Synthesis of azacridone A

Article abstract of DOI:10.1021/ol048382r

The first total synthesis of the only known naturally occurring azaacridone alkaloid (1) has been achieved in 10 steps from phloroglucinol. A variety of ortholithiation reactions are described, and a method for overcoming the originally unfavorable regiochemistry of one of them is provided.

Full text of DOI:10.1021/ol048382r

ORGANIC
LETTERS
2004
Vol. 6, No. 21
3869-3871
Synthesis of Azacridone A
Alex Scopton and T. Ross Kelly*
E. F. Merkert Chemistry Center, Boston College, Chestnut Hill, Massachusetts 02467
ross.kelly@bc.edu
Received August 13, 2004
ABSTRACT
The first total synthesis of the only known naturally occurring azaacridone alkaloid (1) has been achieved in 10 steps from phloroglucinol. A
variety of ortholithiation reactions are described, and a method for overcoming the originally unfavorable regiochemistry of one of them is
provided.
In 1993, a team of Japanese scientists reported the isolation
of the first naturally occurring azaacridone alkaloid.1 The
compound, aptly named azacridone A, was assigned structure
1 on the basis of its spectroscopic properties. Since azacri-
done A is the only known representative of the [1]-
benzopyrano[5,6-b][1,7]naphthyridine ring system and since
no independent confirmation of the structure determination
has been recorded, the synthesis of 1 was undertaken. We
now report the first synthesis of azacridone A and the
affirmation of the original structure assignment.
synthon could be incorporated initially attached to either the
pyridine or the benzopyran unit. While both possibilities were
examined, in practice initial attachment to the benzopyran
entity provided the solution.
A specific embodiment (5) of differentially protected
benzopyran 4 was prepared as shown in Scheme 1. Thus,
Scheme 1
Retrosynthetic analysis suggested (eq 1) that 1 might be
secured by cyclization of 2. The latter should be accessible
from 3, 4, and a carbonyl synthon. In principle, the carbonyl
the known annelation of phloroglucinol (6) with â,â-
dimethylacrylic acid afforded 7.2 The two hydroxyl groups
10.1021/ol048382r CCC: $27.50
© 2004 American Chemical Society
Published on Web 09/17/2004

were selectively protected (7 f 8 f 9) by exploiting the
hydrogen-bonding interaction between the carbonyl oxygen
and the peri hydroxyl. Reduction to 10 and dehydration then
provided 5.
pected, achieve both deprotections, but unfortunately also
led to regiospecific cyclization to the unwanted 18 (Scheme
2).
It was anticipated that the greater chelating properties of
the methoxymethoxy (MOMO) group would cause it to
dominate the regiochemical outcome of ortholithiation of 5.3
That expectation was borne out: D2O-quench studies com-
bined with NOESY measurements revealed that lithiation
of 5 is regiospecific, giving 11. Reaction of 11 with N,N-
dimethylformamide (DMF) gave aldehyde 12 in 73% yield
from 5 (Scheme 2).
Scheme 2
The preferentialsbut undesiredsformation of 18 rather
than 21 was ascribed to hydrogen bonding (see 22), which
by that argument, imposes a conformation on 19 that ordains
the regiochemical outcome.
If the hydrogen bonding interaction (e.g., 22) is, in fact,
the determining factor, then a simple cure is available: swap
the Me and MOM protecting groups in 17. It is hardly
surprising to one skilled in the art of organic synthesis that
accomplishing such a seemingly insignificant change proved
nontrivial.
Switching the two protecting groups themselves at the
bicyclic level to give 24 was straightforward (Scheme 3).
Scheme 3
As expected based on the work of Hands et al.,4 and
confirmed by the 1H NMR spectrum of the product, double
lithiation of the t-BOC derivative (13)5 of 3-aminopyridine
gives 14, which affords 15 upon reaction with aldehyde 12.
Oxidation6 of the benzhydryl alcohol followed by N-
methylation provides 17.
It had been anticipated that treatment of 17 with acid
would deprotect both the t-BOC and MOM groups, giving
19. Subsequent cyclization of 19, perhaps via its minor keto
tautomer 20, should then generate 21. Exposure of 17 to
trifluoroacetic acid (TFA) in dichloromethane did, as ex-
(1) Takemura, Y.; Isono, Y.; Ju-Ichi, M.; Omura, M.; Ito, C.; Furukawza,
H. Chem. Pharm. Bull. 1993, 41, 789.
(2) (a) Sowmithran, D.; Prasad, K. J. R. Synthesis 1985, 545. (b) Timar,
T. J. Heterocycl. Chem. 1988, 25, 871.
(3) For a recent review, see: Whisler, M. C.; MacNeil, S.; Snieckus,
V.; Beak, P. Angew. Chem., Int. Ed. 2004, 43, 2206.
(4) Hands, D.; Bishop, B.; Cameron, M.; Edwards, J. S.; Cottrell, I. F.;
Wright, S. H. B. Synthesis 1996, 877.
(5) Kelly, T. A.; McNeil, D. W. Tetrahedron Lett. 1994, 35, 9003.
(6) Frigerio, M.; Santagostino, M. Tetrahedron Lett. 1994, 35, 8019.
However, lithiation employing the conditions used in Scheme
2 (5 f 11) gave predominantly 25, not the desired 26, as
established by D2O-quench and NOESY studies, as before.
Apparently, while the MOMO group is presumably still the
3870
Org. Lett., Vol. 6, No. 21, 2004

dominant director, the conformationally constrained pyran
oxygen wins out against the conformationally mobile meth-
oxy oxygen. An extensive (see the Supporting Information)
examination of the role of additives,7 solvents, and reaction
conditions revealed that in the presence of 2 equiv of HMPA
Scheme 4
n
(per equivalent of BuLi), lithiation of 24 generates pre-
dominantly 26, rather than 25. We are not certain why the
presence of HMPA changes the outcome, but ortholithiations
can be complicated, with regioselectivity8 being a function
of the complex interactions of kinetics, thermodynamics,
chelation, and aggregation.9 In the event, a change in reaction
conditions transforms a 3:1 regioisomer ratio to a 1:4 ratio.
With the ability to effect the desired lithiation of 24,
completion of the synthesis soon fell into place. Reaction of
26 with Fe(CO)510,11 gave 27 (Scheme 4) in 53% yield (along
with 18% of the 25-derived regioisomer). A three-step
sequence similar to before (Scheme 2), converted aldehyde
27 to ketone 30. Treatment of 30 with TFA resulted in a
cascade of reactions analogous to those in Scheme 2, but
this time leading directly to the desired target, azacridone
A. Spectra for synthetic 1 are in excellent agreement with
those reported for naturally occurring azacridone A, confirm-
ing the structure originally assigned.
In conclusion, we report the first synthesis of azacridone
A, a synthesis that affirms the initially assigned structure.
(7) Fraser, R. R.; Mansour, T. S. Tetrahedron Lett. 1986, 27, 331.
(8) For cases where lithiation regioselectivity has been reversed, see:
(a) Shimano, M.; Meyers, A. I. J. Am. Chem. Soc. 1994, 116, 10815. (b)
Maggi, R.; Schlosser, M. J. Org. Chem. 1996, 61, 5430.
(9) (a) Winkle, M. R.; Ronald, R. C. J. Org. Chem. 1982, 47, 2101. (b)
Ronald, R. C.; Winkle, M. R. Tetrahedron 1983, 39, 2031. (c) Saa, J. M.;
Martorell, G.; Frontera, A. J. Org. Chem. 1996, 61, 5194. (d) Rennels, R.
A.; Maliakal, A. J.; Collum, D. B. J. Am. Chem. Soc. 1998, 120, 421. (e)
Chadwick, S. T.; Rennels, R. A.; Rutherford, J. L.; Collum, D. B. J. Am.
Chem. Soc. 2000, 122, 8640.
(10) (a) Yamashita, M.; Miyoshi, K.; Nakazono, Y.; Suemitsu, R. Bull.
Chem. Soc. Jpn. 1982, 55, 1663. (b) Cornella, I.; Kelly, T. R. J. Org. Chem.
2004, 69, 2191.
(11) Anion 25 was largely unreactive toward DMF at 25 °C. We have
previously shown (ref 10b) that Fe(CO)5 exhibits greater reactivity toward
aryllithium species than DMF.
Acknowledgment. We thank Mr. Jack Beierle for
preliminary studies.
Supporting Information Available: Experimental pro-
cedures and characterization data for selected compounds.
This material is available free of charge via the Internet at
http://pubs.acs.org.
OL048382R
Org. Lett., Vol. 6, No. 21, 2004
3871