Synthesis of enantiopure diamine ligands related to sparteine, via scandium triflate-catalyzed imino Diels-Alder reactions

Article abstract of DOI:10.1016/S0040-4039(02)01671-4

Imino Diels-Alder reactions have been investigated as a new route to sparteine analogues. The first enantioselective synthesis of two diastereoisomeric tricyclic diamines, structurally equivalent to the ABC and BCD rings of the naturally occurring alkaloid, is reported, starting from enantiopure intermediates. The effectiveness of the diamines in the lithiation of N-Boc-pyrrolidine is discussed.

Full text of DOI:10.1016/S0040-4039(02)01671-4

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 7155–7158
Synthesis of enantiopure diamine ligands related to sparteine, via
scandium triflate-catalyzed imino Diels–Alder reactions
Bruno Danieli, Giordano Lesma,* Daniele Passarella, Paola Piacenti, Alessandro Sacchetti,
Alessandra Silvani* and Andrea Virdis
Dipartimento di Chimica Organica e Industriale, Universita` degli Studi di Milano, via Venezian 21, I-20133 Milano, Italy
Received 18 June 2002; revised 8 August 2002; accepted 12 August 2002
Abstract—Imino Diels–Alder reactions have been investigated as a new route to sparteine analogues. The first enantioselective
synthesis of two diastereoisomeric tricyclic diamines, structurally equivalent to the ABC and BCD rings of the naturally occurring
alkaloid, is reported, starting from enantiopure intermediates. The effectiveness of the diamines in the lithiation of N-Boc-pyrro-
lidine is discussed. © 2002 Elsevier Science Ltd. All rights reserved.
The development of efficient stereodirecting ligands
readily available in both enantiomeric forms is nowa-
days of paramount importance to the field of enantiose-
lective catalysis. Particularly interesting is the complex
of sec-butyllithium and (−)-sparteine (1) (Fig. 1), which
Hoppe1 first showed could induce high enantioenrich-
ments in lithiation–electrophilic quench reactions.2 (−)-
ity of sparteine and preparable in both enantiomeric
forms.4 Within our long-term program of enantioselec-
tive synthesis of some quinolizidine alkaloids, we
attempted to address the need of sparteine analogues.
In particular, we focused our attention to find out an
original, stereoselective synthetic methodology for the
preparation of chiral diamines, such as 2 and 3, which
embody, respectively, two different portions (ABC and
BCD rings) of the sparteine framework. Being aware of
difficulties in the synthesis of sparteine itself encoun-
tered by previous workers in the field,5 we thought that
these simplified structures would have been more easily
achievable, moreover in both the enantiomeric forms,
than the natural alkaloid. Most of the chiral architec-
ture of sparteine is retained in diamines 2 and 3, but, at
the same time, the lack of A or D ring, respectively,
significantly modifies their steric hindrances. So, the
efficacy of these diamines appears to be hardly pre-
dictable as sparteine surrogates. Some preparations of
racemic 2 and 3 have been previously reported, based
on different synthetic strategies. For instance, the com-
plete control of the ABC and BCD ring relative stereo-
chemistry in the synthesis of sparteine analogues has
been achieved by O’Brien.6 Quite recently, the same
Sparteine is
a
lupine alkaloid extracted from
papilionaceous plants such as Scotch Broom. It bears a
bis-quinolizidine skeleton, with trans- and cis-quino-
lizidine moieties joined through a common bridge. With
regard to the origin of enantioselectivity during the
deprotonation process, steric effects in the diamine/
organolithium complex have been proved to be impor-
tant. For instance, the use of isosparteine, the
C2-symmetric structural analogue of sparteine consist-
ing of two trans-quinolizidine moieties, appears to slow
down the deprotonation reaction significantly and pro-
vides low enantioselectivity.
Since (+)-sparteine is also naturally occurring, but far
less easily obtained,3 this alkaloid is only commercially
available as its (−)-antipode. This limitation on the use
of sparteine as a chiral ligand prompts to find other
chiral diamines capable of matching the enantioselectiv-
Figure 1.
* Corresponding authors. Tel.: +39 2 50314080; fax: +39 2 50314078; e-mail: alessandra.silvani@unimi.it
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)01671-4

7156
B. Danieli et al. / Tetrahedron Letters 43 (2002) 7155–7158
author developed a resolution protocol which furnishes
chiral 2 with moderate yield and 60% enantiomeric
eccess.7 In this paper, we report the application of an
original, imino Diels–Alder8-based methodology to the
first enantioselective synthesis of tricyclic diamines 2
and 3, preparable in both the enantiomeric forms start-
ing from the chiral cis-piperidine-3,5-dimethanol
monoacetate 4. This precursor is readily available,
together with its enantiomer, by mean of biocatalytic
asymmetrization of the corresponding CS-symmetric
diols or diacetates.9 Recently, we succeeded in develop-
ing some convenient enantioselective entries to highly
functionalized 3,7-diazabicyclo[3.3.1]nonane (bispidine)
derivatives,10 which allowed us to set up an original
route to chiral, non-racemic functionalized piperido–
quinolizidine ring systems.11 Along the line of our
previous results, we devised an innovative strategy for
the preparation of 2 and 3, relied on the application of
imino-Diels–Alder reactions as the key step. In Scheme
1, two different paths are outlined, both of them based
on an imino-Diels–Alder route, but differing in the
order of construction of rings B and C. As a conse-
quence of our previous work on the enantiosynthesis of
3,7-diazabicyclo[3.3.1]nonane derivatives,10 the path a
involves the cycloaddition between the Danishefsky’
diene and the chiral bicyclic hydroxyamine 7 or hydroxy-
lactam 8, from which an imine or acylimine function-
ality can be generated, respectively, by Lewis acid-
catalyzed elimination. In path b, the dissection of ring
B in compounds 2 and 3 yields to the 2-piperidinyl-
2,3-dihydropyridin-4-one 9, which in turn could be
accessed by cycloaddition of imine 10 with the same
diene. For the preparation of imine 10, the chiral cis-
piperidine-3,5-dimethanol monoacetate 4, which was
already described by us,9 can be employed.
5 and 6. In path b, the question about the stereochem-
ical consequences of addition remained to be addressed,
due to the lack of information about the geometry of
the reacting imine and the steric hindrance at the two
faces of the imine moiety.
Our initial work was directed to study the feasibility of
path a. Preliminary results have been obtained on the
reactivity of 7 as a dienophile. By using the Danishef-
sky’ diene under the catalysis of TiCl4, the cycloaddi-
tion product 5 has been isolated, but only in traces.
Many attempts to improve the yield, using a variety of
Lewis acid catalysts, including lanthanide triflates, and
different oxygenated dienes, were unsuccessful. In some
cases, using the (1-tert-butoxy-buta-1,3-dienyloxy)-
trimethylsilane as diene, we could only observe the
competing vinilogue Mannich addition, also in this case
in low yield. Because of the LUMO-dienophile control
to which the imino Diels–Alder reactions are subjected,
we expected an enhanced reactivity for the acylimine
derived from the 2-oxo analogue 8. We found that on
TiCl4-catalyzed treatment with Danishefsky’ diene, 8
gave the cycloadduct 6, still in very low yield. A possi-
ble explanation of the scarce reactivity of 7 and 8 as
dienophiles in the Lewis acid-catalyzed reactions is that
the cation of the Lewis acid might give a bidentate
binding through N3 and N7 of the diazabicycle, thus
preventing the formation of the planar reactive imine or
acylimine moiety. Actually, when 8 was refluxed with
Danishefsky’ diene in toluene for two days, without any
acidic catalysis, the tricycle 6 was formed in some
higher yield (35%). Longer reaction times, higher tem-
peratures and larger excesses of diene had little effect
upon the amount of adduct formed.
The structure of 6 was established by NMR spectro-
scopic analysis12 which demonstrated, to the limits of
detection by 1H NMR, that the reaction is highly
stereoselective, affording only the (6S) stereoisomer,
derived from an endo attack on the si-face of the imino
moiety. The stereochemistry of 6 was deduced from the
observation of an almost negligible coupling (ca. 0.5
Hz) between 6-H (at l=3.93) and 7-H (at l=1.92) and
then supported by NOE interactions of 6-H at l=3.93
with 13e-H at l=4.16, with 5e-H at l=2.30 and with
7-H at l=1.92 (Fig. 2).
In path a, the stereochemical outcome of the cycloaddi-
tion should be easily predictable, since imino
dienophiles in which configuration is fixed by virtue of
ring restraints (such as those derivable from 7 and 8)
usually give stereochemically defined adducts resulting
from endo addition. Moreover, we expected the
cycloaddition to occur predominantly from the less-hin-
dered, convex, methylene bridgehead face of the bis-
pidine nucleus to give ultimately a (6S) configuration
(lupine alkaloids numbering) as indicated in structures
Scheme 1.

B. Danieli et al. / Tetrahedron Letters 43 (2002) 7155–7158
7157
metric, we also examined the one-pot procedure, reling
on a three-component coupling reaction between alde-
hyde, amine and alkene via imine formation and imino
Diels–Alder reaction by using lanthanide triflate as a
catalyst.14 In the presence of 10 mol% of Sc(OTf)3 and
MgSO4, aldehyde 11 was treated with benzylamine and
Danishefsky’ diene in acetonitrile at room temperature.
The reaction proceeded smoothly to afford 9 in very
high yield. Also in this case, 9 was obtained as an
inseparable mixture of adducts 9a and 9b in about a 1:1
ratio (estimated by 1H NMR). Evidently, the chiral
nature of aldehyde 11 does not play a crucial role in the
stereochemical course of the reaction, since there is not
any facial differentiation in the endo approach of the
diene to the imine group. However, this result turned
out to be favourable to us, in that it allows for a
synthetic access to both the diastereoisomeric com-
pounds 2 and 3, at the same time. The mixture of
adducts 9a and 9b was submitted to hydrolysis of the
acetate group and then treated with MsCl to afford
quantitatively the corresponding mesylates 12a and 12b.
Finally, catalytic hydrogenation was performed, in
order to remove the Cbz and benzyl protecting groups
and at the same time to reduce the C5ꢀC6 double bond.
Subsequent heating in THF afforded tricycles 13a and
13b in high yield. Lastly, reduction of the carbonyl
group followed by a reductive methylation at the N12
atom by means of NaBH3CN in the presence of formal-
dehyde yielded diastereoisomeric diamines 2 and 3, at
last separable by careful chromatography. Diamine
(6R,7S,9R)-2 and diamine (6S,7S,9R)-3 show [h]D=
−20 (c 1, CHCl3) and [h]D=−12 (c 1, CHCl3), respec-
tively. Their NMR and MS spectral data resulted in full
agreement with those published for the corresponding
racemic compounds.6b,c
Figure 2. Selected NOE interactions detected by NOE differ-
ence studies (300 MHz, CDCl3, 50°C) of 6.
As a consequence of these unsatisfactory results, which
did not allow to complete the synthesis efficiently, we
considered the alternative strategy (path b) depicted in
Scheme 1, based on the construction of the ring B in
the last steps of the synthetic sequence, by formation of
the N1ꢀC10 bond.
We hoped that carrying out the imino Diels–Alder
cycloaddition before constructing the diazabicy-
clo[3.3.1] nonane framework would have been more
effective, even if predictions on the stereochemical
course of the reaction were more difficult to make in
this case. The aldehyde 11 was converted to the
required imine 10 by treatment with benzylamine in
CH2Cl2 in the presence of MgSO4 or, alternatively, of 4
,
A molecular sieves (Scheme 2).
Due to its lability, this intermediate was not isolated.
After exchange of the solvent, it was immediately sub-
jected to the Lewis acid-mediated reaction with Dan-
ishefsky’ diene. The cycloaddition reaction takes place
from −80°C to room temperature, typically over a
period of 16 h in the presence of 2 M equiv. of ZnCl2
as Lewis acid in dry THF. The intermediate tetra-
hydropyridine derivative could not be isolated, as it
underwent hydrolysis during the aqueous work up
required to eliminate the Lewis acid catalyst. The dihy-
dropyridone 9 was thus obtained directly in moderate
yield, as a 1:1 mixture of R and S diastereoisomers at
C2.13 Bearing in mind that one synthetic problem in the
imino Diels–Alder reactions is the imines’ stability
under the influence of Lewis acids more than stoichio-
A preliminary study was performed, in which the s-
BuLi complexes with 2 and 3 were tested for reactivity
in the enantioselective deprotonation of N-Boc pyrro-
lidine 14 and subsequent trapping with Me3SiCl (Fig.
3).15 Since almost no lithiation occurs in the absence of
a diamine, this case is especially significant for assaying
the utility of new ligands. Using diamine 2, we obtained
a 63% isolated yield of silylated pyrrolidine (S)-15 of
94% ee (by chiral GC), practically comparable to the
Scheme 2. Reagents and conditions: (a) DMSO, (COCl)2, Et3N, 90%; (b) benzylamine, Danishefsky’ diene, MgSO4, 10 mol%
Sc(OTf)3, CH3CN, rt, 89%; (c) NaOH, MeOH, then MsCl, Et3N, CH2Cl2, 76%; (d) Pd/C, H2, EtOH, 2 equiv. HClaq, then Et3N,
THF, reflux, 65%; (e) TsNHNH2, EtOH, reflux, then NaBH4, THF/H2O, reflux, 75%; (f) NaBH3CN, CH2Oaq, THF, 81%, then
chromatographic separation (eluant: CH2Cl2/MeOH 95/5, NH3 1%), 38% (2), 44% (3).

7158
B. Danieli et al. / Tetrahedron Letters 43 (2002) 7155–7158
References
1. For a review, see: Hoppe, D.; Hense, T. Angew. Chem.,
Int. Ed. Engl. 1997, 36, 2282.
2. For recent developments, see: (a) Deiters, A.; Fro¨lich, R.;
Hoppe, D. Angew. Chem., Int. Ed. 2000, 39, 2105; (b)
Johnson, T. A.; Curtis, M. D.; Beak, P. J. Am. Chem.
Soc. 2001, 123, 1004.
3. Ebner, T.; Eichelbaum, M.; Fisher, P.; Meese, C. O.
Arch. Pharm. (Weinheim) 1989, 322, 399.
4. (a) Gallagher, D. J.; Wu, S.; Nikolic, N. A.; Beak, P. J.
Org. Chem. 1995, 60, 8148; (b) Li, X.; Schenkel, L. B.;
Kozlowski, M. C. Org. Lett. 2000, 2, 875.
5. (a) Otomasu, H.; Takatsu, N.; Noguchi, M.; Ohmiya, S.
Chem. Pharm. Bull. 1987, 35, 4990; (b) Wanner, M. J.;
Koomen, G.-J. J. Org. Chem. 1996, 61, 5581; (c) Wendt,
J. A.; Aube´, J. Tetrahedron Lett. 1996, 37, 1531.
6. (a) Harrison, J. R.; O’Brien, P.; Porter, D. W.; Smith, N.
M. J. Chem. Soc., Perkin Trans. 1 1999, 3623; (b) Har-
rison, J. R.; O’Brien, P. Tetrahedron Lett. 2000, 41, 6161;
(c) Harrison, J. R.; O’Brien, P. Tetrahedron Lett. 2000,
41, 6167.
Figure 3.
76% yield and 97% ee which we achieved by (−)-
sparteine. This good result entirely confirms the prelim-
inary encouraging evidences provided by O’Brien et
al.,7 in their studies using a partially resolved diamine
rac-2 (55% ee).
7. Harrison, J. R.; O’Brien, P.; Porter, D. W.; Smith, N. M.
Chem. Commun. 2001, 1202.
Examination of diamine 3 in the same reaction led to a
lower reactivity (31% yield of (S)-15) and to a disap-
pointingly low enantioselectivity (21% ee), providing
anyway insight into the important features relevant to
the stereoselectivity in this deprotonation–substitution
reaction. The ABC ring system of sparteine seems to be
required in order to maintain the capacity of bidentate
coordination of a cation, while the D ring not only
appears unnecessary, but even it works against the
favourable conformation for coordination in the BCD
ring system. More interestingly, it has to be noted that
the tricyclic diamine ent-2, easily accessible from ent-49
as described in Scheme 2, would represent a valuable
surrogate of the commercially unavailable (+)-sparteine,
for enantioselective lithiation–electrophilic quench
reactions.
8. For reviews, see: (a) Weinreb, S. M.; Staib, R. R. Tetra-
hedron 1982, 38, 3087–3128; (b) Waldmann, H. Synthesis
1994, 6, 535–551; (c) Jørgensen, K. A. Angew. Chem., Int.
Ed. 2000, 39, 3558–3588.
9. Danieli, B.; Lesma, G.; Passarella, D.; Silvani, A. J. Org.
Chem. 1998, 63, 3492.
10. Danieli, B.; Lesma, G.; Passarella, D.; Silvani, A.; Vivi-
ani, N. Tetrahedron 1999, 55, 11871.
11. Consonni, A.; Danieli, B.; Lesma, G.; Passarella, D.;
Piacenti, P.; Silvani, A. Eur. J. Org. Chem. 2001, 1377.
12. Selected data for 6: [h]D=−5 (c 1, CHCl3). 1H NMR (300
MHz, CDCl3): l 8.30 (d, 1H, J=8.1 Hz), 7.32 (m, 5H),
5.52 (d, 1H, J=8.1 Hz), 5.07 (s, 2H), 4.43 (br d, 1H,
J=13.4 Hz), 4.34 (br d, 1H, J=13.4 Hz), 3.93 (dd, 1H,
J=15.3, 3.6 Hz), 3.01 (br d, 2H, J=13.8 Hz), 2.70 (t, 1H,
J=15.5 Hz), 2.30 (dd, 1H, J=15.5, 3.3 Hz), 2.12 (br d,
J=13.5 Hz), 1.92 (m, 1H), 1.71 (br d, J=13.5 Hz); 13C
NMR (CDCl3, 75.4 MHz): l 205.7, 169.1, 156.0, 142.5,
136.0, 128.4, 128.0, 109.7, 62.2, 60.5, 58.9, 56.4, 43.7,
39.2, 33.1, 25.6; FAB+MS m/z: 341 [MH+].
In summary, the first enantioselective synthesis of tri-
cyclic diamines 2 and 3, bearing the same absolute
stereochemistry as the ABC and BCD rings of the
naturally occurring alkaloid (−)-sparteine, has been
reported. The key step of the route is a lanthanide
triflate-catalyzed imino Diels–Alder reaction. The
availability of both the enantiomeric forms (4 and
ent-4) of the chiral starting material would allow for an
access to these diamine ligands as both antipodes.
Work is now in progress in order to provide a totally
stereoselective route to diamine 2 and to its enantiomer,
in view of its high capacity of matching the enantiose-
lectivity of sparteine.
13. Selected data for 9a,b: 1H NMR (300 MHz, CDCl3): l
7.50–7.20 (m, 10H), 7.12 (d, 1H, J=8.0 Hz), 5.07 (s, 2H),
4.96 (d, 1H, J=8.0 Hz), 4.55–4.10 (m, 4H), 3.99 (dd, 1H,
J=11.4, 5.1 Hz), 3.84 (dd, 1H, J=11.4, 6.8 Hz), 3.34 (m,
1H), 2.76–2.25 (m, 4H), 2.03 (s, 3H), 2.20–1.85 (m, 3H),
1.15 (q, 0.5H, J=12.4 Hz), 0.94 (q, 0.5H, J=12.4 Hz);
13C NMR (CDCl3, 75.4 MHz, signals in brackets refer to
the same carbon in a and b diastereoisomers): l (189.9,
189.8), 170.7, 155.0, (153.6, 152.8), (136.5, 136.2), 129.1,
128.4, 127.9, 127.7, 127.2, (98.0, 97.8), 67.2, 66.0, 59.3,
58.3, (47.1, 46.9), 45.8, 38.8, (37.0, 36.4), (36.0, 35.6),
(30.4, 29.9), 20.6; FAB+MS m/z: 477 [MH+].
Acknowledgements
14. Kobayashi, S.; Ishitani, H.; Nagayama, S. Synthesis 1995,
1195.
15. (a) Beak, P.; Kerrick, S. T.; Wu, S.; Chu, J. J. Am. Chem.
Soc. 1994, 116, 3231; (b) Gallagher, D. J.; Wu, S.;
Nikolic, N. A.; Beak, P. J. Org. Chem. 1995, 60, 8148.
This work was supported by Ministero dell’Universita` e
della Ricerca Scientifica e Tecnologica (MURST).