Carbocyclic Amino Ketones by Bredt's Rule-Arrested Kulinkovich-de Meijere Reaction

Article abstract of DOI:10.1002/anie.201509983

The Ti^II-mediated formation of cyclopropylamines from alkenes and amides, the Kulinkovich-de Meijere reaction, involves two carbon-carbon bond-forming steps. Strategic use of a tricyclic intermediate can arrest the process if the second step requires formation of a bridgehead double bond. Use of this Bredt's rule constraint results in the production of carbocyclic amino ketones, key alkaloid building blocks. Frustrated by Bredt's rule: An inability to consummate a Kulinkovich-de Meijere cyclopropanation that would involve a bridgehead double bond transforms the reaction into a transannular cyclization of an unsaturated lactam yielding an amino ketone product with potential for alkaloid synthesis.

Full text of DOI:10.1002/anie.201509983

Angewandte
Communications
Chemie
International Edition: DOI: 10.1002/anie.201509983
Synthetic Methods
German Edition:
DOI: 10.1002/ange.201509983
Carbocyclic Amino Ketones by Bredtꢀs Rule-Arrested
Kulinkovich–de Meijere Reaction
Paul B. Finn, Brenden P. Derstine, and Scott McN. Sieburth*
Abstract: The TiII-mediated formation of cyclopropylamines
from alkenes and amides, the Kulinkovich–de Meijere reaction,
involves two carbon–carbon bond-forming steps. Strategic use
of a tricyclic intermediate can arrest the process if the second
step requires formation of a bridgehead double bond. Use of
this Bredtꢀs rule constraint results in the production of
carbocyclic amino ketones, key alkaloid building blocks.
B
redtꢀs rule, the observation that double bonds cannot
include bridgehead atoms, has been a cornerstone of organic
chemistry since its enumeration in 1924.[1,2] This rule is
routinely used to illustrate the concepts of orbital overlap and
to explain reactivity in polycyclic molecules, but it has rarely
been employed to steer reaction pathways outside of elimi-
nation and enolization chemistry.[3]
Scheme 2. Interrupted Kulinkovich–de Meijere reactions. A) Cha’s
imide substrates.[8] B) This work. Reagents and conditions: a) ClTi(Oi-
Pr)3, c-C5H9MgBr, THF, rt, 44%; b) Ti(Oi-Pr)4, c-C5H9MgBr, THF, rt,
85%.
The extraordinary coupling of an alkene and an ester to
give cyclopropanols, mediated by low-valent titanium, has
been extensively studied since its discovery in 1989 by
Kulinkovich, and was extended to tertiary amides in 1996
by de Meijere to yield cyclopropylamines (Scheme 1).[4–6]
electron density stabilizes intermediate 5, averting cyclo-
propane formation.
An alternative and unexplored strategy involves use of
a substrate that would create a bridged intermediate incapa-
ble of titanium alkoxide departure. We have found that the
readily prepared 6-allyl-2-piperidinone 3 will undergo trans-
annular cyclization under standard Kulinkovich conditions to
yield tricyclic 6 (Scheme 2B). Advancement of intermediate 6
to give the normal cyclopropane product would require
formation of a bridgehead double bond to nitrogen, but such
a nitrogen-assisted elimination of the titanium alkoxide would
violate Bredtꢀs rule. Thus confined, intermediate 6 gave, on
workup, the cis-4-amino-2-methyl cycloheptanone 4 in good
yield and as the only product. No trace of a corresponding
cyclopropylamine was detected. The ability to transform
amides to amino ketones has clear potential in synthesis.[9] We
describe here our initial investigation of this new approach to
producing these important building blocks.
Following the demonstration that ene-amide 3 would
undergo the initial carbon–carbon bond-formation of the
Kulinkovich–de Meijere reaction, despite the need for trans-
annular attack, we investigated three additional examples in
which the alkene was attached at the other positions around
the ring (Scheme 3). Each of these reactions would form
a tricyclic intermediate incorporating a five-membered carbo-
cycle through coupling of the internal alkene carbon and the
amide carbonyl. For 5- and 4-substituted piperidinones 8 and
12, two products were formed, the now expected amino
ketones 10 and 14, as well as unexpected aminocyclopropa-
nols 11 and 15. With the 3-(3-butenyl) substitution in 16,
cyclopropylamine 18 was isolated in good yield.
Scheme 1. Kulinkovich (X=OEt) and Kulinkovich–de Meijere
(X=NMe2) reactions.[4]
These powerful transformations have been pivotal in synthe-
ses of natural products and pharmaceutical agents.[7] Two
sequential carbon–carbon bond forming steps occur in this
reaction, and interception of the intermediate can allow the
reaction to take alternative pathways, including use of the
remaining carbon–titanium bond. The Cha group demon-
strated that the Kulinkovich reaction path could be inter-
rupted if an imide substrate was employed, e.g., 1 (Sche-
me 2A).[8] In this example, following the initial imide
carbonyl addition, delocalization of the amide nitrogen
[*] P. B. Finn, B. P. Derstine, Prof. S. McN. Sieburth
Department of Chemistry, Temple University
1901 N. 13th St., Philadelphia, PA 19122 (USA)
E-mail: scott.sieburth@temple.edu
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201509983.
2536
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 2536 –2539

Angewandte
Communications
Chemie
bridge y has a length of zero and the iminium ion is therefore
easily formed and leads to product 18.
To test the limits of this approach to interception of the
Kulinkovich–de Meijere intermediate we have probed alter-
native S values in homologues of 3 and 12 (Table 1 and
Table 2). For the w-allyl lactams, varying the lactam ring size
from five to nine, 21a to 21e, changes the S value for the
intermediate 22 from five to nine (Table 1). With the
Table 1: Kulinkovich–de Meijere reaction with w-allyl lactams.
n
S[b]
Yield [%][c]
Ratio 23:24:25[d]
a
b
c
0
1
2
5
6
7
65
85
60
5:1:0
1:0:0
2:1:0
Scheme 3. Reaction of alkenes at other positions around the piperidi-
none. Reagents and conditions: a) Ti(Oi-Pr)4, c-C5H9MgBr, THF, rt,
(33% for 10, 40% for 14+15, 73% for 18).
d
e
3
4
8
9
73
34
0:0:1
0:0:1
These examples further illustrate the Bredtꢀs rule effect.
Substrate 16 leads to intermediate 17 that can readily
eliminate the titanium alkoxide and form an unstrained
iminium ion, leading to the normal Kulinkovich–de Meijere
product 18 in good yield and with a high degree of
stereocontrol.[10] In contrast, the bridged, polycyclic inter-
mediates 9 and 13 from 8 and 12, cannot eliminate the
alkoxide, yielding amino ketones 10 and 14 on workup. The
surprising formation of aminocyclopropanol products 11 and
15 are also consequences of Bredtꢀs rule. These are among the
first examples of aminocyclopropanols formed during a Kulin-
kovich–de Meijere reaction and the proportion of cyclopro-
panol products in Scheme 3 are some of the highest for the
substrates we have examined.[11] Departure of the anionic
amine group from 9 and 13 was unexpected but the
intermediate oxonium ions, leading to cyclopropanols 11
and 15, are relatively unstrained and therefore accessible. The
diminished yields for the reactions of 8 and 12 are likely
a consequence of the strain of the intermediates 9 and 13.
In his seminal review of Bredtꢀs rule, Fawcett defined the
term S as the sum of the bridging atoms of a bicyclic system,
x + y + z (Figure 1).[3] Using that analysis, he determined that
when the sum of the bridging atoms totalled less than nine,
providing that none of the bridges had a length of zero,
a bridgehead alkene would be too strained to form. For
bicyclic 19, the iminium ion potentially derived from inter-
mediate 13, the ring system is too small to accommodate the
double bond. In contrast, for iminium ion 20 derived from 17,
[a] Reagents and conditions: Ti(Oi-Pr)4, c-C5H9MgBr, THF, rt. [b] Total
number of bridging atoms in 22.[3] [c] Isolated yields. [d] 21b, 23b=3, 4.
exception of the nine-membered lactam substrate 21e, the
yields are uniformly good. For lactam ring sizes between five
and seven-membered, 21a–c with S = 5–7, only 23 and 24 are
produced, with the amino ketone 23 making up 66–100% of
the product mixture, without formation of cyclopropylamine
25. When the S value of 22 reaches 8 however, Bredtꢀs rule
can be violated and there is an abrupt changeover, with 21d
and 21e giving only aminocyclopropanes 25d and 25e as
products.
With an allyl group beta to the carbonyl, 26, cyclization
between the alkene and the amide carbonyl also forms a five-
membered ring (27, Table 2). This cyclization leads to a cyclo-
pentanone product 28 with a side chain carrying a secondary
amine. In the example of intermediate 27a with S = 6, Bredtꢀs
rule prevents formation of a cyclopropylamine; only amino
ketone 28a and cyclopropanol 29a are observed, with the
former comprising 50% of the mixture. With 26c and 26d,
S = 8 and 9 respectively, the S value is compatible with
a bridgehead double bond and iminium ion formation and
cyclopropylamines 30c and 30d were the only products
formed. For the intermediate case of 26b where S = 7, the
product mixture is difficult to purify and appears to be largely
cyclopropanol 29b. This may be a consequence of a strained
transition structure in these transannular reactions.
Two additional examples with S = 7 intermediates are
shown in Scheme 4. These Bredtꢀs rule-arrested transforma-
tions utilize the smaller lactams with longer alkenyl chains.
For piperidinone 31 carrying a 4-(1-butene) beta to the
carbonyl, cyclization was expected to lead to an azabicyclo-
[3.3.1]nonane framework. The result was formation of cyclo-
hexanone 33 and cyclopropanol 34, favouring the former.
Similarly, a pentenyl chain attached adjacent to the nitrogen
Figure 1. S=x+y+z.
Angew. Chem. Int. Ed. 2016, 55, 2536 –2539
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2537

Angewandte
Communications
Chemie
Table 2: Kulinkovich–de Meijere reaction with b-allyl lactams.
n
S[b]
Yield [%][c]
Ratio 28:29:30[d]
a
b
1
2
6
7
40[e]
–
1:1:0
See text
c
d
3
4
8
9
64
75
0:0:1
0:0:1
[a] Reagents and conditions: Ti(Oi-Pr)4, c-C5H9MgBr, THF, rt. [b] Total
number of bridging atoms in 27.[3] [c] Isolated yields. In some cases the
amines were derivatized before isolation, see Supporting Information for
details. [d] 26a, 28a, 29a=12, 14, 15. [e] Isolated yield of 28a.
Scheme 5. Polycyclic substrate. Reagents and conditions: a) Ti(Oi-Pr)4,
c-C5H9MgBr, THF, rt, 93%; b) 3,5-(O2N)2C6H3COCl, Et3N, 75%.
Scheme 6. Intramolecular Mannich reaction of 23a and 23b.
a bicyclic structure incapable of the requisite elimination. This
Bredtꢀs rule constraint allows the reaction to convert alkene-
substituted lactams into amino-substituted carbocyclic
ketones under mild conditions and in some cases with
a high degree of stereocontrol. The examples here have
even greater potential in the form of the second and unused
carbon–titanium bond.
Scheme 4. Changes in side chain length. Reagents and conditions:
a) Ti(Oi-Pr)4, c-C5H9MgBr, THF, rt, b) TsCl, pyridine (40% for 33: 27%
for 37 as a 3:2 mixture of diastereomers).
of a 2-pyrrolidinone was expected to form an azobicyclo-
[4.2.1]nonane intermediate. This transformation yields only
the cyclooctanone 37, albeit in modest yield.
Complex structures are also compatible with these
reaction conditions and may benefit from the additional
structural rigidity (Scheme 5). Tetracyclic substrate 38,[12]
under standard conditions, yields the pentacyclic product 39
in near quantitative yield. In this case the product is stable as
the aminal, a structure confirmed by X-ray crystallography.
Treatment of this aminal with 2,5-dinitrobenzoyl chloride
gave the corresponding ketoamide 40 in quantitative yield,
also confirmed by an X-ray structure analysis (see Supporting
Information).
Acknowledgements
We gratefully acknowledge support from the National
Science Foundation (CHE-1152159). We also thank Professor
Michael J. Zdilla, Clifton R. Hamilton and Charles M.
Stockdale for the crystal structures.
Keywords: amino ketones · Bredt’s rule · Kulinkovich reaction ·
Mannich reaction · titanium
These amino ketone products can be elaborated using the
Mannich reaction. Microwave heating of aminocyclohexa-
none 23a with paraformaldehyde resulted in a high yield of
Mannich products 41 and 42 (Scheme 6), despite the inher-
ently poor orbital overlap of a 5-endo-trig cyclization.[13,14]
Homologous ketone 23b under identical conditions yields
bicyclics 43 and 44, favouring quaternary 43.
The intermediate formed during the two-step Kulinko-
vich–de Meijere cyclopropanation can be prevented from
consummating the second step if the substrate leads to
How to cite: Angew. Chem. Int. Ed. 2016, 55, 2536–2539
Angew. Chem. 2016, 128, 2582–2585
[1] J. Bredt, Justus Liebigs Ann. Chem. 1924, 437, 1 – 13.
[2] E. H. Krenske, C. M. Williams, Angew. Chem. Int. Ed. 2015, 54,
10608 – 10612; Angew. Chem. 2015, 127, 10754 – 10758.
[3] Reviews: F. S. Fawcett, Chem. Rev. 1950, 47, 219 – 274; V. Prelog,
J. Chem. Soc. 1950, 420 – 428; G. Kçbrich, Angew. Chem. Int. Ed.
Engl. 1973, 12, 464 – 473; Angew. Chem. 1973, 85, 494 – 503; G. L.
Buchanan, Chem. Soc. Rev. 1974, 3, 41 – 63; R. Keese, Angew.
2538
www.angewandte.org
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2016, 55, 2536 –2539

Angewandte
Communications
Chemie
Chem. Int. Ed. Engl. 1975, 14, 528 – 538; Angew. Chem. 1975, 87,
Madelaine, O. Buriez, B. Crousse, I. Florent, P. Grellier, P.
568 – 578; K. J. Shea, Tetrahedron 1980, 36, 1683 – 1715; P. M.
Warner, Chem. Rev. 1989, 89, 1067 – 1093; B. R. Bear, S. M.
Sparks, K. J. Shea, Angew. Chem. Int. Ed. 2001, 40, 820 – 849;
Angew. Chem. 2001, 113, 864 – 894; Ref. [2].
Retailleau, Y. Six, Org. Biomol. Chem. 2010, 8, 5591 – 5601.
[8] J. Lee, J. D. Ha, J. K. Cha, J. Am. Chem. Soc. 1997, 119, 8127 –
8128; S.-H. Kim, Y. Park, H. Choo, J. K. Cha, Tetrahedron Lett.
2002, 43, 6657 – 6660; S. Santra, N. Masalov, O. L. Epstein, J. K.
Cha, Org. Lett. 2005, 7, 5901 – 5904.
[9] See: A. Gomtsyan, Org. Lett. 2000, 2, 11 – 13; S. Baktharaman,
R. Hili, A. K. Yudin, Aldrichimica Acta 2008, 41, 109 – 119, and
references therein.
[10] The cis fusion of 18 was determined by nOe studies.
[11] Reactions of N-acylated Oppolzerꢀs sultam have been found to
yield cyclopropanols. R. Mizojiri, H. Urabe, F. Sato, Angew.
Chem. Int. Ed. 1998, 37, 2666 – 2668; Angew. Chem. 1998, 110,
2811 – 2814.
[4] J. K. Cha, O. G. Kulinkovich, Org. React. 2012, 77, 1 – 159; J. C.
Lee, M. J. Sung, J. K. Cha, Tetrahedron Lett. 2001, 42, 2059 –
2061; J. Lee, J. K. Cha, J. Org. Chem. 1997, 62, 1584 – 1585.
[5] O. G. Kulinkovich, S. V. Sviridov, D. A. Vasilevskii, T. S. Prityt-
skaya, Zh. Org. Khim. 1989, 25, 2244 – 2245; O. G. Kulinkovich,
S. V. Sviridov, D. A. Vasilevskii, A. I. Savchenko, T. S. Prityt-
skaya, Zh. Org. Khim. 1991, 27, 294 – 298.
[6] V. Chaplinski, A. de Meijere, Angew. Chem. Int. Ed. Engl. 1996,
35, 413 – 414; Angew. Chem. 1996, 108, 491 – 492; A. De Meijere,
V. Chaplinski, A. Kourdioukov (Bayer AG), DE-19647615,
1998.
[12] P. B. Finn, S. Kulyk, S. McN. Sieburth, Tetrahedron Lett. 2015, 56,
3567 – 3570.
[13] J. E. Baldwin, J. Cutting, W. Dupont, L. Kruse, L. Silberman,
R. C. Thomas, J. Chem. Soc. Chem. Commun. 1976, 736 – 738.
[14] C. P. Johnston, A. Kothari, T. Sergeieva, S. I. Okovytyy, K. E.
Jackson, R. S. Paton, M. D. Smith, Nat. Chem. 2015, 7, 171 – 177.
[7] Review: I. Haym, M. A. Brimble, Org. Biomol. Chem. 2012, 10,
7649 – 7665; Selected examples: J. S. Kingsbury, E. J. Corey, J.
Am. Chem. Soc. 2005, 127, 13813 – 13815; F. LecornuØ, R.
Paugam, J. Ollivier, Eur. J. Org. Chem. 2005, 2589 – 2598; O.
Lack, R. E. Martin, Tetrahedron Lett. 2005, 46, 8207 – 8211; C. A.
Faler, M. M. JoulliØ, Org. Lett. 2007, 9, 1987 – 1990; K. A.
Keaton, A. J. Phillips, Org. Lett. 2007, 9, 2717 – 2719; C.
Received: October 26, 2015
Published online: January 14, 2016
Angew. Chem. Int. Ed. 2016, 55, 2536 –2539
ꢀ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
2539