Total synthesis of (±)-α-isosparteine, (±)-β- isosparteine, and (±)-sparteine from a common tetraoxobispidine intermediate

Article abstract of DOI:10.1021/jo8013512

(Chemical Equation Presented) The three title alkaloids were separately prepared in stereocontrolled fashion from a common tetraoxobispidine precursor, 3,7-diallyl-2,4,6,8-tetraoxo-3,7-diazabicyclo[3.3.1]nonane (16). Bisimide 16 was generated from malonate via acid promoted cyclization of the Knoevenagel condensation adduct 1,1,3,3-propanetetracarboxamide. (±)-α- Isosparteine (dl-2) was elaborated from 16 in 28% overall yield by a two-directional synthetic sequence composed of four reactions: double addition of allylmagnesium bromide, ring-closing olefin metathesis (RCM), hydrogenation, and borane mediated reduction. (±)-β-Isosparteine (dl-3) was targeted along similar lines by a strategic reversal in allylation and reduction operations on the core synthon. Thus, 16 was advanced to dl-3 in five steps and 12% overall yield by a reaction sequence commencing with sodium borohydride mediated reduction and followed by double Sakurai-type allylation of the resulting bishemiaminal. The synthesis of dl-3 was concluded by RCM and then global reduction (H₂, Pd/C; LiAlH₄). The final target, (±)-sparteine (dl-1), was secured in six steps and 11% overall yield from 16 by monoreduction and Sakurai allylation, followed by allyl Grignard addition and then RCM and global reduction as before. Reasons for the inherent C ₂-type regioselectivity of net double nucleophilic additions to tetraoxobispidines are discussed and enantioselective oxazaborolidine mediated reduction of the N,N′-dibenzyl congener of 16 is reported.

Full text of DOI:10.1021/jo8013512

Total Synthesis of (()-r-Isosparteine, (()-ꢀ-Isosparteine, and
(()-Sparteine from a Common Tetraoxobispidine Intermediate†
Neil R. Norcross,§ John P. Melbardis,‡ Margarita Ferris Solera,§ Mark A. Sephton,‡
Colin Kilner,§ Lev N. Zakharov,‡ Peter C. Astles, Stuart L. Warriner,§ and
Paul R. Blakemore*,‡
Department of Chemistry, Oregon State UniVersity, CorVallis, Oregon 97331-4003, School of Chemistry,
UniVersity of Leeds, Leeds, West Yorkshire LS2 9JT, United Kingdom, and Eli Lilly & Co. Ltd., Lilly
Research Centre, Erl Wood Manner, Windlesham, Surrey, GU20 6PH, United Kingdom
paul.blakemore@science.oregonstate.edu
ReceiVed June 25, 2008
The three title alkaloids were separately prepared in stereocontrolled fashion from a common
tetraoxobispidine precursor, 3,7-diallyl-2,4,6,8-tetraoxo-3,7-diazabicyclo[3.3.1]nonane (16). Bisimide 16
was generated from malonate via acid promoted cyclization of the Knoevenagel condensation adduct
1,1,3,3-propanetetracarboxamide. (()-R-Isosparteine (dl-2) was elaborated from 16 in 28% overall yield
by a two-directional synthetic sequence composed of four reactions: double addition of allylmagnesium
bromide, ring-closing olefin metathesis (RCM), hydrogenation, and borane mediated reduction. (()-ꢀ-
Isosparteine (dl-3) was targeted along similar lines by a strategic reversal in allylation and reduction
operations on the core synthon. Thus, 16 was advanced to dl-3 in five steps and 12% overall yield by a
reaction sequence commencing with sodium borohydride mediated reduction and followed by double
Sakurai-type allylation of the resulting bishemiaminal. The synthesis of dl-3 was concluded by RCM
and then global reduction (H2, Pd/C; LiAlH4). The final target, (()-sparteine (dl-1), was secured in six
steps and 11% overall yield from 16 by monoreduction and Sakurai allylation, followed by allyl Grignard
addition and then RCM and global reduction as before. Reasons for the inherent C2-type regioselectivity
of net double nucleophilic additions to tetraoxobispidines are discussed and enantioselective oxazaboro-
lidine mediated reduction of the N,N′-dibenzyl congener of 16 is reported.
Introduction
by a naturally occurring alkaloid (e.g., 1-3, Figure 1).3 The
eponymous and most abundant sparteine alkaloid (1), prevalent
as the levorotatory enantiomorph in many common papiliona-
ceous plants (e.g., Scotch broom Cytius scoparius), is distin-
guished by a nonsymmetric exo-endo arrangement of hydrogen
atoms at C6 and C11 (sparteine numbering). Stenhouse de-
scribed the original isolation of (-)-sparteine (lupinidine, l-1)
in 1851 and was the first to determine its molecular formula
(as C15H26N2).4 Building on the work of Ing,5 Winterfeld,6 and
others,2 the intriguing tetracyclic structure of sparteine was
correctly deduced by Clemo and Raper some 82 years later,7
The lupine alkaloids constitute a structurally diverse group
of quinolizidine bases found in a variety of leguminous plant
and tree species, including broom, lupin, gorse, and laburnum.1
The sparteine subgroup of lupine alkaloids are identified by a
shared 3,11-diazatetracyclo[7.7.1.03,8.011,16]heptadecane ring
system characterized by a bispidine nucleus with peripheral
quinolizidine motifs flanking its core.2 Three diastereomeric
forms for this inherently chiral tetracyclic array are geometrically
accessible, and each possible isomeric variation is represented
† Dedicated to the memory of Professor Nelson J. Leonard (1916-2006) on
the occasion of the 60th anniversary of the Leonard-Beyler synthesis of sparteine.
‡ Oregon State University.
(2) For definitive accounts of the early history of the sparteine subgroup,
see: (a) Leonard, N. J. In The Alkaloids; Manske, R. H. F., Ed.; Academic Press:
New York, 1953; Vol. 3, pp 119-200. (b) Leonard, N. J. In The Alkaloids;
Manske, R. H. F., Ed.; Academic Press: New York, 1960; Vol. 7, pp 253-317.
(3) Wink, M.; Meibner, C.; Witte, L. Phytochemistry 1995, 38, 139–153.
(4) Stenhouse, J. Ann. 1851, 78, 1.
§ University of Leeds.
Eli Lilly & Co. Ltd.
(1) Reviews: (a) Michael, J. P. Nat. Prod. Rep. 2007, 24, 191–222. (b)
Ohmiya, S.; Saito, K.; Murakoshi, I. In The Alkaloids; Cordell, G. A., Ed.;
Academic Press: New York, 1995; Vol. 47, pp 1-114. See also earlier reports
in both of these regular series.
(5) (a) Ing, H. R. J. Chem. Soc. 1933, 504–510. (b) Ing, H. R. J. Chem. Soc.
1932, 2778–2780.
10.1021/jo8013512 CCC: $40.75
Published on Web 09/18/2008
2008 American Chemical Society
J. Org. Chem. 2008, 73, 7939–7951 7939

Norcross et al.
FIGURE 2. Significant conformations of sparteine (1) and as bidentate
metal ligand.
identified it as the final stereoisomer in the sparteine series.17
The equivalence of pusilline and l-spartalupine was established
soon thereafter by Marion and Greenhalgh thus removing earlier
confusion in the literature.18 More recently, (+)-ꢀ-isosparteine
(d-3) has been found in a variety of Lupinus species.19 Various
chemical methods for interconversion between sparteine dias-
tereoisomers within a given enantiomeric series have been
reported.20,21
FIGURE 1. Three naturally occurring levorotatory lupine alkaloids of
the sparteine subgroup (l-1, l-2, l-3) and embedded structural motifs
(4, 5) found within these molecules.
Applications. The biological activity of the sparteine alka-
loids is limited, but not insignificant.22 (-)-Sparteine (l-1) is a
cardiac agent with moderate toxicity23 and it has been inves-
tigated as a potential drug for the management of arrhythmias24
and as an agent for the induction of uterine contractions.25
Although of little medicinal importance, (-)-sparteine (l-1), and
to a lesser extent (-)-R-isosparteine (l-2), nevertheless elicit
intense interest because of their deployment as uniquely effective
chiral ligands in a wide variety of metal-mediated enantiose-
lective synthetic methods.26 It was recognized early on that the
“chair-chair” conformer of sparteine (1cc), calculated (by DFT)
to be 3.4 kcal mol-1 higher in energy than its ground state
“chair-boat” conformation 1cb,27 could behave as a bidentate
metal ligand (Figure 2). Within the resulting metal complexes
1cc · M, the metal ion is partially encapsulated by the chiral
alkaloid framework and provided with a dissymmetric coordina-
tion sphere suitable for the promotion of enantioselective
transformations. In this regard, now rarely cited seminal reports
while stereochemical detail for this compound (and related C15
lupine alkaloids) was only finally clarified by Marion and
Leonard a century after its first isolation.8,9 It is a popular
misconception that the dextrorotatory enantiomorph of sparteine
is not a natural product; however, although more narrowly
distributed than its better known and commercially available
antipode, (+)-sparteine (pachycarpine, d-1) is biosynthesized
by numerous plant species (e.g., Cytisus caucasicus).10,11
Traditionally, (+)-sparteine (d-1) has been most conveniently
obtained from natural (()-lupanine (2-oxosparteine), extracted
from the seeds of Lupinus albus,12 by classical resolution
followed by reduction of the resulting (-)-lupanine.13
The remaining two diastereomeric forms of the parent
sparteine ring system, R-isosparteine (2) and ꢀ-isosparteine (3),
are C2-symmetric and differ by a respective exo-exo and
endo-endo configuration of hydrogen atoms at C6 and C11.
(-)-R-Isosparteine (genisteine, l-2) was first unequivocally
obtained by Winterfeld via semisynthesis from (-)-sparteine
(l-1),14 and only later found to be a naturally occurring substance
(in Lupinus caudatus) by Marion and co-workers.15 Interestingly,
(+)-R-isosparteine (d-2) has yet to be identified as a natural
plant alkaloid. (-)-ꢀ-Isosparteine (l-spartalupine, pusilline, l-3)
belongs to the same enantiomorphic series as (+)-sparteine (d-
1) and was originally isolated (as “pusilline”) from Lupinus
pusillus (also a source of d-1 and (-)-lupanine) by Marion and
Fenton but mis-identified at that time as a dihydrosparteine.16
Carmack and co-workers later obtained the same compound
from Lupinus sericeus (as “l-spartalupine”) and correctly
(17) Carmack, M.; Douglas, B.; Martin, E. W.; Suss, H. J. Am. Chem. Soc.
1955, 77, 4435.
(18) Greenhalgh, R.; Marion, L. Can. J. Chem. 1956, 34, 456–458.
(19) (a) El-Shazly, A.; Ateya, A.-M. M.; Wink, M. Z. Naturforsch. C:
J. Biosci. 2001, 56, 21–30. (b) Van Wyk, B.-E.; Greinwald, R.; Witte, L. Biochem.
Syst. Ecol. 1995, 23, 533–537.
(20) Mercuric ion oxidation of (-)-sparteine (l-1) leads successively to
(-)-∆5-dehydrosparteine and (-)-∆5,11-didehydrosparteine. Reduction of the
former enamine regenerates l-1, while reduction of the latter affords (-)-R-
isosparteine (l-2). (+)-ꢀ-Isosparteine (d-3) is similarly oxidized to the same
enamine intermediates and is therefore a viable synthetic progenitor to either
l-1 or l-2. See ref 14 and: Leonard, N. J.; Thomas, P. D.; Gash, V. W. J. Am.
Chem. Soc. 1955, 77, 1552–1558.
(6) (a) Winterfeld, K.; Ipsen, W. Arch. Pharm. 1930, 268, 372–380. (b)
Winterfeld, K. Arch. Pharm. 1929, 267, 433–455. (c) Winterfeld, K. Arch. Pharm.
1928, 266, 299–325.
(7) The tetracyclic ring system of sparteine was correctly identified by Clemo
and Raper at this time, but not its relative stereochemistry, which was erroneously
designated to be that now attributed to R-isosparteine, see: (a) Clemo, G. R.;
Raper, R. J. Chem. Soc. 1933, 644–645. (b) Clemo, G. R.; Raper, R.; Tenniswood,
C. J. Chem. Soc. 1931, 429–437.
(21) A mixture of (-)-R-isosparteine (l-2) (major) and (+)-ꢀ-isosparteine
(d-3) (minor) is obtained by heating (-)-sparteine (l-1) with AlCl3 at 180-230
°C, see: (a) Galinovsky, F.; Knoth, P.; Fischer, W. Monatsh. Chem. 1955, 86,
1014–1023. (b) Winterfeld, K.; Bange, H.; Lalvani, K. S. Ann. Chem. 1966,
698, 230–234.
(22) Schmeller, T.; Wink, M. In Alkaloids: Biochemistry, Ecology, and
Medicinal Applications; Roberts, M. F., Wink, M., Eds.; Plenum Press: New
York, 1998; p 435.
(8) Marion, L.; Leonard, N. J. Can. J. Chem. 1951, 29, 355–362.
(9) For the assignment of absolute configuration to (-)-sparteine and related
compounds, see: Okuda, S.; Kataoka, H.; Kyosuke, T. Chem. Pharm. Bull. 1965,
13, 491–500.
(10) Orekhov, A. P.; Norkina, S. S.; Maksimova, T. Arch. Pharm. 1935,
273, 369–372.
(11) At the time of writing, at least two U.S. chemical companies list (+)-
sparteine (as pachycarpine) in their catalogues (100 g quantities) at a price ca.
three times greater than that typical for (-)-sparteine.
(12) More recently, (+)-sparteine itself has been discovered to be a minor
alkaloid of Lupinus albus, see: Wysocka, W. Sci. Legumes 1995, 2, 137–140.
(13) Ebner, T.; Eichelbaum, M.; Fischer, P.; Meese, C. O. Arch. Pharm.
1989, 322, 399–403.
(23) Seeger, R.; Neumann, H. G. Inst. Pharmakol. Toxikol. 1992, 132, 1577–
1581.
(24) (a) Ruenitz, P. C.; Mokler, C. M. J. Med. Chem. 1977, 20, 1668–1671.
(b) Senges, J.; Ehe, L. Arch. Pharmacol. 1973, 280, 265–274. (c) Reuter, N.;
Heeg, E.; Haller, U. Arch. Pharmacol. 1971, 268, 323–333.
(25) (a) Van Voorhus, L. W.; Dunn, L. J.; Heggen, D. Am. J. Obstet. Gynecol.
1966, 94, 230–233. (b) Gawecka, I.; Szonert, M. Acta Phys. Pol. 1969, 20, 165–
172.
(26) For authoritative reviews of sparteine alkyllithium reagent pairs in
enantioselective synthesis, see: (a) Hoppe, D.; Christoph, G. In Chemistry of
Organolithium Compounds; Rappoport, Z., Marek, I., Eds.; Wiley: Chichester,
UK, 2004; Vol. 2, pp 1055-1164. (b) Hoppe, D.; Hense, T. Angew. Chem., Int.
Ed. Engl. 1997, 36, 2282–2316.
(14) Winterfeld, K.; Rauch, C. Arch. Pharm. 1934, 272, 273–290.
(15) Marion, L.; Turcotte, F.; Ouellet, J. Can. J. Chem. 1951, 29, 22–29.
(16) Marion, L.; Fenton, S. W. J. Org. Chem. 1948, 13, 780–781.
(27) (a) Galasso, V.; Asaro, F.; Berti, F.; Kovac, B.; Habus, I.; Sacchetti, A.
Chem. Phys. 2003, 294, 155–169. (b) Wiberg, K. B.; Bailey, W. B. J. Mol.
Struct. 2000, 556, 239–244.
7940 J. Org. Chem. Vol. 73, No. 20, 2008

Total Synthesis of Sparteine Group Alkaloids
by Nozaki et al.,28 and others,29 first highlighted a wealth of
possibilities for asymmetric induction using sparteine diamines
in the presence of all manner of organometallic reagents.
Pioneering contributions from Hoppe30,31 and Beak32,33 have
since led to the introduction of highly enantioselective trans-
formations employing complexes of (-)-sparteine (l-1) and
alkyllithium reagents. These complexes are typically used to
effect enantioselective deprotonation (via either kinetic control
or “dynamic thermodynamic” resolution)34 leading to scalemic
carbanionic species that are quenched with electrophilies to
generate enantioenriched products of interest.35 Many significant
implementations of this concept have been reported, including
homoaldol30a and related reactions of allylic carbamates,31c
enantioselective transformations of oxiranes36 and N-Boc aza-
cycles,37 synthesis of enantioenriched ferrocenes,38 and asym-
metric chain extension of boronic esters.39 Other recent devel-
opments have seen the use of (-)-sparteine (l-1) as a ligand
for palladium in the Sigman-Stoltz enantioselective oxidative
resolution of 2° alcohols,40,41 and as a ligand for magnesium in
the asymmetric ring-opening of cyclic meso anhydrides by aryl
SCHEME 1. O’Brien’s (+)-Sparteine Surrogate (7) and a
BCD-Ring Analogue of (-)-Sparteine (8)
Grignard reagents.42 (-)-Sparteine (l-1) is also the ligand of
choice in the Crimmins variant of the Ti-mediated Evans aldol
reaction; however, in this application it does not play a
stereodirecting role.43
(-)-R-Isosparteine (l-2) is also employed in enantioselective
synthesis but this C2-symmetric diamine is only rarely found
to be a superior chiral ligand to its nonsymmetric counterpart
(-)-sparteine (l-1).44 Metal complexes of l-2 (i.e., l-2cc·M
analogous to l-1cc·M above) are more sterically encumbered
than those of l-1 and this facet compromises their kinetic
competence and stability.45
The limited availability of both (+)-sparteine (d-1) and (+)-
R-isosparteine (d-2) restricts the scope of asymmetric method-
ologies which rely on sparteine alkaloids as stereoinduction
elements. Accordingly, there has been a long standing interest
in either a concise asymmetric synthesis of d-1, or else, a readily
accessible equivalent to this privileged ligand. O’Brien’s recently
introduced diamine 7 (Scheme 1), a structurally simplified
tricyclic analogue of (+)-sparteine (d-1) that mimics its ABC-
ring motif, offers a practical solution to this problem.46 Diamine
7, obtained via semisynthesis from the abundant lupine alkaloid
(-)-cytisine (6),47,48 has now been evaluated as a complemen-
tary pseudoenantiomeric ligand to (-)-sparteine (l-1) in many
asymmetric transformations and shown to offer similar enan-
tioselectivity but in the opposite sense.49 Interestingly, an
isomeric diamine (8) that mimics instead the BCD-ring system
of (-)-sparteine (l-1) was significantly inferior to 7 in the
enantioselective lithiation/silylation of N-Boc-pyrrolidine.50
These results suggest that the sparteine D-ring plays a relatively
minor role in stereoinduction; moreover, recent work by
Kozlowski and co-workers has clarified the importance of an
intact A-ring moiety for high enantioselectivity in the application
of sparteine-like diamines.51
(28) (a) Nozaki, H.; Aratani, T.; Noyori, R. Tetrahedron Lett. 1968, 9, 2087–
2090. (b) Nozaki, H.; Aratani, T.; Toraya, T. Tetrahedron Lett. 1968, 9, 4097–
4098. (c) Aratani, T.; Gonda, T.; Nozaki, H. Tetrahedron Lett. 1969, 10, 2265–
2268. (d) Aratani, T.; Gonda, T.; Nozaki, H. Tetrahedron 1970, 26, 5453–5464.
(e) Nozaki, H.; Aratani, T.; Toraya, T.; Noyori, R. Tetrahedron 1971, 27, 905–
913.
(29) For a selection of other early examples of (-)-sparteine and (-)-R-
isosparteine as metal ligands in synthesis, see: (a) Eberhardt, G. G.; Butte, W. A.
J. Org. Chem. 1964, 29, 2928–2932. (b) Kretchmer, R. A. J. Org. Chem. 1972,
37, 2744–2747. (c) Guette´, M.; Capillon, J.; Guette´, J.-P. Tetrahedron 1973, 29,
3659–3667. (d) Trost, B. M.; Dietsche, T. J. J. Am. Chem. Soc. 1973, 95, 8200–
8201. (e) Reetz, M. T.; Westermann, J. Synth. Commun. 1981, 11, 647–654. (f)
Okamoto, Y.; Suzuki, K.; Kitayama, T.; Yuki, H.; Kageyama, H.; Miki, K.;
Tanaka, N.; Kasai, N. J. Am. Chem. Soc. 1982, 104, 4618–4624.
(30) Key early work: (a) Hoppe, D.; Zschage, O. Angew. Chem., Int. Ed.
Engl. 1989, 28, 67–69. (b) Zschage, O.; Schwark, J. R.; Hoppe, D. Angew. Chem.,
Int. Ed. Engl. 1990, 29, 296–297. (c) Hoppe, D.; Hintze, F.; Tebben, P. Angew.
Chem., Int. Ed. Engl. 1990, 29, 1422–1424.
(31) Recent work with leading references: (a) Becker, J.; Fro¨hlich, R.;
Salorinne, K.; Hoppe, D. Eur. J. Org. Chem. 2007, 3337–3348. (b) Chedid, R. B.;
Fro¨hlich, R.; Hoppe, D. Org. Lett. 2006, 8, 3061–3064. (c) Martinez, M. M.;
Hoppe, D. Org. Lett. 2004, 6, 3743–3746. (d) Beckmann, E.; Desai, V.; Hoppe,
D. Synlett 2004, 2275–2280.
(32) Key early work: (a) Kerrick, S. T.; Beak, P. J. Am. Chem. Soc. 1991,
113, 9708–9710. (b) Beak, P.; Basu, A.; Gallagher, D. J.; Park, Y. S.;
Thayumanavan, S. Acc. Chem. Res. 1996, 29, 552–560.
(33) Recent work with leading references: (a) Park, Y. S.; Yum, E. K.; Basu,
A.; Beak, P. Org. Lett. 2006, 8, 2667–2670. (b) Lee, S. J.; Beak, P. J. Am.
Chem. Soc. 2006, 128, 2178–2179. (c) Kim, D. D.; Lee, S. J.; Beak, P. J. Org.
Chem. 2005, 70, 5376–5386.
(34) Beak, P.; Anderson, D. R.; Curtis, M. D.; Laumer, J. M.; Pippel, D. J.;
Weisenburger, G. A. Acc. Chem. Res. 2000, 33, 715–727.
To the best of our knowledge, neither enantiomorph of
ꢀ-isosparteine (3) has received documented attention as a chiral
(35) (a) Clayden, J. Organolithiums: SelectiVity for Synthesis; Pergamon: New
York, 2002. (b) Basu, A.; Thayumanavan, S. Angew. Chem., Int. Ed. 2002, 41,
716–738.
(36) (a) Vrancken, E.; Alexakis, A.; Mangeney, P. Eur. J. Org. Chem. 2005,
1354–1366. (b) Hodgson, D. M.; Maxwell, C. R.; Miles, T. J.; Paruch, E.;
Matthews, I. R.; Witherington, J. Tetrahedron 2004, 60, 3611–3624. (c) Hodgson,
D. M.; Cameron, I. D. Org. Lett. 2001, 3, 441–444.
(42) Shintani, R.; Fu, G. C. Angew. Chem., Int. Ed. 2002, 41, 1057–1059.
(43) Crimmins, M. T.; King, B. W.; Tabet, E. A.; Chaudhary, K. J. Org.
Chem. 2001, 66, 894–902.
(44) For recent applications of (-)-R-isosparteine in enantioselective synthesis
and analysis, see: (a) Allen, B. D.; Cintrat, J.-C.; Faucher, N.; Berthault, P.;
Rousseau, B.; O’Leary, D. J. J. Am. Chem. Soc. 2005, 127, 412–420. (b)
Hodgson, D. M.; Galano, J.-M.; Christlieb, M. Tetrahedron 2003, 59, 9719–
9728. (c) Mu¨ller, P.; Patrice, N.; Bernardinelli, G. Eur. J. Org. Chem. 2001,
4137–4147. (d) Kang, J.; Cho, W. O.; Cho, H. G. Tetrahedron: Asymmetry 1994,
5, 1347–1352.
(45) Wu¨rthwein, E.-U.; Hoppe, D. J. Org. Chem. 2005, 70, 4443–4451.
(46) Dearden, M. J.; Firkin, C. R.; Hermet, J.-P. R.; O’Brien, P. J. Am. Chem.
Soc. 2002, 124, 11870–11871.
(47) Dixon, A. J.; McGrath, M. J.; O’Brien, P.; Holle, S.; Fu¨rstner, A. Org.
Synth. 2006, 83, 141–154.
(48) A de novo synthesis of ent-7 was recently reported: Hermet, J.-P. R.;
Viterisi, A.; Wright, J. M.; McGrath, M. J.; O’Brien, P.; Whitwood, A. C.; Gilday,
J. Org. Biomol. Chem 2007, 5, 3614–3622.
(49) (a) Dearden, M. J.; McGrath, M. J.; O’Brien, P. J. Org. Chem. 2004,
69, 5789–5792. (b) Hermet, J.-P. R.; Porter, D. W.; Dearden, M. J.; Harrison,
J. R.; Koplin, T.; O’Brien, P.; Parmene, J.; Tyurin, V.; Whitwood, A. C.; Gilday,
J.; Smith, N. M. Org. Biomol. Chem. 2003, 1, 3977–3988.
(50) Danieli, B.; Lesma, G.; Passarella, D.; Piacenti, P.; Sacchetti, A.; Silvani,
A.; Virdis, A. Tetrahedron Lett. 2002, 43, 7155–7158.
(37) Selected recent examples: (a) Coldham, I.; Patel, J. J.; Raimbault, S.;
Whittaker, D. T. E.; Adams, H.; Fang, G. Y.; Aggarwal, V. K. Org. Lett. 2008,
10, 141–143. (b) Coldham, I.; Dufour, S.; Haxell, T. F. N.; Patel, J. J.; Sanchez-
Jimenez, G. J. Am. Chem. Soc. 2006, 128, 10943–10951. (c) Kocienski, P. J.;
Christopher, J. A.; Bell, R.; Otto, B. Synthesis 2005, 75–84. See also ref 33b.
(38) (a) Metallinos, C.; Szillat, H.; Taylor, N. J.; Snieckus, V. AdV. Synth.
Catal. 2003, 345, 370–382. (b) Tsukazaki, M.; Tinkl, M.; Roglans, A.; Chapell,
B. J.; Taylor, N. J.; Snieckus, V. J. Am. Chem. Soc. 1996, 118, 685–686.
(39) (a) Stymiest, J. L.; Dutheuil, G.; Mahmood, A.; Aggarwal, V. K. Angew.
Chem., Int. Ed. 2007, 46, 7491–7494. (b) Besong, G.; Jarowicki, K.; Kocienski,
P. J.; Sliwinski, E.; Boyle, F. T. Org. Biomol. Chem. 2006, 4, 2193–2207.
(40) Discovery: (a) Jensen, D. R.; Pugsley, J. S.; Sigman, M. S. J. Am. Chem.
Soc. 2001, 123, 7475–7476. (b) Ferreira, E. M.; Stoltz, B. M. J. Am. Chem. Soc.
2001, 123, 7725–7726. Selected mechanistic investigations: (c) Mueller, J. A.;
Sigman, M. S. J. Am. Chem. Soc. 2003, 125, 7005–7013. (d) Trend, R. M.;
Stoltz, B. M. J. Am. Chem. Soc. 2004, 126, 4482–4483. (e) Mueller, J. A.; Cowell,
A.; Chandler, B. D.; Sigman, M. S. J. Am. Chem. Soc. 2005, 127, 14817–14824.
(41) For discourse on this type of “stereoablative” process, see: Mohr, J. T.;
Ebner, D. C.; Stoltz, B. M. Org. Biomol. Chem. 2007, 5, 3571–3576.
(51) Phuan, P.-W.; Ianni, J. C.; Kozlowski, M. C. J. Am. Chem. Soc. 2004,
126, 15473–15479.
J. Org. Chem. Vol. 73, No. 20, 2008 7941

Norcross et al.
analogous to the Leonard-Beyler intermediate 9 have featured
in two stereocontrolled elaborations of sparteine.57,58
The biosynthetic theory of Robinson59 inspired a number of
early “biomimetic” approaches to the sparteine nucleus in which
Mannich-type processes were used to forge C6-C7/C9-C11
or C9-C10/C7-C17 bonds from suitable iminium ion precur-
sors. The first ostensibly viable route to (()-sparteine (dl-1)
along these lines, and analogous to Robinson’s landmark
tropinone synthesis,60 was reported by Anet, Hughes, and
Ritchie in 1950;61 however, the veracity of the synthesis was
called into question.2b,62 An unambiguous “biogenetic-type”
route to dl-1 was later reported by van Tamelen and Foltz, in
which double Mannich adduct 10 was further converted to (()-
8-oxosparteine, and hence the target alkaloid itself, by oxidative
cyclization followed by Wolff-Kishner reduction (eq 2 in
Figure 3).63 Similar approaches have been disclosed,64,65 but
the more recent work of Koomen and Wanner, in which both
(()-sparteine (dl-1) and (()-ꢀ-isosparteine (dl-3) were elabo-
rated from tetrahydroanabasine,66 represents the most truly
biomimetic synthesis of these alkaloids described to date.67
The C2-symmetric isomers of sparteine have rarely been the
deliberate targets of total synthesis. A notable exception is a
remarkably concise synthesis of (()-R-isosparteine (dl-2)
reported by Kakisawa and co-workers in 1983 (eq 3 in Figure
3).68 Again, an “outside-in” strategy was employed and pen-
tacycle 11, formed diastereoselectively via consecutive nitrone-
olefin cycloadditions between 2 equiv of 2,3,4,5-tetrahydropy-
ridine-N-oxide and 4H-pyran, gave the target molecule directly
upon reductive cyclization (in an unspecified yield).
FIGURE 3. Salient features of four notable synthetic approaches to
sparteine group alkaloids.
ligand for enantioselective synthesis. However, given that (+)-
ꢀ-isosparteine (d-3) is configurationally related to the ineffectual
BCD-tricyclic analogue of (-)-sparteine (8), it is unlikely to
offer useful levels of enantioselectivity in typical applications
employing chiral diamine ligands.
In 2002, Aube´ and co-workers described an enantioselective
total synthesis of (+)-sparteine (d-1) that highlighted the utility
Existing Total Syntheses. The first total synthesis of (()-
sparteine (dl-1) was described by Leonard and Beyler in
1948.52,53 Since that time, many other efforts toward sparteine
and its isomers have been disclosed; however, none match the
synthetic brevity of the original two-step synthesis. In common
with the Leonard-Beyler approach, a significant majority of
known sparteine syntheses may be loosely categorized as
“outside-in” strategies, wherein the central B- and C-rings of
the tetracyclic nucleus are fabricated from compounds replete
with pre-existing A- and D-ring substructure (Figure 3). For
example, in a second generation variant of their original
procedure, Leonard and Beyler transformed 2,4-dipyrid-2-yl
glutarate 9 (available from ethyl 2-pyridylacetate via Knoev-
enagel condensation with paraformaldehyde) directly into a
mixture of (()-sparteine (dl-1) and (()-R-isosparteine (dl-2)
by high-pressure hydrogenative reductive cyclization over a
copper chromite catalyst (eq 1 in Figure 3).52b Closely related
contemporaneous efforts by other groups led to similarly
nonstereoselective syntheses of the sparteine bases,53,54 including
the first recognized preparation of (()-ꢀ-isosparteine (dl-3).55,56
Recently, stereodefined 2,4-dipiperid-2-yl glutarate precursors
(56) Carmack and co-workers later repeated the Sorm-Keil synthesis of
sparteine (ref 54) and succeeded in converting one of the 10,17-dioxosparteine
isomers produced by that approach into (()-ꢀ-isosparteine via LiAlH4 reduction,
see ref 17.
(57) O-Brien’s recently reported elaboration of (-)-sparteine (l-1) is the
second enantiocontrolled synthesis of this important alkaloid, see: Hermet, J.-
P. R.; McGrath, M. J.; O’Brien, P.; Porter, D. W.; Gilday, J. Chem. Commun.
2004, 1830–1831.
(58) (a) Buttler, T.; Fleming, I. Chem. Commun. 2004, 2404–2405. (b) Buttler,
T.; Fleming, I.; Gonsior, S.; Kim, B.-H.; Sung, A.-Y.; Woo, H.-G. Org. Biomol.
Chem. 2005, 3, 1557–1567.
(59) (a) Robinson, R. J. Chem. Soc., Trans. 1917, 111, 876–899. (b)
Robinson, R. The Structural Relations of Natural Products; Oxford University
Press: London, UK, 1955.
(60) Robinson, R. J. Chem. Soc., Trans. 1917, 111, 762–768.
(61) (a) Anet, E.; Hughes, G. K.; Ritchie, E. Nature 1950, 165, 35–36. (b)
Anet, E. F. L. J.; Hughes, G. K.; Ritchie, E. Aust. J. Sci. Res. 1950, 3, 635–641.
(62) Scho¨pf, C. L.; Benz, G.; Braun, F. R.; Hinkel, H.; Rokohl, R. Angew.
Chem. 1953, 65, 161–162.
(63) (a) van Tamelen, E. E.; Foltz, R. L. J. Am. Chem. Soc. 1960, 82, 2400.
(b) van Tamelen, E. E.; Foltz, R. L. J. Am. Chem. Soc. 1969, 91, 7372–7377.
(64) (a) Bohlmann, F.; Mu¨ller, H.-J.; Schumann, D. Chem. Ber. 1973, 106,
3026–3034. (b) Takatsu, N.; Noguchi, M.; Ohmiya, S.; Otomasu, H. Chem.
Pharm. Bull. 1987, 35, 4990–4992.
(65) Somewhat related recent elaborations of (()-anagyrine and (()-
thermopsine by Gallagher and Gray constitute formal syntheses of (()-sparteine,
(()-R-isosparteine, and (()-ꢀ-isosparteine, see: Gray, D.; Gallagher, T. Angew.
Chem., Int. Ed. 2006, 45, 2419–2423.
(66) Tetrahydroanabasine is biosynthesized from cadaverine, the known
natural precursor of sparteine, see: (a) Golebiewski, W. M.; Spenser, I. D. Can.
J. Chem. 1985, 63, 2707–2718. (b) Golebiewski, W. M.; Spenser, I. D. Can.
J. Chem. 1988, 66, 1734–1748. (c) Brown, A. M.; Rycroft, D. S.; Robins, D. J.
J. Chem. Soc., Perkin Trans. 1 1991, 2353–2355.
(52) (a) Leonard, N. J.; Beyler, R. E. J. Am. Chem. Soc. 1948, 70, 2298–
2299. (b) Leonard, N. J.; Beyler, R. E. J. Am. Chem. Soc. 1950, 72, 1316–1323.
(53) Clemo and co-workers succeeded in preparing 17-oxosparteine by a
related multi-step route as early as 1936, but could not convert this material
into sparteine itself with the reducing reagents available at the time. The same
group later completed a true total synthesis of sparteine following the introduction
of LiAlH4, see: (a) Clemo, G. R.; Morgan, W. McG.; Raper, R. J. Chem. Soc.
1936, 1025–1028. (b) Clemo, G. R.; Raper, R.; Short, W. S. J. Chem. Soc. 1949,
663–665.
(67) Wanner, M. J.; Koomen, G.-J. J. Org. Chem. 1996, 61, 5581–5586.
(68) (a) Oinuma, H.; Dan, S.; Kakisawa, H. Chem. Commun. 1983, 654–
655. (b) Oinuma, H.; Dan, S.; Kakisawa, H. J. Chem. Soc., Perkin Trans. 1
1990, 2593–2597.
(69) (a) Smith, B. T.; Wendt, J. A.; Aube´, J. Org. Lett. 2002, 4, 2577-
2579. For preliminary studies, see: (b) Wendt, J. A.; Aube´, J. Tetrahedron Lett.
1996, 37, 1531–1534.
(54) Sorm, F.; Keil, B. Collect. Czech., Chem. Commun. 1948, 13, 544–
556.
(55) Tsuda, K.; Satoh, Y. Pharm. Bull. 1954, 2, 190–193.
(70) For a commentary on this work, see: Iyengar, R.; Gracias, V. Chemtracts
2004, 17, 92–96.
7942 J. Org. Chem. Vol. 73, No. 20, 2008

Total Synthesis of Sparteine Group Alkaloids
SCHEME 2. A Retrosynthetic Analysis of the Sparteine
Group Alkaloids (1-3)
SCHEME 3. Synthesis of
2,4,6,8-Tetraoxo-3,7-diazabicyclo[3.3.1]nonane (22) and
Related Studies
FIGURE 4. Orthographic views of the ORTEP diagram for 2,4,6,8-
tetraoxo-3,7-diazabicyclo[3.3.1]nonane (22) with parameters of interest.
Indicated “exo” and “endo” angles are appropriate intersection angles
betweenthethreeleast-squaresfittedplanescontaining{C1,C2,N3,C4,C5},
{C5,C6,N7,C8,C1}, and {C1,C9,C5}. Fifty percent probability el-
lipsoids are plotted for non-hydrogen atoms.
opportunities for novel discoveries. Retrosynthetic analysis
guided by this consideration led to the identification of a
tetraoxobispidine (such as 16) as an ideal platform from which
to elaborate all principal lupine alkaloids of the sparteine group
(Scheme 2). Thus, it was recognized that bisimide 16, represent-
ing the central sparteine BC-rings, is potentially amenable to
the controlled annulation of rings A and D via nucleophilic
allylation processes followed by ring-closing olefin metathesis
(RCM) and thence exhaustive reduction from intermediate
tetraallyl compounds 15 of appropriate oxidation level and
stereochemistry. From the outset, exo face selective additions
to the bicyclic and tetracyclic intermediates of interest were
anticipated; however, the regiochemical outcome of successive
nucleophilic additions to bisimides such as 16 was far less
certain. In the event, reduction of this plan to practice was
successful and accounts of our studies concerning the syntheses
of (()-R-isosparteine (dl-2)73 and (()-ꢀ-isosparteine (dl-3)74
were previously communicated. Herein, we report full details
for these two efforts with new improvements, the first disclosure
of a synthesis of (()-sparteine (dl-1) along similar lines, and
also preliminary work directed at rendering this kind of approach
enantioselective.75
of the alkyl Schmidt reaction.69,70 Remarkably, this 21st century
work constituted the first ever asymmetric synthesis of any
alkaloid of the sparteine subgroup and, prior to our studies (vide
infra), it was further distinguished by being the only known
effort to deviate from the traditional “outside-in”-type strategy.
Rather, the outer A- and D-rings were successively annulated
on to a bicyclic template representing the central BC-rings of
the target in an “inside-out” fashion. Thus, scalemic C2-
symmetric diketone 12, wrought from achiral norbornadiene by
enantioselective desymmetrization via a three-step hydrosily-
lation/oxidation sequence,71 was advanced to alkyl azide 13 as
a prelude to a stereospecific Schmidt-type nitrenoid insertion
yielding tricycle 14 (eq 4 in Figure 3).72 Stereochemistry at C6
(sparteine numbering) was set by hydrogenation of an exocyclic
olefin en route to 13, and ketone 14was further converted into
(+)-sparteine (d-1) by an 8-step sequence incorporating dias-
tereoselective exo face alkylation at C11 and an unusual photo-
Beckmann rearrangement to effect a second net formal nitrene
insertion into C11-C17.
Results and Discussion
Synthetic Plan. Given the historical and practical relevance
of the sparteine alkaloids, we sought a concise synthetic scheme
capable of targeting any member of this special group of
molecules in a stereocontrolled fashion. As highlighted above,
many “outside-in” approaches to the sparteine bases had been
previously evaluated and we felt that an “inside-out” strategy
conceptually akin to the Aube´ synthesis of (+)-sparteine (d-
1),69 and progressing via sequential desymmetrization of an
initially achiral C2V-symmetric core synthon, would offer greater
Synthesis of Tetraoxobispidines. The viability of our
proposed synthetic scheme hinged on securing access to pivotal
tetraoxobispidine intermediates such as 16. Tetraoxobispidines
(i.e., 2,4,6,8-tetraoxo-3,7-diazabicyclo[3.3.1]nonanes) substituted
at the methylene bridge are well described in the literature, and
have been used as precursors to bispidines with antiarrhythmic
(73) Blakemore, P. R.; Kilner, C.; Norcross, N. R.; Astles, P. C. Org. Lett.
2005, 7, 4721–4724.
(74) Blakemore, P. R.; Norcross, N. R.; Warriner, S. L.; Astles, P. C.
Heterocycles 2006, 70, 609–617.
(75) Blakemore, P. R.; Norcross, N. R.; Astles, P. C. Presented in part at the
232nd ACS National Meeting, San Francisco, CA, September 10-14, 2006:
Abstracts of Papers; American Chemical Society: Washington, DC, 2006;
ORGN#751.
(71) Hayashi, T. Acta Chem. Scand. 1996, 50, 259–266.
(72) For a notable recent application of the intramolecular alkyl Schmidt
reaction, see: (a) Tani, K.; Stoltz, B. M. Nature 2006, 441, 731–734. (b) Clayden,
J.; Moran, W. J. Angew. Chem., Int. Ed. 2006, 45, 7118–7120.
J. Org. Chem. Vol. 73, No. 20, 2008 7943

Norcross et al.
TABLE 1. N,N′-Dialkylation of Tetraoxobispidines 20 and 22
starting
entry
material
X
reaction conditions
product
R
yield, %
1
2
3
4
5
6
7
8
9
22
22
22
22
22
22
22
22
22
20
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CH2
CMe2
allyl-Br, K2CO3, acetone, ∆
allyl-Br, K2CO3, MeCN, ∆
allyl-Br, K2CO3, MeOH, ∆
allyl-Br, NaOMe, MeOH, rt
allyl-OH, DIAD, PPh3, THF, rt
allyl-Br, NaH, DMF, rt
16
16
16
16
16
16
25
26
27
28
allyl
allyl
allyl
allyl
allyl
allyl
Bn
crotyl
prenyl
Bn
29
43
dec
dec
61
83
83
46
51
63
BnBr, NaH, DMF, rt
crotyl-Br, NaH, DMF, rt
prenyl-Br, NaH, DMF, rt
BnBr, K2CO3, acetone, ∆
10
properties;76 however, at the outset of our studies, the parent
compound 22 had featured in only a single century old report
by Guthzeit and Jahn.77 The bisimides of interest are tradition-
ally prepared by acidic hydrolysis of R,R′-dicyanoglutarimides
(“Guareschi imides” e.g., 18), which are in turn conveniently
synthesized by the Guareschi condensation: a Knoevenagel-type
three-component coupling reaction involving coalescence of a
cyanoacetate, a carbonyl compound, and ammonia.78 We found
this procedure entirely acceptable for the synthesis of known
9,9-dimethyltetraoxobispidine (20),76 via meso Guareschi imide
18,79 but attempts to extend the same process to the synthesis
of 22 where wholly unsuccessful (Scheme 3). Thus, the
previously undescribed Guareschi imide 19 was not an identifi-
able component of the reaction mixtures that were obtained from
addition of either formalin or paraformaldehyde to ethyl
cyanoacetate (17) in ethanolic ammonia. The failure of this
transformation was attributed to the high reactivity of the
putative R-cyanoacrylic intermediates which readily polymerize.
Other attempts to synthesize 19 along similar lines via the
Clemens variant of the Guareschi condensation were equally
fruitless,80 as was a related direct cyano-group hydrolysis route
to 22 from the malononitrile derived tetracyanide 23.81 In this
case, exposure of 23 to acidic media (including aq AcOH, HCl,
or H2SO4) gave no reaction at low-to-moderate temperatures,
but caused decomposition to volatile byproducts under more
forcing conditions (g100 °C) with simple ammonium salts
making up the reaction residue.
not encouraging. Following the originally prescribed protocol,77
brief heating of 21 beyond its melting point at subambient
pressure (20-50 mmHg) instigated the vigorous evolution of a
gaseous stream clearly containing both NH3 and H2O, the latter
indicating that competitive undesired cyclization modes to
isoimide forms of 22 were potentially in operation. Upon
cessation of bubbling, the darkly colored resinous pyrrolyte was
allowed to cool and subsequent trituration and recrystallization
from EtOH allowed small quantities of the crystalline bisimide
22 to be isolated (4-6%). Glutarimide (24), the undoubted result
of retro-Claisen-type bond scissions en route to 22, was
consistently identified as a significant component of the mass
balance (ca. 10-40% depending on duration of heating), but
the formation of other simple molecules (e.g., urea) during this
crude transformation cannot be discounted. Notwithstanding the
poor efficiency of the cyclocondensation, it proved impossible
to reliably execute beyond single gram quantities of 21, either
by conventional heating or in a focused microwave reactor.
Reasoning that the presence of liberated free ammonia would
be deleterious to the survival of 22, and that some sort of
electrophilic activation would facilitate cyclization of 21 in the
desired manner, an acid mediated variant of the Guthzeit
cyclization was explored. Sulfuric acid proved unsatisfactory
for this purpose, but reasonable results were obtained by fusing
21 with toluenesulfonic acid (7-10% 22), or (better) by heating
a mechanically stirred paste prepared from this amide and
methanesulfonic acid (23-33% 22). The still modest yield of
the transformation in MsOH is offset by its operational
simplicity, reproducibility over a range of scales (essentially
identical results at 5, 100, and 300 mmol), and the ease with
which the desired bisimide product may be conveniently isolated
from the reaction mixture in pure form by simple trituration
with MeOH.
Guthzeit’s long neglected synthesis of 22 via direct pyrrolysis
of tetraamide 21 presented an obvious solution to the problem
at hand. Indeed, given the extreme ease with which 21 may be
prepared from dimethyl malonate,82 we had initially regarded
this approach to 22 as attractive, but preliminary results were
(76) Scho¨n, U.; Antel, J.; Bru¨ckner, R.; Messinger, J. J. Med. Chem. 1998,
41, 318–331.
As one might anticipate given its highly symmetrical cage-
like structure, the parent tetraoxobispidine (22) is a beautifully
crystalline material and good quality colorless needles are readily
generated upon recrystallization from H2O (mp 295 °C, dec).
X-ray crystallographic analysis of 22 was performed for the first
time and revealed only a slight deviation from perfect C2V-
symmetry (Figure 4). Near planar imide subunits with well-
differentiated exo and endo faces are clearly evident, with the
(77) Guthzeit, M.; Jahn, C. J. Prakt. Chem. 1902, 66, 1–15. For an English
language abstract, see: J. Chem. Soc., Abstr. 1902, 82 (I), 658-659.
(78) The early discoveries of Guareschi were later followed and more
extensively explored by Thorpe and Vogel, see: (a) Guareschi, I. Gazz. Chim.
Ital. 1919, 49, 124–133. (b) Kon, G. A. R.; Thorpe, J. F. J. Chem. Soc. 1919,
686–704. (c) Vogel, A. I. J. Chem. Soc. 1934, 1758–1765.
(79) Holder, R. W.; Daub, J. P.; Baker, W. E.; Gilbert, R. H.; Graf, N. A. J.
Org. Chem. 1982, 47, 1445–1451.
(80) The Clemens variant of the Guareschi reaction calls for base mediated
condensation of a pre-formed (ꢀ-substituted) R-cyanoacrylate with cyanoaceta-
mide; however, given the instability of ꢀ-unsubstituted R-cyanoacrylates, the
isolable dicyclohexylamine salt of R-cyanoacrylic acid (ref 80c) was utilized in
our attempts to access 19 via this route, see: (a) McElvain, S. M.; Clemens,
D. H. J. Am. Chem. Soc. 1958, 80, 3915–3923. (b) McElvain, S. M.; Clemens,
D. H.; Parham, W. E.; Kirklin, P. W.; Noland, W. E. Org. Synth. 1959, 39,
52–53. (c) Krawczyk, H. Synth. Commun. 2000, 30, 657–664.
(81) Bell, R. A.; Brown, B. E.; Duarte, M.; Howard-Lock, H. E.; Lock,
C. J. L. Can. J. Chem. 1987, 65, 261–270.
(82) See the Experimental Section for details concerning an improvement
to our earlier method (ref 73) for the preparation of propane-1,1,3,3-tetracar-
boxamide (21). See also: Gogoll, A.; Johansson, C.; Axe´n, A.; Grennberg, H.
Chem. Eur. J. 2001, 7, 396.
7944 J. Org. Chem. Vol. 73, No. 20, 2008

Total Synthesis of Sparteine Group Alkaloids
SCHEME 4. Synthesis of (()-r-Isospartine (dl-2) from
Tetraoxobispidine 16
former some 16° more open than the latter. Of further interest,
the four carbonyl carbon atoms of 22 lie in a common plane
and constitute a virtually flawless molecular square (as il-
lustrated); the C9 bridge atom lies 1.82 Å above the exact center
of this square. The NMR spectral signature of 22 (in d6-DMSO)
is in accord with expectation: the 13C NMR spectrum exhibits
three signals only (δC 166.9, 47.2, and 22.8 ppm), while the
small magnitude of mutual scalar coupling (t, J3 ) 2.8 Hz)
between anisochronous methylene bridge (δH 2.59 ppm) and
1
methine ring-junction (δH 3.64 ppm) protons in the H NMR
spectrum is consistent with geometrically imposed fixed dihedral
angles of 60((0.3)°.
The conversion of bisimide 22 into various N,N′-dialkyl
derivatives of potential utility en route to the sparteine alkaloids
was next investigated (Table 1). These studies further high-
lighted the propensity of tetraoxobispidines unsubstituted at the
(C9) bridge position to suffer base-promoted decarboxylative
fragmentation. For example (entry 1), treatment of 22 with allyl
bromide in the presence of K2CO3 in refluxing acetone gave a
poor yield of the desired product 16 (29%), which was
accompanied (among other unidentified minor decomposition
products) by glutarimide 29 (10%). The identity of this
byproduct reveals that once the bicyclic bisimide core has been
opened, enolization of any remaining ꢀ-dicarbonyl moieties may
readily occur. Analogous C-alkylation of intact tetraoxobis-
pidines is precluded because formation of the necessary enolate
species would violate Bredt’s rule.83,84
attention was next directed to advancing these materials into
the three different diastereomeric forms of the sparteine ring
system.
Synthesis of (()-r-Isosparteine. According to the prescribed
plan (Scheme 2), a successful synthesis of (()-R-isosparteine
(dl-2) would necessitate regiocontrolled double nucleophilic
allylation of bisimide 16, followed at some later juncture by
exo face selective reduction of a bishemiaminal intermediate
(i.e., 15, X ) Y ) OH, or an RCM derived adduct thereof).
Pleasingly, after some experimentation, it was discovered that
addition of excess allylmagnesium bromide to bisimide 16
generated a bishemiaminal product (30) of the desired symmetry
type in an efficient manner (Scheme 4). Unwanted Cs-type
regioisomers were not significant components of the crude
product mixture (<5% by NMR analysis), and bicycle 30 was
directly isolated as a single diastereoisomer by recrystalliza-
tion.87 Since the first report of this transformation,73 an X-ray
crystallographic analysis has confirmed our original assumption
that both hydroxyl groups within 30 are endo (R) configurated
(see the Supporting Information). It should be noted, however,
that the now established stereochemical assignment for 30 does
not necessarily imply that allylation of 16 occurred with exo
face kinetic diastereoselectivity. Epimerization of the anomeric
stereogenic centers in 30 is a mechanistic possibility, and indeed,
later compounds in the synthetic sequence en route to R-isos-
parteine showed a propensity for such behavior.
Given the low solubility of 22 in acetone, similar alkylations
were conducted in more polar solvents; however, it was evident
that methoxide was inimical to the desired process (entries 2-4).
A Mitsunobu dehydrative coupling85 between 22 and allyl
alcohol in THF gave better results (entry 5), but ultimately it
was discovered that allylation of 22 was best and most simply
performed in the traditional manner by employing NaH as base
and DMF as solvent (entry 6). Following this procedure, key
synthon 16 was isolated in pure form by trituration and without
recourse to chromatography in 83% yield.86 Benzyl (25), crotyl
(26), and prenyl (27) derivatives of 22 were similarly prepared
(entries 7-9). By contrast to the parent tetraoxobispidine (22),
alkylation of the 9,9-dimethyl congener (20) was less problem-
atic and proceeded as indicated without significant decomposi-
tion (entry 10). Presumably, nucleophilic attack on the carbonyl
groups of bisimide 20 is retarded by the presence of flanking
methyl substituents on the bridge.
Double ring-closing olefin metathesis (RCM) of tetraene 30
with Grubbs’ first generation catalyst (“Grubbs 1”)88 occurred
without incident to afford the tetracyclic sparteine congener 31
as an insoluble solid. Spontaneous precipitation of 31 from the
reaction mixture allowed for its easy isolation by means of
filtration. NMR spectral analysis of 31 suggested that the
compound retained the C2-symmetry of its precursor;89 however,
With a satisfactory and scalable route to N,N′-diallyltetraoxo-
bispidine (16) and related bisimides successfully developed,
(83) (a) Ko¨brich, G. Angew. Chem. 1973, 85, 494–503. (b) Buchanan, G. L.
Chem. Soc. ReV. 1974, 3, 41–63.
(84) Successful R-metalation at bridge-head positions in related [3.3.1]
systems is possible in the presence of stronger bases, see: Giblin, G. M. P.; Kirk,
D. T.; Mitchell, L.; Simpkins, N. S. Org. Lett. 2003, 5, 1673–1675, and references
cited therein.
(85) (a) Ahn, C.; Correia, R.; DeShong, P. J. Org. Chem. 2002, 67, 1751–
1753. (b) Hughes, D. L. Org. React. 1992, 42, 335–656. (c) Mitsunobu, O.
Synthesis 1981, 1–28.
(86) For X-ray crystallographic analyses of compounds 16 and 31, see ref
73.
(87) Attempted purification of 30 by column chromatography (SiO2) led to
significantly reduced yields, possibly due to the generation of acyl iminium ions
and thence decomposition.
(88) Herein, “Grubbs 1” ) (Cy3P)2Cl2RudCHPh; “Grubbs 2” ) (Cy3P)(H2-
IMes)Cl2RudCHPh. For recent reviews of olefin metathesis with these catalysts,
see: (a) Grubbs, R. H. Tetrahedron 2004, 60, 7117–7140. (b) Grubbs, R. H.
Handbook of Metathesis; Wiley-VCH: Weinheim, Germany, 2003.
J. Org. Chem. Vol. 73, No. 20, 2008 7945

Norcross et al.
an X-ray diffraction analysis of the crystalline form of 31
indicated an exo/endo arrangement of hydroxyl groups in the
solid state.86 These data are not incompatible and are consistent
with the existence of a solution phase dynamic equilibrium
between 31 anomers that is perturbed upon selective deposition
of the illustrated nonsymmetrical isomer during crystallization.
Lending credence to this hypothesis, subsequent hydrogenative
saturation of the olefinic moieties within 31 revealed the
possibility for interconversion between epimeric 6,11-dihydroxy-
10,17-dioxosparteines (32) under mild conditions. Thus, the
polarity of the hydrogenation reaction medium was noted to
markedly affect the final isomeric ratio for the isolated reduction
product. In a very polar solvent mixture (MeOH-H2O, 5:1),
hydrogenation of 31 over Pd/C (at 1 atm) gave the nonsymmetric
bishemiaminal a-32 as a single isomer in near quantitative yield
following its isolation by trituration. By contrast, execution of
an otherwise identical procedure in a less polar solvent
combination (EtOH-EtOAc, 2:1) afforded the symmetric
product s-32 with endo configurated hydroxyl groups (as
supported by both NMR and X-ray crystallographic analysis,
see the Supporting Information). Yet another hydrogenation
reaction of 31 in EtOH (not illustrated) gave a 1:1 mixture of
a-32 and s-32. In no example were alkoxy groups incorporated
at the anomeric positions of the product and it is therefore
evident that epimerization involved ring-opened amido keto
tautomers rather than the formation of acyl iminium ions. The
latter species were, however, implicated in the unexpectedly
facile dehydration of diol s-32 to enamide 33 upon its exposure
to silica gel during a (superfluous) chromatographic purification
operation. Attempts to reduce this new compound to 10,17-
dioxo-R-isosparteine by analogy to the known conversion of
(-)-∆5,11-didehydrosparteine to l-220 did not meet with im-
mediate success and were not further pursued.90 The more direct
conversion of 30 to 32 by a tandem RCM/hydrogenation
procedure was also briefly entertained, but failed due to the
insolubility of diene 31 in 1,2-dichloroethane.91
Conclusion of the total synthesis of (()-R-isosparteine (dl-
2) required deoxygenation of a-32 or s-32 with stereocontrolled
exomode introduction of hydride at C6 and C11. In the event,
this objective was successfully accomplished from either anomer
by the action of a large excess of borane tetrahydrofuran
complex. Presumably, identical transient acyl iminium ions were
generated during the course of both transformations, with
exogenous hydride delivery occurring in the desired and
anticipated manner.92 The reduction was significantly more
efficient from a-32 than s-32, but in each case, the free base of
dl-2 spontaneously crystallized as its hydrate from alkaline
solution during workup, greatly facilitating isolation of the
racemic alkaloid target. The identity of the crystalline material
obtained was confirmed by comparison to spectroscopic and
physical data previously recorded from both natural and
synthetic samples of R-isosparteine.93
Synthesis of (()-ꢀ-Isosparteine. A two-directional approach
to (()-ꢀ-isosparteine (dl-3) was pursued along similar lines to
the synthesis of its symmetry related congener, (()-R-isos-
parteine (dl-2). To target the alternate endo-endo configuration
of C6/C11 hydrogen atoms presented by dl-3, a strategic reversal
in the order of introduction of allyl and hydride nucleophiles
onto a tetraoxobispidine template was called for. Accordingly,
following a preliminary survey of potentially suitable reducing
reagents, it was observed that reaction of N,N′-dibenzyl bisimide
25 with an excess of lithium triethylborohydride (Super Hy-
dride)94 proceeded in the desired regiochemical mode to afford
bishemiaminal 34 as the double R-anomer. Other hydride
sources, including diisobutylaluminum hydride (DIBAL-H) and
sodium bis(methoxyethoxy)aluminum hydride (Red-Al), gave
inferior results and intractable byproduct, while simple boranes
(9-BBN and catecholborane) failed to react. Exposure of bishem-
iaminal 34 to boron trifluoride etherate in the presence of
allyltrimethylsilane resulted in an exo-face selective double
Sakurai-type allylation process affording bislactam 35. The stereo-
chemical outcome of this transformation, evident from
the low magnitude of scalar coupling between vicinal bridgehead
and CHN methine protons (<1 Hz, as indicated),95 is a further
testament to the predictable nature of nucleophilic addition to
acyl iminium cations derived from constrained bispidine systems.
With a viable strategy in place for setting the desired regio-
and stereochemistry, extension of the same two-step protocol
to N,N′-diallyl bisimide 16, a precursor better suited for eventual
advancement to the akaloid of interest, was attempted. In the
event, 16 proved to be a significantly less robust substrate than
25 (possibly due to the absence of sterically encumbering groups
on imide N-atoms) and its exposure to excess lithium triethyl-
borohydride gave complex mixtures from which only traces of
the desired N,N′-diallyl bishemiaminal 38 could be identified
(e11% yield). As previously outlined, this problem was initially
solved by resorting to a circuitous four-step sequence (two
iterations of LiBHEt3 monoreduction/mono-Sakurai allylation)
that yielded 39 in 9-14% overall yield from 16.74 It has since
been discovered that 39 can be obtained from bisimide 16 more
directly via sodium borohydride mediated reduction followed
by Sakurai allylation of the resulting cocktail of unpurified
hemiaminals 36 and 38 (Scheme 5). This step-saving protocol
afforded a mixture of triene 37 and tetraene 39 in higher net
yield (33% and 16%, respectively) than the less convenient four-
step sequence. Accordingly, that the mixture of 37 and 39
(93) For compiled 13C NMR spectral data for sparteine alkaloids 1-3, see:
(a) Galasso, V.; Asaro, F.; Berti, F.; Kovac, B.; Habus, I.; Sacchetti, A. Chem.
Phys. 2003, 294, 155–169. (b) Mikhova, B.; Duddeck, H. Magn. Reson. Chem.
1999, 36, 779–796.
(94) Brown, H. C.; Krishnamurthy, S. J. Am. Chem. Soc. 1973, 95, 1669–
1671.
(95) The Karplus relation predicts a very low J value in this case as a
consequence of a near 90° dihedral angle between the bridgehead proton and
endo-configurated CHN methine proton in bislactam 35. A significantly larger
J value is expected for isomeric compounds wherein the same CHN methine
proton is exo-configurated, e.g., as observed for bishemiaminal 34. For a review
of the Karplus relation, see: Minch, M. J. Concept. Maget. Reson. 1994, 6, 41–
56.
(89) The 13C NMR (75 Hz, d6-DMSO) spectral signature of 31 comprises
eight signals and is compatible with a (static or time-averaged) C2-symmetric
bishemiaminal structure. The H NMR (300 MHz, d6-DMSO) spectrum for 31
1
exhibits significant line broadening, but is also suggestive of net C2-symmetry.
(90) Bisenamide 33 proved resistant to hydrogenation as catalyzed by either
Pd/C or Pd(OH)2/C (albeit at 1 atm pressure of H2), and was likewise recovered
unchanged following treatment with NaBH(OAc)3 or NaBH3CN in AcOH.
(91) Louie, J.; Bielawski, C. W.; Grubbs, R. H. J. Am. Chem. Soc. 2001,
123, 11312–11313.
(92) Exo-face selective addition of Grignard reagents to acyl iminium ions
derived from N-acyl bispidines has been previously observed, see: Harrison, J. R.;
O’Brien, P. Tetrahedron Lett. 2000, 41, 6167–6170.
7946 J. Org. Chem. Vol. 73, No. 20, 2008

Total Synthesis of Sparteine Group Alkaloids
SCHEME 5. Synthesis of (()-ꢀ-Isosparteine (dl-3) from
Tetraoxobispidine 16
SCHEME 6. Synthesis of (()-Sparteine (dl-1) from
Imidolactam 37
via essentially the same sequence of steps used previously to
convert 39 to (()-ꢀ-isosparteine (dl-3). In this case, however,
the double RCM reaction could not be coaxed to completion
by the action of the first generation Grubbs metathesis catalyst
(up to 30 mol% of “Grubbs 1”), presumably due to a product
inhibition effect. The problem was circumvented by use of the
more reactive second generation Grubbs catalyst instead (“Grubbs
2”),88 which gave the desired tetracyclic diene 44 in an excellent
yield. An X-ray crystallographic analysis secured the structural
assignment for 44 (see the Supporting Information) and
subsequent hydrogenation led to the expected oxygenated
sparteine derivative 45. The illustrated R-anomeric configuration
of 45 was established by further X-ray diffraction work that
also revealed a hydrogen-bonded heteroenantiomeric dimer for
the crystalline racemic amido alcohol (see the Supporting
Information). Hemiaminal 45 showed similar acid catalyzed
instability to s-32 and upon prolonged exposure to silica gel
gave a dehydration adduct (46) analogous to 33. Even brief
contact of 45 with SiO2 resulted in anomerization (leading to
R:ꢀ ) 2:3) and so chromatographic purification was best
avoided. Finally, reduction of 45 with lithium aluminum hydride
gave (()-sparteine (dl-1) in 80% yield (after Kugelrohr distil-
lation), indistinguishable from a commercial sample of natural
(-)-sparteine (l-1) by IR and NMR spectral analysis.
Enantioselective transformations. During the course of our
studies, it was discovered that sequential net double nucleophilic
additions to C2V-symmetric tetraoxobispidine intermediates
typically afford (inherently chiral) C2-type rather than (poten-
tially meso) Cs-type regioisomers; however, a priori no such
intrinsic substrate bias was anticipated. In fact, it was initially
believed that chiral reagent control would be required to target
the desired C2-type addition adducts (49) by repeated
recognition of the shared local prochirality of a given pair
of homotopic carbonyl groups within bisimides 47 (Figure
5). The tactic of employing chiral reagent control to influence
regioselectiVity in this manner has been but rarely applied
in synthesis.97 On the basis of our observations, the two
remaining imide carbonyl groups in imidolactam intermedi-
ates of type 48 (and even those lacking a formal alkoxide
moiety, e.g., 37) are evidentally sufficiently electronically
differentiated to allow for their regioselective transformation
generated in this manner proved difficult to separate, was viewed
as only a minor detraction of the new process. Exposure of a
two component mixture of alkenes 37 and 39 to Grubbs’ first
generation metathesis catalyst88 gave the expected (and easily
separated) RCM adducts 41 and 40. Exhaustive reduction of
the latter resulted in its high-yielding transformation to the target
alkaloid (dl-3) of immediate interest. The identity of the
synthetic (()-ꢀ-isosparteine (dl-3) generated by the described
1
five-step route from 16 was confirmed by comparison of H
and 13C NMR spectral data with those previously reported for
this relatively obscure member of the sparteine family.67,74,93
Synthesis of (()-Sparteine. With a hoard of knowledge
gained from the successful tetraoxobispidine based elaborations
of (()-R-isosparteine (dl-2) and (()-ꢀ-isosparteine (dl-3), an
assault on the parent alkaloid (dl-1) was approached with
confidence. The nonsymmetrical nature of sparteine precluded
an obvious wholly two-directional approach from bisimide 16,
although one such strategy was devised and briefly entertained
before its abandonment.96 A way forward presented itself when
it was discovered that imidolactam 37, the intermediate en-
countered en route to dl-3,74 received allylmagnesium chloride
in a regioselective manner (Scheme 6). The resulting pseudo-
C2-symmetric addition adduct 42 was obtained as a single
stereoisomer, presumably with an endo configurated hydroxyl
group (cf. 56, vide infra), following its isolation by chroma-
tography. Tetraene 42 was converted to the final target molecule
(96) It was envisioned that ionic reduction of a bis(diethylsilyl) hydride (2
× OSiEt2H) derivative of diol 31 (or related compounds 32) would proceed via
a successive combination of intramolecular (endo selective) and intermolecular
(exo selective) hydride additions to acyl iminium ions generated in the presence
of a Lewis acid and external silane (Et3SiH). This plan stalled when the requisite
silyl ether derivatives could not be generated.
J. Org. Chem. Vol. 73, No. 20, 2008 7947

Norcross et al.
SCHEME 7. Enantioselective Reduction of
Tetraoxobispidine 25
FIGURE 5. Conceptualized enantioselective addition to C2V-symmetric
tetraoxobispidines 47. Nu* represents a generic enantioselective nu-
cleophilic addition process that preferentially recognizes imide carbonyl
groups with local pro S enantiotopicity.
viability of this precedented desymmetrization tactic for the
enantioselective reduction of our tetraoxobispidine synthons.
Initial studies focused on borane reduction of N,N′-diallyl
bisimide 16 catalyzed by Corey’s (S)-diphenylprolinol derived
oxazaborolidine 51 (the “CBS” catalyst).104b,c Hydroboration
of the alkenyl moieties in 16 competed with imide reduction
when borane tetrahydrofuran complex was used as a terminal
reductant and application of catecholborane in the same role
proved ineffectual.105 Comparable reactions from the crotyl and
prenyl analogues of 16 were also unsuccessful: alkene hydrobo-
ration and decomposition plagued attempted CBS-reduction of
dicrotyl bisimide 26, while the sterically encumbered diprenyl
derviative 27 was wholly unreactive. More promising results
were obtained from dibenzyl bisimide 25, a compound that had
earlier been shown to give isolable hemiaminal derivatives upon
its reduction (e.g., 25 f 34). However, it became immediately
evident that double reduction of 25 as mediated by BH3 ·THF
and promoted by oxazaborolidine 51 led to the undesired Cs-
type regiochemistry (Scheme 7). Compounding this problem,
the potentially valuable monoreduction product 52 could not
be secured in a truely useful yield because further reduction of
(the borane complex of) this compound occurred at a rate greater
than that for initial hydride delivery to bisimide 25. Thus, after
considerable optimization experiments, imidolactam 52 was
obtained in at best a 13% isolated yield, accompanied by 15%
of the essentially meso bishemiaminal 53, following the
treatment of bisimide 25 with equimolar quantities of CBS
“catalyst” 51 and BH3 ·THF.
without the need for subtle absolute stereorecognition effects.
The apparent enhanced electrophilicity of the imide carbonyl
group diagonally opposite from the first addition point is
tentatively attributed to the inductive effect of the proximal
lactam carbonyl group. Even a cursory inspection of molec-
ular models reveals that steric factors are highly unlikely to
be the cause of the substrate regioselectivity bias.
While chiral reagent control was ultimately not necessary to
gain access to bisimide addition adducts of the desired regio-
chemistry, this paradigm (ideally in a catalytic enantioselective
guise) would provide the most efficient way to target nonracemic
sparteine alkaloid targets from tetraoxobispidines. For example,
an enantioselective addition of a single hydride anion equivalent
to bisimide 16 would be sufficient to obtain scalemic samples
of both ꢀ-isosparteine (3) and sparteine (1) via the routes already
described. Likewise, R-isosparteine (2) could be secured from
16 in optically active form via an initial enantioselective allyl
anion addition event (or alternatively by isomerization of 1 or
3).20 Numerous methods for the enantioselective allylation98 and
reduction99 of aldehydes and ketones are available but very few
of these protocols have been successfully extended to transfor-
mation of carbonyl groups at the less electrophilic carboxyl
oxidation level.100 Notable contributions from Speckamp and
Hiemstra,101 among others,102,103 have shown that chiral ox-
azaborolidine catalyzed borane reduction104 is an effective tool
for the differentiation of enantiotopic carbonyl groups within
prochiral cyclic imides. Accordingly, we elected to evaluate the
The opposite regiochemical outcomes for bisimide double
reduction with use of a monohydride donor (such as LiBHEt3)
versus a multiple hydride donor (such as BH3 ·THF) is readily
understood when considering the possibility of intramolecular
delivery of the second hydride equivalent in the latter case.
Following exo face selective monoreduction of 25 with
BH3 ·THF (or with BH3 ·51; the enantiodetermining step), the
resulting borate complex 57 is perfectly poised to add hydride
to the adjacent imide carbonyl group from the endo face (i.e.,
(97) Hayashi’s enantioselective hydrosilylation of norbornadiene (ref 71),
utilized by Aube´ to access diketone 12 en route to (+)-sparteine (Figure 3, eq
4), owes its (C2-symmetric) regioselectivity to this phenomena. For elaboration
on the concept as it relates to the regiodivergent conversion of enantiomers within
a racemic mixture by a chiral reagent, see: Kagan, H. B. Croat. Chem. Acta
1996, 69, 669–680.
(98) Denmark, S. E.; Fu, J. Chem. ReV. 2003, 103, 2763–2793.
(99) Itsuno, S. Org. React. 1998, 52, 395–576.
(100) Carboxylic acid anhydrides represent the best studied substrate class,
see: Atodiresei, I.; Schiffers, I.; Bolm, C. Chem. ReV. 2007, 107, 5683–5712.
(101) (a) Romagnoli, R.; Roos, E. C.; Hiemstra, H.; Moolenaar, M. J.;
Speckamp, W. N. Tetrahedron Lett. 1994, 35, 1087–1090. (b) Ostendorf, M.;
Romagnoli, R.; Pereiro, I. C.; Roos, E. C.; Moolenaar, M. J.; Speckamp, W. N.;
Hiemstra, H. Tetrahedron: Asymmetry 1997, 8, 1773–1789.
(103) For related work using a chiral thiazazincolidine catalyst, see: Kang,
J.; Lee, J. W.; Kim, J. I.; Pyrun, C. Tetrahedron Lett. 1995, 36, 4265–4268.
(104) (a) Hirao, A.; Itsuno, S.; Nakahama, S.; Yamazaki, N. J. Chem. Soc.,
Chem. Commun. 1981, 315–317. (b) Corey, E. J.; Bakshi, R. K.; Shibata, S.
J. Am. Chem. Soc. 1987, 109, 5551–5553. Review: (c) Corey, E. J.; Helal, C. J.
Angew. Chem., Int. Ed. 1998, 37, 1986–2012.
(102) (a) Shimizu, M.; Nishigaki, Y.; Wakabayashi, A. Tetrahedron Lett.
1999, 40, 8873–8876. (b) Jones, S.; Dixon, R. A. Tetrahedon: Asymmetry 2002,
13, 1115–1119. (c) Barker, M. D.; Dixon, R. A.; Jones, S.; Marsh, B. J.
Tetrahedron 2006, 62, 11663–11669. (d) Barker, M. D.; Dixon, R. A.; Jones,
S.; Marsh, B. J. Chem. Commun. 2008, 2218–2220.
(105) Corey, E. J.; Bakshi, R. K. Tetrahedron Lett. 1990, 31, 611–614.
7948 J. Org. Chem. Vol. 73, No. 20, 2008

Total Synthesis of Sparteine Group Alkaloids
57 f 58). The exact regio- and stereochemical identity of 53
is predicated on this basis, and supported by the fact that the
isolated material was optically active when the reduction was
controlled by 51. If thermodynamic configurational equilibration
of anomeric stereogenic centers in 53 was the principle origin
of relative stereochemistry, then the product would be be
racemic.
the controlled fabrication of other types of nitrogenous hetero-
cycles (e.g., piperidines and bispidines, or tricyclic compounds
such as cytisine (6) and sparteine surrogate 7). Various tactics
for differentiating the two pairs of enantiotopic carbonyl groups
within prochiral C2V-symmetric tetraoxobispidines are potentially
available and one such method (chiral oxazaborolidine mediated
borane reduction) was shown to offer a useful level of
enantioselectivity, albeit in low yield. An entirely satisfactory
asymmetric rendering of the alkaloid syntheses described herein
awaits a more thorough evaluation of enantioselective additions
to bisimides.
Finally, it is sobering to reflect on the fact that in the 60 years
since the publication of Leonard and Beyler’s total synthesis
of (()-sparteine,52 no other purely artificial preparation of this
important alkaloid (including the work herein) has come close
to matching the simple elegance and brevity of that seminal
effort. Despite continuing advances,107 sparteine retains its
position as the ligand of choice for a great many metal mediated
enantioselective methods and yet a genuinely practical asym-
metric synthesis of this long known diamine system that can
compete with natural supply remains elusive.
Sakurai-type allylation of 52 and 53 with excess allyltrim-
ethylsilane converted each compound to the expected product
in good yield. Bishemiaminal 53 gave the true meso compound
1
55, the Cs-symmetry of which was obvious from its H NMR
spectral signature, while imidolactam 52 gave an allylated adduct
(54) analogous to the intermediate (37) used earlier to access
(()-sparteine (dl-1). HPLC analysis of 54 on a chiral stationary
phase (Daicel Chiralcel OD-RH column) revealed an enantio-
meric excess of 71%, the absolute sense of which was not
established.106 Addition of allylmagnesium bromide to 54
proceeded in a like manner to that observed previously for 37
(Scheme 6), and an X-ray crystallographic analysis confirmed
the exact nature of the resulting pseudo-C2-symmetric product
56. While synthetic possibilities clearly exist for the conversion
of intermediates 54 and 56 to sparteine diamines, given the low
efficiency of the conversion of 25 to 52, generation of the
optically active target alkaloids along these lines was deemed
to hold no practical value and was not further pursued.
Nevertheless, the above studies serve a useful purpose in
establishing proof-of-principle for reagent controlled enanti-
oselective addition to a tetraoxobispidine platform.
Experimental Section
Propane-1,1,3,3-tetracarboxamide (21). 21 was synthesized by
a higher yielding modification of an earlier method.73 A stirred
mixture of dimethyl malonate (300 mL, d ) 1.15, 345 g, 2.61 mol)
and paraformaldehyde (19.8 g, 0.660 mol) at 75 °C was treated
with KOH (3.0 mL, 10 wt % in MeOH) and the resulting solution
heated to reflux (ca. 95 °C) for 6 h. The mixture was allowed to
cool to rt and shaken with saturated aq NH4Cl (50 mL). The organic
phase was separated and dried (Na2SO4), and excess dimethyl
malonate was removed by distillation at reduced pressure (water
aspirator, bp 80-90 °C, ca. 30 mmHg). The residual tetramethyl
1,1,3,3-propanetetracarboxylate was treated with conc aq NH4OH
(400 mL) and the initially biphasic mixture stirred vigorously for
15 h at rt. During this time a colorless precipitate formed. The
insoluble solid was removed by filtration, triturated with EtOAc-
MeOH (2:1, 500 mL), and dried in vacuo to afford tetraamide 21
(102.5 g, 0.474 mol, 72%) as a colorless powder: mp 256-257 °C
Conclusion
(MeOH); IR (KBr) 3385, 3188, 1656, 1419, 1383, 1346, 615 cm-1
;
1H NMR (300 MHz, d6-DMSO) δ 7.18 (4H, s), 7.05 (4H, s), 2.95
(2H, t, J ) 7.5 Hz), 2.00 (2H, t, J ) 7.4 Hz) ppm; 13C NMR (75
MHz, d6-DMSO) δ 170.7 (4C, 0), 50.3 (2C, 1), 28.3 (2) ppm; MS
(ES) m/z 239 (M + Na)+, 217 (M + H)+; HRMS (ES) m/z 239.0755
(calcd for C7H12N4O4Na 239.0756).
In summary, a new “inside-out” strategy for the preparation
of lupine alkaloids of the sparteine subgroup has been success-
fully demonstrated. All three diastereomeric variants of the
parent sparteine tetracyclic skeleton were individually targeted
by appropriate synthetic manipulations of a pivotal N,N′-
diallyltetraoxobispidine intermediate (16) prepared from mal-
onate via a modified Guthzeit cyclization. Racemic alkaloids,
(()-R-isosparteine (dl-2), (()-ꢀ-isosparteine (dl-3), and (()-
sparteine (dl-1), were each concisely elaborated from this
common bisimide precursor in good overall yield (i.e., 28%,
12%, and 11%, respectively). Throughout, nucleophilic additions
to tetraoxobispidine derived bicyclic electrophiles were found
to be exo face selective, and sequential net double additions to
such templates gave predominantly C2-type rather than Cs-type
regiomers. The predictable nature of these reactions bodes well
for the wider deployment of bisimide synthons such as 16 for
2,4,6,8-Tetraoxo-3,7-diazabicyclo[3.3.1]nonane (22). 22 was
synthesized by a higher yielding modification of an earlier method.73
A 1 L round-bottomed flask was charged with tetraamide 21 (22.8
g, 105 mmol) and MsOH (21.4 mL, d ) 1.48, 31.7 g, 330 mmol)
and the resulting paste was manually stirred with a metal spatula
while a “cool” (yellow-blue) Bunsen burner flame was applied for
10 min. During heating, an initial period of intense effervescence
was observed, which subsided to leave a gently boiling homoge-
neous yellow liquid (internal temp 170-180 °C). The liquid was
allowed to cool to rt and triturated with MeOH (100 mL) to induce
crystallization, then the resulting suspension stirred for 12 h. The
solid was filtered off, pulverized, then suspended in fresh MeOH
(100 mL) and heated at reflux for 1 h to dissolve additional residual
ammonium mesylate. The suspension was allowed to cool to rt,
then filtered to yield bisimide 22 (6.77 g, 95 wt % with NH4OMs,
35.3 mmol, 33%) as a colorless solid: mp 295 °C dec (H2O); IR
(KBr) 3257, 2838, 1740, 1714, 1339, 1274, 1201, 833, 800, 768
(106) The illustrated enantiomer of 52 is that anticipated on the basis of the
standard Corey model (ref 104c) assuming that only exo-face addition of hydride
to 25 is possible, and that the N-benzyl moiety is the more sterically demanding
group (as advocated by Speckamp and Hiemstra for related bicyclic mono-imides,
ref 101b). Almost identical results were obtained with Jones’ aminoindanol
derived oxazaborolidine (ref 102c) in place of the more common CBS-catalyst
51. In this case, reduction of 25 with the B-methyl oxazaborolidine derived from
(1R,2S)-cis-1-amino-2-indanol gave the opposite enantiomer of 52 to that
generated from the (S)-prolinol derived catalyst 51, but also in ca. 70% ee.
1
cm-1; H NMR (300 MHz, d6-DMSO) δ 11.44 (2H, s), 3.64 (2H,
t, J ) 2.8 Hz), 2.59 (2H, t, J ) 2.8 Hz) ppm; 13C NMR (75 MHz,
(107) Stead, D.; O’Brien, P.; Sanderson, A. Org. Lett. 2008, 10, 1409–1412.
J. Org. Chem. Vol. 73, No. 20, 2008 7949

Norcross et al.
d6-DMSO) δ 166.9 (4C, 0), 47.2 (2C, 1), 22.8 (2) ppm; MS (EI)
m/z 182 M+•. Crystals suitable for X-ray diffraction analysis (Figure
4) were obtained by recrystallization from H2O.
(()-(1R*,2R*,5R*,6S*)-4,8-Dioxo-2-hydroxy-2,3,6,7-tetraallyl-
3,7-diazabicyclo[3.3.1]nonane (42). A vigorously stirred solution
of triene 37 (405 mg, 1.40 mmol) in anhydrous THF (11 mL) at
-78 °C under Ar was treated dropwise with allylmagnesium
bromide (2.47 mL, 0.74 M in Et2O, 1.83 mmol). After 25 min at
-78 °C, saturated aq NH4Cl (5 mL) was added and the quenched
reaction mixture was partitioned between EtOAc (20 mL) and H2O
(20 mL). The layers were then separated and the aqueous phase
extracted with EtOAc (2 × 20 mL). The combined organic phases
were washed with brine (10 mL), dried (Na2SO4), and concentrated
in vacuo. The residue was further purified by column chromatog-
raphy (SiO2, eluting with 0-50% EtOAc in hexanes) to yield
tetraene 42 (240 mg, 0.727 mmol, 52%) as a colorless oil: IR (neat)
3,7-Diallyl-2,4,6,8-tetraoxo-3,7-diazabicyclo[3.3.1]nonane (16).
16 was synthesized by a higher yielding modification of an earlier
method.73 A vigorously stirred suspension of bisimide 22 (17.2 g,
94.4 mmol) in anhydrous DMF (180 mL) at 0 °C under Ar was
treated portionwise with NaH (9.11 g, 60 wt % disp. in oil, 228
mmol). The ensuing gas evolution ceased within 2 min. The
resulting solution was stirred for 3 min and then treated dropwise
with neat allyl bromide (19.3 mL, d ) 1.43, 27.6 g, 228 mmol).
The mixture was warmed to rt and stirred for an additional 2 h.
After this time, saturated aq NH4Cl (50 mL) was added and the
quenched reaction mixture was partitioned between H2O (160 mL)
and EtOAc (240 mL). The layers were separated and the aqueous
phase extracted with EtOAc (2 × 80 mL). The combined organic
extracts were washed successively with H2O (2 × 80 mL) and brine
(35 mL) and then dried (Na2SO4) and concentrated in vacuo. The
resulting solid residue was triturated with hexanes (65 mL), filtered-
off, and sucked dry to afford the pure diallylated product 16 (20.5
g, 78.2 mmol, 83%) as a colorless crystalline solid: mp 130-132
°C (EtOAc); IR (KBr) 3006, 1699, 1361, 1328, 1190, 982, 928
1
3329, 3071, 2972, 1643, 1613, 1432, 1157, 916 cm-1; H NMR
(300 MHz, CDCl3) δ 5.80-5.43 (4H, m), 5.32 (1H, d, J ) 1.9
Hz), 5.07-4.88 (8H, m), 4.25 (1H, ddt, J ) 15.2, 4.9, 1.6 Hz),
4.01 (1H, ddm, J ) 14.9, 6.6 Hz), 3.76 (1H, ddm, J ) 14.9, 5.2
Hz), 3.55-3.43 (2H, m), 2.75-2.62 (3H, m), 2.54 (1H, dm, J )
14.6 Hz), 2.24 (1H, ddd, J ) 14.3, 9.7, 1.8 Hz), 2.13 (1H, dt, J )
14.4, 9.6 Hz), 1.99 (2H, t, J ) 3.1 Hz) ppm; 13C NMR (75 MHz,
CDCl3) δ 169.7 (0), 169.1 (0), 134.7 (1), 133.0 (1), 132.6 (1), 131.5
(1), 118.8 (2C, 2), 118.1 (2), 116.4 (2), 87.2 (0), 58.8 (1), 47.6 (2),
43.1 (2), 42.8 (2), 42.5 (1), 39.0 (1), 35.8 (2), 16.8 (2) ppm; MS
(EI) m/z 330 (M•+, <5%), 289 (100%), 206 (20%); HRMS (EI)
m/z 330.1946 (calcd for C19H26N2O3 330.1944).
(()-∆3,13-Didehydro-10,17-dioxo-6-hydroxy-ꢀ-isosparteine
(44). A solution of tetraene 42 (465 mg, 1.41 mmol) in anhydrous
CH2Cl2 (5 mL) at rt under N2 was treated with
(Cy3P)(H2IMes)Cl2RudCHPh (22 mg, 0.026 mmol)88 and the
resulting mixture was stirred at reflux for 7 h. After this time, the
mixture was cooled to rt and concentrated in vacuo, then the residue
was purified by column chromatography (SiO2, eluting with 1%
MeOH in CH2Cl2) to afford the unsaturated tetracycle 44 (354 mg,
1.29 mmol, 92%) as a colorless solid: mp 223-225 °C (EtOAc-
MeOH); IR (KBr) 3270, 3036, 2916, 1620, 1429, 1255, 1201, 988,
896, 672 cm-1; 1H NMR (300 MHz, d6-DMSO) δ 5.83-5.74 (1H,
m), 5.72-5.58 (3H, m), 5.54 (1H, s), 4.75 (1H, dm, J ) 18.7 Hz),
4.54 (1H, dm, J ) 18.0 Hz), 3.56 (1H, dd, J ) 11.2, 3.8 Hz),
3.47-3.27 (2H, m), 2.70 (1H, dt, J ) 3.7, 1.9 Hz), 2.64 (1H, dm,
J ) 17.8 Hz), 2.58-2.53 (1H, m), 2.35 (1H, tm, J ) 13.9 Hz),
2.20-1.99 (4H, m) ppm; 13C NMR (75 MHz, d6-DMSO) δ 169.0
(0), 165.4 (0), 124.7 (1), 124.3 (1), 123.5 (1), 122.7 (1), 82.9 (0),
54.8 (1), 47.3 (1), 41.6 (2), 40.6 (1), 38.1 (2), 37.1 (2), 31.1 (2),
18.3 (2) ppm; MS (EI) m/z 274 (M•+, 6%), 256 (28%), 146 (24%),
84 (100); HRMS (EI) m/z 274.1311 (calcd for C15H18N2O3
274.1317). The identity of 44 was confirmed by X-ray diffraction
analysis.
(()-10,17-Dioxo-6-hydroxy-ꢀ-isosparteine (45). A mixture of
unsaturated tetracycle 44 (125 mg, 0.456 mmol) and 10 wt % Pd/C
(13 mg) in MeOH-H2O (3:1, 8 mL) was stirred vigorously under
1 atm of H2 at rt for 8 h. The active gas was then purged with N2,
and the mixture was diluted with CH2Cl2 (20 mL) and filtered
through a celite pad. The filter cake was washed with CH2Cl2 (2 ×
10 mL) and the filtrate and combined washings dried (Na2SO4)
and concentrated in vacuo to yield tetracycle 45 (116 mg, 0.417
mmol, 91%) as a colorless solid: mp 198-200 °C (Et3N); IR (KBr)
3183, 2943, 1642, 1429, 1190, 1005, 634, 503 cm-1; 1H NMR (300
MHz, CDCl3) δ 4.71 (1H, ddt, J ) 13.1, 4.3, 1.9 Hz), 4.70 (1H, d,
J ) 1.0 Hz), 4.40 (1H, ddt, J ) 12.8, 4.5, 2.2 Hz), 3.50 (1H, dm,
J ) 11.3 Hz), 3.18 (1H, td, J ) 13.0, 3.1 Hz), 2.73 (1H, dt, J )
3.3, 2.7 Hz), 2.62-2.58 (1H, m), 2.52 (1H, td, J ) 12.9, 2.8 Hz),
2.23-2.05 (3H, m), 2.00-1.93 (1H, m), 1.90-1.55 (8H, m),
1.47-1.30 (2H, m) ppm; 13C NMR (75 MHz, CDCl3) δ 168.8 (0),
167.4 (0), 84.2 (0), 59.9 (1), 47.0 (1), 43.8 (2), 42.6 (1), 37.7 (2),
37.5 (2), 32.0 (2), 25.4 (2), 25.1 (2), 24.9 (2), 20.2 (2), 19.4 (2)
ppm; MS (EI) m/z 278 (M•+, <5%), 260 (100%); HRMS (EI) m/z
278.1635 (calcd for C15H22N2O3: 278.1631). Crystals suitable for
X-ray diffraction analysis (see the Supporting Information) were
obtained by recrystallization from Et3N.
1
cm-1; H NMR (300 MHz, CDCl3) δ 5.73 (2H, ddt, J ) 16.8,
10.5, 5.9 Hz), 5.14 (2H, dq, J ) 10.5, 1.2 Hz), 5.13 (2H, dq, J )
16.6, 1.2 Hz), 4.36 (4H, dt, J ) 5.9, 1.3 Hz), 4.07 (2H, t, J ) 2.9
Hz), 2.57 (2H, t, J ) 2.9 Hz) ppm; 13C NMR (75 MHz, CDCl3) δ
164.8 (4C, 0), 130.7 (2C, 1), 119.0 (2C, 2), 48.5 (2C, 1), 42.5 (2C,
2), 22.6 (2) ppm; MS (ES) m/z 263 (M + H)+; HRMS (ES) m/z
263.1027 (calcd for C13H15N2O4 263.1032).
Direct Synthesis of (()-(1R*,5S*,6R*)-3,6,7-Triallyl-2,4,8-
trioxo-3,7-diazabicyclo[3.3.1]nonane(37)and(()-(1R*,4S*,5R*,8S*)-
2,6-Dioxo-3,4,7,8-tetraallyl-3,7-diazabicyclo[3.3.1]nonane (39).
A well-stirred solution of bisimide 16 (2.70 g, 10.3 mmol) in
anhydrous THF (45 mL) at 0 °C was treated with NaBH4 (274
mg, 7.21 mmol, 0.70 equiv). The resulting mixture was stirred for
4.5 h and then aq HCl (8 mL, 4 M) was added to quench excess
borohydride reagent. Following cessation of effervescence (5 min),
the mixture was warmed to rt and partitioned between EtOAc (40
mL) and H2O (20 mL). The layers were separated and the aqueous
phase extracted with EtOAc (2 × 40 mL). The combined organic
phases were then washed with saturated aq NaHCO3 (10 mL), dried
(Na2SO4), and concentrated in vacuo to afford 2.30 g of a residual
pale yellow oil containing a complex mixture of hemiaminals. A
stirred solution of the residue (2.30 g) in anhydrous CH2Cl2 (25
mL) at rt under Ar was treated with allyltrimethylsilane (4.16 mL,
d ) 0.719, 2.99 g, 26.2 mmol), followed by the dropwise addition
of BF3 ·OEt2 (1.64 mL, d ) 1.15, 1.89 g, 13.3 mmol). The mixture
was stirred for 40 h at rt, and then diluted with CH2Cl2 (30 mL),
washed with H2O (2 × 20 mL), dried (Na2SO4), and concentrated
in vacuo. The residue (2.55 g) was purified by column chroma-
tography (SiO2, eluting with 0-2% MeOH in CH2Cl2) to afford,
in order of elution, pure triene 37 (829 mg, 2.88 mmol, 28%) and
a mixture of 37 and tetraene 39 [640 mg, 37:39 ) 22:78 mol:
effectively, 37 (134 mg, 0.465 mmol, 5%), 39 (506 mg, 1.61 mmol,
16%)], as colorless oils. Data for 37: IR (neat) 3082, 2949, 1739,
1684, 1455, 1357, 1194, 995, 922, 630 cm-1; 1H NMR (400 MHz,
CDCl3) δ 5.78-5.66 (2H, m), 5.65-5.54 (1H, m), 5.24-5.06 (6H,
m), 4.51 (1H, dm, J ) 15.2 Hz), 4.37-4.25 (2H, m), 3.70 (1H,
dm, J ) 9.6 Hz), 3.67-3.64 (1H, m), 3.49 (1H, dd, J ) 15.3, 7.4
Hz), 3.13-3.09 (1H, m), 2.71 (1H, dm, J ) 14.5 Hz), 2.37 (1H,
ddd, J ) 13.8, 3.2, 2.1 Hz), 2.29-2.20 (2H, m) ppm; 13C NMR
(100 MHz, CDCl3) δ 173.1 (0), 168.1 (0), 163.0 (0), 132.8 (1),
132.0 (1), 131.4 (1), 120.0 (2), 119.0 (2), 118.2 (2), 58.7 (1), 48.3
(1), 47.8 (2), 42.1 (2), 39.9 (1), 36.7 (2), 19.8 (2) ppm; MS (EI)
m/z 288 (M•+, 6%), 247 (100%), 193 (20%), 164 (32%), 136 (36%);
HRMS (EI) m/z 288.1478 (calcd for C16H20N2O3 288.1474). The
authenticity of 39 was confirmed by comparison to spectral data
collected previously from a pure sample obtained via the multistep
conversion sequence.74
7950 J. Org. Chem. Vol. 73, No. 20, 2008

Total Synthesis of Sparteine Group Alkaloids
(()-Sparteine (dl-1). dl-1 was synthesized by a modification of
O’Brien’s method.57 A stirred solution of bislactam 45 (55 mg,
0.198 mmol) in anhydrous THF (1.3 mL) at 0 °C under N2 was
treated with excess LiAlH4 (79 mg, 2.08 mmol). The resulting
suspension was heated to a gentle reflux and stirred for 16 h. The
mixture was then cooled to rt and moistened Na2SO4 added
portionwise until effervescence ceased. After being stirred for a
further 30 min, the quenched reaction mixture was filtered through
a shallow celite pad and the solids washed with MeOH-CH2Cl2
(1:9, 20 mL). The filtrate and combined washings were then dried
(Na2SO4) and concentrated in vacuo. The residue was subjected to
Kuglerohr distillation (130-140 °C, <5 mmHg) to yield the title
alkaloid dl-1 (37 mg, 0.158 mmol, 80%) as a colorless oil: IR (neat)
235.2169). Essentially identical IR and NMR spectra were obtained
from a commercial (Aldrich) sample of natural (-)-sparteine (l-1)
recorded under the same conditions (nb. NMR signals show a slight
concentration dependence).93
Acknowledgement. Oregon State University, the University
of Leeds, and Eli Lilly & Co. Ltd. are thanked for generous
financial support. N.R.N. thanks the University of Leeds for a
Henry Ellison Scholarship and Eli Lilly & Co. Ltd. for a CASE
award. James Titchmarsh (University of Leeds) is gratefully
acknowledged for performing HPLC analyses for ee determina-
tions.
3395 (H2O), 2927, 2856, 2758, 1647, 1446, 1353, 1288, 1113 cm-1
;
1H NMR (300 MHz, CDCl3) δ 2.76 (1H, dm, J ) 12.2 Hz),
2.70-2.62 (2H, m), 2.51 (1H, dm, J ) 10.9 Hz), 2.32 (1H, dd, J
) 11.2, 3.5 Hz), 2.08-1.90 (4H, m), 1.82-1.77 (1H, m), 1.74-1.63
(3H, m), 1.60-1.15 (12H, m), 1.04 (1H, dt, J ) 12.0, 2.4 Hz) ppm;
13C NMR (75 MHz, CDCl3) δ 66.7 (1), 64.5 (1), 62.1 (2), 56.4
(2), 55.5 (2), 53.6 (2), 36.2 (1), 34.5 (2), 33.2 (1), 29.5 (2), 27.8
(2), 26.0 (2), 25.8 (2), 24.9 (2), 24.8 (2) ppm; MS (ES) m/z 235
(M + H)+; HRMS (ES) m/z 235.2161 (calcd for C15H27N2
Supporting Information Available: General methods and
all other experimental procedures not previously described,
characterization data, and 1H and 13C NMR spectra for all new
compounds, as well as CIF files for compounds 22, 30, s-32,
44, 45, and 56.86 This material is available free of charge via
the Internet at http://pubs.acs.org.
JO8013512
J. Org. Chem. Vol. 73, No. 20, 2008 7951