The First and Second Cinchona Rearrangement. Two Fundamental Transformations of Alkaloid Chemistry

Article abstract of DOI:10.1021/jo030363s

Stereochemistry, products, and driving forces of the "first and second Cinchona rearrangement" have been investigated and a unified theory is presented. The first cage expansion affords [3.2.2]azabicyclic α-amino ether and is formulated via a configurationally stable bridgehead iminium ion and quasiequatorial nucleophilic attack. The second cage expansion affords β-functionalized [3.2.2]azabicycles. In this case a nonclassical nitrogen-bridged cation is postulated to account for retention of configuration and potential reversibility of the cage expansion. The second rearrangement is favored for the so-called cinch bases (6′-R = H) in trifluoroethanol. Stereoelectronic factors, electron demand at C9, ground state conformation, and solvent type are crucial in all cases. A two-step protocol for preparing 9-epi-configured Cinchona alkaloids from 9-nat precursors is described.

Full text of DOI:10.1021/jo030363s

Th e F ir st a n d Secon d Cin ch on a Rea r r a n gem en t. Tw o
F u n d a m en ta l Tr a n sfor m a tion s of Alk a loid Ch em istr y
M. Heiko Franz,† Stefanie Ro¨per,† Rudolf Wartchow,‡ and H. M. R. Hoffmann*,†
Department of Organic Chemistry, University of Hannover, Schneiderberg 1 B,
30167 Hannover, Germany, and Department of Inorganic Chemistry, University of Hannover,
Callinstrasse 9, 30167 Hannover, Germany
hoffmann@mbox.oci.uni-hannover.de
Received November 26, 2003
Stereochemistry, products, and driving forces of the “first and second Cinchona rearrangement”
have been investigated and a unified theory is presented. The first cage expansion affords [3.2.2]-
azabicyclic R-amino ether and is formulated via a configurationally stable bridgehead iminium ion
and quasiequatorial nucleophilic attack. The second cage expansion affords â-functionalized [3.2.2]-
azabicycles. In this case a nonclassical nitrogen-bridged cation is postulated to account for retention
of configuration and potential reversibility of the cage expansion. The second rearrangement is
favored for the so-called cinch bases (6′-R ) H) in trifluoroethanol. Stereoelectronic factors, electron
demand at C9, ground state conformation, and solvent type are crucial in all cases. A two-step
protocol for preparing 9-epi-configured Cinchona alkaloids from 9-nat precursors is described.
SCHEME 1. “F ir st Cin ch on a Rea r r a n gem en t” of
9-epi-Con figu r ed Br om oqu in in e (epi-1-Br )
Med ia ted by Silver Ion a n d Dr iven by F or m a tion
of Im in iu m Ion
In tr od u ction
Solvolysis and nucleophilic displacement reactions at
saturated carbon are among the most studied and best
known reactions of organic chemistry.1 Although Cin-
chona alkaloids are very useful templates in asymmetric
synthesis,2,3 chemical transformations of these natural
products have been little studied. For solvolyses of
Cinchona alkaloids with a C9-leaving group not only the
basic amino nitrogen but also the 6′-substituent (R′ )
H, OMe) and stereochemistry add further complexity. We
have earlier reported on the silver benzoate assisted
reaction of 9-epi-quinine with bromide as the leaving
group (epi-1-Br) in methanol to afford cage-expanded
1-azabicyclo[3.2.2]nonane with a methoxy group R to the
bridgehead nitrogen (cf. 2-OMe, Scheme 1). The rear-
rangement was shown to be stereoelectronically favor-
able, involving shift of the (nonprotonated) nitrogen lone
pair, nucleophilic shift of carbon C7 to C9, and stereo-
controlled external capture of a strained, nonplanar
bridgehead iminium ion intermediate. In the observed
† Department of Organic Chemistry.
‡ Department of Inorganic Chemistry.
(1) For the solvolysis of Cinchona alkaloids see: Rabe, P. Liebigs
Ann. Chem. 1949, 561, 132. Rabe, P. Chem. Ber. 1941, 74, 725 and
references therein. For general reviews of solvolytic displacements
see: Streitwieser, A. Chem. Rev. 1956, 56, 71. Ingold, C. K. Structure
and Mechanism in Organic Chemistry, 2nd ed.; Cornell University
Press: Ithaca, NY, 1969; Chapter VII. March, J . Advanced Organic
Chemistry, 4th ed.; Wiley: New York, 1992; pp 292-307. Carey, F.
A.; Sundberg, R. J . Advanced Organic Chemistry, 4th ed.; Plenum
Press: New York, 1993.
(2) Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem. Rev.
1994, 94, 2483. J ohnson, R. A.; Sharpless, K. B. Catalytic Asymmetric
Dihydroxylation. In Catalytic Asymmetric Synthesis; Ojima, I., Ed.;
VCH: Weinheim, Germany, 1993.
product 2-OMe the quinolyl and methoxy groups adopt
a quasi trans-2,3-diequatorial orientation. In an analo-
gous sequence quinidine with a C9-epi-configured chlo-
ride leaving group provided the pseudoenantiomeric4
R-methoxy amine.5
(3) Nelson, A. Angew. Chem. 1999, 111, 1685. Corey, E. J .; Bo, Y.;
Busch-Petersen, J . J . Am. Chem. Soc. 1998, 120, 13000. Lygo, B.;
Crosby, J .; Lowdon, T. R.; Peterson, J . A.; Wainwright, P. G. Tetra-
hedron 2001, 57, 2403 and references therein.
(4) Replacement of a vinyl group by hydrogen generates enanti-
omers.
(5) Braje, W.; Wartchow, R.; Hoffmann, H. M. R. Angew. Chem., Int.
Ed. 1999, 38, 2540.
10.1021/jo030363s CCC: $27.50 © 2004 American Chemical Society
Published on Web 03/27/2004
J . Org. Chem. 2004, 69, 2983-2991
2983

Franz et al.
SCHEME 3. Meth a n olysis of O-Tosyla ted 1-OTs
SCHEME 2. F ir st Rea r r a n gem en t of QCI a n d QCD
Iod in a ted a t C9: Distin ct P seu d oen a n tiom er ic4
Br id geh ea d Im in iu m Ion s a + a n d b+ a n d Th eir
Ster eosp ecific Nu cleop h ilic Ca p tu r e6
baffling molecular rearrangements is essential for further
work.
Resu lts
The first rearrangement was also realized for Quinco-
rine QCI and Quincoridine QCD,6 after the introduction
of iodide as the leaving group (Scheme 2).
Cage expansion and introduction of the 3-hydroxy
group were accomplished as follows: The cinch bases
(6′-R ) H) with C9-nat configuration were mesylated and
submitted to trifluoroethanolysis in the presence of added
nucleophile. With added NaOBz the benzoic esters 5-OBz
(43%) and 6-OBz (56%) were formed, isolated, and
hydrolyzed to the alcohol (see Scheme 4 above). Alter-
natively, 6-OH was prepared in a single flask by nucleo-
philic displacement with n-Bu4NONO and acidic hydrol-
ysis (Table 1, entry 1). Furthermore, â-amino alcohols
5-OH and 6-OH were prepared directly in pure water (see
Scheme 10 below).
Of the various nucleophiles investigated in solvent
trifluoroethanol thiocyanate ion was found to be most
effective for capturing and “freezing out” the cage-
expanded structure giving covalent thiocyanate 6-SCN
(89%, entry 2).
Azide ions in the trifluoroethanol furnished cage
conserved [2.2.2]- and [3.2.2]azabicycles 3-N3 and 5-N3
side-by-side (entry 3). Covalent azide 3-N3 was formed
with retention at C9 consistent with azabridged inter-
mediate i (Scheme 4).8 (In contrast, O-mesylated cincho-
nine 4-OMs and NaN3 in trifluoroethanol enter into
cascade reactions involving intramolecular 1,3-dipolar
cycloaddition with formation of triazolines.8)
In the reaction with KCN/CF3CH2OH (entry 4) we
isolated four different products (76%). The displacement
allows the one-carbon homologation to â-amino nitrile
5-CN (40%). The two products 3-CN and 3-OCH2CF3 with
intact azabicyclic cage were formed with retention at C9.
Ring-expanded trifluoroethyl ether 5-OCH2CF3 was also
By comparison to the Ag+ ion mediated rearrangement
of epi-1-Br in Scheme 1, the methanolysis of O-tosylqui-
nine 1-OTs with natural configuration at C9 appears
hopelessly complex (Scheme 3).
Remarkably, the change of solvent to the more ionizing
trifluoroethanol and the change of substrate to O-
mesylated cinchonidine 3-OMs and cinchonine 4-OMs
have simplified the product spectrum and brought to light
a “second Cinchona rearrangement” giving 1-azabicyclo-
[3.2.2]nonanes functionalized at carbon C3 rather than
C2 (Scheme 4). Nitrogen-bridged cation i (and the
pseudoenantiomeric species) is formulated as the reactive
intermediate: Four ligands around the bridging nitrogen
are constrained in one hemisphere. The nitrogen lone
pair is equivalent to a fifth valence and in the second
imaginary hemisphere. The intermediacy of cation i
accounts for retention of configuration via double inver-
sion on nucleophilic attack of carbon C9.7 Once generated
the cation of the “second rearrangement” manifold is
capable of sustaining both cage expansion and cage
conservation.
A major goal of our work is the synthesis of a variety
of functionalized 1-azabicyclo[3.2.2]nonanes with control
over the two cage-expansion routes and stereochemistry.
Added nucleophiles are expected to afford new function-
ality and new homologues. An understanding of the
(6) (a) Ro¨per, S.; Frackenpohl, J .; Schrake, O.; Wartchow, R.;
Hoffmann, H. M. R. Org. Lett. 2000, 2, 1661. (b) Ro¨per, S.; Wartchow,
R.; Hoffmann, H. M. R. Org. Lett. 2002, 4, 3179.
(7) Ro¨per, S.; Franz, M. H.; Wartchow, R.; Hoffmann, H. M. R. J .
Org. Chem. 2003, 68, 4944.
(8) See also: Ro¨per, S.; Franz, M. H.; Wartchow, R.; Hoffmann, H.
M. R. Org. Lett. 2003, 5, 2773.
2984 J . Org. Chem., Vol. 69, No. 9, 2004

The First and Second Cinchona Rearrangement
SCHEME 4. “Secon d Cin ch on a Rea r r a n gem en t” of C9-Activa ted , So-Ca lled Cin ch Ba ses (6′)R ) H) in
Tr iflu or oeth a n ol: Ca ge Exp a n sion to â-Am in o Ben zoa tes 5-OBz a n d 6-OBz Is Accom p a n ied by
Ster eor eten tive Solvolysis a t C9 w ith In ta ct Aza bicyclic Ca gea
a
Q′ ) 4′-quinolyl; ‘a’ and ‘b’ indicate nucleophilic trajectories.
formed, although in only 5%.9 For comparison trifluoro-
ethanolysis of 3-OMs in the presence of NaOBz instead
of KCN provided benzoate 5-OBz as the only cage-
expanded product (see Scheme 4).
served [2.2.2]azabicyclic bromides 3-Br and 4-Br with
retention of configuration! This efficient conversion of
cinchonine and cinchonidine into the C9-nat-configured
bromides in only two steps is currently without precedent
(Scheme 5). We suggest that in trifluoroethanol both
O-mesylated cinchonidine 3-OMs and cinchonine 4-OMs
enter the second rearrangement manifold (Scheme 4)
with generation of bridged ion i as an intermediate. In
the presence of chloride and fluoride ion cage expanded
5-Cl an 5-F could, however, be isolated. The two bromides
3-Br and 4-Br with intact [2.2.2]azabicyclic moiety
(Scheme 5) appear to be favored over their ring-expanded
isomers under the experimental conditions. In principle,
the second rearrangement is reversible as we demon-
strated in another reaction (see guideline 4 below).8
Attempted preparation of tosylated quinine (6′-R )
OMe; C9-nat configuration) at elevated temperature
furnished a minor amount of cage-expanded chloride 5a -
Cl (see the Supporting Information), which was isolated
as beautiful crystals (4%). The structure and 3R config-
uration were confirmed by X-ray crystallography.10 In-
troduction of chloride was more efficient starting with
activated cinch base 3-OMs (6′-R ) H) and n-Bu4NCl in
trifluoroethanol. Cage-expanded chloride 5-Cl with 3R
configuration was isolated (46%), and both unrearranged
3-Cl and 3-OCH3CF3 were formed with retention at C9
(entry 6).
Similar to the reaction with nitrite ion (entry 1) the
hard fluoride ion was introduced via highly hygroscopic
n-Bu4N+F- (4 equiv of n-Bu4NF, entries 7 and 8).
Surprisingly, treatment of 3-OMs or 4-OMs with
Et4N+Br- in trifluoroethanol furnished only cage con-
Silver Ion Med ia ted “Secon d Cin ch on a Rea r -
r a n gem en t”. As shown 9-epi-configured bromide epi-1-
Br and AgOBz undergo the first Cinchona rearrangement
(Scheme 1). Therefore, the behavior of C9-nat-configured
substrates was of interest. An X-ray crystal study and
NOE showed that in 3-Br (C9-nat configuration) the C9-
Br and N-C8 σ-bonds are antiperiplanar. In fact, silver
ion assisted heterolysis in methanol (Scheme 6) provided
both 5-OMe (16%) and 5-OBz (21%) with a change of
product type and mechanism, although the reaction was
sluggish (complete after 3 d).
(9) For examples of the ionizing and solvating power of trifluoro-
ethanol see: Mayr, H.; Minegishi, S. Angew. Chem., Int. Ed. 2002, 41,
4493. McClelland, R. A. Tetrahedron 1996, 52, 6823. Presumably,
trifluoroethanol is partially deprotonated in the presence of KCN: CF3-
CH2OH + KCN f CF3CH2OK + HCN. For the effect of trifluoroethanol
on protein folding (TFE effect) see: D´ıaz, M. D.; Fiorini, M.; Burger,
K.; Berger, S. Chem. Eur. J . 2002, 8, 1663 and references therein.
(10) Thus, even quinine (6′-R ) OMe) undergoes the second rear-
rangement, although in much lower yield. See also Scheme 3.
Both 3-OMe and 3-OBz were formed with retention of
configuration. We were pleased to find that a solvent
change to trifluoroethanol provided just two products
J . Org. Chem, Vol. 69, No. 9, 2004 2985

Franz et al.
TABLE 1. Tr iflu or oeth a n olysis in th e P r esen ce of Ad d ed Nu cleop h ile
conditionsa
products
entry
substrate
added nucleophile (equiv)
solvent trifluoroethanol
cage expansion
cage conservation
1
2
3
4
4-OMs
4-OMs
3-OMs
3-OMs
n-Bu4N+ONO- (2); then 2 N HCl
n-Bu4NSCN (2)
NaN3 (2)
reflux, 6 h
reflux, 4 h
reflux, 18 h
reflux, 9 h
6-OH (41%)b
6-SCN (89%)
5-N3 (53%)
3-N3 (44%)
3-CN (11%)
KCN (4)
5-CN (40%)
5-OCH2CF3 (5%)
6-CN (56%)c
6-OCH2CF3 (9%)
5-Cl (46%)
3-OCH2CF3 (20%)
5
6
7
4-OMs
3-OMs
4-OMs
KCN (3)
reflux, 5.5 h
reflux, 6.5 h
reflux, 6 h
n-Bu4NCl (1.6)
n-Bu4NFd (4)
3-Cl (28%)
3-OCH2CF3 (2%)
4-F (16%)
6-F (64%)
6-OCH2CF3 (traces)
6-OH (traces)
5-F (61%)
4-OCH2CF3 (6%)
8
3-OMs
n-Bu4NFd (4)
reflux, 24 h
3-F (18%)
5-OH (3%)
3-OCH2CF3 (5%)
3-OMs (11%)
a
b
O-Mesylated cinchonine 4-OMs reacted faster than O-mesylated cinchonidine 3-OMs. See Scheme 4: hydrolysis of the resulting
benzoic acid ester with Ba(OH)2 afforded 3-OH and 4-OH, respectively (80%); see ref 7. c The products with an intact [2.2.2]azabicyclic
d
cage were not isolated in this case. 1 M in THF, contains water (5%).
SCHEME 5. P r ep a r a tion of Un r ea r r a n ged 3-Br
a n d 4-Br fr om 3-OMs a n d 4-OMs, Resp ectively, by
Reten tive Su bstitu tion
SCHEME 6. Solvolysis of C9-n a t Con figu r ed 3-Br
(6′-R ) H) w ith AgOBz in Meth a n ol: “Secon d
Cin ch on a Rea r r a n gem en t”
with an increase in cage expansion from 37% to 61%
(Scheme 7)! Retentive solvolysis is favored on stereoelec-
tronic grounds.
Th e P r ep a r a tion of 9-epi-Con figu r ed Cin ch on a
Alk a loid s: Aqu eou s Hyd r olysis in th e P r esen ce of
Ta r ta r ic Acid a s a P r oton Don or . Hydrolysis of 9-nat-
configured quinine and quinidine mesylates in water in
the presence of tartaric acid has opened an efficient route
to 9-epi-quinine epi-1-OH and 9-epi-quinidine, respec-
tively, in a clean reaction. The azabicyclic skeleton
remains intact in this case and hydrolysis entails nu-
cleophilic attack at carbon C9 with complete inversion
of configuration (Scheme 8). In the presence of tartaric
acid the bridgehead nitrogen is at least partially proto-
nated, so the possibility of the first rearrangement
diminishes (cf. ref 19 below). Further hydrogen-bonding
2986 J . Org. Chem., Vol. 69, No. 9, 2004

The First and Second Cinchona Rearrangement
SCHEME 7. Solvolysis of 3-Br w ith AgOBz in
CF 3CH2OH11
SCHEME 9. Hyd r olysis of 3-OMs a n d 4-OMs in
Wa ter /Ta r ta r ic Acid a
SCHEME 8. Hyd r olysis of 9-n a t-Qu in in e Mesyla te
in th e P r esen ce of Ta r ta r ic Acid 12
contacts between tartaric acid and the alkaloid are
feasible and should help to dissolve the alkaloid in
water.12
The hydrolysis has been extended to the C9-activated
cinchonine and cinchonidine (6′-R ) H). Although the
parent alkaloids are “practically insoluble in water”13 we
observed that O-mesylated cinchonidine 3-OMs first
melts, then dissolves and hydrolyzes on heating in water.
Unlike 9-mesylated quinine 1-OMs (Scheme 8) the analo-
gous cinchonidine substrate 3-OMs furnished a minor
new cage-expanded product 5-OH (2%), which was de-
tectable even in the presence of tartaric acid (1.1 equiv;
Scheme 9, top entry). After an increase of added tartaric
acid to 4.7 equiv, cage expansion was suppressed and
[3.2.2]azabicycle 5-OH was no longer discernible. Com-
plete inversion of configuration to 9-epi-cinchonidine epi-
3-OH was observed (43%), but with diminished conver-
sion. On the other hand, after refluxing in pure water14
(no tartaric acid) the amount of cage-expanded 1,2-amino
alcohol 5-OH rose to 35% (not shown in Scheme 9, see
also Scheme 10 and ref 19 below).
a
O-Mesylated cinch bases (C9-nat configuration) undergo "sec-
ond Cinchona expansion" even in the presence of tartaric acid.
Traces of cinchene are formed (for X-ray crystal structure, see the
Supporting Information).
SCHEME 10. Hyd r olysis of 3-OMs a n d 4-OMs in
P u r e Wa ter : “Secon d Cin ch on a Rea r r a n gem en t”
(11) Reaction rates in MeOH and CF3CH2OH were roughly equal
(Schemes 6 and 7). Inspection of the X-ray crystal structure suggests
that generation of a silver dication is feasible from 3-Br. Complexation
of the bridgehead nitrogen is expected to slow migration of nitrogen.
(12) X-ray crystal structure of quininium hydrogen (S,S)-tartrate
hemihydrate: Ryttersgard, C.; Larsen, S. Acta Crystallogr. Sect. C
1998, C54, 1701. Quinine has three hydrogen bond acceptors excluding
the quinoline nitrogen. Tartaric acid is present in an extended chain
configuration and has four hydrogen bond donor and six potential
acceptor atoms. Cooperative binding seems feasible, in the presence
of 6′-R ) OMe more so than in 6′-R ) H.
(13) Merck Index, 11th ed.; Merck & Co, Inc.: Rahway, NJ , 1989.
(14) For reactivity in an aqueous environment see: Candler, D.
Nature 2002, 417, 491. Ball, P. Nature 2003, 423, 25. See also: Grieco,
P. A. Organic Synthesis in Water; Blackie Academic; London, UK, 1998.
Li, C.-J .; Chan, T.-H. Organic Reactions in Aqueous Media; Wiley: New
York, 1997. Breslow, R. Acc. Chem. Res. 1991, 6, 159.
Hydrolysis of pseudoenantiomeric O-mesylated cincho-
nine 4-OMs furnished up to 20% of cage-expanded 1,2-
amino alcohol 6-OH, even in the presence of tartaric acid
(1.1 equiv)! The inverted product with intact azabicyclic
cage (epi-4-OH, 55%) was also formed under these
conditions.15 This is in agreement with earlier work on
the hydrolysis (1.1 equiv of tartaric acid) of O-mesylated
quinine (C9-nat configuration), which was shown to
J . Org. Chem, Vol. 69, No. 9, 2004 2987

Franz et al.
F IGURE 1. Stabilization versus eclipsing strain in a C9
Cinchona cation (6′-R ) OMe).
F IGURE 2. Preferred conformation of epi-1-OMs.
deliver 9-epi-quinine with 100% inversion (see Scheme
8).
SCHEME 11. P r ep a r a tion of epi-Su lfon a tes
In pure water (no tartaric acid) the yield of cage-
expanded 1,2-amino alcohol 6-OH from O-mesyl cincho-
nine 4-OMs rose to 51% (Scheme 10). The structure of
3S-configured 6-OH was confirmed by single-crystal
X-ray diffraction (Supporting Information). The prepara-
tion of 5-OH and 6-OH in water14 represents green
chemistry. Methanesulfonic acid is inexpensive, bio-
degradable, and toxicologically harmless.
The vinyl and quinolyl group in major 3-hydroxyamine
6-OH are opposite to each other and less crowded than
in minor pseudoenantiomeric 5-OH (cf. also X-ray crystal
structure). The mesylates of the cinch bases (6′-R ) H)
are readily soluble in CF3CH2OH while pure 3-OMs and
4-OMs, without additives, only melt in water. In H2O-
CF3CH2OH (1:1) the yield of cage-expansion rose to 68%
(6-OH, 59%; 6-OCH2CF3 as byproduct, 9%). At higher
temperature (ca. 110 °C) elimination to cinchene was
facilitated (Scheme 9).16,17
Clearly, the second cage expansion (and neighboring
nitrogen participation) is favored for cinch bases (6′-R )
H). In contrast, for quinine and quinidine the C9 cation
is stabilized by the remote methoxy group (ii T iii),
despite eclipsing strain along the quinoline perimeter.
Electron demand at C9 and participation by nitrogen
recede (Figure 1).
SCHEME 12. P r ep a r a tion of r-Am in o Eth er
2-OMe: “F ir st Rea r r a n gem en t”
“F ir st Cin ch on a Rea r r a n gem en t" with ou t Ag+ Is
Also F ea sible: Str in gen t Ster eoch em ica l a n d Ex-
p er im en ta l Con d ition s. Previous studies (conforma-
tional analysis by NOE) showed that the antiperiplanar
requirement of the first rearrangement is fulfilled in epi-
1-Br, which was rearranged in the presence of Ag+
(Scheme 1). Generally 9-epi-quinidine and 9-epi-cinchoni-
dine are tosylated reluctantly, the desired sulfonate epi-
7-OTs and epi-3-OTs being formed with an excess of TsCl
(5 equiv) under forcing conditions. However, tosylation
under modified conditions (addition of DMAP) and es-
pecially mesylation with MeSO2Cl/NEt3 proceed under
mild conditions at room temperature.18 Mesylates were
used preferentially because they can be handled more
readily and are also more soluble than tosylates in
solvents such as methanol.
TABLE 2
NaOBz
[equiv]
t
yield
[%]
yielda
[%]
entry
[h]
T
1
2
3
0
0.1
2
1
4
3
reflux
reflux
reflux
30
55
84
50
68
84
a
Based on recovered starting material.
[2.2.2]octyl moiety faces the simple hydrogen atom at-
tached to neighboring C9.
Antiperiplanarity of the C9-OMs bond and migrating
C7-C8 σ-bond is not only encountered in epi-1-Br, but
also in epi-configured O-mesyl quinine epi-1-OMs. In the
conformation shown (Figure 2) the bulky 1-azabicyclo-
We were pleased to find that methanolysis was com-
plete after 2 h at reflux and provided R-amino ether
2-OMe (81% yield, Scheme 12).
O-Mesylated epi-quinidine epi-7-OMs furnished 8-OMe.
Ring expansion was optimized to 84% (see Table 2). These
experiments also demonstrate the beneficial effect of
added NaOBz, which acts inter alia as a buffer for
liberated methanesulfonic acid.
Under optimized conditions (Table 2, entry 3, OTs and
OMs leaving groups) epi-configured cinch bases were also
ring expanded to 9-OMe and 10-OMe in respectable yield
(Scheme 14).
(15) As another route to 9-epi configured cinch bases with intact
cage the Mitsunobu reaction can be used with moderate success (ca.
50% by standard protocol). See: Hughes, D. L. Org. React. 1992, 42,
335-656. Castro, B. R. Org. React. 1983, 29, 1-162.
(16) Ko¨nigs, W. A. Chem. Ber. 1881, 28, 1854.
(17) See the X-ray crystal structure.
(18) Mesylation with MeSO2Cl/NEt3 involves the sterically less
encumbered sulfene CH2dSO2 as intermediate; see: King, J . F. Acc.
Chem. Res. 1975, 8, 10 and references therein.
2988 J . Org. Chem., Vol. 69, No. 9, 2004

The First and Second Cinchona Rearrangement
SCHEME 13. P r ep a r a tion of r-Am in o Eth er 8-OMe
azabicycle is irreversible and driven by bridgehead
iminium resonance en route to R-amino ethers. The “first
rearrangement” has been realized for the near-prototype
of quincorine (QCI) and quincoridine (QCD) with iodide
as the leaving group and on treatment with silver
benzoate. The nucleophile is introduced quasiequatori-
ally. The “second rearrangement” is currently unknown
in the QCI-QCD area.
3. The (nonplanar) cationic intermediate of the first
Cinchona expansion shows a substantial barrier to
flipping of the three-carbon bridge, in a formal violation
of Bredt’s rule. Usually there is no crossover or leakage
between pseudoenantiomeric systems during the discrete
QCI-QCD expansion (Scheme 2).
SCHEME 14. P r ep a r a tion of r-Am in o Eth er s
9-OMe a n d 10-OMe fr om Cin ch Ba ses: “F ir st
Cin ch on a Rea r r a n gem en t”
4. The second rearrangement can be reversed (e.g.
Scheme 5 and text). For example, we have ring contracted
a 1-azabicyclo[3.2.2]nonane with a quasiequatorial C3
mesyloxy leaving group to the 1-azabicyclo[2.2.2]octane
derivative, as shown by azide trapping.8
5. In general, the “second cage expansion” of C9-nat-
configured substrates is more favorable for cinchonine
than for cinchonidine (e.g. Schemes 4 and 10). Relief of
nonbonded repulsion in the initial state and product is
thought to contribute (see also Table 1 and X-ray crystal
structures).
6. O-Mesylated cinch bases undergo the “second cage
expansion” in water even in the presence of tartaric acid
(1.1 equiv, Scheme 9). Overall, the second cage expansion
has the greater driving force for the two O-mesylated
cinch bases (C9-nat and 6′-R ) H) than for quinine and
quinidine substrates (6′-R ) OMe) (see Figure 1). There
is increased electron demand at carbon C9 of cinch bases
in the transition state.
SCHEME 15. P r ep a r a tion of 2-OCH2CF 3
7. Complete retention of configuration for nucleophilic
attack on the O-mesylated cinch bases 3-OMs and 4-OMs
(C9-nat configuration, 6′-R ) H) is formulated via double
inversion involving N-bridged cation i. Currently, this
retentive solvolysis is unique among C9-activated cinch
bases under neutral conditions (cf. Schemes 4 and 5).
8. Solvolysis of Cinchona substrates with the 6′-
methoxy donor is faster and more prone to E1-like
enamine formation than that of the methoxy-free sub-
strates from cinch bases. On steric grounds a 9-epi-
configured leaving group facilitates formation of enam-
ine.7,8 The second rearrangement of O-mesylated cinch
bases (C9-nat, 6′-R ) H) is faster in CF3CH2OH than in
EtOH (4 h vs 3.5 d).
9. Addition of a salt such as NaOBz is helpful for
intercepting product cations and to buffer liberated
sulfonic acid with partial precipitation of NaOMs.
In conclusion, the set of four activated Cinchona
substrates and the further preparation of the four epi
compounds (C9-nat, C9-epi, 6′-R ) H, OMe) have enabled
us to investigate the product spectrum of nucleophilic
displacement and solvolysis. The unified theory developed
is in excellent agreement with experience accumulated
so far.
In an attempt to prepare R-amino benzoate 2-OBz we
used the solvent trifluoroethanol instead of methanol.
Only the more stable R-amino ether 2-OCH2CF3 was
isolated (73%, Scheme 15).
We propose to rationalize the experimental facts and
previous experience7,8 along a set of general guidelines.
1. In the solvent water with added tartaric acid the
solubility of quinine and quinidine mesylates is enhanced
due to protonation of the bridgehead nitrogen. Further
supramolecular contacts with tartaric acid and water
involving 6′-OMe oxygen and the C9 leaving group as
hydrogen bond donors seem feasible. Simultaneously,
N-protonation (which is not necessarily complete) blocks
cage expansion by the first and perhaps somewhat less
so by the second rearrangement: The [2.2.2]azabicyclic
skeleton stays intact (for cinch bases see guideline 6).
Aqueous hydrolysis of 9-epi-quinine substrates in the
presence of tartaric acid entails 100% retention of con-
figuration.5,19
(19) Protonation of bridgehead nitrogen and molecular recognition
of tartaric acid by further hydrogen bonding contacts are considered
to generate
a hydrophilic pocket guiding attack of C9 by water
molecule(s). The pocket is open for retentive hydrolysis in the 9-epi-
quinine series (see Chart 1), but not so in the 9-nat series: In this
case preorientation favors an SN2-like attack by water; see: Braje, W.
M.; Holzgrefe, J .; Wartchow, R.; Hoffmann, H. M. R. Angew. Chem.,
Int. Ed. 2000, 39, 2085.
2. Under neutral or weakly acidic conditions two
different Cinchona rearrangements have been estab-
lished. Overall, the “first cage expansion” to [3.2.2]-
J . Org. Chem, Vol. 69, No. 9, 2004 2989

Franz et al.
CHART 1. Gen er a l Cin ch on a Ca ge Exp a n sion a n d Solvolysis Sch em e
Regarding the accessibility of the various reaction
channels one must bear in mind that Cinchona alkaloids,
unlike traditional organic substrates, generally have
well-developed conformational minima with compara-
tively high energy barriers. Diastereomeric transition
states are evidently very different. Even pseudoenantio-
meric reaction paths can be distinguished when reactions
at carbon C9 are involved. Some remarkable solvent
effects have also come to light. It is obvious that a host
of new molecular scaffolds with multiple diversity con-
nectors are now accessible from Cinchona alkaloids via
the “first cage expansion” and the “second cage expan-
sion” and their variants. The two rearrangements and
the solvolysis are much more than mechanistic curiosi-
ties. The enantiopure materials are of interest for asym-
metric syntheses and combinatorial and medicinal chem-
istry and set the stage for a fruitful development of the
Cinchona alkaloid area.
2990 J . Org. Chem., Vol. 69, No. 9, 2004

The First and Second Cinchona Rearrangement
chonine 4-OMs (437 mg, 1.17 mmol) was allowed to react
according to general procedure A with n-Bu4NSCN (706 mg,
2.35 mmol) to afford 6-SCN (342 mg, 89%). 1H NMR (400 MHz,
CDCl3): δ 8.91 (d, J ) 4.5 Hz, H-2′), 8.20-8.06 (m, H-8′, H-5′),
7.77-7.69 (m, H-7′), 7.68-7.60 (m, H-6′), 7.31 (d, J ) 4.6 Hz,
H-3′), 5.99 (ddd, J ) 6.9, 10.4, 17.2 Hz, H-10), 5.23-5.01 (m,
H-11, H-11), 4.71 (d, J ) 11.4 Hz, H-2), 4.08-3.98 (m, H-3),
3.65 (dd, J ) 9.6, 15.0 Hz, H-7), 3.27-3.20 (m, H-8), 2.93-
2.83 (m, H-7), 2.63-2.41 (m, H-8. H-4, H-4), 2.33-2.26 (m,
H-6), 1.98-1.88 (m, H-5, H-9), 1.82-1.71 (m, H-9). 13C NMR
(100 MHz, CDCl3): δ 149.3 (CH, C-2′), 149.2 (C, C-10′), 144.4
(C, C-4′), 139.5 (CH, C-10), 130.5 (CH, C-8′), 129.3 (CH, C-7′),
127.4 (C, C-9′), 127.2 (CH, C-5′), 123.6 (CH, C-6′), 116.8 (CH,
C-3′), 115.9 (CH2, C-11), 110.4 (C, C-12, SCN), 69.6 (CH, C-3),
56.5 (CH2, C-7), 44.5 (CH, C-2), 42.9 (CH, C-6), 39.1 (CH2, C-8),
34.1 (CH2, C-4), 32.2 (CH, C-5), 27.8 (CH2, C-9). IR (Golden
Gate ATR) υ 2961 m, 2873 m, 2147 w, 2052 s, 1593 w, 1507
w, 1464 m, 1381 m, 1162 w, 995 w, 881 w, 733 m. MS-MAT
(160 °C): m/z 337 (M+ + 2, 11.86), 335 (M+, 38), 309 (6), 302
(100), 295 (47), 278 (66), 250 (19), 236 (17), 222 (24), 210 (11),
195 (24), 183 (33), 167 (42), 155 (20), 136 (24), 115 (15), 108
(8), 95 (6), 81 (19), 67 (11). HRMS (C20H21N3S) calcd 335.1456,
found 335.1454.
Exp er im en ta l Section
Gen er a l. The atom numbering follows the Cinchona alka-
loid convention. The new ring expanded Cinchona alkaloids
are numbered according to the Beilstein AutoNom.
1H NMR and 13C NMR spectra (peaks assigned by DEPT,
HMQC, HMBC, H, H-COSY) were recorded at 400 and 500
MHz in deuterated chloroform unless otherwise stated, with
tetramethylsilane as internal standard. Mass spectra were
recorded at 70 eV. Preparative column chromatography was
performed on silica gel (particle size 30-60 µm). Analytical
TLC was carried out on aluminum-backed 0.2-mm silica gel
60 F254 plates. Petrol ether (PE) and methyl tert-butyl ether
(MTBE) were distilled before use. Methanol was dried over
calcium hydride and citric acid and distilled over magnesium
before use. The quality of the Cinchona alkaloids may vary
from batch to batch. Usually, the cinch bases (6′-R ) H) are
accompanied by 10,11-dihydro and 6′-OMe derivatives (<10%).
Gen er a l P r oced u r e A for P r ep a r a tion of Rin g-Ex-
p a n d ed 1-Aza bicyclo[3.2.2]n on a n es w ith C3 Su bstitu -
en ts. To a solution of O-mesylated cinchonine or cinchonidine
in trifluoroethanol (5 mL/mmol) was added the nucleophile
(2-3 equiv) and the reaction mixture was refluxed at 80-90
°C for 4-8 h. Progress of the reaction was monitored by thin-
layer chromatography (TLC). Saturated aq NaHCO3 was
added and the aqueous layer was extracted with CH2Cl2. The
combined organic layer was dried (Na2SO4) and the solvent
was removed in vacuo. The crude product was purified by
column chromatography (MTBE-MTBE/MeOH 5:1) to furnish
the ring-expanded products and products with intact cage.
Illu str a tive Solvolysis w ith Th iocya n a te a s Nu cleo-
p h ile: (1S,2R,3S,5S,6R)-2-Qu in olin -4-yl-3-th iocya n a to-6-
vin yl-1-a za -bicyclo[3.2.2]n on a n e 6-SCN. O-Mesylated cin-
Ack n ow led gm en t. We thank Buchler GmbH for
Cinchona alkaloids, InnoChemTech Braunschweig and
the Fonds der Chemischen Industrie for support, Amy
Stockwell for experimental assistance, Dr. E. Hofer for
NMR assistance, and Ulrike Eggert and Lars Ole
Haustedt for their help.
Su p p or tin g In for m a tion Ava ila ble: General procedures
and spectroscopic data for 23 azabicycles with expanded and
intact cage; crystallographic data for 3-Br, 5a -Cl, 6-OH, and
E-cinchene as CIF files. This material is available free of
charge via the Internet at http://pubs.acs.org.
J O030363S
J . Org. Chem, Vol. 69, No. 9, 2004 2991