Determining the Scope of the Organolanthanide-Catalyzed, Sequential Intramolecular Amination/Cyclization Reaction: Formation of Substituted Quinolizidines, Indolizidines, and Pyrrolizidines

Article abstract of DOI:10.1021/jo035205f

The scope of the lanthanide-mediated, intramolecular amination/cyclization reaction was determined for the formation of substituted quinolizidines, indolizidines, and pyrrolizidines. A methyl group was installed at diverse positions in the substrates to d

Full text of DOI:10.1021/jo035205f

Deter m in in g th e Scop e of th e Or ga n ola n th a n id e-Ca ta lyzed ,
Sequ en tia l In tr a m olecu la r Am in a tion /Cycliza tion Rea ction :
F or m a tion of Su bstitu ted Qu in olizid in es, In d olizid in es, a n d
P yr r olizid in es
Gary A. Molander* and Shawn K. Pack
Roy and Diana Vagelos Laboratories, Department of Chemistry, University of Pennsylvania,
Philadelphia, Pennsylvania 19104-6323
gmolandr@sas.upenn.edu
Received August 16, 2003
The scope of the lanthanide-mediated, intramolecular amination/cyclization reaction was determined
for the formation of substituted quinolizidines, indolizidines, and pyrrolizidines. A methyl group
was installed at diverse positions in the substrates to determine the sense and magnitude of
diastereoselection. High diastereoselectivity (>20:1) was achieved for the formation of some
quinolizidines and indolizidines. The sense of relative asymmetric induction was contrary to
previously studied systems, and although some questions remain, a rationalization for these results
is put forward.
In tr od u ction
much remained undone. Additionally, while the synthesis
of pyrrolizidines and indolizidines was studied to a
limited extent (eqs 1-5), bicyclic systems from the
quinolizidine family were not examined.
Cascade reactions involving the formation of multiple
ring systems in a single operation have proven to be
among the most powerful in the arsenal of synthetic
organic chemists. However, the utility of these reactions
cannot be fully realized until the scope of the reaction in
question is outlined. The lanthanide-catalyzed sequential
reaction of amino dienes is one such potentially useful
transformation. Li and Marks first developed this elegant
process.1 In a thorough study, they have shown that
lanthanide complexes possess a unique reactivity. Unlike
other metal-catalyzed intramolecular amination reac-
tions, the organometallic formed after the initial nitrogen-
carbon bond-forming event can insert into a carbon-
carbon multiple bond to form indolizidines and pyrroli-
zidines (Scheme 1).
Although a few reactions were performed that shed
light on the stereochemical aspects of these reactions,
(1) (a) Li, Y.; Marks, T. J . J . Am. Chem. Soc. 1998, 120, 1757. For
other aspects of the hydroamination reaction, see: (b) Gagne´, M. R.;
Marks, T. J . J . Am. Chem. Soc. 1989, 111, 4108. (c) Gagne´, M. R.;
Nolan, S. P.; Marks, T. J . Organometallics 1990, 9, 1716. (d) Gagne´,
M. R.; Stern, C. L.; Marks, T. J . J . Am. Chem. Soc. 1992, 114, 275. (e)
Gagne´, M. R.; Brard, L.; Conticello, V. P.; Giardello, M. A.; Stern, C.
L.; Marks, T. J . Organometallics 1992, 11, 2003. (f) Li, Y.; Fu, P.-F.;
Marks, T. J . Organometallics 1994, 13, 439. (g) Li, Y.; Marks, T. J . J .
Am. Chem. Soc. 1996, 118, 707. (h) Li, Y.; Marks, T. J . Organometallics
1996, 15, 3770. (i) Li, Y.; Marks, T. J . J . Am. Chem. Soc. 1996, 118,
9295. (j) Giardello, M. A.; Conticallo, V. P.; Brard, L.; Gagne´, M. R.;
Marks, T. J . J . Am. Chem. Soc. 1994, 116, 10241. (k) Tian, S.;
Arredondo, V. M.; Stern, C. L.; Marks, T. J . Organometallics 1999,
18, 2568. (l) Bu¨rgstein, M. R.; Berberich, H.; Roesky, P. W. Organo-
metallics 1998, 17, 1452. (m) Kim, Y. K.; Livinghouse, T.; Bercaw, J .
E. Tetrahedron Lett. 2001, 42, 2933. (n) Gilbert, A. T.; Davis, B. L.;
Emge, T. J .; Broene, R. D. Organometallics 1999, 18, 2125. (o) Kim,
Y. K.; Livinghouse, T.; Horino, Y. J . Am. Chem. Soc. 2003, 125, 9560
and references therein. For reviews on the application of lanthanocenes
in organic synthesis, see: (p) Molander, G. A.; Dowdy, E. D. In
Lanthanides: Chemistry and Use in Organic Synthesis; Kobayashi,
S., Ed.; Springer-Verlag: Berlin, 1999. (q) Molander, G. A.; Romero,
J . A. C. Chem. Rev. 2002, 102, 2161.
Because numerous biologically active alkaloids contain
the pyrrolizidine and indolizidine motif (Figure 1), the
application of this method toward their synthesis would
be very appealing.2 From relatively simple starting
(2) Dewick, P. M. Medicinal Natural Products: A Biosynthetic
Approach; J ohn Wiley & Sons Ltd.: Chichester, 1997.
10.1021/jo035205f CCC: $25.00 © 2003 American Chemical Society
Published on Web 10/29/2003
9214
J . Org. Chem. 2003, 68, 9214-9220

Formation of Substituted Quinolizidines
used to make substrate 14.4 Purification by flash chro-
matography of the HCl salts, using 2-propanol/EtOAc as
F IGURE 1. Representative indolizidine and pyrrolizidine
alkaloids.
SCHEME 1
eluent, also allowed recovery of the allylamine or ho-
moallylamine hydrochloride.
materials, one can form two fused rings in a single
synthetic operation. However, to apply this method, its
reliability and stereochemical outcomes given a diverse
array of substrate substitution patterns needed to be
determined. In addition, this potentially important method
would further benefit by its application to the previously
unexamined, but more prevalent, quinolizidine ring
system.2
Herein, we describe our efforts to outline the scope and
limitations of the Marks process as applied to the
synthesis of a variety of nitrogen fused bicyclic systems,
with particular focus on the stereochemical aspects of the
reaction in the formation of diverse bicyclics.
One-carbon homologation of the tosylates with NaCN
generated nitriles 36-38 (eq 9).5 The known nitrile 35
Resu lts a n d Discu ssion
To initiate our studies, an extensive set of amino dienes
was constructed to examine the scope of the process. A
methyl group was installed at various positions in the
cyclization substrates (eqs 6 and 7) to determine the
diastereoselectivity of the sequential reaction. The methyl
group was chosen because of the ease of preparation of
the substrates. Furthermore, if steric effects proved to
be the major factor in establishing the diastereoselectiv-
ity, the methyl group would provide a “worse case
scenario” because of its relatively small size.
was prepared via a literature procedure from the com-
mercially available bromide.6 A modified procedure out-
The synthesis of the HCl salts of substrates 10-13 and
15 (eq 8) was accomplished by modifying a procedure
outlined by Fu¨rstner.3 The workup and yields of this
method proved to be superior to the reported procedure
(4) Martin, S. F.; Benage, B.; Williamson, S. A.; Brown, S. P.
Tetrahedron 1986, 42, 2903.
(5) Ishiyama, H.; Takemura, T.; Tsuda, M.; Kobayashi, J . Tetrahe-
dron 1999, 55, 4583.
(6) Glover, S. A.; Hammond, G. P.; Harman, D. G.; Mills, J . G.;
Rowbottom, C. A. Aust. J . Chem. 1993, 46, 1213.
(3) Fu¨rstner, A.; Langemann, K. Synthesis 1997, 7, 792.
J . Org. Chem, Vol. 68, No. 24, 2003 9215

Molander and Pack
TABLE 1. Or ga n ola n th a n id e-Ca ta lyzed F or m a tion of Qu in olizid in es a n d In d olizid in es
entry
substrate
product
methyl position
catalysta
T (oC)
drb
% isolated yield
n ) 2
1
2
3
4
5
6
7
8
9
1
2
16
17
I
I
IV
III
I
I
IV
I
rt
>50:1
nr
85c
nr
d
91
90
90e
nr
86
89
R1
R2
70
70
40
50
50
90
rt
>50:1d
>2.5:1
>20:1
>20:1e
nr
3
18
4
5
19
20
R3
R4
1:1.3
1:1.9
I
rt
n ) 1
10
11
12
13
14
15
16
17
6
7
21
22
R1
R2
II
I
I
II
I
II
I
50
40
40
40
rt
rt
rt
rt
8.3:1.3:1
27:1
87
86
80e
f
86
f
27:1e
23:1
8
9
23
24
R3
R4
8.5:6.5:1
2.8:2.3:1
5:2.3:1:1
4:3:1.4:1
81
f
II
a
b
I ) Cp*2NdCH(TMS)2, II ) Cp*2SmCH(TMS)2, III ) Me2SiCp*2NdCH(TMS)2, IV ) [CpTMS2NdCH3]2. The diastereomeric ratio was
determined by GC analysis of the filtered crude reaction mixture. c Isolated as the MeI monohydrate salt. The reaction was 70% complete,
d
and the product was not isolated. e 60-100 mg scale reaction. f Product was not isolated.
SCHEME 2 a
requisite purity for catalytic reactions could be achieved
on a milligram scale. In comparison, gram quantities of
material are desired to obtain comparably pure material
by distillation.
Unfortunately, the HCl salt of the amine destroys the
catalyst. Consequently, a simple free base protocol was
required. An anhydrous, easily filtered, polymer base that
is active enough to neutralize the HCl salt in a triphasic
system (HCl salt, benzene, and polymer) would be ideal.
However, to our knowledge, such a polymer does not
exist. Several anhydrous bases such as the metal hy-
drides and metal carbonates were unsuccessful or the
process too complicated. Eventually, the free amine was
liberated with aqueous NaOH and extracted into ben-
zene. Drying the benzene solution of the amine over
activated molecular sieves proved, to our delight, ad-
equate enough to remove the residual moisture without
absorbing significant amounts of the free amine. The
benzene solution was then degassed and taken into a
glovebox where it was treated with the catalyst.
a
Key: (a) (i) Mg, dibromoethane, 30-40 °C, THF, (ii) CuCN,
propylene oxide, -60 to -50 °C, THF, (iii) TsCl, pyridine, CH2Cl2
(85%); (b) (i) equiv of allylamine (n ) 1) or homoallylamine
(n ) 2), THF, 60-70 °C; (ii) HCl in diethyl ether (49-80%).
lined by Brussee7 was used to prepare the HCl salts of
substrates 1, 3-5, and 7-9. The modification involved
quenching the iminium solution into a slurry of the
allylamine or homoallylamine hydrochloride followed by
reduction with LAH at -78 °C. This modification simpli-
fied the quench and minimized an impurity formed in
the reaction. Again, flash chromatography with 2-pro-
panol/EtOAc not only purified the substrates but also
allowed for recovery of the allylamine or homoallylamine
hydrochlorides.
The HCl salts of substrates 2 and 6 were prepared from
tosylate 34 (Scheme 2) as described above. The alcohol
precursor to 34 was generated via a copper-promoted
reaction of 3-butenylmagnesium bromide with propylene
oxide.8
Formation of the quinolizidine ring system was best
achieved at a concentration of 0.06 M with 10% Cp*2-
NdCH(TMS)2 (Table 1). The monocyclic hydroamination
side product was not observed by GC or by overnight
carbon-13 NMR. The diastereoselectivity of the reaction
ranged from 1.3:1 to greater than 50:1 with predomi-
nantly two diastereomers observed. It should be noted
that with three stereogenic centers there are four possible
diastereomers.
Several advantages were realized by isolating the HCl
salts of the substrates. First, handling solids is much
easier than handling liquids, especially in milligram
quantities. Second, oxidation of the amines was greatly
hindered, allowing for easy, prolonged storage. Last, the
The high selectivity found in the formation of 16, 18,
and 22 (Table 1, entries 1, 5, 6, and 11-13, respectively)
was confirmed by carbon-13 NMR spectra taken of the
reaction mixtures prior to workup. Formation of the HCl
salts resulted in the isolation of two or more conformers.
NMR analysis of the free base confirmed the diastereo-
(7) Zandbergen, P.; van den Niewendijk, A. M. C. H.; Brussee, J .;
van der Gen, A.; Kruse, C. G. Tetrahedron 1992, 48, 3977.
(8) Kobayashi, Y.; Nakano, M.; Kumar, G. B.; Kishihara, K. J . Org.
Chem. 1998, 63, 7505.
9216 J . Org. Chem., Vol. 68, No. 24, 2003

Formation of Substituted Quinolizidines
F IGURE 3. Stereochemistry of 25 as determined by single-
crystal X-ray diffractometry.
entries 10-17), as evidenced by the appearance of a third
diastereomer in the latter.
F IGURE 2. Stereochemistry of the major diastereomers as
A slight trend related to the catalyst is also observed
in the formation of the indolizidine ring system. The Cp*2-
NdCH(TMS)2 catalyst appears to provide better selectiv-
ity than the Cp*2SmCH(TMS)2 catalyst for substrates 7
and 8 (entries 11-15). However, substrate 9 yielded a
better selectivity with the samarium catalyst (entries 16-
17). Although substrate 2 did not go to completion with
the [CpTMS2NdCH3]2 catalyst (entry 3), formation of the
indolizidine ring system with the methyl at the R1
position (entry 10) did proceed smoothly. The more rapid
formation of the five-membered ring allowed for lower
temperatures and a shorter reaction time. Unfortunately,
the more accessible coordination sphere of the catalyst
and/or the elevated temperature resulted in poorer
selectivity. An apparent anomaly was observed for sub-
strate 3 and the [CpTMS2NdCH3]2 catalyst (entry 7). Even
at 90 °C, no product was observed, again indicating the
low reactivity in these systems of this particular catalyst.
determined by single-crystal X-ray diffractometry.
meric purity. Further, the free base of 22 was converted
1
back to the HCl salt, resulting in H NMR spectra similar
to that of the originally isolated product. The relative
stereochemistry of the products was elucidated by X-ray
diffraction analysis and is shown in Figure 2. X-ray
quality crystals were obtained by the formation of the
MeI (16), HCl (18), and picrate (22) salts, respectively,
and GC analysis of the corresponding free bases con-
firmed the isolation of the major diastereomer.
Previous results from numerous laboratories1 had
established that lanthanide-mediated cyclization reac-
tions (both intramolecular hydroamination and carbon-
carbon bond formation via olefin insertion) transpired
through chairlike transition structures. Consequently,
isolation of 16 as the major diastereomer in the double
cyclization of 1 was expected. In contrast, the formation
18 and 22, wherein an axial substituent is established,
was quite surprising. A further discussion of this inter-
esting and unanticipated result is presented later.
In these preliminary studies, interesting features of the
various catalysts were revealed. The low selectivity
obtained with 2 and the ansa-bridged catalyst Me2SiCp*2-
NdCH(TMS)2 (entry 4) is most likely due to the more
open coordination sphere of the catalyst.9 As a conse-
quence, this catalyst was not applied to any further
systems. A higher diastereoselectivity was achieved with
the [CpTMS2NdMe]2 catalyst that has a more restricted
approach to the metal center (entry 3). However, in this
case, the reaction could not be forced to completion. The
difficulties encountered with this catalyst convinced us
to drop it from broad application in our studies.
In terms of substrate control of stereochemistry, those
starting materials with a methyl group at R3 result in
the lowest selectivity (entries 8, 14, and 15). A slight
increase is observed by installing a substituent at the
R4 position (entries 9, 16, and 17). These results are an
indication of the preorganization involved during the
insertion of the Ln-N bond into the olefin, resulting in
a significant steric influence by the substituted Cp
rings.10-14 The second clear trend is the decrease in
selectivity in going from the quinolizidine ring system
(n ) 2, entries 1-9) to the indolizidine system (n ) 1,
The indolizidine ring system could also be generated
by first forming the pyrrolidine ring followed by the
piperidine portion (Table 2).
High selectivity was obtained with substrate 10 (Table
2, entries 1-3), and X-ray diffraction analysis of the HCl
salt indicated the expected stereochemistry shown in
Figure 3.
A reversal in catalyst selectivity was observed for
substrate 10 where the neodymium catalyst was less
selective than the samarium catalyst (compare entries
1-2 and 3). However, an increase in selectivity was
observed for substrate 12 (entries 6 and 7). The opposite
results were observed for substrates 6-9 as reported
above (Table 1, entries 10-17). It should also be noted
that substrates 8 (Table 1, entries 14 and 15) and 11
(Table 2, entries 4 and 5) generate the same indolizidine
ring system with considerably different stereoselectivi-
ties. Substrates 13-15 (entries 9 and 10) formed only
two diastereomers in the pyrrolizidine ring system, albeit
in low ratios. Last, initial trials showed that the sa-
marium catalyst was much less reactive than the neody-
mium catalyst for the formation of the quinolizidine ring
system. Conversely, the neodymium catalyst did not
perform as well as the samarium catalyst during the
formation of the pyrrolizidine ring system. Stereoelec-
tronic factors could play a role in this observation,
wherein the larger ionic radius of the neodymium catalyst
favors the formation of the larger six-membered ring and
the smaller ionic radius of the samarium catalyst favors
the formation of the smaller five-membered ring.
(9) (a) Molander, G. A.; Dowdy, E. D. J . Org. Chem. 1998, 63, 8983.
(b) Ryu, J .; Marks, T. J .; McDonald, F. E. Org. Lett. 2001, 3, 3091.
(10) Casey, C. P.; Carpenetti, D. W., II.; Sakurai, H. J . Am. Chem.
Soc. 1999, 121, 9483.
(11) Casey, C. P.; Klein, J . F.; Fagan, M. A. J . Am. Chem. Soc. 2000,
122, 4320.
(12) Casey, C. P.; Lee, T.; Tunge, J . A.; Carpenetti, D. W., II. J . Am.
Chem. Soc. 2001, 123, 10762.
(13) Casey, C. P.; Tunge, J . A.; Lee, T.; Carpenetti, D. W., II.
Organometallics 2002, 21, 389.
As alluded to previously, high diastereoselectivities in
the formation of cis-2,6-disubstituted piperidines and
trans-2,5-disubstituted pyrrolidines from primary amine
precursors have been previously established (eqs 10 and
(14) Casey, C. P.; Tunge, J . A.; Lee, T.; Fagan, M. A. J . Am. Chem.
Soc. 2003, 125, 2641.
J . Org. Chem, Vol. 68, No. 24, 2003 9217

Molander and Pack
TABLE 2. Or ga n ola n th a n id e-Ca ta lyzed F or m a tion of In d olizid in es a n d P yr r olizid in es
entry
substrate
product
methyl position
R1
catalysta
T (oC)
drb
% isolated yield
n ) 2
10
1
2
3
4
5
6
7
25
II
II
I
II
I
50
50
50
rt
rt
rt
25:1
25:1c
10:1
1:1
91
85c
d
84
d
11
12
26
27
R2
R3
1:1e
6:1
II
I
91
d
rt
7:1
n ) 1
13
14
8
9
10
28
29
30
R1
R2
R2
II
II
II
50
rt
rt
2:1
1.3:1
1.5:1
83
83
88
15
a
b
I ) Cp*2NdCH(TMS)2, II ) Cp*2SmCH(TMS)2. The diastereomeric ratio was determined by GC of the filtered crude reaction mixture.
c 80 mg scale reaction. Product was not isolated.
d
inspection and molecular modeling studies offer a poten-
tial explanation.
In an attempt to approximate the transition structures
by modeling the product geometries, the Merck Molecular
Force Field17 (substituting Cp*2Sc for Cp*2Nd) was
employed to provide insight into the observed stereo-
chemistry. Interestingly, this analysis in fact correctly
predicts the sense of diastereoselection observed for all
of these substrates (Figure 5). For the cyclization of a
primary amine, e.g., 6-amino-1-hexene (I), reaction through
a chair transition structure with the olefin oriented
pseudoequatorially (IA) is predicted to be favored over
the axial orientation (IB). The long carbon-lanthanide
bonds (∼2.4-2.7 Å) aid the former in overcoming what
would otherwise be a highly strained trans-bicyclo[4.2.0]
ring system in the transition structure. The relatively
normal chair conformation of the six-membered ring in
IA attests to the relative lack of strain in the system.
The increased strain in IB manifests itself in a greater
degree of distortion in the six-membered ring.
F IGURE 4. Transition structure leading to the formation of
25.
11).15,16 The diastereomers observed result from a transi-
tion structure wherein an energy-minimized chair con-
formation of the incipient ring places all of the substit-
uents either equatorially or pseudoequatorially.
Molecular modeling of the cyclization of 6-N-methyl-
amino-1-hexene, a secondary amine, reveals a surprise.
Here, cyclization through a transition structure with the
olefin pseudoaxial (IIB) is predicted to be more favorable
than the pseudoequatorial alternative (IIA), in line with
our observations. The same issues concerning strain in
the bicyclic ring system pertain to this system, but among
other factors a severe A1,3 interaction (closest contact 1.76
Å) between the axial N-methyl group and the developing
metallomethyl group in IIA stands out in differentiating
this structure from that of IIB as well as that of both
des-methyl models IA and IB. This interaction may
account for a significant fraction of the unfavorable
interactions that mandate a switch in the transition
structures leading to the observed products.
A priori, these results could be used to explain the
sense and magnitude of diastereoselection in the forma-
tion of 16. In addition, these results explain the high
selectivity found in the construction of 25 (Figure 4). In
this case, the initial intramolecular amination step leads
to the formation of a trans-2,5-disubstituted pyrrolidine
ring (eq 11). The transition structure of the subsequent
olefin insertion step is similar to the formation of cis-
2,6-disubstituted piperidines.
However, barring a boat transition structure, the
relative stereochemistry found in 18 and 22 requires that
either the methyl group on the incipient ring be oriented
axially or the olefin engaged in ring formation must be
situated pseudoaxially in the transition structure leading
to the monocyclic intermediate. On the face of it, neither
of these options is particularly appealing, and the latter
appears contrary to all extant examples. However, closer
We sought further evidence for these observations. We
reasoned that if the chair transition structure was
correct, then reaction of a substrate with appropriately
(15) Molander, G. A.; Dowdy, E. D.; Pack, S. K. J . Org. Chem 2001,
66, 4344.
(16) Gagne´, M. R.; Stern, C. L.; Marks, T. J . J . Am. Chem. Soc. 1992,
114, 275.
9218 J . Org. Chem., Vol. 68, No. 24, 2003

Formation of Substituted Quinolizidines
F IGURE 5. Calculated transition structure models for the lanthanide-catalyzed intramolecular amination reaction of primary
(I) and secondary (II) aminoalkenes in which the alkene is oriented pseudoequatorially (A) and pseudoaxially (B). Methyl groups
on the Cp* rings have been removed for visual clarity.
placed gem-dimethyl substituents should be slowed or
inhibited altogether because of the unavoidable, ad-
ditional 1,3-diaxial interaction (eq 12).
Substrate 40 was synthesized from the known nitrile
39 via the commercially available bromonitrile in ca. 55%
overall yield using procedures described above (eq 13).18
not react with substrates 1 and 3. Another observation
not readily explained by these models is the low diaste-
reoselectivity seen in the conversion of 4 to 19. Given the
analysis provided above, high diastereoselectivity was
anticipated. Thus, although some new light has been
shed on the factors involved in lanthanocene catalyzed
reactions, a truly comprehensive understanding of all the
critical features involved in controlling reactivity and
selectivity of this class of catalysts is obviously multi-
variate and not readily deconvoluted.
Con clu sion
This substrate, along with substrates 1 and 3, was
treated with 8-9 mol % of catalyst in a 0.06 M benzene-
d6 solution, and the reactions, performed side-by-side,
were monitored by NMR. The results are displayed in
eqs 14-16.
The stereoselectivity achieved during the formation of
methyl-substituted quinolizidines, indolizidines, and pyr-
rolizidines via a sequential intramolecular amination/
cyclization reaction was determined. A new stereocontrol
element has been detected that appears to differentiate
the transition structures of N-alkyl substituted (second-
ary) amino dienes from primary amino dienes. Addition-
ally, a neodymium-based catalyst performed better for
the initial formation of six-membered rings, and a sa-
marium catalyst performed better for the initial forma-
tion of five-membered rings. Still, many questions remain
concerning the reactivity and selectivity of this class of
catalysts in sequential amination/cyclization reactions.
Exp er im en ta l Section
N-3-Bu ten yl-5-h exen -1-am in e, Hydr och lor ide (1). (Rep-
r esen ta tive P r oced u r e for th e F or m a tion of Secon d a r y
Am in es fr om Th eir Cor r esp on d in g Nitr iles.) Adapted
from Brussee’s procedure.7 A 250 mL round-bottomed flask,
equipped with a thermocouple, stir bar, and a nitrogen line,
was charged with 890 mg (9.36 mmol) of nitrile (35) and 70
mL of HPLC-grade diethyl ether. The clear solution was cooled
Indeed, both substrates substrate 1 and 3 reacted
faster than 40, which was virtually unreactive using the
Cp* catalyst. Substrate 40 did react with the ansa-
bridged neodymium catalyst in which steric access to the
metal is more facile. Curiously, this same catalyst did
(17) Halgren, T. A. J . Comput. Chem. 1996, 17, 490.
(18) Gilbert, A. T.; Davis, B. L.; Emge, T. J .; Broene, R. D.
Organometallics 1999, 18, 2125.
J . Org. Chem, Vol. 68, No. 24, 2003 9219

Molander and Pack
to -78 °C and charged with 22.5 mL (22.5 mmol) of 1 M
DIBALH in hexane while maintaining an internal temperature
less than -65 °C. The hazy white solution was allowed to stir
at -78 °C for ca. 3 h. Meanwhile, a separate round-bottomed
flask, equipped with a thermocouple, stir bar, and a nitrogen
line, was charged with 2.45 g (22.8 mmol, 2.4 equiv) of the
HCl salt of homoallylamine and 20 mL of HPLC grade diethyl
ether. The cold iminium solution was then transferred, via
cannula, to the HCl salt slurry over a period of ca. 30 min.
The hazy white solution was allowed to stir at rt overnight.
The mixture was then cooled to -78 °C and charged with 13.5
mL (13.5 mmol) of 1 M LAH in diethyl ether while maintaining
an internal temperature less than -70 °C. The white solution
was allowed to stir at -78 °C for ca. 2 h and quenched with
75 mL of 1 M aqueous HCl. The aqueous layer was separated,
washed with 2 × 75 mL of diethyl ether and neutralized with
75 mL of 10 wt % aqueous NaOH. The free amine was
extracted with 3 × 75 mL of diethyl ether and acidified with
16 mL of 2M HCl in diethyl ether. The solvents were removed
under vacuum and the resulting mixture was separated by
flash chromatography using 2-propanol/EtOAc as eluent.
Crystallization was achieved by solvent exchanging the com-
bined fractions into heptane. The white crystalline solids were
filtered, providing 1.23 g (6.50 mmol, 69%) of the title
compound. The HCl salt of homoallylamine was also recovered
layered with an additional 380 mg of C6D6 and 440 mg (12.9
mg, 68 µmol) of a 2.94 wt % solution of the free amine of 1 in
C6D6. The layered mixture was allowed to sit at rt and
monitored by NMR. Upon reaction completion, ca. 96 h, the
green solution was charged with 2 mL of heptane and allowed
to oxidize for ca. 2 h. The yellow slurry was filtered through
Celite and the clear, pale yellow solution was analyzed by GC
(>50:1 dr). The mixture was then treated with excess MeI,
and the resulting solids were filtered and dried under high
vacuum to yield 18.2 mg (58 µmol, 85%) of a pale yellow,
monohydrate, crystalline solid. Recrystallization using acetone/
heptane provided anhydrous crystals that were suitable for
X-ray analysis: mp (monohydrate) 202-210 °C; 1H NMR
(monohydrate, 500 MHz, CDCl3) δ 4.02-3.88 (m, 5H), 3.11 (s,
3H), 2.21-2.05 (m, 2H), 1.95-1.90 (m, 4H), 1.74-1.68 (m, 4H),
1.58 (s, 2H), 1.38-1.35 (dt, J ) 15.1, 12.2 Hz, 1H), 1.03-1.02
(d, J ) 6.3 Hz, 3H); 13C NMR (125 MHz, CDCl3) δ 69.5, 64.9,
64.8, 38.3, 35.1, 29.2, 28.5, 26.8, 21.8, 21.3, 20.4; IR (CDCl3)
2957.8, 2253.2, 2203.4, 1459.7, 1380.3, 1094.4 cm-1; HRMS
calcd for C11H22N (M - I; ESI+) 168.1752, found 168.1757;
LRMS (ESI+) m/z 168 (100).
Ack n ow led gm en t. We thank Dr. Pat Carroll for the
X-ray structure determination of 16, 18, 22, and 25.
Special thanks to Professor Marisa Kozlowski for per-
forming the calculations, and both Professor Kozlowski
and Professor Patrick Walsh for critical discussions
concerning the stereochemical aspects of this study. We
gratefully acknowledge the National Institutes of Health
(Grant No. GM48580) and Merck Research Laboratories
for their generous support of this research. S.K.P.
thanks the Process R&D group of Bristol-Myers Squibb
for their financial support and a study leave.
1
(1.0 g, 9.3 mmol, 1.0 equiv): mp 222-230 °C; H NMR (500
MHz, CDCl3) δ 9.59 (s, 2H), 5.82-5.72 (m, 2H), 5.19-5.12 (d,
J ) 17.2 Hz, 1H), 5.19-5.12 (d, J ) 10.3 Hz, 1H), 5.03-4.96
(d, J ) 16.9 Hz, 1H), 5.03-4.96 (d, J ) 10.3 Hz, 1H), 2.99-
2.94 (m, 4H), 2.71-2.66 (dt, J ) 7.9, 7.2 Hz, 2H), 2.11-2.06
(m, 2H), 1.94-1.88 (m, 2H), 1.51-1.45 (m, 2H); 13C NMR (125
MHz, CDCl3) δ 137.7, 132.7, 118. 6, 115.5, 47.8, 47.0, 33.2,
30.2, 26.2, 25.4; IR (CDCl3) 3412.1, 3079.9, 2938.9, 2768.1,
2444.0, 2251.4, 2220.3, 1642.6, 1589.8, 1469.2, 1381.3, 1301.5
cm-1; HRMS calcd for C10H20N (M - Cl; CI+) 154.1595, found
154.1600; LRMS (ESI+) m/z 112 (30), 154 (100).
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures for all compounds not described within the text.
N-Meth yl-2-m eth ylqu in olizid in e, Iod id e (16). (Rep r e-
sen ta tive P r oced u r e for th e Ta n d em Hyd r oa m in a tion /
C-C Cycliza tion Rea ction .) A sealable NMR tube was
charged with 380 mg (3.8 mg, 6.6 µmol) of a 1 wt % solution of
Cp*2NdCH(TMS)2 in C6D6. The catalyst solution was then
1
HRMS, LRMS, and H and 13C NMR spectra of representative
compounds. X-ray data for 16, 18, 22, and 25. This material
is available free of charge via the Internet at http://pubs.acs.org.
J O035205F
9220 J . Org. Chem., Vol. 68, No. 24, 2003