entries 2 and 3). Only in the case of electron-withdrawing
substrate 1c, product 2g was afforded in moderate yield
(55%, Table 3, entry 4). With 2e, it was possible to remove
the p-nosyl group by treatment with thiophenol/K2CO3 and
afford the unprotected hydrazine 3 (Scheme 2, 75% yield).17
The reaction was then performed with enantiopure
(þ)-1a. Unfortunately, product 2a was isolated in racemic
form using optimized conditions for Cu(II) and Rh(II)
catalysts and variants performed at lower temperatures.
To understand the origin of the racemization, 2a was
separated into single enantiomers using a CSP-HPLC
resolution protocol (see the Supporting Information).
The compound was found to be configurationally stable
in neutral media. However, by subjecting one of the
enantiomers to the reaction conditions for 4 h, a decrease
of the enantiomeric excess of 2a from 99% to 50% in the
case of Rh2(esp)2 and a complete racemization in the pre-
sence of Cu(OTf)2 were observed. This indicates that the
racemization can occur notonlyduringthe formation of2a
but also after the product has been formed.
planarization of the core. Intermediates 5 can be formed
either directly from 4 or from products 2 through a ring
opening induced by the Lewis acid metal catalysts.19
This proposal also explains the lower reactivity of
electron-poor bridgehead nitrogen atoms (e.g., in 1c and
1d) which was further confirmed by an experiment using
€
unsymmetrically substituted Troger base 6 (Scheme 3:
X = NO2, Y = OMe) carrying both electron-donating and
-withdrawing substituents on the two aromatic rings.20
A
single regioisomer was obtained (7, 45%). It results clearly
from an initial attack on the electron-rich nitrogen atom (para
to the methoxy group), the structure being ascertained by
X-ray crystallographic analysis (Supporting Information).21
Finally, to address the lack of configurational stability
of rearranged products 2, the introduction of substituents
ortho to the nitrogen atoms was pursued (Table 4).22
Treatment of 8a (R1 = Me, R2 = H, R3 = Me) with
PhINTs and Rh2(esp)2 afforded desired product 9a in 24%
yield along with novel CꢀH amination product 10a (22%)
as a single diastereomer (Table 4, entry 1). The reaction
required 3 equiv of PhINTs and a longer duration for
completion. Similar results were obtained with Cu(OTf)2
(Table 4, entry 2). The structure of 10a was confirmed by
X-ray analysis. Clearly, in this present case, the hindrance
around the nitrogen atoms23 favors an alternative amina-
tion on the benzylic position. A similar reactivity was
observedwith8band 8ccarryingortho-methyl substituents
(Table 4, entries 3 and 4).24 A stronger inclination toward
the CꢀH insertion reaction was noticed in the presence of
bromine atoms at meta and ortho positions (Table 4,
entries 4 and 5). Substrate 8a was then resolved into single
enantiomers.25 Satisfyingly, under Cu(OTf)2 catalysis, the
reaction of (ꢀ)-8a yielded rearranged product (ꢀ)-9a with
a complete transfer of chirality (8a g 99% ee).26 In the
presenceofRh2(esp)2, using(þ)-8a asthe substrate, (þ)-9a
was obtained in 92% ee along with racemic 10a.
The racemization and the insertion reactivity as a whole
can be rationalized in terms of mechanism (Scheme 3).
It involves the catalytic generation of electrophilic metal
€
nitrenes and additions of Troger bases 1 to these inter-
mediates. Sulfonyl aminimide moieties of type 4 result.
€
Then, by a CꢀN bond cleavagethatistraditionalinTroger
base chemistry upon quaternization of one of the nitrogen
atoms,18 zwitterionic ring-opened species of type 5 are
formed that collapse subsequently to form the imino-
€
methano Troger bases 2. Under this hypothesis, the ob-
served loss of enantiomeric purity results probably
from the formation of monocyclic intermediates 5 which
have the possibility to racemize through a complete
(17) (a) Fukuyama, T.; Jow, C. K.; Cheung, M. Tetrahedron Lett.
1995, 36, 6373–6374. (b) Mao, H.; Joly, G. J.; Peeters, K.; Hoornaert,
G. J.; Compernolle, F. Tetrahedron 2001, 57, 6955–6967.
€
(18) Trapp, O.; Trapp, G.; Kong, J. W.; Hahn, U.; Vogtle, F.;
Schurig, V. Chem.;Eur. J. 2002, 8, 3629–3634.
€
In conclusion, methano-Troger bases 1 react with aryl-
sulfonyl iminophenyliodinanes (Cu(II) or Rh(II) catalysis,
€
(19) The Troger base catalyzed aziridination of chalcones with
€
20 to 40 °C) to form imino-methano Troger bases 2 in
O-mesitylenesulfonylhydroxylamine as reagent has been reported
(Shen, Y.-M.; Zhao, M.-X.; Xu, J.; Shi, Y. Angew. Chem., Int. Ed.
2006, 45, 8005–8008). Under these conditions, an aminimide intermedi-
ate could formally form. However, if it is the case, it does not undergo the
subsequent rearrangement observed herein.
(20) Pardo, C.; Ramos, M.; Fruchier, A.; Elguero, J. Magn. Reson.
Chem. 1996, 34, 708–710.
good yields (55ꢀ88%). Sulfonyl aminimides are probably
formed that insert in adjacent CꢀN bonds via a two-step
€
mechanism. While classic enantiopure Troger bases afford
racemic products, bis-ortho substituted derivatives can
react with a full transfer of chirality (ee g 99%).
(21) Interestingly, this reaction contradicts the regioselectivity ob-
served in the reaction of 6 with R-diazo diester reagents under Rh(II)-
catalysis at elevated temperatures (ref 5a). It suggests that the first step of
these nitrene-mediated reactions is strictly under kinetic rather than
thermodynamic control.
Acknowledgment. We thank the University of Geneva,
the Swiss National Science Foundation, the NCCR Chem-
ical Biology for financial support. We acknowledge the
contributions of the Sciences Mass Spectrometry (SMS)
platform at the Faculty of Sciences, University of Geneva.
(22) It has been established that compounds of type 8 (R3 = Me or Br),
€
unlike Troger bases 1, are highly stable as single enantiomers, even in
strongly acidic media: Lenev, D. A.; Lyssenko, K. A.; Golovanov, D. G.;
Buss, V.; Kostyanovsky, R. G. Chem.;Eur. J. 2006, 12, 6412–6418.
(23) (a) Gao, X.; Hampton, C. S.; Harmata, M. Eur. J. Org. Chem.
2012, 7053–7056. (b) Pereira, R.; Cvengros, J. J. Organomet. Chem.
2013, 729, 81–85.
(24) All products 10 are afforded as single diastereomers. The
proposed configuration is based on the X-ray diffraction analysis of
10a (see Supporting Information).
(25) Didier, D.; Tylleman, B.; Lambert, N.; Velde, C.; Blockhuys, F.;
Collas, A.; Sergeyev, S. Tetrahedron 2008, 64, 6252–6262.
(26) Insertion product 10a is again obtained as a single diastereomer
but only in 19% ee.
Supporting Information Available. Synthesis and spec-
€
tral caracterization imino-methano Troger bases 2. CSP-
HPLC determination of the enantiomeric purity of 9a.
This material is available free of charge via the Internet at
The authors declare no competing financial interest.
Org. Lett., Vol. 15, No. 15, 2013
3933