Synthesis of nitrogen bridgehead bicyclic heterocycles via ring-closure of β-ammonio 5-hexenyl radicals

Article abstract of DOI:10.1021/jo960886i

The 2-(3-methylenepiperidinyl)ethyl radical (6) displays considerable reluctance to ring-closure under conditions which its carbocyclic analog, the 2-(3-methylenecyclohexyl)ethyl radical (2), cyclizes essentially completely. Molecular mechanics calculations suggest that the increased activation barrier associated with ring-closure of 6 is the result of a higher than expected transition state energy. A study of the behavior of β-ammonio-substituted 5-hexenyl radicals, such as the 3,3-dimethyl-3-azonia-5-hexenyl radical (22), reveals that cyclization occurs readily. Treatment of 1-methyl-1-(2-(phenylselenyl)ethyl)-3-methylenepiperidinium iodide (20) with tributyltin hydride in tert-amyl alcohol yields the bridgehead nitrogen bicyclic heterocycle, 1,5-dimethyl-1-azoniabicyclo-[3.2.1]octane iodide (26), in excellent yield and without contamination, thus providing an attractive synthetic route to this hitherto unknown heterocyclic system.

Full text of DOI:10.1021/jo960886i

J . Org. Chem. 1996, 61, 7529-7533
7529
Syn th esis of Nitr ogen Br id geh ea d Bicyclic Heter ocycles via
Rin g-Closu r e of â-Am m on io 5-Hexen yl Ra d ica ls
Ernest W. Della* and Andrew M. Knill
Department of Chemistry, Flinders University, Bedford Park, South Australia 5042
Received May 14, 1996X
The 2-(3-methylenepiperidinyl)ethyl radical (6) displays considerable reluctance to ring-closure
under conditions which its carbocyclic analog, the 2-(3-methylenecyclohexyl)ethyl radical (2), cyclizes
essentially completely. Molecular mechanics calculations suggest that the increased activation
barrier associated with ring-closure of 6 is the result of a higher than expected transition state
energy. A study of the behavior of â-ammonio-substituted 5-hexenyl radicals, such as the 3,3-
dimethyl-3-azonia-5-hexenyl radical (22), reveals that cyclization occurs readily. Treatment of
1-methyl-1-(2-(phenylselenyl)ethyl)-3-methylenepiperidinium iodide (20) with tributyltin hydride
in tert-amyl alcohol yields the bridgehead nitrogen bicyclic heterocycle, 1,5-dimethyl-1-azoniabicyclo-
[3.2.1]octane iodide (26), in excellent yield and without contamination, thus providing an attractive
synthetic route to this hitherto unknown heterocyclic system.
In tr od u ction
decided from the outset, therefore, to commence with an
examination of the 2-(3-methylenepiperidinyl)ethyl radi-
cal (6), the aza analog of the 2-(3-methylenecyclohexyl)-
ethyl radical (2). As a member of the â-amino radical
family, 6 was not expected to possess the same kind of
stabilization displayed by 5 and thus be more amenable
toward cyclization. In fact, it has been suggested6 that
3-aza-5-hexenyl radicals are better suited to radical
cyclization than the parent species.
We considered it essential to conduct a preliminary
investigation of the behavior of the parent open-chain
species, the 3-methyl-3-aza-5-hexenyl radical (8). It was
found7 that 8, produced by treatment of the selenide 7
with Bu3SnH, undergoes ring-closure with complete
regioselectivity to give 1,3-dimethylpyrrolidine (10) via
the exo-trig product 9; significantly, the rate of cyclization
is faster than the corresponding reaction of the 5-hexenyl
radical. This result was encouraging for the planned
extension of the investigation to a study of the 2-(3-
methylenepiperidinyl)ethyl radical 6. We now disclose
the results of our work on this system.
As part of a program of studies directed toward the
synthesis of bridgehead-substituted bicycloalkanes, we
recently reported1-3 that the modified 5-hexenyl radicals
1 and 2 undergo facile ring-closure to the isomeric species
3 and 4 and that such cyclizations give very good yields
of the bicycles with useful functionality at the bridgehead
positions. One of the long-range objectives of this work
was to determine whether this kind of transformation
could be usefully employed for the synthesis of hetero-
cyclic compounds, in particular bicyclic systems with
nitrogen at the bridgehead.
Resu lts a n d Discu ssion
We were precluded from employing the popular pre-
cursor for generating radicals, the organobromide or
iodide, because of the lability of the required â-amino
halide toward internal displacement of the halogen;
accordingly we elected to use the phenylselenide 11
instead. The synthesis of 11 was accomplished (Scheme
1) using commercially-available 3-(hydroxymethyl)pip-
eridine (12) as starting material. All the steps depicted
in Scheme 1 represent essentially standard procedures;
a noteworthy feature is that optimum yields of the amide
13 were obtained using the exchange process illustrated8
rather than by selective acylation using acetic anhydride,
for example, because the latter was found to give a
product that was difficult to purify . The final step is a
modification of the process of reductive amination de-
scribed by Danishefsky and Panek.9
Previous observations by Padwa and his colleagues4
that impinge directly on this work reveal that R-amino
radicals such as the 2-benzyl-2-aza-5-hexenyl radical (5)
are reluctant to cyclize, presumably as a result of the
stabilizing influence of the nitrogen atom on the radical
center5 which raises the expected barrier to rearrange-
ment of 5 by lowering its ground-state energy. We
X Abstract published in Advance ACS Abstracts, September 15, 1996.
(1) Della, E. W.; Knill, A. M.; Pigou, P. E. J . Org. Chem. 1993, 58,
2110.
(2) Della, E. W.; Knill, A. M. Aust. J . Chem. 1994, 47, 1833.
(3) Della, E. W.; Knill, A. M. J . Org. Chem. 1995, 61, 3518.
(4) Padwa, A.; Nimmersgen, H.; Wong, G. S. K. Tetrahedron Lett.
1985, 26, 957. Padwa, A.; Nimmersgen, H.; Wong, G. S. K. J . Org.
Chem. 1985, 50, 5620.
Treatment of 1-(2-(phenylselenyl)ethyl)-3-methylene-
piperidine (11) with tributyltin hydride was attempted
(6) Beckwith, A. L. J . Tetrahedron 1981, 37, 3073.
(7) Della, E. W.; Knill, A. M. Aust. J . Chem. 1995, 48, 2047.
(8) Ferna´ndez, S.; Mene´ndez, E.; Gotor, V. Synthesis 1991, 713.
(9) Danishefsky, S. J .; Panek, J . S. J . Am. Chem. Soc. 1987, 109,
917.
(5) Coolidge, M. B.; Borden, W. T. J . Am. Chem. Soc. 1988, 110,
2298.
S0022-3263(96)00886-9 CCC: $12.00 © 1996 American Chemical Society

7530 J . Org. Chem., Vol. 61, No. 21, 1996
Della and Knill
Sch em e 1
Our experimental data do not appear to support this
possibility. For example, in the earlier investigation7
involving Bu3SnH-mediated reduction of the acyclic
amine 7, we found no evidence for such an inhibiting
effect toward ring-closure, and the product of cyclization,
1,3-dimethylpyrrolidine, was the sole constituent. Fur-
thermore, when a sample of the precursor 11, which had
been purified by rigorous chromatography to remove any
residual diphenyl diselenide, was treated under the above
conditions, no improvement in the yield of bicyclic
product was detected. Moreover, it was observed that
heating the purified sample of 11 in the absence of Bu3-
SnH did not lead to production of (PhSe)2.
In connection with the alternative possibility, viz., an
increase in activation energy to ring-closure, we began
to suspect that the presence of a neighboring nitrogen
atom may be responsible for conferring enhanced ther-
modynamic stability on the radical 6. Although as
mentioned above such stabilizing effects are well-known
for radicals containing R-nitrogen substituents,5 there has
been no suggestion that â-nitrogen exerts a similar effect.
There is, nevertheless, some precedence for this concept
in the postulate by Barton11 and Giese12 that oxygen
attached to a â-carbon exerts a marked stabilizing effect
on a radical (termed11 the “â-oxygen effect”). While this
view has aroused some controversy,13 it is now commonly
accepted that the operation of such an effect is unlikely.
Interestingly, J enkins14 has asserted that a â-oxygen can
have either an activating or deactivating effect on the
rate of radical formation, depending upon the mode of
generation. Quite recently, a definitive article on the
origin of the “â-oxygen effect” in radicals has appeared.15
In an attempt to shed some light on the possibility that
a â-nitrogen might exert a similar effect, we16 have
performed ab initio calculations at the MP2/6-31G** level
on the energy (∆E) associated with the isodesmic reaction
represented by eq 1. Unlike the situation encountered
under conditions (110 °C, AIBN, toluene) found earlier3
to be conducive for cyclization of the carbocyclic analog,
the 2-(3-methylenecyclohexyl)ethyl radical (2), in the
expectation that the radical 6 would behave in a similar
fashion. However, the extent of ring-closure in this case
was disappointing; GC analysis of the product showed it
to consist of a 9:1 mixture of the isomeric amines 18 and
19. Modifying conditions, such as the rate of addition of
the reagent, did not have any appreciable effect on the
extent of cyclization; even under more drastic conditions
(150 °C) an unacceptably large amount of reduced
product was still evident. For assistance in identification
of the products, an authentic specimen of the monocyclic
amine 18 was prepared by reduction of the amide 16 with
lithium aluminum hydride.
•CH2CH2NH2 + CH3CH2CH3 f
•
CH3CH2NH2 + CH3CH2CH2 (1)
with an R-amino-substituted radical,5 the stability of
which is enhanced as a result of interaction of the
nitrogen lone pair with the radical center, the calcula-
tions predict that the reaction depicted in (eq 1) is
essentially energy-neutral. Accordingly, in view of the
lack of theoretical support for the concept of a â-nitrogen
stabilizing effect, we are inclined to doubt its existence.
We believe that the underlying cause for the reluctance
of radical 6 to undergo ring-closure compared with its
carbocyclic analogue 2 is associated with an increased
energy of the transition state for cyclization of the former.
There is strong support for the larger activation barrier
for reaction of 6 according to MM2 data we have obtained
on the calculated energy profile for the two cyclizations.
What is the explanation for the dramatic difference
between the facility for ring-closure of the radicals 2 and
6? Several possibilities were considered: (i) the presence
of adventitious diphenyl diselenide, and (ii) an increased
activation barrier to cyclization.
Initially, we felt that the use of the phenylselenyl
functional group may have actually introduced an un-
wanted complication. For example, Crich and Yao10 have
demonstrated that in the tributyltin hydride reduction
of oxygen-based heterocyclic substrates containing the
phenylselenide group, the nature of the product is af-
fected by the presence of diphenyl diselenide produced
by decomposition of the substrate. Tributyltin hydride
rapidly converts (PhSe)2 into phenylselenol which is a far
more effective hydrogen atom transfer agent than
(Bu)3SnH, and it therefore traps the initially-formed
radical much more efficiently. Crich and Yao10 have
suggested that only a catalytic quantity of diphenyl
diselenide is required as the phenylselenol is continually
regenerated.
(11) Barton, D. H. R.; Hartwig, W.; Motherwell, W. B. J . Chem. Soc.,
Chem. Commun. 1982, 447.
(12) Dupois, J .; Giese, B.; Ruegge, D.; Fischer, H.; Korth, H.-G.;
Sustmann, R. Angew. Chem., Int. Ed. Engl. 1984, 23, 896.
(13) (a) Beckwith, A. L. J .; Brumby, S. J . Chem. Soc., Perkin Trans.
2 1987, 1801. (b) Beckwith, A. L. J . Chem. Soc. Rev. 1993, 22, 144. (c)
Motherwell, W. B.; Crich, D. In Free Radical Chain Reactions in
Organic Synthesis; Academic Press: London, 1992; Chapter 3.
(14) J enkins, I. J . J . Chem. Soc., Chem. Commun. 1994, 1227.
(15) Crich, D.; Beckwith, A. L. J .; Chen, C.; Yao, Q.; Davison, I. G.
E.; Longmore, R. W.; de Parrodi, C. A.; Quintero-Cortes, L.; Sandoval-
Ramirez, J . J . Am. Chem. Soc. 1995, 117, 8757.
(10) Crich, D.; Yao, Q. J . Org. Chem. 1995, 60, 84.
(16) Della, E. W.; Elsey, G. M. Unpublished observations.

Synthesis of Nitrogen Bridgehead Bicyclic Heterocycles
J . Org. Chem., Vol. 61, No. 21, 1996 7531
The force field calculations were performed following the
procedure described by Beckwith and Schiesser17 that we
employed in earlier work.1,3 They predict a barrier of 16.0
kcal mol-1 for ring-closure of radical 6 which is clearly
larger than the value, 12.6 kcal mol-1, derived previously3
for cyclization of 2. This observation was unexpected
because it has been shown experimentally7 in the case
of the corresponding acyclic radicals that the activation
energy for ring-closure of the 3-aza-5-hexenyl radical is
lower than that for the 5-hexenyl radical as predicted.6
We have no explanation for the increased barrier to ring-
closure of the radical 6 and hence its reluctance to
undergo cyclization. Undoubtedly, the use of 1-(2-(phe-
nylselenyl)ethyl)-3-methylenepiperidine (11) as a precur-
sor to the 1-azabicyclo[3.2.1]octyl system is impractical.
In an attempt to circumvent this problem, we decided
to investigate the potential for ring-closure of the corre-
sponding ammonium salt 20. In this case, the expecta-
tion was that the extra substituent attached to nitrogen
in 20 would enhance the rate of cyclization of the derived
radical 25 by analogy with the beneficial effect1-3 of
added ester groups at C1, for example, in promoting ring-
closure of radicals 1 and 2. Thus, in the transition state
associated with cyclization of 25, the additional methyl
group adopts the equatorial configuration and, accord-
ingly, it is not expected to influence appreciably the
energy of the transition state relative to that of the
neutral species 6. On the other hand, whereas radical 6
(through either nitrogen inversion or ring flip) can adopt
a favorable ground state in which the substituent at-
tached to nitrogen is equatorial, rotational interconver-
sion in the case of the charged species 25 requires that
it must always exist in a conformation in which one of
the substituents attached to nitrogen has the axial
orientation. The anticipated net effect of the increased
ground state energy of radical 25 is a diminished activa-
tion barrier for its ring-closure relative to that of the
neutral species 6.
temperatures in excess of 80 °C would be required for
optimal cyclization, it was decided to employ tert-amyl
alcohol as solvent; this alcohol is inexpensive and has
the added requirement that it is expected to be inert to
radicals. In the event, treatment of a solution of 21 in
tert-amyl alcohol at 80 °C with Bu3SnH gave a product
1
which, by careful H and 13C NMR analysis, was found
to consist of 1,1,3-trimethylpyrrolidinium iodide (23)
exclusively. NMR signals that could be ascribed to other
products were not detected and, in particular, there was
no evidence for the formation of N-ethyl-N,N-dimethyl-
N-allylammonium iodide (24), the product of reduction.
The identity of the salt 23 was confirmed by demethy-
lation with DABCO in DMF as described,18 giving the
known parent amine, 1,3-dimethylpyrrolidine (10).7 The
yield of cyclized material (92%) obtained under these
conditions confirms that the presence of positively-
charged nitrogen has no deleterious effect on the cycliza-
tion reaction. In fact, the intermediate radical 22 is seen
to ring-close rapidly and at a rate comparable with that
of the neutral species 8. The real test, of course, was
whether this situation could be duplicated in the case of
the salt 20.
Conversion of the monocyclic amine 11 into 1-methyl-
1-(2-(phenylselenyl)ethyl)-3-methylenepiperidinium io-
dide (20) was achieved by reaction with iodomethane.
Exposure of a solution of 20 in tert-amyl alcohol to Bu3-
SnH at 104 °C, but otherwise under conditions specified
above, afforded a crystalline product whose identity was
established by 1H and 13C NMR analysis as 1,5-dimethyl-
1-azoniabicyclo[3.2.1]octane iodide (26). Signals attribut-
able to other compounds, including 1-ethyl-1-methyl-3-
As far as we are aware, there are no reports in the
literature on the use of ammonium salts in radical
cyclizations of this type, and we therefore chose to initiate
the study by an examination of the simpler acyclic
system, 1-(phenylselenyl)-3,3-dimethyl-3-azonia-5-hexene
iodide (21). We had access to 1-(phenylselenyl)-3-methyl-
3-aza-5-hexene (7) previously,7 and the synthesis of the
salt 21 was accomplished simply by treating the amine
with neat iodomethane.
methylenepyrrolidinium iodide (27), the noncyclized
product of reduction, were not detected in the NMR
spectra.
This unprecedented transformation possesses several
outstanding features. For example, it demonstrates
conclusively that cyclization of these heterocyclic-based
â-ammonio-substituted radicals occurs with relative ease,
in contrast with their parent amines; presumably, as
suggested above, the added methyl substituent does have
the desired effect in terms of raising the ground state of
the radical. Secondly, the reaction can be seen to proceed
with high regioselectivity; as the sole product of the
reaction, the salt 26 is seen to be produced in a high state
of purity and in a spectacular yield (87%). Finally, as
far as we are aware, this report represents the first
disclosure of a synthesis of the 1-azoniabicyclo[3.2.1]-
octane system, which can be achieved, significantly, via
a route involving several relatively easy steps.
Exp er im en ta l Section
Molecular mechanics calculations were performed as de-
scribed previously;1 spectral, chromatographic, and other
standard experimental procedures adopted were as specified.19
1-Acetyl-3-(h yd r oxym eth yl)p ip er id in e (13). The pro-
cedure developed by Ferna´ndez and her colleagues8 was used
In view of the limited solubility of the salt 21 in
hydrocarbon solvents coupled with the expectation that
(17) Beckwith, A. L. J .; Schiesser, C. H. Tetrahedron 1985, 41, 3925.
Beckwith, A. L. J .; Cliff. M. D.; Schiesser, C. H. Tetrahedron 1992, 48,
4641.
(18) Ho, T. L. Synthesis 1972, 702.
(19) Della, E. W.; Pigou, P. E.; Schiesser, C. H.; Taylor, D. K. J .
Org. Chem. 1991, 56, 4659.

7532 J . Org. Chem., Vol. 61, No. 21, 1996
Della and Knill
to convert 3-(hydroxymethyl)piperidine (12) (10 g, 0.087 mol)
into the title compound 13 (11.7 g, 86%) after distillation
(Kugelrohr: 120 °C (0.01 mm)); its 1H NMR spectrum was
consistent with that reported;20 13C NMR (CDCl3) δ 169.12,
63.93, 63.83, 49.49, 47.20, 44.46, 42.14, 38.92, 37.87, 26.84,
26.63, 24.60, 23.83, 21.10.
1-E t h yl 3-m et h ylen ep ip er id in e (18). A solution of
1-acetyl-3-methylenepiperidine (16) (1 g, 7.2 mmol) in ether
(2 mL) was slowly added to a stirred solution of LiAlH4 (1.09g,
0.029 mol) in dry ether (20 mL), and the solution was heated
under reflux for 3 h. The cooled mixture was quenched with
saturated aqueous sodium sulfate and then filtered, and the
solids were washed thoroughly with ether. The combined
filtrate was dried (Na2SO4) and the solvent carefully removed
to afford the amine 18 (0.77g, 86%) as a colorless liquid: 1H
NMR (CDCl3) δ 4.78 (d, 2H), 2.95 (s, 2H), 2.56-2.48 (m, 2H),
2.45 (t, 2H), 2.2-2.12 (m, 2H), 1.75-1.65 (m, 2H), 1.1 (t, 3H),
13C NMR (CDCl3) δ 144.16, 109.50, 59.95, 53.10, 52.18, 32.73,
26.14, 11.97. The amine was converted into its picrate salt
mp: 121-123 °C. Anal. Calcd for C14H18N4O7: C, 47.46; H,
5.12; N, 15.81. Found: C, 47.46; H, 5.09; N, 15.84.
Red u ction of 1-(2-(P h en ylselen yl)eth yl)-3-m eth ylen e-
p ip er id in e (11) w ith Bu 3Sn H. (i) At 110 °C. A solution of
tributyltin hydride (0.62 g, 2.1 mmol) in dry toluene (3 mL)
containing a few crystals of AIBN was slowly added over a 3
h period to 1-(2-(phenylselenyl)ethyl)-3-methylenepiperidine
(0.2 g, 0.71 mmol) (11) in refluxing toluene (25 mL).The extent
of the reaction was monitored by GC/MS analysis and, upon
completion, the mixture was cooled and extracted with 5% HCl
(3x). The combined aqueous extracts were washed with
hexane before being basified (pH10) and reextracted with CH2-
Cl2 (2×). The combined organic extracts were dried (Na2SO4)
and concentrated carefully. GC/MS analysis of the residue
(0.065 g, 73%) showed it to consist of a 9:1 mixture of 1-ethyl
3-methylenepiperidine (18) and 1-aza-5-methylbicyclo(3.2.1)-
octane (19).
1-Acet yl-3-m et h ylen ep ip er id in e (16). Treatment of
1-acetyl-3-(hydroxymethyl)piperidine (13) with mesyl chloride
as described21 gave the mesylate 14 (90%) which was used
without further purification. Sodium iodide (25 g, 0.17 mol)
was added to a stirred solution of 14 (13.7 g, 58 mmol) in
dimethoxyethane (100 mL), and the mixture was heated at
65 °C for 4 h. Solvent was removed under vacuum and the
residue taken up into CH2Cl2 (150 mL) which was washed
twice with saturated sodium chloride. The organic layer was
dried (Na2SO4) and evaporated leaving 1-acetyl-3-(iodomethyl)-
piperidine (15) (15 g, 96%) as a light yellow oil; 1H NMR
(CDCl3) δ 4.60-3.65 (m, 2H), 3.28-3.00 (m, 4H), 2.15 (d, 3H),
1.90-1.25 (m, 5H); 13C NMR (CDCl3) δ 168.04, 167.97, 51.50,
46.49, 46.08, 41.27, 38.36, 36.81, 30.77, 30.48, 24.53, 23.30,
20.87, 9.59, 8.82. Potassium tert-butoxide (9.5 g, 0.84 mol) was
introduced slowly into a solution of the iodide (15) (15 g, 0.56
mol) in dimethoxyethane (250 mL) and the mixture allowed
to stir at room temperature for 2 h. The mixture was quenched
with saturated ammonium chloride (20 mL) and extracted (3×)
with CH2Cl2. The combined organic extracts were washed
with water (2×) and dried (MgSO4), and the solvent was
removed under vacuum. The resulting light yellow residue
was distilled (Kugelrohr: 100 °C (1 mm)) to give the title
compound 16 (7.0 g, 90%): 1H NMR (CDCl3) δ 4.95-4.65 (s,
2H), 3.95 (d, 2H), 3.80-3.35 (m, 2H), 2.50-2.12 (m, 2H), 2.05
(s, 3H), 1.95-1.38 (m, 2H); 13C NMR (CDCl3) δ 168.94, 168.84,
142.89, 142.21, 110.83, 110.21, 53.60, 48.30, 46.74, 42.22,
32.78, 27.27, 26.49, 21.71, 21.51. Anal. Calcd for C8H13NO:
C, 69.03; H, 9.41; N, 10.06. Found: C, 68.97; H, 9.20; N, 10.19.
3-Meth ylen ep ip er id in e (17). 1-Acetyl-3-methylenepip-
eridine (16) (1 g, 7.2 mmol) was added to 10% NaOH (10 mL)
and heated under reflux for 3 h. The cooled solution was
diluted with saturated sodium chloride and extracted with
ether (3x). The combined organic extracts were washed with
water and then dried (Na2SO4). Removal of the solvent
afforded the title product 17 as a clear liquid (0.5 g, 71%): 1H
NMR (CDCl3) δ 4.65 (d, 2H), 3.30 (s, 2H), 2.90-2.85 (m, 2H),
2.25 (t, 2H), 1.78 (s, 1H), 1.70-1.60 (m, 2H); 13C NMR (CDCl3)
δ 146.36, 107.24, 52.96, 46.15, 32.98, 29.30. The amine was
converted into its hydrochloride salt which crystallized from
ether/CH2Cl2 as colorless needles, mp 210-213 °C. Anal.
Calcd for C6H12NCl: C, 53.90; H, 9.05; N, 10.48. Found: C,
53.54; H, 8.72; N, 10.30.
(ii) At 150 °C. Analysis of the product obtained from
treatment of the selenide 11 with a solution of Bu3SnH in tert-
butylbenzene at reflux under conditions described above
showed the amines 18 and 19 to be formed as a 3:1 mixture.
3,3-Dim eth yl-1-(p h en ylselen yl)-3-a zon ia h ex-5-en e Io-
d id e (21). 1-Allyl-1-methyl-1-(2-(phenylselenyl)ethyl)amine
(7) (0.2 g, 0.78 mmol)7 was stirred with iodomethane (2 mL)
overnight. Removal of excess MeI under vacuum afforded a
yellow solid which upon recrystallization (benzene/CH2Cl2)
gave colorless needles (0.27 g, 86%) of the ammonium salt 21;
mp 85-87 °C; 1H NMR (acetone-d6) δ 7.75-7.65 (m, 2H), 7.40-
7.30 (m, 3H), 6.20-6.05 (m, 1H), 5.88-5.61 (m, 2H), 4.48 (d,
2H), 3.93-3.83 (m, 2H), 3.60-3.52 (m, 2H), 3.41 (s, 6H); 13C
NMR (acetone-d6) δ 133.85, 130.29, 128.88, 128.97, 128.43,
126.28, 66.16, 64.73, 50.72, 19.70. Anal. Calcd for C13H20
-
INSe: C, 39.41; H, 5.09; N, 3.54. Found: C, 39.29; H, 4.80;
N, 3.37.
Red u ction of 3,3-Dim eth yl-1-(p h en ylselen yl)-3-a zon ia -
h ex-5-en e Iod id e (21) w ith Bu 3Sn H:1,1,3-Tr im eth ylp yr -
r olid in iu m Iod id e (23). A solution of 3,3-dimethyl-1-
(phenylselenyl)-3-azoniahex-5-ene iodide (21) (0.25 g, 0.63
mmol) in dry tert-amyl alcohol (25 mL) was maintained at 80
°C and treated with a solution of tributyltin hydride (0.22 g,
0.75 mmol) and a few crystals of AIBN in dry tert-amyl alcohol
(2 mL). After addition was complete, heating was continued
for a further 30 min after which the mixture was cooled and
the solvent removed in vacuo leaving a light yellow liquid.
Trituration with dry ether and filtration afforded 1,1,3-
trimethylpyrrolidinium iodide (23) as a colorless solid (0.13 g,
1-(2-(P h en ylselen yl)eth yl)-3-m eth ylen ep ip er id in e (11).
(Phenylselenyl)acetaldehyde (2.0 g, 10.3 mmol) was added to
a solution of 3-methylenepiperidine (17) (0.5 g, 5.15 mmol) in
a 50:50 mixture of methanol and tetrahydrofuran (25 mL) and
the mixture left to stir for 15 min at room temperature.
Sodium cyanoborohydride (0.65 g, 10.3 mmol) was then added,
and stirring was continued for 24 h after which the mixture
was quenched with saturated ammonium chloride (20 mL) and
extracted (3x) with CH2Cl2. The combined organic extracts
were dried (MgSO4), and solvent was removed under vacuum.
The residue was dissolved in ether (40 mL) which was
extracted (2×) with 5% HCl. The combined acid washes were
basified (pH 10) with NaOH and then extracted with CH2Cl2.
The organic extract was washed with saturated sodium
chloride and then dried (MgSO4) and evaporated. The result-
ing yellow oil was distilled (Kugelrohr: 125 °C (0.1 mm)) to
yield the selenide 11 (0.61 g, 42%) as a light yellow oil: 1H
NMR (CDCl3) δ 7.53-7.46 (m, 2H), 7.35-7.20 (m, 3H), 4.75
(s, 2H), 3.10-3.00 (m, 2H). 2.95 (s, 2H), 2.77-2.66 (m, 2H),
2.53 (t, 2H), 2.15 (t, 2H), 1.70-1.60 (m, 2H); 13C NMR (CDCl3)
δ 143.91, 133.44, 131.97, 128.73, 126.4, 109.04, 59.88, 58.14,
53.1, 32.45, 25.97, 24.49; mass spectrum m/ z (relative inten-
1
92%): mp 180-182 °C; H NMR (DMSO-d6) δ 3.71-3.61 (m,
1H), 3.58-3.50 (m, 2H), 3.18 (s, 3H), 3.10-3.02 (m, 1H), 3.08
(s, 3H), 2.69-2.52 (m, 1H), 2.38-2.21 (m, 1H), 1.78-1.63 (m,
1H), 1.08 (d, 3H). 13C NMR (DMSO-d6) δ 70.92, 70.88, 65.12,
65.08, 52.73, 52.68, 51.73, 51.62, 30.44, 30.23, 17.77.
1,3-Dim eth ylp yr r olid in e (10). Heating 1,1,3-trimeth-
ylpyrrolidinium iodide (23) (0.1 g), with DABCO in DMF as
described by Ho18 yielded the amine 10 almost quantitatively.
The product was identical with an authentic specimen.7
1-Met h yl-1-(2-(p h en ylselen yl)et h yl)-3-m et h ylen ep ip -
er id in iu m Iod id e (20). 1-(2-(Phenylselenyl)ethyl)-3-meth-
ylenepiperidine (11) (0.5 g, 1.78 mmol) was allowed to stir with
iodomethane (5 mL) overnight. Removal of the excess io-
domethane under vacuum and recrystallization (benzene/CH2-
Cl2) of the residue afforded the ammonium salt 20 (0.68 g, 90%)
as a pale yellow solid: mp 102-104 °C; 1H NMR (CDCl3) δ
7.69-7.58 (m, 2H), 7.39-7.25 (m, 3H), 5.15 (d, 2H), 4.37-4.13
(m, 2H), 3.95-3.65 (m, 4H), 3.42-3.20 (m, 2H), 3.31 (s, 3H),
sity); 202 (12.5), 110 (85), 69 (100); HRMS calcd for C14H19N78
Se 279.0690, found 279.0680.
-
(20) Toone, E. J .; Bryan J ones, J . Can. J . Chem. 1987, 65, 2772.
(21) Crossland, R. K.; Servis, K. L. J . Org. Chem. 1970, 35, 3195.

Synthesis of Nitrogen Bridgehead Bicyclic Heterocycles
J . Org. Chem., Vol. 61, No. 21, 1996 7533
2.50-2.30, (m, 2H), 2.02-1.63 (m, 2H); 13C NMR (CDCl3) δ
piperidinium iodide (20) (0.2 g, 0.47 mmol) in dry tert-amyl
alcohol (18 mL) was maintained at 104 °C while a solution of
tributyltin hydride (0.16 g, 0.58 mmol) and a few crystals of
AIBN in dry tert-amyl alcohol (2 mL) was added over 15 min.
After addition, heating was continued for a further 30 min
after which the mixture was cooled and the solvent removed
in vacuo. Trituration of the light residue with dry ether
afforded a white solid which was filtered and washed with
ether. Recrystallization from benzene/CH2Cl2 yielded 1,5-
dimethyl-1-azoniabicyclo[3.2.1]octane iodide (26) (0.11 g, 87%);
1H NMR (CDCl3) δ 4.08-3.65 (m, 6H), 3.55 (s, 3H), 2.30-1.89
(m, 4H), 1.78-1.55 (m, 2H), 1.25 (s, 3H); 13C NMR (CDCl3) δ
72.43, 63.50, 62.35, 51.19, 41.06, 35.03, 34.37, 22.84, 19.54.
Anal. Calcd for C9H18NI: C, 40.46; H, 6.79; N, 5.24. Found:
C, 40.35; H, 7.02; N, 5.15.
134.06, 133.24, 129.26, 127.81, 127.0, 118.82, 65.51, 61.95,
59.93, 48.01, 29.10, 20.26, 18.78. Anal. Calcd for C15H22
-
INSe: C, 42.67; H, 5.25; N, 3.32. Found: C, 41.92; H, 5.31;
N, 3.21.
1-Eth yl-1-m eth yl-3-m eth ylen epiper idin iu m iodide (27).
1-Ethyl-3-methylenepiperidine (18) (0.5 g, 4.0 mmol) was
allowed to stir with iodomethane (5 mL) overnight. Removal
of the excess iodomethane under vacuum and recrystallization
(benzene/CH2Cl2) of the residual solid afforded 1-ethyl-1-
methyl-3-methylenepiperidinium iodide (27) as a pale yellow
1
solid (0.98 g, 92%): mp 116-118 °C; H NMR (CDCl3) δ 5.3
(d, 2H), 4.3 (s, 2H), 3.95-3.70 (m, 4H), 3.25 (s, 3H), 2.52 (t,
2H), 2.10-1.96 (m, 2H), 1.43 (t, 3H); 13C NMR (CDCl3) δ
134.37, 118.54, 65.14, 58.76, 57.56, 46.54, 28.75, 20.02, 7.45.
Anal. Calcd for C9H18IN: C, 40.46; H, 6.79; N, 5.24. Found:
C, 40.09; H, 6.99; N, 5.13.
Red u ction of 1-Meth yl-1-(2-(p h en ylselen yl)eth yl)-3-
m eth ylen ep ip er id in iu m Iod id e (20) w ith Bu 3Sn H:1,5-
Ack n ow led gm en t. We thank the Australian Re-
search Council for financial support of this work.
Dim eth yl-1-a zon ia bicyclo[3.2.1]octa n e Iod id e (26).
A
solution of 1-methyl-1-(2-(phenylselenyl)ethyl)-3-methylene-
J O960886I