A new synthetic method for an acromelic acid analog, a potent neuroexcitatory kainoid amino acid, via photoinduced benzyl radical cyclization

Article abstract of DOI:10.1016/S0040-4039(02)01817-8

A new synthetic method for an acromelic acid analog, MFPA, was developed. The key step is the photoinduced benzyl radical cyclization with excellent stereoselectivity.

Full text of DOI:10.1016/S0040-4039(02)01817-8

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 7777–7780
A new synthetic method for an acromelic acid analog, a potent
neuroexcitatory kainoid amino acid, via photoinduced benzyl
radical cyclization
Satoshi Itadani, Shigeyuki Takai, Chieko Tanigawa, Kimiko Hashimoto* and Haruhisa Shirahama†
School of Science, Kwansei Gakuin University, Uegahara, Nishinomiya 662-8501, Japan
Received 8 July 2002; revised 22 August 2002; accepted 23 August 2002
Abstract—A new synthetic method for an acromelic acid analog, MFPA, was developed. The key step is the photoinduced benzyl
radical cyclization with excellent stereoselectivity. © 2002 Published by Elsevier Science Ltd.
Acromelic acids A and B (1 and 2) are potent neuroex-
citatory amino acids isolated as toxic compounds from
the poisonous mushroom Clitocybe acromelalga by us
(Fig. 1).1 These amino acids are members of the kainoid
amino acids which act on excitatory amino acid (EAA)
receptors in mammalian central nervous systems as
potent agonists.2 Among the EAA receptors, the
kainate subtype, one of the subtypes of the EAA
ionotropic receptors, are specifically depolarized by
these amino acids.3 We have developed various syn-
thetic methods for the kainoid amino acids in order to
find a more efficient method.4 During the course of
these studies, we synthesized the most potent analog,
the methoxyphenyl derivative (MFPA, 3), among the
kainate type agonists.4a,b The potency of the depolariza-
tion depends on the nature of the C4-substituent of the
kainoid.3c The compound showing the more potent
depolarization activity has a C4-substituent with the
higher HOMO energy.5 Accordingly, we developed a
new synthetic method suitable for synthesizing the
analogs with a C4-substituent having the higher
HOMO energy, which is exemplified in the synthesis of
MFPA (3).
The tosylate 4 prepared from 2-(2-methoxyphenyl)-
acetic acid by reduction with LiAlH4 followed by tosyl-
ation with TsCl and pyridine was coupled with vinyl-
oxazolidinone 56 prepared from
-methionine to give 6
L
in 83% yield (Scheme 1). The vinyl group of 6 was
converted to the unsaturated ester 7 by successive
ozonolysis and Horner–Emmons reactions. An acetoni-
trile or benzene solution of 7 was irradiated using a
high-pressure mercury lamp in a quartz test tube at
room temperature in the presence of a-trifluoroace-
tophenone and KF to afford the cyclized products 8
and 9 in a 1:1 ratio in 50% combined yield. In the
model studies of the photoreaction in acetonitrile, some
acyclic ethers are cyclized into THF rings with trans-
Figure 1.
Keywords: benzyl radical; radical cyclization; photoreaction; salt effect; kainoid.
* Corresponding author. Present address: Laboratory of Biochemical Resources, Plant Science Center, RIKEN, 2-1 Hirosawa, Wako-shi, Saitama
351-0198, Japan. Tel.: +81-48-467-5893; fax: +81-48-467-5407; e-mail: kimikoh@postman.riken.go.jp
† Present address: 2-16, Minami 18 Nishi 8, Chuo-ku, Sapporo, 064-0918, Japan.
0040-4039/02/$ - see front matter © 2002 Published by Elsevier Science Ltd.
PII: S0040-4039(02)01817-8

7778
S. Itadani et al. / Tetrahedron Letters 43 (2002) 7777–7780
Scheme 1. Reagents and conditions: (a) KH/DMF, 0°C, 83%; (b) O3/MeOH, then DMS; (c) (MeO)2P(O)CH2CO2Me, NaH/THF,
96% (two steps); (d) hw, CF3COPh, KF/CH3CN or benzene, 50% (8:9=1:1); (e) 1 M KOH/MeOH, reflux, then Boc2O; (f) CH2N2,
72% (two steps); (g) Jones’ oxidation, then CH2N2, 76%; (h)
1 M KOH/EtOH, quant.; (i) TFA, quant.; (j)
(EtO)2P(O)CH2CONMeOMe, NaH/THF, 88% (two steps); (k) hw, CF3COPh, KF/benzene, 41%; (l) HClaq, 66%, then CH2N2.
substituents as the major product without KF.8 On the
other hand, the presence of KF or TBAF was essential
for the reaction of 7 otherwise the yield was negligible.9
Thus the obtained cis-isomer was converted to MFPA
(3, [h]D+9.37°(c 0.25, H2O) [lit.:4a [h]D +4.99° (c 0.30,
H2O)]) by our previous synthetic reactions with slight
modifications.4a To improve the stereoselectivity in the
cyclization reaction, the radical acceptor was changed
from the methyl ester to the Weinreb amide using the
amide-type Horner–Emmons reagent prepared accord-
ing to a previous report.10 An acetonitrile solution of
the amide 10 was irradiated under the same conditions
as in the case of 7 to give 11 and 12 in a 1:1 ratio in
41% combined yield. On the other hand, a benzene
solution of the amide 10 was exposed to the photoreac-
tion to give cyclized products 11 and 12 in a ratio of
98:2.11 The yield was 41% with recovery of the starting
material (32%), which was rather low but the stereose-
lectivity was remarkably improved.12 The amide 11 was
hydrolyzed to the acid and the resulting compound was
esterified to yield 8 which can be converted to MFPA
(3) by the same methods as already described.
The plausible mechanism of the photoreaction drawn
by consulting Wagner’s studies7 is shown in Scheme 2.
The photoexcited a-trifluoroacetophenone [B] produces
a charge transfer complex (exciplex) with [A]. Through
the proton transfer reaction, the complex produces two
Scheme 2.

S. Itadani et al. / Tetrahedron Letters 43 (2002) 7777–7780
7779
Res. 1988, 474, 386–389; (c) Ishida, M.; Shinozaki, H. Br.
J. Pharmacol. 1991, 104, 873–878.
4. (a) Hashimoto, K.; Horikawa, M.; Shirahama, H. Tetra-
hedron Lett. 1990, 31, 7047–7050; (b) Hashimoto, K.;
Shirahama, H. Tetrahedron Lett. 1991, 32, 2625–2628; (c)
Horikawa, M.; Hashimoto, K.; Shirahama, H. Tetra-
hedron Lett. 1993, 34, 331–334; (d) Horikawa, M.; Shira-
hama, H. Synlett 1996, 95–96; (e) Kamabe, M.;
Miyazaki, T.; Hashimoto, K.; Shirahama, H. Heterocy-
cles 2002, 56, 105–111.
5. (a) Shirahama, H. In Organic Synthesis in Japan, Past,
Present, and Future; Noyori, R., Ed. Principles of a
poisonous mushroom working as neurotransmitters. A
dynamic aspect of natural product chemistry; Tokyo
Kagaku Dozin: Tokyo, 1992; pp. 373–384; (b) Hashi-
moto, K.; Matsumoto, T.; Nakamura, K.; Ohwada, S.;
Ohuchi, T.; Horikawa, M.; Konno, K.; Shirahama, H.
Bioorg. Med. Chem. 2002, 10, 1373–1379.
Figure 2.
radical species, [C] and [D]. The radical [D] cyclizes to
[E]. The radical [E] then gets an electron from [C] to
produce [F] which abstracts a proton probably from [G]
to produce [H]. These types of radical cyclizations
usually produce a cis-substituted five-membered ring,
which is explained by the transition state bearing the
equatorial substituents at the two carbons of the partly
formed bond (Fig. 2).13
In the case of 10, the product ratio was 98:2 in benzene
solution giving the cis-isomer as the major product for
which the above explanation should be applicable. The
photoreaction of the same amide 10 in acetonitrile
produced a mixture of 1:1 stereoisomers. The anion [F]
would be more stable in acetonitrile than in benzene
because of the solvent polarity; therefore, the anion [F]
would equilibrate with [I] before abstracting a proton
from [G], resulting in the production of the mixture.
6. Ohfune, Y.; Kurokawa, N. Tetrahedron Lett. 1984, 25,
1071–1074.
7. (a) Wagner, P. J.; Leavitt, R. A. J. Am. Chem. Soc. 1970,
92, 5806–5808; (b) Wagner, P. J.; Leavitt, R. A. J. Am.
Chem. Soc. 1973, 95, 3669–3677; (c) Wagner, P. J.;
Puchalski, A. E. J. Am. Chem. Soc. 1978, 100, 5948–5949;
(d) Wagner, P. J.; Puchalski, A. E. J. Am. Chem. Soc.
1980, 102, 6177–6178.
8. Itadani, S.; Hashimoto, K.; Shirahama, H. Heterocycles
1998, 49, 105–108.
On the other hand, in the case of 7, the cyclization
reactions in both acetonitrile and benzene produce cis-
and trans-isomers in a ratio of 1:1. The stereoselectivity
might be attributable to the stability of the anion [F] of
the ester independent of the solvent.
In the THF ring cyclization, the photoreaction smoothly
proceeded without salt. This might be due to the desir-
able conformation of the starting ether for the cycliza-
tion, which resulted in faster cyclization. On the other
hand, the amide derivative contained a nitrogen with sp2
hybridization, which might be disadvantage to adopt the
suitable conformation for the cyclization. Accordingly,
the amide derivative fell into equilibration between [A]
and [D]. The salt added might prevent [D] to [A] by some
effects (cf. Ref. 9), which promotes further cyclization.
In the ether series, the major product had 3,4-trans
configuration. This might be attributable to the structure
of the radical acceptor, i.e. ester, and also the solvent
used, i.e. acetonitrile.
Comparing the stability of the anions [F] of the ester
and of the amide, the latter must be rather destabilized
by the resonance effect of the nitrogen.
In addition, the cyclized cis-isomers could not be con-
verted into the corresponding trans-isomers and vice
versa under our photoreaction conditions by the
epimerization at the benzylic position through the radi-
cal intermediate [J].14
9. The role of KF and TBAF (salt effect) might be either or
some of the followings: (i) stabilization of radical ion pair
(exciplex) (ii) inhibition of back electron transfer ([D] to
[A]) (iii) acceleration of charge separation. See: Santa-
maria, J. In Photoinduced Electron Transfer; Fox, M. A.;
Chanon, M., Eds. Solvent and salt effects. Elsevier:
Amsterdam, 1988; Chapter 3.1, pp. 483–540.
References
1. (a) Konno, K.; Shirahama, H.; Matsumoto, T. Tetra-
hedron Lett. 1983, 24, 939–942; (b) Konno, K.; Hashi-
moto, K.; Ohfune, Y.; Shirahama, H.; Matsumoto, T.
Tetrahedron Lett. 1986, 27, 607–610; (c) Hashimoto, K.;
Konno, K.; Shirahama, H.; Matsumoto, T. Chem. Lett.
1986, 1399–1400; (d) Konno, K.; Hashimoto, K.; Ohfune,
Y.; Shirahama, H.; Matsumoto, T. J. Am. Chem. Soc.
1988, 110, 4807–4815.
2. For recent reviews on excitatory amino acids, see: (a)
Moloney, M. G. Nat. Prod. Rep., 1998, 15, 205–219; (b)
Moloney, M. G. Nat. Prod. Rep. 1999, 16, 485–498; (c)
Braeuner-Osborne, H.; Egebjerg, J.; Nielsen, E. Ø.; Mad-
sen, U.; Krogsgaard-Larsen, P. J. Med. Chem. 2000, 43,
2609–2645.
We tried some salts in the photoreaction, such as, KBr,
LiCl, and LiBr. However, no effect was observed for
these salts in the photoreaction. Accordingly, the desir-
able effect in the reaction probably arose from the
fluoride ion. The salt effect was well documented for
organic compounds and some inorganic compounds,
such as, LiClO4, Mg(ClO4)2, and Bu4NClO4. The effects
of fluoride ion to this kind of organic synthesis would be
the first example. Fluoride-promoted dye-sensitized pho-
tooxidation was reported by Wasserman; however, the
role of the fluoride ion seemed to be different from that
of our reaction. See Wasserman, H. H.; Pickett, J. E. J.
Am. Chem. Soc. 1982, 104, 4695–4696.
3. (a) Shinozaki, H.; Ishida, M.; Okamoto, T. Brain Res.
1986, 399, 395–398; (b) Ishida, M.; Shinozaki, H. Brain

7780
S. Itadani et al. / Tetrahedron Letters 43 (2002) 7777–7780
10. Nuzillard, J.-M.; Boumendjel, A.; Massiot, G. Tetra-
hedron Lett. 1989, 30, 3779–3780.
of a-trifluoroacetophenone and/or KF. This might be due
to the formation of the pinacol [K], which disturbs the
formation of [F]. In the photoreaction, the irradiation for
long time made the decrease in the yields both of the
products and recovery of the starting material.
11. Typical experimental: To a solution of 10 (55 mg, 0.16
mmol) in acetonitrile (2.2 ml) in a quartz test tube were
added KF (10.0 mg, 0.17 mmol) and a-trifluoroacetophe-
none (0.12 ml, 0.82 mmol). The solution was irradiated
with stirring under argon atmosphere using high-pressure
mercury lamp at rt for 1 h. The mixture was concentrated
and the residue was chromatographed on silica gel (40%
EtOAc–hexane) to afford a mixture of 11 (22. 5 mg, 41%)
and 12 (0.8% from 1H NMR), and recovered 10 (17.5 mg,
32%). The stereochemistries of the products were deter-
mined after conversion to the known 3.
13. Fossey, J.; Lefort, D.; Sorba, J. Free Radicals in Organic
Chemistry; John Wiley & Sons: Chichester, 1995; p. 154.
14. The carbonyl group of a-trifluoroacetophenone in the
excited state [B] is reported to be hard to abstract hydro-
gen. Accordingly, the direct production of the radical
species [J] (and also [E]) from [H] could be excluded. The
exciplex formation between [B] and [H] followed by pro-
ton transfer would produce [J]; however, such a reaction
was not observed.
12. The yield was not improved by increasing the quantities