Stereocontrolled syntheses of kainoid amino acids from 7-azabicyclo[2.2.1] heptadienes using tandem radical addition-homoallylic radical rearrangement

Article abstract of DOI:10.1021/jo0513865

N-Boc syn-7-(2-hydroxyethyl)-4-(alkyl or aryl)sulfonyl-2-azabicyclo[2.2.1] hept-5-enes serve as precursors in syntheses of the neuroexcitants 3-(carboxymethyl)pyrrolidine-2,4-dicarboxylic acid 43, α-kainic acid 12, α-isokainic acid 14, and α-dihydroalloka

Full text of DOI:10.1021/jo0513865

Stereocontrolled Syntheses of Kainoid Amino Acids from
7-Azabicyclo[2.2.1]heptadienes Using Tandem Radical
Addition-Homoallylic Radical Rearrangement
David M. Hodgson,*,† Shuji Hachisu,† and Mark D. Andrews‡
Department of Chemistry, University of Oxford, Chemistry Research Laboratory, Mansfield Road, Oxford,
OX1 3TA, United Kingdom, and Pfizer Global Research and Development, Ramsgate Road, Sandwich,
Kent CT13 9NJ, United Kingdom
david.hodgson@chem.ox.ac.uk
Received July 5, 2005
N-Boc syn-7-(2-hydroxyethyl)-4-(alkyl or aryl)sulfonyl-2-azabicyclo[2.2.1]hept-5-enes serve as precur-
sors in syntheses of the neuroexcitants 3-(carboxymethyl)pyrrolidine-2,4-dicarboxylic acid 43,
R-kainic acid 12, R-isokainic acid 14, and R-dihydroallokainic acid 77. The key step in these syntheses
is the intermolecular radical addition of 2-iodoethanol to a N-Boc 2-(alkyl or aryl)sulfonyl-7-
azabicyclo[2.2.1]heptadiene 7 to induce nitrogen-directed homoallylic radical rearrangement.
Oxidative cleavage of the resulting 2-azabicyclo[2.2.1]hept-5-enes provide straightforward access
to polysubstituted pyrrolidines and, in particular, an efficient entry to the kainoid amino acids.
SCHEME 1. Radical Addition of Thiols to
7-Azabicyclo[2.2.1]heptadiene 12
Introduction
Radical cyclizations and rearrangements are versatile
processes for the construction of ring systems.1 We
recently demonstrated that the addition of alkyl or aryl
thiols to 7-azabicyclo[2.2.1]heptadiene 1 resulted in a
tandem intermolecular radical addition-homoallylic radi-
cal rearrangement (RA-HRR) reaction, which ultimately
led to 7-thio-substituted 2-azabicyclo[2.2.1]hept-5-enes 2
(Scheme 1).2 This rearrangement (3 f 5) is considered
to be facilitated by the stabilization imparted to the
intermediate radical 5 by the R-nitrogen.3 The lone pair
of electrons on the nitrogen atom can stabilize the
adjacent radical via a HOMO-SOMO orbital interac-
tion.4
In the present paper, we detail a full account of
extension and adaptation of the aforementioned meth-
odology to the addition of carbon-based radicals to
7-azabicyclo[2.2.1]heptadienes 7 (Scheme 2).5 This radical
addition process was accompanied by homoallylic radical
rearrangement leading to the formation of 2-azabicyclo-
[2.2.1]hept-5-enes 10. The latter could in turn serve as
templates for fashioning 2,3,4-trisubstituted pyrrolidines
* Corresponding author. Phone: 440-186-527-5697; fax: 440-186-
528-5002.
† University of Oxford.
‡ Pfizer Global Research and Development.
(1) (a) Beckwith, A. L. J.; Ingold, K. U. In Rearrangements in Ground
and Excited States; de Mayo, P., Ed.; Academic Press: New York, 1980;
Vol. 1, pp 161-310. (b) Giese, B.; Kopping, B.; Gobel, T.; Dickhaut, J.;
Thoma, G.; Kulicke, K. J.; Trach, F. Org. React. 1996, 48, 301-856.
(c) Radicals in Organic Synthesis; Renaud, P., Sibi, M. P., Eds.; Wiley-
VCH: Weinheim, 2001.
(2) (a) Hodgson, D. M.; Bebbington, M. W. P.; Willis, P. Chem.
Commun. 2001, 889-890. (b) Hodgson, D. M.; Bebbington, M. W. P.;
Willis, P. Org. Biomol. Chem. 2003, 1, 3787-3798.
(3) For other nitrogen-directed radical rearrangements, see: (a)
Rigby, J. H.; Pigge, F. C. Tetrahedron Lett. 1996, 37, 2201-2204. (b)
Hodgson, D. M.; Galano, J.-M. Org. Lett. 2005, 7, 2221-2224.
(4) Wayner, D. D. M.; Clark, K. B.; Rauk, A.; Yu, D.; Armstrong, D.
A. J. Am. Chem. Soc. 1997, 119, 8925-8932.
(5) (a) Hodgson, D. M.; Hachisu, S.; Andrews, M. D. Org. Lett. 2005,
7, 815-817. (b) Hodgson, D. M.; Hachisu, S.; Andrews, M. D. Synlett
2005, 1267-1270.
10.1021/jo0513865 CCC: $30.25 © 2005 American Chemical Society
Published on Web 09/24/2005
8866
J. Org. Chem. 2005, 70, 8866-8876

Stereocontrolled Syntheses of Kainoid Amino Acids
SCHEME 2. Strategy to Trisubstituted Pyrrolidines (Kainoids)
11. To demonstrate the utility of this novel strategy,
syntheses of several kainoid amino acids were accom-
plished.
latter is initiated by exo-selective conjugate addition of
(-)-methyl lactate to dienyl sulfone 16 to furnish two
diastereomeric Michael adducts that can be separated by
an efficient fractional recrystallization. The requisite
dienyl sulfone 16 enantiomer can then be liberated by
base-induced â-elimination of the original (-)-methyl
lactate portion (following reduction of the methyl ester
with NaBH4).
R-Kainic acid 12, R-allokainic acid 13, and R-isokainic
acid 14 are members of the kainoid family of natural
products (Figure 1).6 There are also several more complex
kainoids, whose structural diversity originates from
variation of the C-4 side chain, such as isodomoic acid G
15.7 The kainoids have been the focus of considerable
synthetic activity, primarily due to their challenging
structures combined with the marked excitatory neu-
rotransmitting activities they induce in the mammalian
nervous system.8 In particular, syntheses of R-kainic acid
12 and R-allokainic acid 13 have been accomplished by
several approaches;6,9 however, we considered that our
strategy would have the benefit of brevity and enough
flexibility to address the various challenging substitution
patterns and stereochemical issues found in several
kainoid amino acids.
SCHEME 3. RA-HRR Reaction of Azadiene 16
Using Various Alkyl Iodides
Unlike the symmetrical 7-azabicyclo[2.2.1]heptadiene
1 (Scheme 1), dienyl sulfone 16 could potentially undergo
radical addition at four distinct sites at the double bonds.
There is also a stereochemical issue (exo vs endo attack)
and whether the radical addition would lead on to any
homoallylic rearrangement. Consideration of these issues
indicates that radical addition to dienyl sulfone 16 could
potentially lead to up to 16 distinct products. Hence, it
was apparent from the outset that attaining high selec-
tivity in the radical addition and the subsequent re-
arrangement would be crucial for future synthetic utility.
Initially, the additions of electrophilic radicals to dienyl
sulfone 16 were investigated. These were attempted
using ethyl bromoacetate, ethyl iodoacetate, and diethyl
bromomalonate by atom transfer (Karasch-type) process
and reductive addition employing Bu3SnH or Ni(OAc)2/
FIGURE 1. Kainoid Amino Acids.
Results and Discussion
Before embarking upon the syntheses of various kain-
oids, a study of the viability of the key intermolecular
RA-HRR reaction was carried out. Dienyl sulfone 16 was
chosen as the radical acceptor (Scheme 3) since it is
available in one step from commercial materials (N-Boc
pyrrole and tosyl ethyne) and the enantiomers are also
accessible by a straightforward resolution protocol.10 The
(9) For selected recent syntheses, see: (a) Chevliakov, M. V.;
Montgomery, J. J. Am. Chem. Soc. 1999, 121, 11139-11143. (b)
Campbell, A. D.; Raynham, T. M.; Taylor, R. J. K. J. Chem. Soc., Perkin
Trans. 1 2000, 3194-3204. (c) Xia, Q.; Ganem, B. Org. Lett. 2001, 3,
485-487. (d) Clayden, J.; Menet, C. J.; Tchabanenko, K. Tetrahedron
2002, 58, 4727-4733. (e) Trost, B. M.; Rudd, M. T. J. Am. Chem. Soc.
2005, 127, 4763-4776. (f) Martinez, M. M.; Hoppe, D. Eur. J. Org.
Chem. 2005, 1427-1443. (g) Scott, M. E.; Lautens, M. Org. Lett. 2005,
7, 3045-3047. (h) Morita, Y.; Tokuyama, H.; Fukuyama, T. Org. Lett.
2005, 7, 4337-4340.
(6) (a) Parsons, A. F. Tetrahedron 1996, 52, 4149-4174. (b) Moloney,
M. G. Nat. Prod. Rep. 2002, 19, 597-616. (c) Clayden, J.; Read, B.;
Hebditch, K. R. Tetrahedron 2005, 61, 5713-5724. (d) Calaf, R.;
Barlatier, A.; Maillard, C.; Balansard, G.; Garcon, D. Plant. Med.
Phytother. 1989, 23, 24-32. (e) Calaf, R.; Barlatier, A.; Maillard, C.;
Ollivier-Vidal, E.; Garcon, D. Plant. Med. Phytother. 1989, 23, 16-23.
(7) Zaman, L.; Arakawa, O.; Shimosu, A.; Onoue, Y.; Nishio, S.;
Shida, Y.; Noguchi, T. Toxicon 1997, 35, 205-212.
(8) (a) Shinozaki, H.; Konishi, S. Brain Res. 1970, 24, 368-371. (b)
Olney, J. W.; Rhee, V.; Ho, O. L. Brain Res. 1974, 77, 507-512. (c)
Bleakman, D.; Lodge, D. Neuropharmacology 1998, 37, 1187-1204.
(10) Leung-Toung, R.; Liu, Y.; Muchowski, J. M.; Wu, Y.-L. J. Org.
Chem. 1998, 63, 3235-3250.
J. Org. Chem, Vol. 70, No. 22, 2005 8867

Hodgson et al.
borohydride-exchange resin.11 All these attempts to effect
selective addition at the electron rich (nonsulfone bear-
ing) double bond of dienyl sulfone 16 led to the formation
of complex mixtures.
interest in such systems as analogues of the potent
nonopioid analgesic nicotinic acetylcholine receptor ago-
nist epibatidine.13
To investigate the origins of the selectivity for the
isomeric azabicycles in the radical addition reaction,
xanthate 29 was synthesized for radical deoxygenation
(Scheme 4). Treatment of epoxide 27 (available via regio-
and stereoselective epoxidation of dienyl sulfone 16) with
t-BuMgCl (3 equiv) led to the formation of tricyclic alcohol
28 (92%).14 Attempted radical deoxygenation of the
derived xanthate 29 at room temperature using Et3B and
Bu3SnH led only to recovery of the xanthate 29. There-
fore, the deoxygenation was carried out with AIBN and
Bu3SnH at 110 °C, which led to the sole formation of the
rearranged 2-azabicycle 18 (45%, Scheme 4).
The addition of nucleophilic alkyl radicals to dienyl
sulfone 16 was then explored to test selectivity for
preferential reaction at the electron poor (tosyl-bearing)
double bond. In an effort to maximize facial selectivity
(cf. Scheme 1), a tertiary radical was examined. Reductive
radical addition using t-BuI and Bu3SnH (3 equiv each)
with Et3B/O2 as initiator1c in CH2Cl2 led to the formation
of 7-azabicycle 17 and 2-azabicycle 18 in good combined
yield (83%, Scheme 3), when the stannane was slowly
added to the iodide and dienyl sulfone 16 (0.01 M) at
room temperature. Other conditions were found to be less
satisfactory. The unrearranged 7-azabicycle 17 predomi-
nated 2-fold over the rearranged 2-azabicycle 18, and this
was initially considered to be principally due to a
stabilizing effect of the tosyl group on the radical
intermediate (cf. 8, Scheme 2). This stabilizing effect
would slow the HRR that ultimately leads to the forma-
tion of 2-azabicycle 18. In support of this argument,
radical addition of thiols (better H-atom donors then
stannanes) to 7-azabicyclo[2.2.1]heptadiene 1 (initially
∼0.1 M in thiol and 1), which does not contain a
potentially radical-stabilizing tosyl group, led to the
exclusive formation of rearranged 2-azabicyclo[2.2.1]hept-
5-enes 2 (Scheme 1).2 Importantly, H-atom transfer from
Bu3SnH to the radical precursors of azabicycles 17 and
18 is clearly preferred relative to dienyl sulfone 16
oligimerization; the latter may be disfavored due to
severe intermolecular steric interactions in the potential
addition steps (and/or capto stabilization for the radical
precursor to 17).
A range of other alkyl iodides was investigated to probe
the effect of the differing electronic and steric character
of the corresponding alkyl radicals and to establish the
scope of the RA-HRR reaction (Scheme 3). In all cases,
good yields (70-83%) of radical addition products were
obtained. Although addition of the various radicals was
efficient, the extent of homoallylic rearrangement varied
considerably. The addition of tertiary radicals led to the
formation of the unrearranged 7-azabicycles 17 and 23
as the major products, a secondary radical led to the
formation of the rearranged 2-azabicycle 20 as the major
product, and primary radicals led to the formation of
rearranged 2-azabicycles (22 and 26) as almost the sole
products. Side products due to ethyl radical incorporation
from the initiator were not detected in these reactions.
Presumably, this process occurs to a small extent (espe-
cially with the primary iodides, where the rate of iodine
atom transfer is slowest),12 but propagation of the desired
reaction must be efficient relative to ethyl radical gen-
eration. It is also noteworthy that exclusive exo-attack
of the alkyl radicals on dienyl sulfone 16 was always
observed, whereas the addition of thiols to diene 1
(Scheme 1) always led to an epimeric mixture of sulfides
2 (favoring the syn-isomer from exo-attack).2 This novel
stereocontrolled entry into 7-substituted 2-azabicyclo-
[2.2.1]hept-5-enes is potentially useful due to the current
SCHEME 4. Radical Deoxygenation of Tricyclic
Alcohol 28a
a (a) t-BuMgCl, THF, -78-20 °C (92%); (b) KH, CS2, MeI, THF,
0-20 °C (92%); (c) Bu3SnH, AIBN, PhMe, 110 °C (45%).
Assuming that radical dexoygenation of xanthate 29
proceeds via tricyclic radical 30 as an intermediate
implies that, once the latter is reached in the RA-HRR
chemistry, then it can also only lead to rearranged
2-azabicycle rather than (back to) 7-azabicycle. In this
analysis of the RA-HRR, the ratio of the 7-aza to 2-aza
products (32:33, Scheme 5) is then determined by the rate
of cyclization of the initially formed intermediate from
radical addition, 34 to give tricycle 37, relative to the rate
of H-atom transfer to 34. The origin of the trend in the
ratio of products is likely not the electronic nature of the
alkyl radicals that are added because the addition of the
adamantyl radical gives an essentially identical result
to the addition of a standard tertiary (t-Bu) radical
(Scheme 3). The electronic nature of an adamantyl radical
is rather different from that of an ordinary tertiary
radical, due to the formers’ inability to participate in
hyperconjugative radical stabilization.15 On the other
hand, the steric demand of an adamantyl group is similar
to that of t-Bu group. Hence, it seems reasonable to
attribute the unrearranged/rearranged ratio to the steric
effect exerted by the alkyl groups once added to 7-azabi-
(13) (a) Malpass, J. R.; White, R. J. Org. Chem. 2004, 69, 5328-
5334. (b) Malpass, J. R.; Handa, S.; White, R. Org. Lett. 2005, 7, 2759-
2762. (c) Hodgson, D. M.; Maxwell, C. R.; Wisedale, R.; Matthews, I.
R.; Carpenter, K. J.; Dickenson, A. H.; Wonnacott, S. J. Chem. Soc.,
Perkin Trans. 1 2001, 3150-3158.
(14) (a) Jin, Z.; Fuchs, P. L. J. Am. Chem. Soc. 1995, 117, 3022-
3028. (b) Hodgson, D. M.; Jones, M. L.; Maxwell, C. R.; Ichihara, O.;
Matthews, I. R. Synlett 2005, 325-327.
(15) (a) Chick, W. H.; Ong, S. H. Chem. Commun. 1969, 216-217.
(b) Humphrey, L. B.; Hodgson, B.; Pincock, R. E. Can. J. Chem. 1968,
46, 3099-3103.
(11) Joung, M. J.; Ahn, J. H.; Lee, D. W.; Yoon, N. M. J. Org. Chem.
1998, 63, 2755-2757.
(12) Newcomb, M.; Sanchez, R. M.; Kaplan, J. J. Am. Chem. Soc.
1987, 109, 1195-1199.
8868 J. Org. Chem., Vol. 70, No. 22, 2005

Stereocontrolled Syntheses of Kainoid Amino Acids
SCHEME 5. Suggested Reaction Pathway of RA-HRR Reaction
cycle 16. The addition of alkyl radicals to 7-azabicycle
16 initially leads to the formation of intermediate 34
(Scheme 5), where the radical is likely to be stabilized
by the sulfone. The radical in intermediate 34 is unlikely
to reside in a sp2 like hybrid orbital (requiring 120° bond
angles), due to the rigidity of the bicyclic system.16
Instead, the radical is likely to be residing in a more sp3-
like hybrid orbital, as suggested in conformations 35 and
36 (which may be in equilibrium). It is expected that
conformation 35 would be preferred, where steric inter-
actions between the sulfone and the alkyl group (R) are
minimized; conformation 36 (which possesses the desired
SOMO-HOMO overlap for cyclopropane formation) would
not be as stable, due to eclipsing of the sulfone and the
alkyl group (R), and this effect would be most pronounced
for the more sterically demanding alkyl groups. Hence,
addition of a sterically demanding tertiary radical leads
preferentially to the formation of the unrearranged
7-azabicycle 32, whereas the addition of sterically less
demanding primary radical leads to the formation of
rearranged 2-azabicycle 33 as the major product.
projected kainoid amino acid syntheses because the
resulting 2-azabicyclo[2.2.1]heptene 26 contains a 7-hy-
droxyethyl substituent that can potentially be converted
into carboxymethyl functionality appended at C-3 of the
kanoid pyrrolidine core. Therefore, the radical addition
of 2-iodoethanol to dienyl sulfone 16 was examined in
more detail at different temperatures, and this produced
an interesting trend (Table 2). The efficiency of the
reaction at -90 °C was excellent (96% yield), but there
was a reversal in selectivity in favor of 7-azabicycle 25
at low temperatures (-78 and -90 °C). A reaction carried
out at above ambient temperature only resulted in a
decrease in yield, with no difference in product ratio being
discernible as compared to that at ambient temperature.
These results indicate that the presence of sufficient
thermal energy is essential for homoallylic rearrange-
ment to occur relative to direct reduction, and that the
RA-HRR reaction is most efficiently (and conveniently)
carried out at ambient temperature.
TABLE 2. RA-HRR Reaction of Azadiene 16 Using
2-Iodoethanol at Various Temperatures
Bu3SnH is a well-established and versatile reducing
agent in radical chemistry, but its toxicity, cost, and the
problematic removal of byproducts make it less preferable
for use in many applications.17 Therefore, we investigated
two alternative hydrogen atom donors: dichlorogallane
(Cl2GaH)18 and 1-ethylpiperidium hypophosphite (EPHP).19
Radical addition of i-PrI to dienyl sulfone 16 employing
Cl2GaH or EPHP with Et3B/O2 as radical initiator led to
high yields of products, comparable to using Bu3SnH
(Table 1). Of note was the efficiency of the intermolecular
radical addition of i-PrI using Cl2GaH, where just over 1
equiv each of Cl2GaH and i-PrI proved sufficient.
reaction
25 + 26
entry
T (°C)
time (min)
yield (%)
25:26
1a
2b
3a
4a
5c
-90
-78
0
20
85
45
80
60
60
35
96
54
67
78
53
1:0.33
1:0.45
1:14
1:13
1:14
a Bu3SnH added over reaction time indicated, Et3B, dry air (O2),
CH2Cl2. b Bu3SnH added in one portion, Et3B, dry air (O2), CH2Cl2.
c AIBN and Bu3SnH in PhMe added over reaction time indicated.
TABLE 1. RA-HRR Reaction of Azadiene 16 Using i-PrI
and Various Hydrogen Atom Donors
H-donor i-PrI
yield
entry H-donor (equiv) (equiv) solvent T (°C) 19:20 (%)
Sulfone functionality is necessary to facilitate the [4
+ 2] cycloaddition, giving the starting 7-azadiene 16,20
but its presence is not required in the target kainoids;
hence, its efficient removal21 is an important goal. Pleas-
ingly, treatment of the inseparable mixture of sulfones
1a
2b
3c
Bu3SnH
Cl2GaH
EPHP
3
1.2
9.0
3
1.2
1.0
CH2Cl2
THF
20
0
20
1:2.5
1:4.8
1:2.3
83
63
67
dioxane
a 0.01 M in 16. b 0.18 M in 16. c 0.16 M in 16.
(17) Curran, D. P. Synthesis 1988, 417-439.
The radical addition of 2-iodoethanol (Scheme 3) is a
particularly important result in the context of our
(18) Mikami, S.; Fujita, K.; Nakamura, T.; Yorimitsu, H.; Shinokubo,
H.; Matsubara, S.; Oshima, K. Org. Lett. 2001, 3, 1853-1855.
(19) Jang, D. O.; Cho, D. H.; Chung, C.-M. Synlett 2001, 1923-1924.
(20) Chen, Z.; Trudell, M. L. Chem. Rev. 1996, 96, 1179-1193.
(21) Najera, C.; Yus, M. Tetrahedron 1999, 55, 10547-10658.
(16) Kawamura, T.; Koyama, T.; Yonezawa, T. J. Am. Chem. Soc.
1973, 95, 3220-3228.
J. Org. Chem, Vol. 70, No. 22, 2005 8869

Hodgson et al.
SCHEME 6. Desulfonylation Reactionsa
a Asterisk: Combined yield of 39 and 40.
25 and 26 (1:13) with 6% sodium amalgam and boric
acid22 in methanol at reflux gave alkene 40 (74%, Scheme
6) and a trace amount of 7-azabicycle 39, separable by
chromatography. With the aim of definitively identifying
7-aza-
bicycle 39, desulfonylation was carried out on a mixture
of sulfones 25 and 26 in which the 7-aza isomer predomi-
nated (1:0.33). Intriguingly, it was observed that the
quantity of 2-azabicycle 40 produced from desulfonylation
was larger than the quantity of 2-azabicyclic sulfone 26
initially present (Scheme 6). This (fortuitous) enhance-
ment in the yield of 2-azabicycle 40 may be rationalized
by recognizing that during desulfonylation of 7-azabicycle
25, a radical could be generated at C-2 that might
undergo partial HRR.13c
Having established a viable RA-HRR reaction and
subsequent desulfonylation, the stage was set to apply
the methodology to kainoid targets. To straightforwardly
demonstrate the viability of the synthetic strategy, triacid
43, which is known to exhibit strong neuroexcitatory
activity,23 was chosen as an initial target (Scheme 7).
Oxidative cleavage of alkene 40 was carried out in a
biphasic mixture of EtOAc and H2O containing a catalytic
amount of RuO2‚H2O with NaIO4 as the stoichiometric
oxidant.24 Prolonged exposure of the resulting hydroxy
diacid 41 to these conditions led to the formation of Boc-
protected triacid 42, but unfortunately, this was ac-
companied by an unacceptable degree of decomposition.
The oxidation of the primary alcohol was therefore
carried out in a separate step with Jones’ reagent to
cleanly yield Boc-protected triacid 42. The final depro-
tection was carried out, following literature precedent,
by dissolving Boc-protected triacid 42 in formic acid,
which gave triacid 43 (77% from alkene 40) with spectral
data in accord with those reported in the literature.23
Application of the RA-HRR methodology to the syn-
thesis of R-kainic acid 12 was considered a more chal-
lenging target. The RA-HRR reaction was readily applied
to the synthesis of triacid 43 (Scheme 7) because oxida-
tive cleavage of the double bond in 2-azabicycle 40 led
directly to the desired stereochemistry at the three
contigious stereogenic centers, and differentiation be-
tween the (originally) alkene termini was not necessary
as they were both manipulated into carboxylic acid
functionality. In comparison, a successful synthesis of
R-kainic acid 12 would necessitate differentiation of the
alkene termini to construct an isopropenyl group, and
inversion of stereochemistry at C-4 (R-kainic acid num-
bering). We considered that if differentiation of the
originally alkene termini after oxidative cleavage of the
double bond could be achieved, then hydration of a
derived enol ether 44 (Scheme 8) might furnish the
desired cis-stereochemistry between C-3 and C-4, driven
by avoidance of the additional ring strain that could
result from a trans-fused bicyclic system.
SCHEME 8. Synthetic Strategy to r-Kainic Acid 12
Ozonolysis of 2-azabicycle 26 followed by a reductive
(Me2S) workup cleanly furnished a mixture of lactols 45
and 46 (100% crude yield, unstable to chromatography,
Scheme 9). The lactols could potentially interconvert
through a transient monocyclic hydroxy dialdehyde. We
envisaged that selective oxidation of one of the lactols
45 or 46 to a lactone might be possible, and this could
lead to just one oxidation product provided that lactols
45 and 46 readily interconvert (the lactol ratio showed a
significant solvent dependence, 45:46, 1:1 in CDCl3, 1:10
in DMSO-d6). However, Br2-mediated oxidation25 of lac-
tols 45 and 46 in a mixture of MeOH and H2O buffered
SCHEME 7. Synthesis of Triacid 43a
(22) Suh, Y.-G.; Jung, J.-K.; Seo, S.-Y.; Min, K.-H.; Shin, D.-Y.; Lee,
Y.-S.; Kim, S.-H.; Park, H.-J. J. Org. Chem. 2002, 67, 4127-4137.
(23) Goldberg, O.; Luini, A.; Teichberg, V. I. Tetrahedron Lett. 1980,
21, 2355-2358.
(24) Arakawa, Y.; Yasuda, M.; Ohnishi, M.; Yoshifuji, S. Chem.
Pharm. Bull. 1997, 45, 255-259.
a (a) RuO2‚H2O, NaIO4 (10% aq), EtOAc, 0 °C; (b) Jones’ reagent,
acetone, 20 °C; (c) HCO2H, 20 °C (77% from 40).
8870 J. Org. Chem., Vol. 70, No. 22, 2005

Stereocontrolled Syntheses of Kainoid Amino Acids
SCHEME 9. Synthesis of Acetals 52 and 53 from 2-Azabicycle 26a
a (a) O3/O2, Me2S, CH2Cl2, -78 °C (100%, crude yield); (b) (COCl)2, DMSO, NEt3, CH2Cl2, -78-20 °C (97%, crude yield); (c) K2CO3,
MeOH-CH2Cl2 (1:1), 20 °C (42% from 26); (d) Br2, NaHCO3, MeOH-H2O (9:1), 20 °C (24% from 26); (e) as part b, then AcCl, MeOH, 0
°C; (f) Ac2O, DMAP, NEt3, CH2Cl2, 20 °C (43% from 26); (g) Boc2O, DMAP, NEt3, CH2Cl2, 20 °C (55% from 26).
with NaHCO3 only gave ring opened diester 47, in 24%
yield from 2-azabicycle 26. Screening of a variety of other
oxidation conditions were studied, but all led to either
mixtures of lactones or ring-opened esters/acids. A break-
through was achieved when lactols 45 and 46 were
subjected to Swern oxidation, and lactone 48 was ob-
tained as a single product (97% crude yield, unstable to
chromatography). The selectivity in this oxidation is
thought to be due to either the larger proportion of lactol
46 present in solution assisted by excess DMSO employed
in the reaction and/or the higher reactivity of less-
hindered lactol 46 (the hydroxyl group in lactol 45 is
adjacent to a quaternary center).
In an effort to establish a bicyclic framework for an
eventual elaboration into enol ether 44, KOMe-catalyzed
methanolysis of lactone 48 was carried out. Unfortu-
nately, although this led to methanolysis of the lactone,
undesired deformylation was also observed to give alcohol
49 (42% from 26). The deformylation could occur by a
base-catalyzed retro-Claisen-type process. Indeed, the
latter is a common problem encountered in Julia-olefi-
nation chemistry when the formation of a trisubstituted
double bond is desired.26 An obvious way to try to
circumvent such a base-catalyzed deformylation process
was to carry out the desired methanolysis under acidic
conditions. Addition of anhydrous HCl (generated from
the addition of AcCl) in MeOH to the Swern oxidation
reaction resulted in the formation of a mixture of lactol
50 and hydroxy aldehyde 51 (50:51, 1.7:1 in CDCl3,
unstable to chromatography). This mixture could be
converged into a stable bicyclic acetate 52 or carbonate
53 (in 43 and 55% yields, respectively). The increase in
yield when using Boc2O as compared to Ac2O to selec-
tively trap and protect lactol 50 is probably due to
reprotection of any secondary amine generated during
the earlier acid-catalyzed methanolysis of lactone 48.
Initially, Julia-type elimination to obtain enol ether 44
was attempted on acetate 52 using standard phosphate-
buffered conditions with 5% Na amalgam in a mixture
of MeOH and THF. Unfortunately, this led to the
exclusive formation of alcohol 49 (Scheme 9), which is
thought to occur via NaOMe-induced deacetylation, fol-
lowed by deformylation as discussed previously. To
circumvent this problem, the Julia-type elimination was
carried out using SmI2 in a mixture of HMPA and THF.26
The latter (essentially neutral) conditions with acetate
52 led to formation of the desired enol ether 44, albeit in
low yield (28%, Scheme 10). Further experimentation
with sodium amalgam as the reductant was therefore
carried out. It was considered essential to decrease the
presence of NaOMe that may form in the reaction
mixture when employing sodium amalgam in MeOH. To
achieve this, B(OH)3 was used as the buffer to decrease
the pH of the solution in comparison to phosphate
buffers.22 Pleasingly, Julia-type elimination with sodium
amalgam and B(OH)3 led to the formation of enol ether
44 in an improved 48% yield. It was also noted at this
point that the yields obtained in these Julia-type reac-
tions were essentially unaffected by the proportion of Na
present (5-20%) in the commercially available amal-
gams, so the more cost-effective 20% Na amalgam was
utilized henceforward. The optimized Julia-type reaction
conditions (Na amalgam and B(OH)3) were then applied
SCHEME 10. Synthesis of Lactol 54a
(25) Collins, P.; Ferrier, R. Monosaccharides: Their Chemistry and
Their Roles in Natural Products; Wiley: Chichester, 1995; pp 126-
127.
(26) Marko, I. E.; Murphy, F.; Kumps, L.; Ates, A.; Touillaux, R.;
Craig, D.; Carballares, S.; Dolan, S. Tetrahedron 2001, 57, 2609-2619.
a (a) SmI2, HMPA, THF (28%); (b) 5% Na-Hg, B(OH)3, MeOH-
THF (1:1), 0-20 °C (48%); (c) 20% Na-Hg, B(OH)3, MeOH-THF
(1:1), 0-20 °C (74%); (d) HCl, THF (73%); (e) 20% Na-Hg, B(OH)3,
MeOH-THF (1:1), 0-20 °C, then HCl (0.4 M, aq) (42%).
J. Org. Chem, Vol. 70, No. 22, 2005 8871

Hodgson et al.
to carbonate 53, and this led to the formation of enol
ether 44 in good yield (74%).
carbon would undergo (anti-Markovnikov) nucleophilic
attack by water, leading to aldehyde formation.29
Acid-catalyzed hydration of enol ether 44 gave lactol
54 (73% yield), with the desired cis-configuration between
C-3 and C-4. Furthermore, this acid-catalyzed hydration
of enol ether 44 could be effected by prolonging exposure
to the acidic workup at the end of the Julia-type elimina-
tion to directly furnish lactol 54 from Boc-acetal 53, albeit
in modest overall yield (42%). The stereochemistry of the
ring fusion was assigned cis by correlation with the
known lactone acid,27 following lactol to lactone oxidation
(TPAP/NMO) and methyl ester hydrolysis (aqueous
KOH) and, ultimately, by conversion of lactol 54 to
R-kainic acid 12.
Initially, attempts to open lactol 54 did not give
satisfactory results: direct methylation using MeMgBr
and MeTi(i-PrO)3 left the lactol moiety intact, and
attempted silyl protection of the open aldehyde-alcohol
tautomer employing TESCl, TESOTf, TBDMSCl, or TB-
DMSOTf led to ring-closed silyl-acetals. Opening of the
lactol ring was finally accomplished via a Wittig olefi-
nation employing PPh3MeBr deprotonated with KHMDS
in PhMe, to yield alkene 55 (72%, Scheme 11). Alcohol
55 was converted into diester 57 in two steps via Jones’
oxidation to the acid 56, followed by esterification using
TMSCHN2. It was initially conceived that diester 57
could be directly converted into ketone 60 by Wacker
oxidation. Terminal olefins generally yield ketones on
Wacker oxidation,28 but unfortunately in the present
case, aldehyde formation was observed as the major
reaction pathway. The reversal of the usual selectivity
in the Wacker oxidation could be due to coordination of
the palladium catalyst to the ester at C-3 in diester 57.
This might then cause the palladium associated to the
double bond to lie closer to the internal olefinic carbon,
leading to a higher concentration of positive charge
residing on the terminal carbon. Hence, the terminal
Because of the failure of the Wacker oxidation to give
the desired ketone 60, an alternative tactic was exam-
ined. Iodolactonization of acid 56 gave iodolactone 58.30
This lactone proved unstable to silica gel column chro-
matography, so it was used in crude form in subsequent
steps. Nevertheless, an analytical sample of the major
iodolactone diastereoisomer could be obtained via recrys-
tallization from CH2Cl2 and Et2O. Treatment of crude
iodolactone 58 with a mild base (NMe4F) in PhH at reflux
led to elimination of HI to form enol-lactone 59, in 51%
yield over three steps from alcohol 55. The use of other
more commonly used bases, such as DBU or the Am-
berlyst A-26 (polymer-supported trimethylammonium) F-
form,31 gave only decomposition products. Mild metha-
nolysis of enol-lactone 59 using Et3N in MeOH led to the
known ketone 6032 in 91% yield. Olefination of ketone
60 to give alkene 61 was carried out using a nonbasic
Nozaki reagent. Epimerization at C-4 of ketone 60 is
known to be facile, and the employment of a basic Wittig
reagent would likely have led to epimerization at this
center.33 The successful generation of alkene 61 consti-
tutes a formal synthesis of R-kainic acid 12,34 but
deprotection was also carried out in our hands to achieve
a total synthesis of R-kainic acid 12, in 80% yield. The
spectroscopic data for our synthetic R-kainic acid 12 was
in accord with literature values, and co-mixing of syn-
thetic and authentic samples of R-kainic acid 12 led to a
1
homogeneous H NMR spectrum.
To further develop the present strategy toward other
kainoids, specifically R-isokainic acid 14, we considered
utilizing a decarboxylative Ramberg-Ba¨cklund reaction
(RBR)35 on sulfone 63 to access the isopropylidene unit
at C-4 in a direct manner (Scheme 12). We envisaged that
SCHEME 12. Synthetic Strategy to r-Isokainic
Acid 14
SCHEME 11. Synthesis of r-Kainic Acid 12a
(27) Baldwin, J. E.; Turner, S. C. M.; Moloney, M. G. Synlett 1994,
925-928.
(28) Tsuji, J. Synthesis 1984, 369-384.
(29) (a) Pellissier, H.; Michellys, P.-Y.; Santelli, M. Tetrahedron
1997, 53, 7577-7586. (b) Pellissier, H.; Michellys, P.-Y.; Santelli, M.
Tetrahedron 1997, 53, 10733-10742.
(30) Goldberg, O.; Luini, A.; Teichberg, V. I. J. Med. Chem. 1983,
26, 39-42.
(31) Cardillo, G.; Orena, M.; Porzi, G.; Sandri, S.; Tomasini, C. J.
Org. Chem. 1984, 49, 701-703.
(32) (a) Baldwin, J. E.; Fryer, A. M.; Pritchard, G. J. J. Org. Chem.
2001, 66, 2588-2596. (b) Ahern, D. G.; Seguin, R. J.; Filer, C. N. J.
Labeled Compd. Radiopharm. 2002, 45, 401-405.
(33) Monn, J. A.; Valli, M. J. J. Org. Chem. 1994, 59, 2773-2778.
(34) Hanessian, S.; Ninkovic, S. J. Org. Chem. 1996, 61, 5418-5424.
(a) PPh3MeBr, KHMDS, PhMe, THF, 20 °C (72%); (b) Jones’
reagent, acetone, 20 °C; (c) TMSCHN2, hexane-MeOH (7:2), 20
°C (81% from 55); (d) I2, NaHCO3, MeCN, 0 °C; (e) NMe4F, PhH,
reflux (51% from 55); (f) NEt3, MeOH, -78 °C (91%); (g) Zn, CH2I2,
TiCl4, CH2Cl2, THF, 20 °C (66%); (h) KOH, H2O, THF, 20 °C; (i)
TFA, CH2Cl2, 20 °C (80% from 61).
8872 J. Org. Chem., Vol. 70, No. 22, 2005

Stereocontrolled Syntheses of Kainoid Amino Acids
sulfone 63 could be generated from oxidative cleavage of
alkene 62, which could potentially be obtained from the
application of a RA-HRR reaction on an appropriate
7-azabicycle.
fact that in CDCl3, the two lactols are present in a 1:1
mixture, whereas in DMSO-d6, only lactol 71 was present,
1
as determined by H NMR spectroscopy. Subjection of
lactols 70 and 71 to oxidation with Br2 buffered by
NaHCO3 in MeOH-H2O led to the formation of mo-
noester 72. Thus, decarboxylation (or deformylation) at
C-4 had occurred spontaneously under the reaction
conditions. This result contrasts with the bromine oxida-
tion of lactols 45 and 46, which led to the formation of
diester 47 (Scheme 9). The difference in these Br2
mediated oxidations is likely due to the more electron-
withdrawing nature of the R-chlorosulfonyl functionality,
which strongly enhances the vicinal decarboxylation (or
deformylation) relative to the tosyl functionality. De-
formylation for the latter required more basic conditions
(K2CO3, MeOH) for the formation of monoester 49 from
lactone 48 (Scheme 9). Pleasingly, treatment of sulfone
72 with t-BuOK at -78 °C led to RBR to produce alkene
73 in 45% yield. Alkene 73 could potentially be trans-
formed into R-isokainic acid 14 after an oxidation, hy-
drolysis, and Boc-deprotection sequence, but it was
decided to carry out further work on the earlier bromine
oxidation step. It was envisaged that simultaneous
oxidation of the primary alcohol (masked as lactol) and
aldehydes in 70/71 would considerably shorten the
synthesis. Furthermore, if the oxidation could be carried
out in basic media, then the RBR might also occur
concurrently. This global oxidation and concurrent RBR
was realized when lactols 70 and 71 were treated with
4-AcNH-TEMPO and Br2 under basic conditions (aq
KOH)38 to give diacid 65. Under these conditions, heating
was necessary for the RBR to proceed at a useful rate.
To facilitate purification, the resulting diacid 65 was
esterified using TMSCHN2 to form protected R-isokainic
acid 74 in 72% yield from 2-azabicycle 69. R-Isokainic
acid 14 was finally obtained after hydrolysis of the methyl
esters (THF and aq KOH), and deprotection of the Boc
group (TFA in CH2Cl2),34 in 81% yield over two steps.
Spectroscopic (1H and 13C NMR) data for R-isokainic
acid 14, required to ratify the present synthesis of
R-isokainic acid 14, do not exist in the literature. There-
fore, our sample of (()-R-isokainic acid 14 was compared
to material obtained from treatment of (-)-R-kainic acid
12 with refluxing aqueous HBr (20%) for 3 h, which is
known to give R-kainic acid δ-lactone 75 and R-isokainic
acid 14 (Scheme 15).39 The mixture of R-kainic acid
δ-lactone 75 and R-isokainic acid 14 were purified with
Dowex-50H+ and then with Amberlite CG-50, and pleas-
ingly, the 1H NMR spectrum of this mixture showed
peaks with identical chemical shifts and coupling con-
stants to synthetic (()-R-isokainic acid 14.
To examine the strategy outlined in Scheme 12,
7-azabicycle 67 was prepared by [4 + 2] cycloaddition of
N-Boc pyrrole and 2-(ethynylsulfonyl)propane 66 (63%,
Scheme 13). 2-(Ethynylsulfonyl)propane 66 has previ-
ously been prepared in five steps from acetylene, but it
was more conveniently made by following a one-pot
protocol for forming S-alkyl ynethiol ethers from trichlo-
roethylene, followed by oxidation with peracetic acid.36
This chemistry uses i-PrSK to effect MeOK-catalyzed
elimination from trichloroethene to give dichloroacety-
lene, which undergoes i-PrSK-catalyzed addition of i-
PrSH to give the dichloro enethiol ether. n-BuLi (2.2
equiv) is then directly added to generate the lithium
acetylide of 2-(ethynylthio)propane, which, on addition
of MeOH, gives 2-(ethynylthio)propane. Following an
aqueous wash, the resulting solution was treated directly
with peracetic acid to give 2-(ethynylsulfonyl)propane 66
(53% from i-PrSH), thus avoiding isolation of the inter-
mediate acetylenic thiol. Pleasingly, treatment of 7-azab-
icycle 67 with 2-iodoethanol, Bu3SnH, and Et3B/O2 led
to RA-HRR to give a mixture of 2-azabicycles 64 and 68
in good yield (82%). The radical addition was exo-
selective, favoring 2-azabicycle 64 (64:68, 13:1). It is
interesting to note that the only minor product obtained
from this reaction was 2-azabicycle 68, the C-7 epimer
of 64, whereas the minor product observed from RA-HRR
reaction of 7-azabicycle 16 was the unrearranged 7-azab-
icycle 25 (Table 2). 2-Azabicycles 64 and 68 are chro-
matographically separable, but it was more convenient
for large-scale preparative purposes to carry the mixture
on to the next step. Chlorination of 2-azabicycle 64 to
give R-chlorosulfone 69 was achieved by deprotonation
R- to the sulfone with n-BuLi, followed by trapping of the
intermediate R-lithiosulfone with C2Cl6.37 When this
reaction is carried out on a large scale with a mixture of
2-azabicycles 64 and 68, chloride 69 is conveniently
isolated by recrystallization.
Ozonolysis of 2-azabicycle 69 with a reductive workup
(Me2S) led cleanly to the formation of lactols 70 and 71
(Scheme 14), which are likely in equilibrium (cf. lactols
45 and 46, Scheme 9). The latter was supported by the
SCHEME 13. Synthesis of 2-Azabicycle 69a
Additional evidence was sought to further support the
synthesis of R-isokainic acid 14, which was via the
(35) (a) Wladislaw, B.; Marzorati, L.; Russo, V. F. T.; Zim, M. H.;
Di Vitta, C. Tetrahedron Lett. 1995, 36, 8367-8370. (b) Taylor, R. J.
K.; Casy, G. Org. React. 2003, 62, 357-475.
(36) (a) Laba, V. I.; Polievktov, M. K.; Prilezhaeva, E. N.; Maira-
novskii, S. G. Izv. Akad. Nauk., Ser. Khim. 1969, 2149-2156; ibid.
Chem. Abstr. 1970, 72, 54654. (b) Nebois, P.; Kann, N.; Greene, A. E.
J. Org. Chem. 1995, 60, 7690-7692.
(37) Vacher, B.; Samat, A.; Allouche, A.; Laknifli, A.; Baldy, A.;
Chanon, M. Tetrahedron 1988, 44, 2925-2932.
a (a) KH, THF, 20 °C, then Cl2CdCHCl, cat. MeOH, -50-20
°C, then n-BuLi, -70 to -40 °C; (b) MeCO3H, MeOH, THF 0-20
°C (53% from i-PrSH); (c) N-Boc pyrrole, 85 °C (63%); (d)
HO(CH2)2I, Bu3SnH, Et3B, dry air (O2), CH2Cl2, 20 °C (82%); (e)
n-BuLi, C2Cl6, THF, -78 to 20 °C (86%).
(38) Merbouh, N.; Bobbitt, J. M.; Bru¨ckner, C. J. Carbohydr. Chem.
2002, 21, 65-77.
(39) Morimoto, H. Yakugaku Zasshi 1955, 75, 916-919; ibid. Chem.
Abstr. 1956, 50, 24089.
J. Org. Chem, Vol. 70, No. 22, 2005 8873

Hodgson et al.
SCHEME 14. Synthesis of r-Isokainic Acid 14a
(a) O3/O2, then Me2S, CH2Cl2, -78 to 20 °C (100%, crude yield); (b) Br2, NaHCO3, MeOH-H2O (9:1), 20 °C (61% from 69); (c) t-BuOK,
THF, -78 °C (45%, 10% recovered 72); (d) 4-AcNH-TEMPO, Br2, KOH (added to maintain pH at 11.5), Na2S2O5, MeCN/H2O (1:1), 0-95
°C; (e) TMSCHN2, PhMe/MeOH (7:2), 20 °C (72% from 69); (f) KOH, H2O, THF, 20 °C; (g) TFA, CH2Cl2, 20 °C (81% from 74).
SCHEME 15. Synthesis of r-Kainic Acid δ-Lactone
75 and r-Isokainic Acid 14
Experimental Procedures
General experimental details are described in the Support-
ing Information.
Synthesis of r-Isokainic Acid 14. 2-(Ethynylsulfonyl)-
propane 66. 2-Propanethiol (6.97 mL, 75.0 mmol) in THF (100
mL) was added dropwise to a suspension of KH (15.1 g, 30%
in mineral oil, 113 mmol) in THF (100 mL) over 10 min and
allowed to stir for 17 h. The reaction mixture was cooled to
-50 °C, and trichloroethylene (7.50 mL, 83.5 mmol) in THF
(75 mL) was added dropwise over 5 min, followed by MeOH
(0.2 mL, 4.9 mmol). The reaction mixture was warmed to 20
°C and stirred until gas evolution was complete. After 4 h,
the reaction mixture was cooled to -70 °C, and n-BuLi (138
mL, 1.2 M in hexane, 166 mmol) was added dropwise over 15
min. After 30 min at -70 °C, the mixture was warmed to -40
°C over 30 min and then treated dropwise with MeOH (9 mL)
over 10 min. The mixture was allowed to warm to 20 °C, and
after 10 min, further MeOH (15 mL) was added. The reaction
mixture was washed with saturated aqueous NH4Cl (300 mL)
and H2O (4 × 100 mL). Peracetic acid (37.2 mL, 36-40% in
acetic acid) was added to the organic layer at 0 °C, and the
mixture was allowed to warm to 20 °C. After 15 h, the organic
phase was separated from the acidic phase and washed with
saturated aqueous NaHCO3 (4 × 300 mL) and H2O (3 × 100
mL). The organic layer was dried and concentrated in vacuo.
The residue was purified by column chromatography (30%
Et2O-petroleum ether) to give ethynyl sulfone 6636a as a
colorless liquid (5.25 g, 53%); Rf (50% Et2O-petroleum ether)
0.3; νmax/cm-1 3236, 2984, 2940, 2067, 1467, 1332, 1262, 1132;
δH (400 MHz, CDCl3) 3.44 (1H, s), 3.25 (1H, sept, J ) 6.9 Hz),
synthesis of R-dihydroallokainic acid 77 (Scheme 16).
Catalytic hydrogenation of protected R-isokainic acid 74
using Adams’ catalyst gave protected R-dihydroallokainic
acid 76 in 98% yield. Hydrogenation occurred highly
stereoselectively, without observation of protected-R-
dihydrokainic acid (C-4 epimer). Hydrolysis (aqueous
KOH) followed by Boc deprotection (TFA in CH2Cl2) of
protected-R-dihydroallokainic acid 76 gave R-dihydro-
allokainic acid 7740 in 91% yield. This latter result
provided further supporting evidence for our synthesis
of R-isokainic acid 14 (as all possible diastereomers of
dihydrokainic acid are known)41 and establishes a diver-
gent approach to R-dihydroallokainic acid 77.
SCHEME 16. Synthesis of r-Dihydroallokainic
Acid 77a
1.48 (6H, d, J ) 6.9 Hz); δC (100 MHz, CDCl3) 81.6, 76.5, 57.6,
15.6; m/z (CI) 150 (M + NH4+, 100%) (Found: M + NH4
150.0588. C5H12NO2S requires M, 150.0589).
,
+
N-Boc 2-[(1-Methylethyl)sulfonyl]-7-azabicyclo[2.2.1]-
heptadiene 67. Ethynyl sulfone 66 (1.32 g, 9.99 mmol) and
N-t-butoxycarbonylpyrrole (8.51 mL, 49.9 mmol) were mixed
neat at 85 °C under Ar and protected from light. After 27 h,
the reaction mixture was cooled to 20 °C and subjected directly
to purification by column chromatography (gradient elution:
10-40% Et2O-petroleum ether) to give 7-azabicycle 67 as a
yellow solid (1.88 g, 63%); Rf (50% Et2O-petroleum ether) 0.2;
mp 65-66 °C; νmax/cm-1 2978, 2935, 1710, 1369, 1338, 1309,
1257, 1163, 1136; δH(250 MHz, DMSO-d6, 90 °C) 7.87-7.84
(1H, m), 7.27-7.22 (1H, m), 7.12-7.08 (1H, m), 5.39-5.36 (1H,
m), 5.31-5.28 (1H, m), 3.29 (1H, sept, J ) 6.8 Hz), 1.39 (9H,
s), 1.25-1.20 (6H, m); δC (63 MHz, DMSO-d6, 90 °C) 156.3,
a (a) PtO2, H2 (1 atm), MeOH, 20 °C (98%); (b) KOH, H2O, THF,
20 °C; (c) TFA, CH2Cl2, 20 °C (91% from 76).
In summary, a tandem intermolecular radical addi-
tion-homoallylic radical rearrangement reaction has
been developed to access 2-azabicyclo[2.2.1]hept-5-enes
that are not readily accessible by other means. These
2-azabicycles can be used as templates to form 2,3,4-
trisubstituted pyrrolidines with excellent stereocontrol.
This methodology has been exemplified in racemic syn-
theses of a known biologically active kainoid analogue
43, R-kainic acid 12, R-isokainic acid 14, and R-dihy-
droallokainic acid 77.
(40) Sugawa, T. Yakugaku Zasshi 1957, 77, 332-333; ibid. Chem.
Abstr. 1957, 51, 62288.
(41) Hashimoto, K.; Konno, K.; Shirahama, H. J. Org. Chem. 1996,
61, 4685-4692.
8874 J. Org. Chem., Vol. 70, No. 22, 2005

Stereocontrolled Syntheses of Kainoid Amino Acids
155.2, 152.8, 143.0, 142.0, 80.3, 67.7, 66.6, 52.5, 27.4, 14.4, 14.0;
m/z (CI) 317 (M + NH4+, 100%) (Found: M + NH4+, 317.1533.
C14H25N2O4S requires M, 317.1535).
N-Boc 3-(Carboxymethyl)-4-(i-propylidene)-2-pyrro-
lidinecarboxylic Acid Dimethyl Ester 74. 2-Azabicycle 69
(270 mg, 0.711 mmol) was dissolved in CH2Cl2 (200 mL), and
the temperature was lowered to -78 °C. A mixture of O3/O2
was bubbled through the solution until it turned blue, followed
by O2 until the solution became colorless. Then, Ar was passed
through the solution for 5 min, and Me2S (5 mL, 68 mmol)
was added dropwise. The mixture was warmed to 20 °C and
stirred for 1 h. The mixture was then concentrated in vacuo,
and the residue was dissolved in Et2O (50 mL, the dissolution
aided by a small amount of CH2Cl2). This solution was washed
with H2O (2 × 30 mL), dried (MgSO4), and concentrated in
vacuo to give a crude mixture of lactols 70 and 71 as a white
foam. This crude mixture was dissolved in MeCN/H2O (17 mL,
1:1 MeCN/H2O), and the temperature was lowered to 0 °C.
4-Acetamido-2,2,6,6-tetramethylpiperidine-1-oxyl (14 mg, 0.066
mmol) was added, followed by aqueous KOH (7.5 M) until the
pH was 11.5. A mixture of Br2 (0.21 mL, 4.10 mmol) and MeCN
(6 mL) was added dropwise over 30 min, together with aqueous
KOH (7.5 M) to maintain the pH at 11.5. Residual Br2 was
washed into the reaction using MeCN/H2O (6 mL, 1:1 MeCN/
H2O), and the reaction was further stirred at pH 11.5 and 0
°C for 2 h. Na2S2O5 (1.35 g, 7.10 mmol) was then added, and
the mixture was stirred vigorously until it turned colorless.
Aqueous KOH (6 mL, 7.5 M, 45.0 mmol) was now added, and
the temperature was raised to 95 °C for 2 h. The mixture was
cooled to 20 °C and acidified to pH 1 with aqueous HCl (2.0
M). The aqueous layer was extracted with EtOAc (4 × 20 mL),
and the combined organic extracts were washed with brine
(100 mL), dried (MgSO4), and concentrated in vacuo. The
residue was dissolved in PhMe/MeOH (18 mL, 7:2 PhMe/
MeOH), (trimethylsilyl)diazomethane (5 mL, 2.0 M in hexane,
10.0 mmol) was added, and the mixture was stirred at 20 °C
for 16 h. AcOH (0.6 mL) was then added, and the mixture was
concentrated in vacuo. Purification of the residue by column
chromatography (25% Et2O-pentane) gave protected R-iso-
kainic acid 74 (174 mg, 72%) as a colorless oil; Rf (25% Et2O-
pentane) 0.3; νmax/cm-1 2976, 1742, 1706, 1394, 1256, 1175,
1104, 1108, 897; δH (400 MHz, CDCl3) 4.33 and 4.23 (1H, s),
4.09-3.85 (2H, m), 3.67 and 3.64 (3H, s), 3.67 and 3.64 (3H,
s), 3.27 (1H, dd, J ) 9.4, 5.5 Hz), 2.45-2.32 (2H, m), 1.60 (3H,
s), 1.56 (3H, s), 1.43 and 1.37 (9H, s); δC (100 MHz, CDCl3)
172.5 and 172.3, 172.0 and 171.8, 155.1 and 154.6, 130.0 and
129.1, 126.4 and 126.3, 80.2 and 80.2, 64.4 and 63.9, 52.4 and
52.3, 51.9 and 51.9, 48.6, 42.7 and 41.9, 38.5, 28.5 and 28.3,
21.1 and 21.1, 20.5 and 20.5; m/z (ES) 705 (2M + Na+, 14%),
364 (M + Na+, 100%), 308 (22%), 242 (M - Boc + H+, 39%)
(Found: M + H+, 342.1927. C17H28NO6 requires M, 342.1917).
r-Isokainic Acid 14. Protected R-isokainic acid 74 (51 mg,
0.15 mmol) was dissolved in a mixture of THF (1.2 mL) and
aqueous KOH (4.6 mL, 2.5%). The reaction mixture was stirred
for 15 h, and then aqueous HCl (2.0 M) was added until pH
∼2. The aqueous layer was extracted with EtOAc (4 × 10 mL),
dried, and concentrated in vacuo. The residue was redissolved
in CH2Cl2 (10 mL), TFA (0.5 mL, 6.5 mmol) was added
dropwise, and the reaction mixture was stirred for 28 h. The
mixture was then concentrated in vacuo, and the residue was
purified by column chromatography (Dowex-50H+, WX8-200,
8% cross linking, 100-200 wet mesh), eluting with NH4OH
(2.0 M). The resulting solution was concentrated in vacuo,
redissolved in H2O, and stirred with Amberlite CG-50 (100-
200 dry mesh) for 1 h, and then filtered. The filtrate was
concentrated in vacuo to give R-isokainic acid 1442 (26 mg, 81%)
as a white solid; mp 240-245 °C (dec) [Lit.42 237-240 °C (dec)];
νmax/cm-1 2927, 2361, 2342, 1564, 1557, 1373, 1297, 1043, 859;
δH (250 MHz, D2O) 4.15 (1H, d, J ) 1.2 Hz), 4.06 (1H, d, J )
14.9 Hz), 3.94 (1H, d, J ) 14.9 Hz), 3.55 (1H, td, J ) 6.8, 1.2
Hz), 2.54 (1H, dd, J ) 14.8, 6.8 Hz), 2.46 (1H, dd, J ) 14.8,
6.8 Hz), 1.71 (3H, s), 1.62 (3H, s); δC(126 MHz, D2O) 178.1,
174.1, 131.0, 126.0, 66.7, 47.4, 42.2, 39.8, 21.2, 20.9; m/z (ES)
N-Boc syn-7-(2-Hydroxyethyl)-4-[(1-methylethyl)sul-
fonyl]-2-azabicyclo[2.2.1]hept-5-ene 64 and N-Boc anti-
7-(2-Hydroxyethyl)-4-[(1-methylethyl)sulfonyl]-2-aza-
bicyclo[2.2.1]hept-5-ene 68. 7-Azabicycle 67 (956 mg, 3.19
mmol) was dissolved in CH2Cl2 (320 mL), and Ar was bubbled
through the solution for 15 min. 2-Iodoethanol (1.50 mL, 19.2
mmol) and Et3B (1.91 mL, 1.0 M in hexane, 1.91 mmol) were
added to the degassed solution. Bu3SnH (5.16 mL, 19.2 mmol)
was added via a syringe pump over 100 min, and dry air (10
mL portions) was injected into the reaction mixture in parallel
at 5 min intervals. The reaction mixture was then washed with
saturated aqueous NaH2PO4 (400 mL), and the aqueous layer
was extracted with CH2Cl2 (3 × 200 mL). The organic extracts
were combined and dried and then concentrated in vacuo.
Purification of the resulting residue by column chromatogra-
phy (Et2O) gave a mixture of azabicycles 64 and 68 (906 mg,
82%) as colorless oil. Further chromatographic purification
(Et2O) led to the isolation of 64 (842 mg, 76%) and 68 (58 mg,
5%) as colorless oils. syn-64: Rf (Et2O) 0.1; νmax/cm-1 3468,
2977, 2936, 2359, 1338, 1689, 1393, 1368, 1303, 1259, 1161,
1119, 1051; δH (500 MHz, DMSO-d6, 90 °C) 6.61-6.57 (2H,
m), 4.56 (1H, s), 4.25 (1H, s), 3.66 (1H, d, J ) 9.2 Hz), 3.51-
3.40 (3H, m), 2.85 (1H, d, J ) 9.2 Hz), 2.31 (1H, dd, J 3.1, J )
9.8 Hz), 1.90-1.84 (1H, m), 1.42 and 1.41 (9H, s), 1.42-1.26
(1H, m), 1.32 and 1.32 (3H, d, J ) 6.8 Hz), 1.31 and 1.31 (3H,
d, J ) 6.8 Hz); δC(125 MHz, DMSO-d6, 90 °C) 155.9, 137.0,
136.9, 80.4, 74.8, 63.9, 59.8, 59.5, 51.8, 43.7, 28.4, 27.6, 15.3,
14.1; m/z (CI) 363 (25%), 346 (M + H+, 100%) (Found: M +
H+, 346.1687. C16H28NO5S requires M, 346.1688). anti-68: Rf
(Et2O) 0.2; νmax/cm-1 3469, 2976, 2361, 2342, 1696, 1394, 1368,
1302, 1163, 1133, 1052, 872; δH (500 MHz, DMSO-d6, 90 °C)
6.49 (1H, dd, J ) 5.8, 2.4 Hz), 6.41 (1H, d, J 5.8), 4.66 (1H, s),
4.18 (1H, s), 3.64 (1H, d, J ) 8.9 Hz), 3.48-3.39 (3H, m), 2.92
(1H, d, J ) 8.9 Hz), 2.68-2.65 (1H, m), 1.83-1.77 (1H, m),
1.43-1.36 (1H, m), 1.41 (9H, s), 1.34 and 1.34 (3H, d, J ) 6.7
Hz), 1.32 and 1.32 (3H, d, J ) 6.7 Hz); δC (125 MHz, DMSO-
d6, 90 °C) 153.8, 133.7, 130.2, 78.9, 76.0, 62.5, 59.5, 59.0, 52.2,
48.5, 27.8, 27.6, 15.5, 14.4; m/z (CI) 713 (2M + Na+, 100%),
368 (M + H+, 50%) (Found: M + H+, 346.1684. C16H28NO5S
requires M, 346.1688).
N-Boc 4-[(1-Chloro-1-methylethyl)sulfonyl]-7-(2-hy-
droxyethyl)-2-azabicyclo[2.2.1]hept-5-ene 69. 2-Azabicycle
64 (829 mg, 2.40 mmol) was dissolved in THF (20 mL), and
the temperature was lowered to -78 °C. n-BuLi (3.52 mL, 1.5
M in hexane, 5.28 mmol) was added dropwise over 10 min,
and then C2Cl6 (1.70 g, 7.20 mmol) dissolved in THF (15 mL)
and precooled to -78 °C was added over 15 min by cannula.
The resulting mixture was stirred for 75 min, warmed to 20
°C, and stirred for a further 50 min. The mixture was then
diluted with Et2O (200 mL) and washed with aqueous HNO3
(200 mL, 0.10 M). The aqueous layer was extracted with Et2O
(3 × 100 mL), and the combined organic extracts were dried
and concentrated in vacuo. The residue was purified by column
chromatography (Et2O) to give chloride 69 (784 mg, 86%) as a
white solid. Recrystallization from Et2O/CH2Cl2 gave an
analytically pure sample (recrystallization at this point is
recommended for large scale preparation, as this makes the
separation of azabicycles 64 and 68 unnecessary in the
previous step); Rf (Et2O) 0.3; mp 115-116 °C; νmax/cm-1 3446,
2977, 2935, 1698, 1393, 1368, 1310, 1167, 1108; δH (500 MHz,
DMSO-d6, 90 °C) 6.70 (1H, d, J ) 6.1 Hz), 6.60 (1H, dd, J )
6.1, 2.7 Hz), 4.58 (1H, s), 3.76 (1H, d, J ) 9.2 Hz), 3.52-3.40
(2H, m), 2.98 (1H, d, J ) 9.2 Hz), 2.49-2.48 (1H, m), 2.02-
1.95 (1H, m), 1.97 (6H, s), 1.42 (9H, s), 1.34-1.26 (1H, m); δC
(125 MHz, DMSO-d6, 90 °C) 154.4, 135.6, 134.9, 83.5, 79.1,
74.7, 61.9, 59.1, 58.4, 44.8, 28.4, 27.6, 26.7, 26.4; m/z (ES) 675
(68%), 402 (M + Na+, 55%), 380 (M + H+, 100%) (Found: M
+ H+, 380.1293. C16H2735ClNO5S requires M, 380.1303).
(42) Honjo, M. Yakugaku Zasshi 1957, 77, 598-603. ibid. Chem.
Abstr. 1957, 51, 90648.
J. Org. Chem, Vol. 70, No. 22, 2005 8875

Hodgson et al.
212 (M - H+, 100%), 168 (35%) (Found: M + H+, 214.1074.
C10H16NO4 requires M, 214.1079).
Supporting Information Available: Full experimental
details of syntheses and characterization of compounds not
1
described in the Experimental Procedures. Copies of H and
13C spectra for all compounds. This material is available free
of charge via the Internet at http://pubs.acs.org.
Acknowledgment. We thank the EPSRC and Pfizer
for a CASE award (S.H.). We also thank the EPSRC
National Mass Spectrometry Service Centre (Swansea)
for mass spectra.
JO0513865
8876 J. Org. Chem., Vol. 70, No. 22, 2005