield using this same procedure.
Moreover, dibromide 16e underwent a group selective Heck
endocyclization to afford mono bromide 17b,10 albeit in lower
yield. Taken together, these results de-emphasize the role
of sterics at the E-ring ortho position and suggest instead
that the system may need to be preorganized for internal
carbopalladation of the 1,1-disubstituted ethylene to proceed
at a useful rate. Preorganization might arise from the
formation of an intermediate cyclic palladium(II) phenoxide
via reaction of a C8 palladium(II) halide with the E-ring
phenol or phenoxide. To our knowledge, there is no prece-
dent for this type of substrate direction in intramolecular
Heck cyclizations although Miura has recently invoked an
intermediate palladium(II) phenoxide in the bimolecular
arylation of naphthols with aryl halides catalyzed by Pd-
(0).11 Mechanistic insight notwithstanding, the efficient
conversion of 16b to 17a demonstrates that considerably
more hindered macrocyclic olefins are available via Heck
cyclization than had been reported previously.12 Whether
this protocol is restricted to our current substrate class or
will prove more broadly applicable is the subject of ongoing
experiments. Methods to convert lactams 5 to diazonamide
fragments 3 and further efforts directed toward the total
synthesis of these unique natural products will be reported
shortly.
Ack n ow led gm en t. Financial support provided by the
Advanced Research Program of the Texas Higher Educa-
tion Coordinating Board (003660-035), The Robert A. Welch
Foundation, and junior faculty awards administered through
the Howard Hughes Medical Institute and the University
of Texas Southwestern Medical Center are gratefully
acknowledged.
(8) Tarnawski, J .; Rzeszotarska, B.; Nadolska, B.; Pawelczak, K. Liebigs
Ann. Chem. 1979, 761-768.
(9) Yields based on analyzing crude HPLC traces and spectroscopic data,
retrospectively, after 17a had been fully characterized in latter experiments.
(10) C6 bromo derivatives provide access to the diazonamide B series.
See refs 1a and 1b.
(11) Satoh, T.; Kawamura, Y.; Miura, M.; Nomura, M. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 1740-1742.
(12) Walters, M. A. Chemtracts Org. Chem. 1998, 11, 291-296 and
references therein.
Su p p or tin g In for m a tion Ava ila ble: Experimental proce-
dures, characterization data, and 1H NMR spectra for 7, 8, 13c,
14, 15, 16b/e, and 17a /b (20 pages).
J O981791E