Asymmetric synthesis of highly substituted azapolycyclic compounds via 2-alkenyl sulfoximines: Potential scaffolds for peptide mimetics

Article abstract of DOI:10.1021/ja057012a

The application of metalated, enantiomerically pure acyclic and cyclic 2-alkenyl sulfoximines for the synthesis of highly substituted aza(poly)cyclic ring systems is described. The method relies on a one-pot combination of a reagent-controlled allyl transfer reaction to α- or β-amino aldehydes, followed by a Michael-type cyclization of the intermediate vinyl sulfoximines generated in the first step. The sulfur-free target compounds are preferentially obtained by samarium iodide treatment of the sulfonimidoyl substituted heterocycles. In addition to this methodological work, initial results on the biological activity of selected examples are reported. Furthermore, a concept for the transformation of peptidic lead structures into non-peptide mimetics is described, and the relevance of the new approach to highly substituted azaheterocycles in this context is discussed.

Full text of DOI:10.1021/ja057012a

Published on Web 03/08/2006
Asymmetric Synthesis of Highly Substituted Azapolycyclic
Compounds via 2-Alkenyl Sulfoximines: Potential Scaffolds
for Peptide Mimetics
Michael Reggelin,* Bernd Junker,‡ Timo Heinrich,§ Stefan Slavik, and Philipp Bu¨hle|
Contribution from the Clemens-Scho¨pf-Institute of Organic Chemistry and Biochemistry,
Technical UniVersity of Darmstadt, Petersenstrasse 22, 64287 Darmstadt, Germany
Received October 21, 2005; E-mail: re@punk.oc.chemie.tu-darmstadt.de
Abstract: The application of metalated, enantiomerically pure acyclic and cyclic 2-alkenyl sulfoximines for
the synthesis of highly substituted aza(poly)cyclic ring systems is described. The method relies on a one-
pot combination of a reagent-controlled allyl transfer reaction to R- or â-amino aldehydes, followed by a
Michael-type cyclization of the intermediate vinyl sulfoximines generated in the first step. The sulfur-free
target compounds are preferentially obtained by samarium iodide treatment of the sulfonimidoyl substituted
heterocycles. In addition to this methodological work, initial results on the biological activity of selected
examples are reported. Furthermore, a concept for the transformation of peptidic lead structures into non-
peptide mimetics is described, and the relevance of the new approach to highly substituted azaheterocycles
in this context is discussed.
Chart 1. Examples for “Type-III Mimetics”
Introduction
In a seminal paper published by Farmer more than 20 years
ago the author defines peptidomimetics as novel scaffolds
designed to replace the entire peptide backbone while retaining
the isosteric topography of the enzyme-bound peptide (or
assumed receptor-bound) conformation.1 Since that time the term
“peptidomimetic” has been given so many different meanings
that there is now a considerable confusion of ideas. In 1998
Rich et al. presented a classification scheme which very much
clarified the situation.2 According to this work three types of
peptidomimetics have to be distinguished (type-I to -III). Among
them “type-III mimetics represent the ideal in peptidomimetics
in that they possess noVel templates which, though appearing
unrelated to the original peptides, contain the necessary groups
positioned on a noVel nonpeptide scaffold to serVe as topo-
graphical mimetics”.2 The synthetic methodology to be de-
scribed in this article aims at the synthesis of this kind of
peptidomimetics. In the meantime quite a number of type-III
mimetics have been synthesized (Chart 13-10) among which the
two piperidines Roche-1 and Roche-2 merit special interest.
These two non-peptide peptidomimetic inhibitors of renin were
‡ Current address: Aventis Pharma Deutschland GmbH, Process De-
velopment Chemistry, Industriepark Ho¨chst, G 838, 65926 Frankfurt,
Germany.
§ Current address: Merck KGaA, Med Chem DA3, Frankfurter Str. 250,
64293 Darmstadt, Germany.
found to stabilize an enzyme conformation not previously
observed for this enzyme.11
| Current address: Merck KGaA, Pigments R&D Cosmetics, Frankfurter
Str. 250, 64293 Darmstadt, Germany.
Although the question whether these non-peptidic leads are
really true mimics of the peptide is a matter of debate,12,13 many
(1) Bridging the gap between bioactiVe peptides and nonpeptides: some
perspectiVes in design; Farmer, P. S., Ed.; Academic Press: New York,
1980; Vol. 11, p 119.
(2) Ripka, A. S.; Rich, D. H. Curr. Opin. Chem. Biol. 1998, 2, 441.
(3) Belanger, P. C.; Dufresne, C. Can. J. Chem. 1986, 64, 1514.
9
10.1021/ja057012a CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 4023-4034
4023

A R T I C L E S
Reggelin et al.
Figure 1. â-Turn structure and definitions of the torsional angles æ and ψ (left). Topographical abstraction and definition of â as the angle between the
planes spanned by the atoms connected in blue with those connected in red (right).
Scheme 1. General Outline of the Method; Sij Denotes the Chiral
of them display high levels of biological activity. Despite these
Sulfonimidoyl Unit as Specified in Chart 3
successes, the structural transition from a bioactive peptide to
a nonpeptidic compound is still a very tedious process, involving
the identification of the crucial amino acid side chains, the
establishment of their spatial relationship, and finally the
selection of an organic template suitable for reproducing the
geometry of the pharmacophoric model.
To solve the latter problem a structural variable connecting
the peptide with the “non-peptide world” would improve the
situation considerably. If one assumes, and there is much
evidence for the validity of this assumption,14-16 that â-turns
play a dominant role in the molecular recognition process, one
relationship between these two conformational subunits. This
such variable has been proposed by Ball in 1993.17 Contrasting
way â is a measure of the relative orientation of the exposed
the common description of â-turn topography based on the
side chains at positions 2 and 3 (RR2 and RR3) which are
backbone angles æ and ψ, he proposed a pseudo-dihedral â as
suspected to be of major importance in the ligand/receptor
a means to describe and to compare the spatial relationship of
interaction. With these ideas in mind it seems promising to
potentially pharmacophoric groups in both peptides and non-
develop a new synthesis of highly functionalized nitrogen
peptides (Figure 1).
heterocycles giving access to a broad range of compounds
The basis for this approach is the observation that the
displaying pharmacophoric groups under a wide range of â. This
conformational relationship between bonds 1 and 2 and atom
entails the necessity to find a synthetic protocol flexible enough
RC3 (as well as bonds 3/4 and RC2) remains approximately
to generate mono- and polycyclic systems with varying side
constant for any turn containing trans peptide bonds. The
chains and maximum control of both the relative and absolute
pseudo-dihedral â [TORS (C1, R2, R3, N4) thus describes the
configuration at the stereogenic centers. In this article we would
like to describe our efforts toward this goal.
(4) Snider, R. M.; Constantine, J. W.; Lowe, J. A., III; Longo, K. P.; Lebel,
W. S.; Woody, H. A.; Drozda, S. E.; Desai, M. C.; Vinick, F. J.; Spencer,
R. W.; Hess, H.-J. Science 1991, 251, 435.
(5) McLean, S.; Ganong, A. H.; Seeger, T. F.; Bryce, D. K.; Pratt, K. G.;
Reynolds, L. S.; Siok, C. J.; Lowe, J. A., III; Heym, J. Science 1991, 251,
437.
(6) Lam, P. Y. S.; Jadhav, P. K.; Eyermann, C. J.; Hodge, C. N.; Ru, Y.;
Bacheler, L. T.; Meek, O. M. J.; Rayner, M. M. Science 1994, 263, 380.
(7) Hirschmann, R.; Nicolaou, K. C.; Pietranico, S.; Salvino, J.; Leahy, E. M.;
Sprengeler, P. A.; Furst, G.; Smith, A. B., III; Strader, C. D.; Cascieri, M.
A.; Candelore, M. R.; Donaldson, C.; Vale, W.; Maechler, L. J. Am. Chem.
Soc. 1992, 114, 9217.
(8) Guller, R.; Binggeli, A.; Breu, V.; Bur, D.; Fischli, W.; Hirth, G.; Jenny,
C.; Kansy, M.; Montavon, F.; Muller, M.; Oefner, C.; Stadler, H.; Vieira,
E.; Wilhelm, M.; Wostl, W.; Marki, H. P. Bioorg. Med. Chem. Lett. 1999,
9, 1403.
(9) Vieira, E.; Binggeli, A.; Breu, V.; Bur, D.; Fischli, W.; Guller, R.; Hirth,
G.; Marki, H. P.; Muller, M.; Oefner, C.; Scalone, M.; Stadler, H.; Wilhelm,
M.; Wostl, W. Bioorg. Med. Chem. Lett. 1999, 9, 1397.
(10) Oefner, C.; Binggeli, A.; Breu, V.; Bur, D.; Clozel, J. P.; D’Arcy, A.; Dorn,
A.; Fischli, W.; Gruninger, F.; Guller, R.; Hirth, G.; Marki, H. P.; Mathews,
S.; Muller, M.; Ridley, R. G.; Stadler, H.; Vieira, E.; Wilhelm, M.; Winkler,
F. K.; Wostl, W. Chem. Biol. 1999, 6, 127.
General Outline of the Method
In 1994 we discovered that titanated open-chain 2-alkenyl
sulfoximines 3, derived from cyclic sulfonimidates 1, are highly
stereoselective allyl transfer reagents (Scheme 1).18 Shortly after
that, Gais et al. introduced N-methyl substituted 2-alkenyl
sulfoximines which he elaborated into useful synthetic tools for
asymmetric synthesis.19
In the time to follow we developed this chemistry further by
the introduction of cyclic 2-alkenyl sulfoximines 2 and the
exploitation of the electron-deficient double bond in the primary
products 5 of the allyl transfer reaction.20-23
(18) Reggelin, M.; Weinberger, H. Angew. Chem., Int. Ed. Engl. 1994, 33, 444.
(19) For excellent related work on allylic sulfoximines see: (a) Gais, H. J.;
Bruns, P. R.; Raabe, G.; Hainz, R.; Schleusner, M.; Runsink, J.; Babu, G.
S. J. Am. Chem. Soc. 2005, 127, 6617. (b) Gais, H.-J.; Babu, G. S.; Guenter,
M.; Das, P. Eur. J. Org. Chem. 2004, 1464. Gais, H.-J.; Hainz, R.; Muller,
H.; Bruns, P. R.; Giesen, N.; Raabe, G.; Runsink, J.; Nienstedt, S.; Decker,
J.; Schleusner, M.; Hachtel, J.; Loo, R.; Woo, C.-W.; Das, P. Eur. J. Org.
Chem. 2000, 3973. Gais, H.-J.; Mueller, H.; Decker, J.; Hainz, R.
Tetrahedron Lett. 1995, 36, 7433.
(11) Bursavich, M. G.; Rich, D. H. J. Med. Chem. 2002, 45, 541.
(12) Fletcher, M. D.; Campbell, M. M. Chem. ReV. 1998, 98, 763.
(13) Muller, G.; Hessler, G.; Decornez, H. Y. Angew. Chem., Int. Ed. 2000, 39,
894.
(14) Hruby, V. J. Life Sci. 1982, 31, 189.
(15) Richardson, J. S. AdV. Protein Chem. 1981, 34, 167.
(16) Rose, G. D.; Gierasch, L. M.; Smith, J. A. AdV. Protein Chem. 1985, 37,
1.
(17) Ball, J. B.; Hughes, R. A.; Alewood, P. F.; Andrews, P. R. Tetrahedron
1993, 49, 3467.
(20) Reggelin, M.; Gerlach, M.; Vogt, M. Eur. J. Org. Chem. 1999, 1011.
(21) Reggelin, M.; Heinrich, T. Angew. Chem., Int. Ed. 1998, 37, 2883.
9
4024 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006

Potential Scaffolds for Peptide Mimetics
A R T I C L E S
Scheme 3. Target Structure 14 and Encoding of the Ring
Chart 2. Oxacyclic Compounds Accessible via 2-Alkenyl
Sulfoximines and Oxygen Substituted Aldehydes; Sij Denotes the
Chiral Sulfonimidoyl Unit as Specified in Chart 3
Systems (a: absent; p: present)
Scheme 2
Chart 3. 2-Alkenyl Sulfoximines Used in This Study, Coding of
the Absolute Configuration, and the O-Substitution in the Sulfur
Moiety
These efforts led to the synthesis of isomerically pure highly
substituted tetrahydrofuranes 7,22 2-oxabicyclo[3.3.0]octanes 8,
and 2-oxabicyclo[4.3.0]nonanes 920 (Chart 2).
As a consequence of this successful elaboration, we began
to realize that these syntheses may be just aspects of a widely
applicable protocol for the asymmetric synthesis of highly
substituted (poly)heterocyclic ring systems (Scheme 2).
The retrosynthetic analysis of the generalized (poly)hetero-
cyclic structure 10 (X ) O, N) should lead to allylic sulfox-
imines 2 and heteroatom-substituted aldehydes 11. Both frag-
ments may or may not contain rings of variable size and
substitution. The sulfoximines 2 can be prepared starting from
the cyclic sulfonimidates 1a or 1b24-26 following published
procedures.18,20,23
A closer look at formula 10 (X ) N) reveals that the structure
is at least quadricyclic if rings 1-3 are present. The central
ring is built up in the course of the reaction sequence, whereas
the other three rings are components of the starting materials.
This analysis of the target structure leads to an obvious
classification: Each ring may be either present (p) or absent
(a) entailing the principal accessibility of 23 ) 8 structural types.
These can be encoded in a binary fashion reaching from type
000 (all rings absent) to 111 (all rings present) (Scheme 3).
In this article we would like to present our results on the
synthesis of the highly substituted aza-(poly)cyclic ring systems
of general formula 14 (Scheme 3).
can take four different values (SS,SC: i ) 1; SS,RC: i ) 2; RS,RC:
i ) 3; RS,SC: i ) 4). The second superscript j can adopt three
different values depending on the nature of the protecting group
Y (Y ) TBS: j ) a; Y ) TMS: j ) b; Y ) H: j ) c). To
facilitate the recognition of the structures, the value of j is also
used in the compound numbers 16-21. A comprehensive
example to clarify the use of this system is also given in Chart
3.
The R-amino aldehydes needed were prepared starting from
the corresponding amino acids by the reduction-reoxidation
sequences depicted in Schemes 4 and 5. To avoid a symmetrical
intermediate, serines 22c and ent-22c had to be protected by
tBu on both oxygens27 prior to lithium aluminum hydride
(LAH)-reduction. For a successful oxidation of the tryptophane
derived alcohol 24f, Boc-protection of the indole NH was found
Synthesis of the Starting Sulfoximines and Aldehydes
The 2-alkenyl sulfoximines used in this study are summarized
in Chart 3. Their synthesis has already been described.18,20,23
To save journal space, the chiral sulfur moiety has been given
the symbol Sij as shown in Chart 3. The first superscript i
encodes the absolute configuration of the auxiliary and therefore
(22) Reggelin, M.; Weinberger, H.; Heinrich, T. Liebigs Ann. Recl. 1997, 1881.
(23) Reggelin, M.; Weinberger, H.; Gerlach, M.; Welcker, R. J. Am. Chem.
Soc. 1996, 118, 4765.
(24) Reggelin, M.; Weinberger, H. Tetrahedron Lett. 1992, 33, 6959.
(25) Reggelin, M.; Welcker, R. Tetrahedron Lett. 1995, 36, 5885.
(26) Reggelin, M.; Junker, B. Chem.sEur. J. 2001, 7, 1232.
(27) Schroeder, E. Liebigs Ann. 1963, 670, 127.
9
J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006 4025

A R T I C L E S
Reggelin et al.
Scheme 4. Synthesis of the R-Aminoaldehydes 25a-25ea
Scheme 6. Synthesis of the â-Amino Aldehydesa
a (a) LiAlH4, THF, 65 °C. (b1) Fmoc-Cl, NaHCO3, dioxane, H2O, rt.
(b2) Phthalic anhydride, 4. (c) DMP, CH2Cl2, 0 °C to rt. (d) Isobutene,
H2SO4, dioxane.
Scheme 5. Synthesis of the Protected Tryptophanal 25fa
a (a1) Fmoc-Cl, NaHCO3, dioxane, H2O, rt. (a2) Phthalic anhydride, 4.
(b) DMP, CH2Cl2, 0 °C to rt.
aldehyde 25k. This aldehyde is also accessible via indoline-7-
carbaldehyde 25m by γ-MnO2 oxidation in quantitative yield.
This alternative pathway is very interesting, because it avoids
the multistep preparation of ester 32 and allows for the
preparation of all four aldehydes 25k-n via Boc-indoline as a
common precursor.
Synthesis of the Azaheterocycles: From Acyclic
Sulfoximines
As already stated, we would like to organize the material
using the binary encoding scheme depicted in Scheme 3. With
acyclic sulfoximines as starting material, ring 1 (see Scheme
3) is always absent and therefore only type-0XY systems are
covered in this section.
Type-000 Azacycles: Substituted Pyrrolidines and Pip-
eridines. From Scheme 3 it is obvious that both the aldehyde
and the sulfoximine have to be acyclic. Pyrrolidines are obtained
if R-amino aldehydes (m ) 0) are used (Scheme 7, Table 1),
whereas â-amino aldehydes (m ) 1) give rise to the synthesis
of piperidines, as will be discussed later.
After deprotonation of the sulfoximines using nBuLi in
toluene, followed by transmetalation with chlorotris(iso-
propoxy)titanium (ClTi(OiPr)3) and addition of the protected
aldehydes 25, the titanated 5-amino-4-hydroxy-vinyl sulfox-
imines 35a-f and 36a-g were obtained. These intermediates
were not isolated but treated with 10 equiv of piperidine
typically after 1 h of reaction time at -78 °C. This treatment
releases the N-nucleophile, which attacks intramolecularly the
acceptor-substituted double bond of the vinyl sulfoximines
(Figure 2).
a (a) LiAlH4, THF. (b) Fmoc-Cl, NaHCO3, dioxane, H2O, rt. (c) Me2NEt,
TMSCl, rt. (d) Boc2O, CH3CN, DMAP. (e) 0.5 N HCl. (f) DMP, CH2Cl2,
0 °C to rt.
to be advantageous (Scheme 5). The whole sequence including
LAH-reduction, protection, and oxidation of the protected amino
alcohols 24 by Dess-Martin periodinane (DMP) proceeds
without notable racemization21,28 as checked by NMR-shift
experiments using Pr(hfc)3 as a chiral shift reagent. The R-amino
aldehydes 25a-f were used without further purification, and
the overall yield starting from the amino acids usually exceeded
70%.
The preparation of the â-aminoaldehydes 25g-m is described
in Scheme 6. Aldehydes 25g, 25h, and 25i were prepared in a
straightforward manner from the commercially available amino
alcohols 28 and 30. For the Boc-protected 7-formyl indole 25l
we started from ester 32,29 whereas for the indoline derived
aldehydes 25n and 25m Boc-indoline30-32 was used as starting
material. LAH-reduction of ester 32 proceeded with 97% yield
to alcohol 33 which was oxidized by DMP in 78% yield to
(28) Myers, A. G.; Zhong, B.; Movassaghi, M.; Kung, D. W.; Lanman, B. A.;
Kwon, S. Tetrahedron Lett. 2000, 41, 1359.
(29) We thank Merck (Darmstadt) for a generous gift of ester 32.
(30) Iwao, M.; Kuraishi, T. Org. Synth. 1996, 73, 85.
(31) Meyers, A. I.; Milot, G. J. Org. Chem. 1993, 58, 6538.
(32) Iwao, M.; Kuraishi, T. Heterocycles 1992, 34, 1031.
After removal of the dibenzofulven-piperidine adduct by
crystallization from methanol and a short column filtration, the
side chain of the auxiliary was desilylated and the ring nitrogen
Boc-protected. The removal of the sulfonimidoyl moiety in the
9
4026 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006

Potential Scaffolds for Peptide Mimetics
A R T I C L E S
Scheme 7. Highly Substituted Pyrrolidines from Acyclic 2-Alkenyl
by the prochirality of the double bond geometry in the 2-alkenyl
sulfoximine and the sense of chirality at sulfur in the auxiliary,
respectively.18,20,21,23
Sulfoximinesa
All pyrrolidines derived from 21b (S-configuration at sulfur)
have absolute configurations at C-3 and C-4 being in accordance
with a mutual Re-face attack of the aldehyde and the allylic
reaction partner. The opposite is true for the products derived
from the RS configured sulfoximines ent-21b.
It is worth mentioning that even in the cases of anti-Cram
combinations (RS-configured sulfoximines-S-configured amino
aldehydes) the configurational preferences of the aldehydes are
completely overcompensated (reagent control). This means that
the induced absolute configuration at C-3/C-4 is correlated to
the absolute configuration of the auxiliary in a predictable
manner. The stereochemical outcome at C-5 (pyrrolidine
numbering), the stereogenic center being created during the
cyclization event, is governed by conformational preferences
of the cyclization precursors, the vinylic sulfoximines. The
experimental results are in accordance with the assumption that
the absolute topicity of attack of the N-nucleophile onto C-5
(pyrrolidine numbering) of 35 and 36 is dominated by the
minimization of allylic strain and 1,5-repulsion (Figure 2).
For the terminally substituted allylic sulfoximines 21b/ent-
21b this is fulfilled best in conformation A. In contrast, for
sulfoximines 20b/ent-20b lacking the terminal substituent,
conformation B is a reasonable alternative leading to an
inversion of the absolute topicity of attack onto the double bond.
However, due to 1,5-repulsion B is still energetically disfavored
compared to A, so that even in the 4-unsubstituted (pyrrolidine
numbering) cases the 3,5-trans pyrrolidines should be obtained
in excess, which is indeed the case (Table 1, 39a-c; 40a-c).
The relative and absolute configuration of the pyrrolidines
and those of the minor diastereomers epimeric at C-5 have been
proven by X-ray crystallographic analysis of 39d,33 2-epi-40g,34
40d,34 5-epi-39a,35 and 5-epi-40a,36 respectively.
a (a) nBuLi, -78 °C. (b) ClTi(OiPr)3, 0 °C. (c) Aldehyde 25, -78 °C.
(d) Piperidine, 0 °C, (NH4)2CO3. (e) K2CO3, MeOH. (f) Boc2O, NaHCO3.
(g) SmI2.
Table 1. Highly Substituted Pyrrolidines from Metalated Acyclic
2-Alkenyl Sulfoximinesa
From the above analysis of the stereochemical outcome of
the cyclization the expectation follows that cyclic 2-alkenyl
sulfoximines with the C-4-substituent (pyrrolidine numbering)
being incorporated into the ring (Scheme 2, formula 2) should
react with a comparable stereoselectivity as 21/ent-21 (Chart
3). This expectation is fulfilled perfectly as will be discussed
later.
Furthermore, the mode of stereocontrol entails the logical
consequence that all enantiomeric pyrrolidines can be prepared
simply by switching both the absolute configuration of the
aldehyde and the sulfoximine.
Following the general outline depicted in Scheme 3 we next
increased m in 14 from 0 to 1 within the subgroup of 000-type
heterocycles, which should lead to substituted piperidines
(Scheme 8).
Indeed, reacting the â-amino aldehyde 25h with the titanated
sulfoximine derived from ent-21b followed by hydrazine-
induced cyclization (the motivation for the development of this
alternative cyclization protocol will be discussed in a later
a Compounds in shaded rows have been characterized by X-ray structural
analysis. For 39a and 40a the structures of their respective minor
diastereomers (5-epi) have been studied. The .cif files can be found in the
Supporting Information (#.cif). b Determined by 1H NMR spectroscopic
analysis of the crude reaction mixture. The ds-value corresponds to the
selectivity of the cyclization. Therefore the minor diastereomer has the
inverted absolute configuration at C-5. c (c in dichloromethane). d N-
Deprotected and isolated as p-toluenesulfonate; value of optical rotation
determined in methanol.
thus obtained pyrrolidine derivatives 37a-f and 38a-g was
accomplished by SmI2 in methanol/THF as described before.21
(33) Bats, J. W.; Heinrich, T.; Reggelin, M. Acta Crystallogr. 1998, C54,
IUC9800033.
The relative and absolute configurations of the newly created
stereogenic centers in the allyl transfer step (at C-3 and C-4,
Scheme 7) are a consequence of the uniform relative topicity
(like) of attack of the reactive partners onto each other governed
(34) Bolte, M. Acta Crystallogr. 1997, C53, IUC9700028.
(35) Bats, J. W.; Heinrich, T.; Reggelin, M. Acta Crystallogr. 1998, C54,
IUC9800039.
(36) Bats, J. W.; Heinrich, T.; Reggelin, M. Acta Crystallogr. 1998, C54,
IUC9800038.
9
J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006 4027

A R T I C L E S
Reggelin et al.
Figure 2. Putative transition state model of the cyclization reaction. Val is used as an abbreviation for the trimethylsilylated side chain of the sulfoximine
moiety.
Scheme 8 a
Scheme 10. Type-010 and -011 Heterocyclesa
a (a) nBuLi, -78 °C. (b) ClTi(OiPr)3, -78 °C to 0 °C. (c) -78 °C, 25h.
(d) Hydrazine (80% in H2O), EtOH. (e) K2CO3, MeOH.
Scheme 9 a
a (a) nBuLi, -78 °C. (b) ClTi(OiPr)3, -78 °C to 0 °C. (c) 25x, -78 °C.
(d) (NH4)2CO3. (e) K2CO3, MeOH. (f) Piperidine, THF, rt.
imine 44. Although the yield was quite disappointing (29%)
the compound was obtained as a single diastereomer. The
subsequent piperidine-induced cyclization delivered the desired
benzopiperidine derivative 45 with moderate 40% yield, again
as a single isomer. The overall yield of 45 could be increased
to 27% if the one-pot protocol was applied. With crotyl
sulfoximine 21b and the indoline derived aldehyde 25m (see
Scheme 6) the tricyclic system 47 (type 011) was obtained by
spontaneous cyclization of the primary addition product 46 with
33% overall yield. It is worth mentioning that it is not necessary
to N-protect the aldehyde 25m due to the very low basicity of
the titanium reagent produced in step (b) of the sequence.
a (a) nBuLi, -78 °C. (b) ClTi(OiPr)3, -78 °C to 0 °C. (c) 25e, -78 °C.
(d) (NH4)2CO3. (e) Piperidine, 0 °C. (f) K2CO3, MeOH. (g) Amount of 43
in the crude reaction mixture of the one-pot sequence as judged from H
1
NMR.
section) led to piperidine 41 with 68% yield under complete
stereocontrol.
Type-001, -010, and -011 Heterocycles. Type-001 hetero-
cycles should be accessible using an open chain sulfoximine
and a heteromonocyclic amino aldehyde. To verify that expecta-
tion we used sulfoximine ent-21b and Fmoc-protected prolinal
25e as starting materials (Scheme 9).
Synthesis of Azaheterocycles: From Cyclic
Sulfoximines
In the first experiment we tried to apply the one-pot protocol
to access 43 without isolating the intermediate vinyl sulfoximine
42 which, in analogous cases, proved to be detrimental in terms
of overall reaction yield. The pyrrolizidine was formed indeed
Type-100 Heterocycles: 2-Azabicyclo[x.y.0]alkanes. The
reaction between the cyclic 2-alkenyl sulfoximines 12a and 12b
and either R-amino aldehydes or â-amino aldehydes 48 (m ) 0
or m ) 1, respectively) should give rise to the formation of
substituted 2-azabicyclic derivatives 49-52 as depicted in
Scheme 11.
2-Azabicyclo[3.3.0]octanes 50. For the synthesis of the
2-azabicyclo[3.3.0]octanes 50 we employed three different
starting materials differing in the relative and absolute config-
uration of the stereogenic centers in the sulfur moiety (Scheme
12, boxed structure). As described earlier, the vinyl sulfoximine
intermediates obtained after the deprotonation, transmetalation,
and γ-hydroxyalkylation sequence (a-c in Scheme 12) were
not isolated but directly converted to the corresponding bicycles
using either piperidine (50a-c, 50h-k) or hydrazine (50d-g)
to induce the cyclization (Scheme 12, Table 2).
1
with a yield around 90% as judged from the H NMR of the
crude reaction mixture, but unfortunately we had difficulties in
purifying the highly polar material. To get 43 in a pure state
the experiment was repeated, this time isolating 42 albeit with
a rather low yield of 35%. To our delight this material underwent
smooth cyclization after the addition of piperidine, and analyti-
cally pure pyrrolizidine 43 could be isolated as a single isomer.
A crystal structural analysis revealed the absolute configuration
of the stereogenic centers as drawn.37
As examples for type-010 and -011 systems, the bicyclic and
tricyclic compounds 45 and 47 were prepared, respectively
(Scheme 10).
To make sure that the γ-hydroxyalkylation of 21a with
aldehyde 25i (see Scheme 6) proceeded with complete stereo-
control as expected, we isolated the intermediate vinyl sulfox-
For the synthesis of the serine-derived 2-aza-bicycles 50d-
50g we found it advantageous to use the Phth-protected
aldehydes 25c and ent-25c (Scheme 4) instead of the corre-
sponding Fmoc derivatives. In these cases the cyclization of
(37) Heinrich, T.; Reggelin, M.; Bats, J. W. Acta Crystallogr. 1999, C55,
IUC9900131.
9
4028 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006

Potential Scaffolds for Peptide Mimetics
A R T I C L E S
Scheme 11. General Scheme for the Synthesis of
Table 2. 2-Azabicyclo[3.3.0]octanesa
2-Azabicyclo[x.y.0]alkanes from Cyclic Sulfoximines
Scheme 12. 2-Azabicyclo[3.3.0]octanesa
a Compounds in shaded rows have been characterized by X-ray structural
analysis. b (c in dichloromethane). c X-ray of Boc-protected and desulfurized
compounds (Boc-72x, see Table 4). d Aldehyde addition at 0 °C.
Scheme 13. Potential Problem with the Hydrazine-Induced
Cyclization and Its Solutiona
a (a) nBuLi, toluene, -78 °C. (b) ClTi(OiPr)3, -78 °C to 0 °C. (c) 25c,
-78 °C. (d) Hydrazine, -78 °C to rt. (e) EtOH, ∆. (f) 5% HCl in EtOH,
rt.
sometimes tedious removal of the titanium hydrolysates by
(NH4)2CO3 solution as is necessary with the original procedure.
Despite these obvious advantages sometimes care must be
taken not to isolate intermediates such as 54 resulting from
incomplete hydrazinolysis. In these cases, simply refluxing the
semihydrazide in EtOH transformed it to the desired product.
All 2-azabicyclo[3.3.0]octanes depicted in Scheme 12 are
produced as enantiomerically pure diastereomers as judged from
the 1H NMR of the crude reaction mixtures. As with the already
described pyrrolidines, the reaction proceeded in a completely
reagent-controlled manner.
a (a) nBuLi, toluene, -78 °C. (b) ClTi(OiPr)3, -78 °C to 0 °C. (c)
Aldehyde (see Table 2), -78 °C. (d) Piperidine, -78 °C to rt, (NH4)2CO3.
(e) Hydrazine, -78 °C to rt. (f) EtOH, ∆. (g) 5% HCl in EtOH, rt. (h)
Boc2O, NaHCO3, dioxane, rt.
the γ-hydroxyalkylation products was initiated by hydrazine as
the deprotecting agent.
The absolute configurations of the newly created stereogenic
centers are a sole consequence of the absolute configuration at
sulfur. Again any stereochemical preference of the chiral
aldehyde employed is overcompensated by the high level of
asymmetric induction exerted by the auxiliary. The absolute
configuration at the induced stereogenic centers remains constant
for a given absolute configuration at sulfur irrespective of the
aldehyde configuration (compare 50a/50b; 50d/50e; 50f/50g).
If, on the other hand, the absolute configuration at sulfur is
changed, the same is true for the induced stereogenic centers
again irrespective of the stereochemical bias exerted by the
aldehyde (compare 50d/50f; 50e/50g).
The hydrazine method was developed to circumvent some
minor workup problems encumbered with the original piperidine
induced deprotection-cyclization sequence using Fmoc-pro-
tected amino aldehydes. The latter protocol entails the generation
of considerable amounts of a piperidine-fulven adduct which
sometimes hampers the isolation of the product. With hydrazine
as a deblocking agent the byproduct is phthalhydrazide 55 which
is almost insoluble in the reaction medium and therefore can
be removed easily by a simple filtration (Scheme 13). Further-
more 55 is able to complex the oxo-titanium compounds
generated with the employed EtOH/H2O/N2H4 mixture. This
greatly facilitates further the workup procedure and avoids the
Furthermore, the dominating inductive power of the centro-
chiral sulfur atom compared to the corresponding influence of
9
J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006 4029

A R T I C L E S
Reggelin et al.
Table 3. Diastereoselectivity of the Reaction Remains Constant
the stereogenic valine derived C-atom in the sulfonimidoyl
moiety is demonstrated by the comparison of 50b with 50c and
50h with 50i.
over a Wide Temperature Rangea
In 50b both stereogenic centers in the auxiliary are S-
configured (lk; SS, SC), whereas in 50c the configuration at
carbon is inverted (ul; SS,RC). From the identical configuration
induced in both compounds the overriding influence of the sulfur
configuration is obvious. The same conclusion can be drawn
from the product pair 50h and 50i. Switching from SS,RC (S2a)
to RS,RC (S3a) entails an inversion of the configuration at the
newly created stereogenic centers demonstrating again the major
influence of the sulfur configuration. It should be mentioned
here, that this stereochemical dominance of the sulfur is
especially pronounced in the five-membered ring 2-alkenyl
sulfoximines. With the six-membered ring sulfoximines the lk-
configuration in the auxiliary (SS,SC or RS,RC) was found to be
necessary to achieve complete stereocontrol23 in reactions with
R-oxy-aldehydes. Due to the fact that this preferable “intra-
molecular matched” situation (accompanied by a mutual en-
hancement of the induction by both stereogenic units) can be
easily established by proper choice of the valine isomer, this
does not entail any restriction in the configurational scope of
the reaction.
a (a) nBuLi, toluene, T, t1. (b) ClTi(OiPr)3, T to 0 °C, 1 h 0 °C. (c)
Aldehyde, T, t2. (d) Piperidine, T to rt.
Scheme 14. Synthesis of 2-Azabicyclo[4.3.0]nonanes 49g and
49ha
Finally the synthesis of three phenyl substituted derivatives
(50h, i, k) in an isomerically pure state, employing the
notoriously configurationally unstable phenylglycinal 25b,
demonstrates the low basicity of the allylic titanium intermediate
preventing the aldehyde from racemization.
The configurations of the products shown in Scheme 12 have
been confirmed by X-ray structural analysis of 50m, 50e and
desulfurized 50i, NBoc-50c, NBoc-50f, and NBoc-50d (Table
4). It is worth mentioning that compounds 50 can be synthesized
on a rather large scale. For example, 13.54 g of 50e have been
obtained in 78% yield starting from 16a without isolation of
intermediates via the hydrazine route.
a (a) nBuLi, toluene, -78 °C. (b) ClTi(OiPr)3, -78 °C to 0 °C. (c1)
25g, -78 °C. (c2) 25h, -78 °C. (d) (NH4)2CO3. (e) Piperidine, THF, 0 °C
to rt. (f) Hydrazine, -78 °C to rt.
In the course of our efforts to develop a variant of the reaction
working on MeOPEG-5000 as a soluble polymeric support,38
we encountered some solubility problems related to the low
reaction temperatures (-78 °C) seemingly necessary for the
deprotonation and γ-hydroxyalkylation steps. For this reason
we studied the temperature dependence of the resulting azacycles
isomeric purity using 50b and 50k as examples (Table 3).
With the â-amino aldehyde 25g (Scheme 6) and sulfoximine
18a we obtained the 2-azabicyclo[4.3.0]nonane 49g in 67%
overall yield using the Fmoc route (Scheme 14).
This time the intermediate vinyl sulfoximine 56 was isolated
and cyclized in a separate step to make sure that the γ-hy-
droxyalkylation of 18a with â-amino aldehyde 25g proceeded
with complete stereocontrol, which was found to be true. To
our delight also the final cyclization was completely stereose-
lective yielding 49g as a single stereoisomer. With the SS,SC-
configured cyclopentenyl methyl sulfoximine 16a and the Phth-
protected aldehyde 25h the diastereomer 49h with identical
configuration at the induced stereogenic centers was obtained
with 70% overall yield without isolation of 57.
To our delight no erosion of isomeric purity occurred up to
0 °C reaction temperature. Furthermore, it was possible to reduce
the reaction time, and we observed an increase in reaction yield
especially for 50b (entries 5 and 6). This turned out to be a
general observation as will be demonstrated later (Scheme 16).
2-Azabicyclo[x.y.0]nonanes and 2-Azabicyclo[4.4.0]-
decane
Encouraged by these results we studied the corresponding
behavior of the ring enlarged sulfoximine 17a in its reaction
with both R-amino- and â-amino aldehydes (Scheme 15).
The resulting 2-azabicycles 51a and 52h were obtained
isomerically pure (without isolation of intermediates) in good
and very good overall yields, respectively. The latter was
prepared on a larger scale (10 g) via the hydrazine route, and
its configuration was confirmed by X-ray crystallographic
analysis.
To broaden the scope of the reaction and to study possible
effects on the stereochemical outcome we modified both the
aldehyde and the 2-alkenyl sulfoximine. To gain access to
bicyclic compounds of type 49, 51, and 52 (Scheme 11) we
employed combinations of â-amino aldehydes and cyclohexenyl
sulfoximines in addition to the already described R-amino
aldehydes and cyclopentenyl sulfoximines.
Type-101, -110, and -111 Azapolycycles. According to
(38) Reggelin, M.; Slavik, S., unpublished results.
Scheme 3, the combination of a cyclic sulfoximine with
9
4030 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006

Potential Scaffolds for Peptide Mimetics
A R T I C L E S
Scheme 15. Synthesis of the 2-Azabicyclo[3.4.0]nonane 51a and
2-Azabicyclo[4.4.0]decane 52ha
Scheme 18. Benzoannelated Piperidine 62 as Example for a
Type-110 Heterocyclea
a Key: (a) nBuLi, toluene, -78 °C. (b) ClTi(OiPr)3, -78 °C to 0 °C.
(c) 25i, -78 °C. (d) Piperidine, (NH4)2CO3.
a Key: (a) nBuLi, toluene, -78 °C. (b) ClTi(OiPr)3, -78 °C to 0 °C.
(c1) 25a, -78 °C. (c2) 25h, -78 °C. (d1) Piperidine, -78 °C to rt. (d2)
Hydrazine, -78 °C to rt.
Scheme 19. Retrosynthetic Scheme to the Quadricyclic
Compounds 63 and 64
Scheme 16 . Synthesis of Two Type-101 Azapolycyclesa
a Key: (a) nBuLi, toluene, -78 °C. (b) ClTi(OiPr)3, -78 °C to 0 °C.
(c1) 25e, 0 °C. (c2) 25e, -78 °C. (d) Piperidine, (NH4)2CO3.
Scheme 17
To complete the set of azapolycycles defined in Scheme 3
we explored the possibility to prepare the quadricyclic com-
pounds 63 and 64 from sulfoximine 17a (Scheme 19) and the
bicyclic aldehydes 25k-n, whose preparation has already been
described (Scheme 6).
Due to the fact that none of the aldehydes 25k-n is Fmoc-
or Phth-protected and due to the peculiarities often encountered
with an indole nitrogen, a new cyclization protocol had to be
developed. Therefore we decided to prepare the target structures
63 and 64 via their corresponding vinyl sulfoximine intermedi-
ates 65-67, which in turn can be used as starting materials for
the new cyclization procedure (Scheme 20).
Being aware of the already mentioned low basicity of the
titanated sulfoximines20 (Scheme 17) we expected the reaction
to work even with the unprotected aldehydes 25m and 25k.
Indeed both reacted smoothly to the corresponding vinyl
sulfoximines 65 and 66, respectively, with complete stereo-
control.
heterocyclic, homocyclic, or heterobicyclic amino aldehydes
should lead to tri- or even tetracyclic systems.
To evaluate the feasibility of such processes we first tried to
react the cyclopentenyl sulfoximines 16a and 18a (differing only
at the C-configuration in the side chain of the auxiliary) with
Fmoc-protected (S)-prolinal 25e (Scheme 16).
Applying the standard protocol (aldehyde addition at -78
°C) to the reaction between sulfoximine 18a and aldehyde 25e,
the tricyclic compound 58 was formed isomerically pure in 43%
yield. In accordance with our results in the type-100 series (see
Table 3) raising the temperature in the allyl transfer step to 0
°C does not diminish the stereoselectivity as can be deduced
from the experiment yielding tricycle 59. Moreover these
reaction conditions entail an increase in yield from 35% to 58%.
Moreover, the former compound cyclized readily in refluxing
ethanol to yield the quadricycle 64 in 52% yield again as a single
isomer. Unfortunately, but not unexpectedly, the indole deriva-
tive 66 failed to cyclize under these conditions. Any attempt to
force the system to cyclize by NH-deprotonation failed due to
retroaddition. To avoid this unproductive reaction, the hydroxy
function in 66 was silylated, and indeed, this time the base
induced cyclization worked and, after solvolysis, delivered the
desired product 63a in 71% overall yield as a single isomer.
Alternatively we succeeded in synthesizing 63c via the Boc-
protected vinyl sulfoximine 67 which cyclized under the
influence of TBAF as the Boc-deprotection agent.40 Interest-
ingly, this cyclization is accompanied by a Boc-group migration
from nitrogen to oxygen, delivering the carbonate 68 as an
intermediate, which can be isolated in 67% yield. Solvolysis
with K2CO3 containing methanol delivered 63c in quantitative
yield. The configuration of 66 and 63a was confirmed by X-ray
structural analysis.
In an attempt to synthesize the euglobals G1 and G2
(terrestrial natural products isolated from Eucalyptus grandis)39
we tried to cyclize vinyl sulfoximines 60 to the advanced
precursor 61 (Scheme 17).20 It is worth mentioning here that a
sterically hindered tetrasubstituted aromatic aldehyde reacted
with the titanated sulfoximine to furnish 60b in a respectable
62% yield. This is even more surprising if one takes into account
that protection of the phenolic hydroxyl was not necessary due
to the low basicity of the titanium reagent.
Unfortunately, even after considerable experimentation we
were not able to obtain 61 in more than trace amounts. With
this disappointing result in mind, we were curious to see whether
it should be possible to accomplish cyclization if we replace
the phenolic oxygen by nitrogen. In the event we found that
the standard Fmoc-based one-pot protocol delivered the
benzoannelated model-system 62 in 35% yield as a pure
diastereomer (Scheme 18).
(39) Takasaki, M.; Konoshima, T.; Kozuka, M.; Haruna, M.; Ito, K.; Shingu,
(40) Routier, S.; Sauge, L.; Ayerbe, N.; Coudert, G.; Merour, J. Y. Tetrahedron
Lett. 2002, 43, 589.
T. Chem. Pharm. Bull. 1994, 42, 2591.
9
J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006 4031

A R T I C L E S
Reggelin et al.
Scheme 20. Two Examples of Type-111 Azapolycyclesa
Scheme 22. LN-Mediated Desulfurizations of
2-Azabicyclo[3.3.0]octanes
Scheme 23. Successful and Unsuccessful Desulfurization with
Raney-Nickel
a Key: (a) nBuLi, toluene, -78 °C. (b) ClTi(OiPr)3, -78 °C to 0 °C.
(c1) 25k, 0 °C. (c2) 25m, 0 °C. (c3) 25l, -78 °C. (d1) Phthalhydrazide,
EtOH, H2O. (d2) (NH4)2CO3. (e) Me2N-SiMe3, ∆, 2 h. (f) tBuOH, tBuOK,
∆, 2 h. (g) K2CO3, MeOH. (h) TBAF, THF, rt.
Scheme 21. Desulfurization of 69 with Raney-Nickel and Lithium
Naphthalenide (LN)
and Boc-72f were produced with 54% and 70% yield, respec-
tively. The former was accompanied by 31% of the ring-opened
homoallylic alcohol 74 reflecting the residual nucleofugacity
of the amine.
The reduced auxiliary 73 was isolated in about 80% yield
with complete retention of the sulfur configuration as was shown
by comparison with an independently prepared sample. Al-
though this result confirmed our initial assumptions concerning
the diminished nucleofugacity of the amide, the reaction was
not really satisfying. This is all the more true because the
unsaturated amine 74, contrasting the behavior of the related
alcohol 71 (Scheme 21), cannot be converted to the target
structure 72c by acid treatment, as expected.
Despite the fact that Raney-nickel (RN) proved to be a rather
unreliable desulfurization agent with the sulfonimidoyl-
substituted pyrrolidines,21 we tried this reagent with the bi- and
tricyclic sulfoximines Boc-50c (derived from 50c) and 75
(derived from 59), respectively (Scheme 23).
Whereas it was possible to desulfurize Boc-50c yielding 68%
Boc-72c, which was characterized by X-ray structural analysis,
the reaction failed to work with 75. Again we encountered
problems with the reliability and the constitutional scope of the
RN-mediated desulfurization, and therefore we decided to
explore the possibility of applying the SmI2 reduction we
developed for the sulfonimidoyl-substituted pyrrolidines some
years ago (Scheme 24).21
Removal of the Auxiliary. Since sulfoximines can be taken
as aza-analogues of sulfones, it is obvious that desulfurization
methods applicable for the latter should also work for the former.
For that reason we started our work on the removal of the
sulfonimidoyl moiety using reagents known to effect the
desulfurization of sulfones.
In the course of our work on the synthesis of 2-oxabicyclo-
[3.3.0]octanes20 we found Raney-nickel to be the reagent of
choice for the desulfurization of these compounds (Scheme 21).
With 69 as a model compound Raney-nickel delivered the
desulfurized 2-oxabicycle 70 with nearly quantitative yield. On
the other hand treatment of 69 with lithium naphthalenide (LN)
produced the sulfur-free, albeit ring-opened homoallylic alcohol
71 in 83% yield. This latter compound can be converted
stereoselectively to the target bicycle 70 in a quantitative
manner, simply by acid treatment.
From these experiments we learned that every desulfurization
method accompanied by an increase in electron density at the
carbon adjacent to the sulfur atom may initiate a â-elimination
destroying the bicyclic framework.
On the other hand for the azacyclic compounds under
consideration here, the LN method may be feasible due to the
reduced nucleofugacity of the nitrogen. Therefore we tried to
desulfurize sulfoximines 50c and 50f with LN (Scheme 22).
Indeed, the derived angular methyl-substituted derivatives 72c
For the SmI2 method to work, it is necessary to have the
N-atom Boc-protected and the side chain of the auxiliary
deprotected, as already discussed.21 Therefore, the substrates
9
4032 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006

Potential Scaffolds for Peptide Mimetics
A R T I C L E S
Chart 4. NK1 Receptor Antagonists Containing the Common
Scheme 24
Structural Subunit 82
a
Table 4. Desulfurization of 2-Azabicyclo[3.3.0]octanes with SmI2
Quite a number of compounds have been synthesized
incorporating this or similar structural motifs.5,42,43 Many of them
are based on either the quinuclidine nucleus (e.g., CP-96,345)5,42
or the 1-azabicyclo[3.3.0]-scaffold (e.g., 83) or are pyrrolidine
(84, n ) 0) or piperidine (84, n ) 1) derivatives.43,44 From these
studies it became obvious that the biological activity of these
compounds not only was dependent on the nature of the scaffold
but also was sensitive to the substitution pattern and the
configuration of the carbon atoms displaying the pharmacophoric
groups. For these reasons and the fact that the compounds
depicted in Chart 4 are all within the constitutional and
configurational scope of the method described above, we
decided to study its applicability within the context of NK1
antagonism. Toward this end we synthesized the substituted
pyrrolidine 85 and the bicyclic derivative 86 (Scheme 25).
The absolute configuration of 85 was confirmed by X-ray
structural analysis of the corresponding p-toluenesulfonic acid
salt which in turn secures the structural assignment of 40g too.
Although both compounds do not reach the high level of activity
of CP-96,345, the pyrrolidine 85 comes quite close. Taking this
result as an encouraging starting point, we are currently
exploring changes in ring sizes, substitution pattern, and
configuration of the side chains presented, thereby exploiting
the potential of the synthetic approach presented in this article
to superimpose structural diversity from the side chains and the
scaffold simultaneously.
a Compounds in shaded rows have been characterized by X-ray structural
analysis. The .cif files can be found in the Supporting Information (#.cif).
b (c in dichloromethane).
49g and 52h were converted to the N-protected and desilylated
derivatives 77 and 79, respectively. To our delight both
compounds were desulfurized with SmI2 in THF with excellent
yields delivering the azabicycles 78 and 80 as single isomers.
Moreover, using this reaction we were able to desulfurize all
2-azabicyclo[3.3.0]octanes (Scheme 12) after the above-
mentioned functional group interconversions (Table 3).
The yields given in Table 4 are not optimized and can be
regarded as the lower limit. This is especially true for entries
4-6 which can be traced back to our lack of experience with
the preparation of the SmI2 reagent at the time the experiments
were made. Nevertheless it should be noted that we were not
able to desulfurize the tricyclic compounds 58 and 59 including
several derivatives thereof. The reasons for these failures are
not clear so far.
Conclusions
Starting from the cyclic sulfonimidates 1a and 1b and their
enantiomers (Scheme 2, 1a and ent-1a are commercially
available; Aldrich # 54099-4 and 54412-4 respectively) the
open chain and cyclic 2-alkenyl sulfoximines 16 to 21 can be
(42) Lowe, J. A., III; Drodzda, S. E.; Snider, R. M.; Longo, K. P.; Zorn, S. H.;
Morrone, J.; Jackson, E. R.; McLean, S.; Bryce, D. K.; Bordner, J.;
Nagahisa, A.; Kanai, Y.; Suga, O.; Tsuchiya, M. J. Med. Chem. 1992, 35,
2591.
NK1 Receptor Antagonists. In the context of research
directed toward the synthesis of nonpeptidic neurokinin subtype
I (NK1) receptor antagonists, the 2-phenyl-3-heterobenzyl and
the 2-phenylalkyl-3-heterobenzyl ethylamine subunits contained
in 82 were identified as important structural motifs (Chart 4).41
(43) Appell, K. C.; Fragale, B. J.; Loscig, J.; Singh, S.; Tomczuk, B. E. Am.
Soc. Pharmacol. Experim. Therapeut. 1992, 772.
(44) Harrison, T.; Williams, B. J.; Swain, C. J.; Ball, R. G. Bioorg. Med. Chem.
Lett. 1994, 4, 2545.
(45) Craig, D. A. Trends Pharmacol. Sci. 1993, 14, 89.
(46) We thank Dr. R. Bru¨ckner (Solvay Pharmaceuticals, Hannover, Germany)
for providing us with the biological data for 85 and 86.
(41) Desai, M. C.; Lefkowitz, S. L.; Thaideio, P. F.; Longo, K. P.; Snider, R.
M. J. Med. Chem. 1992, 35, 4911.
9
J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006 4033

A R T I C L E S
Reggelin et al.
Scheme 25. Synthesis of NK1 Receptor Antagonistsa
bicyclic) eight different types of ring systems (type-000 to -111,
Scheme 3) can be prepared as single isomers.
According to our goal, the development of a highly flexible
entry to a broad range of highly substituted azacyclic compounds
as potential peptide mimetics, we synthesized the pyrrolidine
derivative 85 and the 2-azabicycle 86 (Scheme 25). Both
compounds showed high biological activity in NK1-functional
assays.
The sulfonimidoyl moiety can be removed using Raney-
nickel, lithium naphthalenide, or SmI2. The latter reagent turned
out to be the reagent of choice in most cases, although the
auxiliary is destroyed during the reduction.
Our current work is directed toward a functionalization of
ring 1 (Scheme 3) and toward a functionalizing desulfurization
combining the desulfurization step with the introduction of
functional groups at the carbon atom adjacent to sulfur.
Finally, we explore the possibility of “translating” structural
data from NMR studies on bioactive peptides to non-peptides
of general structure 14 or desulfurized derivatives thereof.
Acknowledgment. We thank the Fonds der Chemischen
Industrie (Kekule´ Fellowship for P.B.) and Solvay Pharmaceu-
ticals (Hannover, Germany) for generous financial support. We
also thank Ms. S. Foro (TU Darmstadt) and Dr. J. Bats
(University of Frankfurt/Main) for their careful work on the
X-ray structures and Prof. H. Fuess (Department of Material
Science, TU-Darmstadt) for measuring time on his diffracto-
meter.
a In vitro binding affinity of 85 and 86 for the NK1 receptor as measured
by the reduction of substance P induced relaxation in precontracted aortic
rings isolated from guinea pigs.45,46 Compare with CP-96,345 under the
same conditions: IC50 ) 10 nM, pA2 ) 8.5.
prepared in high yields and on a multigram scale (e.g., 16a with
78% overall yield starting from 1a, Chart 3). The organotitanium
compounds derived from these allylic sulfoximines can be
hydroxyalkylated with almost complete stereocontrol delivering
vinyl sulfoximines of type 13 (Scheme 3) using R- or â-amino
aldehydes as electrophiles. These can be isolated, but in most
cases it was found to be preferable to cyclize them immediately
from the crude reaction mixture yielding highly substituted aza-
(poly)cyclic compounds of general formula 14.
Supporting Information Available: Experimental procedures
and characterization of the products (PDF). Crystallographic
information files (CIF) for compounds 39d, 5-epi-39a, 5-epi-
40a, 40d, 2-epi-40g, 43, 50e, 50m, 52h, 63a, 66, Boc-72c-f,
Boc-72i, and 85. This material is available free of charge via
the Internet at http://pubs.acs.org.
With two types of sulfoximines (acyclic/cyclic) and four types
of aldehydes (open chain, homocyclic, heterocyclic, hetero-
JA057012A
9
4034 J. AM. CHEM. SOC. VOL. 128, NO. 12, 2006