Directing group enhanced carbonylative ring expansions of amino-substituted cyclopropanes: Rhodium-catalyzed multicomponent synthesis of N-heterobicyclic enones

Article abstract of DOI:10.1021/ja401936c

Aminocyclopropanes equipped with suitable N-directing groups undergo efficient and regioselective Rh-catalyzed carbonylative C-C bond activation. Trapping of the resultant metallacycles with tethered alkynes provides an atom-economic entry to diverse N-heterobicyclic enones. These studies provide a blueprint for myriad N-heterocyclic methodologies.

Full text of DOI:10.1021/ja401936c

Communication
pubs.acs.org/JACS
Directing Group Enhanced Carbonylative Ring Expansions of Amino-
Substituted Cyclopropanes: Rhodium-Catalyzed Multicomponent
Synthesis of N‑Heterobicyclic Enones
Megan H. Shaw,† Ekaterina Y. Melikhova,† Daniel P. Kloer,‡ William G. Whittingham,‡
and John F. Bower*,†
†School of Chemistry, University of Bristol, Bristol BS8 1TS, United Kingdom
‡Syngenta, Jealott’s Hill International Research Centre, Bracknell, Berkshire RG42 6EY, United Kingdom
S
* Supporting Information
Scheme 1. Aminometallacyclopentanones as a Catalysis
Platform for N-Heterocycle Synthesis
ABSTRACT: Aminocyclopropanes equipped with suit-
able N-directing groups undergo efficient and regioselec-
tive Rh-catalyzed carbonylative C−C bond activation.
Trapping of the resultant metallacycles with tethered
alkynes provides an atom-economic entry to diverse N-
heterobicyclic enones. These studies provide a blueprint
for myriad N-heterocyclic methodologies.
odern medicinal chemistry requires more efficient and
1
Mdiverse methods for the synthesis of chiral scaffolds.
Cognizant of both this goal and the ideals of green chemistry,2 we
recently initiated a program aimed at the generation and trapping
of amino-substituted metallacyclopentanones. Here, we envis-
aged that exposure of CO and aminocyclopropanes 1 to an
appropriate catalyst would afford metallacyclic intermediates 2
(Scheme 1A). Intra- or intermolecular trapping of these species
with various π-unsaturates then potentially delivers heterocyclic
products 3. In this scenario, diverse and readily available building
blocks are converted to medicinally valuable chiral heterocycles
with complete atom economy. Moreover, modification of the
catalyst system with appropriate chiral ligands potentially
provides asymmetric processes. In a broader context, this
approach requires the development of a new family of
(3+2+1) cycloadditions involving aminocyclopropanes. Cyclo-
propane-based carbonylative (3+2+1) cycloadditions have been
developed previously under Rh-catalyzed conditions by
Narasaka3 and Yu,4 but examples involving amino-substituted
variants have not been reported.
formation are not applicable to aminocyclopropanes, especially
within the wider context of the general methodology outlined in
Scheme 1A.12 Of particular relevance to this study is the work of
Chirik, who demonstrated phosphinite-directed Rh insertion
into the more hindered C−C bond of alkyl-substituted
cyclopropanes in the absence of CO.13 Based upon this, we
envisaged selective metallacycle formation by employing N-
protecting groups that can serve a dual role of pre-coordinating
the metal catalyst (Scheme 1D).14 Herein, we report the
successful realization of this strategy in the context of processes
involving tethered alkynes. This provides a direct entry to
complex N-heterobicyclic enones and also validates a unique
Key to achieving the goal outlined in Scheme 1A is the
realization of strategies that facilitate selective generation of the
requisite aminometallacyclopentanone 2. In principle, this is
achievable by either (a) the oxidative addition of a metal catalyst
into a cyclobutanone acyl−carbon bond5,6 or (b) the carbon-
ylative insertion of a metal into a cyclopropane C−C bond
(Scheme 1B).3,7−10 This latter approach is more attractive
because parent aminocyclopropane is commercially available and
substituted variants are readily accessed using established
asymmetric technologies (Scheme 1C).11 However, the key
issue of regioselective metallacycle formation remains, and a
strategy to promote selective insertion of the metal and CO into
the more hindered aminocyclopropane C−C bond is required.
Existing strategies for regioselective metallacyclopentanone
Received: February 22, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja401936c | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
approach to the carbonylative C−C bond activation of
aminocyclopropanes.
terized by X-ray diffraction.17 In support of our catalysis design,
the structure of complex 10 reveals axial coordination of the
carbamate directing group and the desired regioselectivity for
both the oxidative addition and CO insertion steps.18 This
establishes that neutral carbonyl-ligated Rh(I) catalysts are
effective for directed rhodacyclopentanone formation.
Our initial studies established that directing group controlled
oxidative addition is feasible under the conditions of Rh catalysis.
Accordingly, carbamate-protected aminocyclopropane 4 was
exposed to various Rh(I) catalysts in the absence of CO (Scheme
2A). Here, we found that neutral Rh(I) salts (e.g., [Rh(cod)Cl]2/
PPh3) slowly delivered branched enamine 5, thereby indicating
(slow) insertion into the less hindered cyclopropane C−C bond.
On the other hand, cationic Rh(I) systems (e.g., [Rh(cod)2]BF4/
PPh3), which are more Lewis acidic, efficiently delivered linear
product 6.15 To confirm the importance of the carbamate moiety,
we performed analogous experiments on alkyl-substituted
cyclopropane 7. Here, both neutral and cationic Rh(I) systems
delivered branched products that result from exclusive insertion
into the less hindered cyclopropane C−C bond. In the case of
[Rh(cod)Cl]2, only the direct product of β-hydride elimination
and reductive elimination, 8a, was observed. [Rh(cod)2]BF4
afforded a mixture of alkene products 8a−c (1:7:12); 8b,c likely
arise by Rh-catalyzed isomerization of the initial adduct 8a.16
Taken together, these results confirm that carbamate-directed
oxidative addition is possible if the Rh catalyst is sufficiently
Lewis acidic. Under carbonylative conditions, we anticipated that
neutral systems may also be effective because CO is a strong π-
acceptor ligand; consequently, carbonyl-ligated neutral Rh(I)
catalysts may possess sufficient Lewis acidity to promote directed
oxidative addition. Probing this possibility was challenging
because treatment of carbamate 4 with neutral Rh(I) systems
under a CO atmosphere did not lead to appreciable quantities of
enamines 5 or 6, presumably because rhodacyclopentanones
formed instead. We therefore undertook studies to isolate and
characterize a model metallacyclic intermediate (Scheme 2B).
Heating carbamate 9 with stoichiometric [Rh(CO)2Cl]2 at 60 °C
generated dimeric rhodacyclopentanone 10, which was charac-
With these preliminary studies in hand, we sought to
incorporate our directing-group-based strategy into a proto-
typical multicomponent carbonylative coupling process involv-
ing a tethered alkyne (Table 1). Seminal work by Narasaka has
demonstrated the feasibility of a related process to provide
carbocyclic systems; in these cases, the alkyne moiety was
proposed to direct Rh insertion.3a Relatively high temperatures
and catalyst loadings and elevated CO pressures, in conjunction
with a favorable Thorpe−Ingold effect, were required to provide
acceptable yields of the targets. In our case, judicious choice of N-
directing group is critical as it must (a) compete for the catalyst
with the alkyne moiety of the starting material but (b) also be
labile enough to then allow alkyne coordination at the stage of
the metallacyclopentanone (see Scheme 1A). Accordingly, we
evaluated the conversion of a range of differentially protected
aminocyclopropanes 11a−d to the corresponding enones 12a−
d.
Table 1. Development of a Prototypical Process
Scheme 2. Studies on Controlling the Regioselectivity of
Metallacyclopentanone Formation
a
7.5 mol% for bidentate ligands and 15 mol% for monodentate
b
1
ligands. Yields were determined by H NMR using 1,4-dinitroben-
zene as a standard. Isolated yields are given in parentheses. Na2SO4
(20 mol%) was employed as desiccant. [Rh(cod)Cl]2 (3.75 mol%)
c
d
and P(3,5-(CF3)2C6H3)3 (15 mol%) were used. [1,2-DCB = 1,2-
dichlorobenzene]
Exposure of amide 11a to a BINAP-ligated neutral Rh(I)-
catalyst system at 160 °C under an atmospheric pressure of CO
generated 12a, but in only 26% yield (Table 1, entry 1).
Anticipating that the amide directing group was not sufficiently
Lewis basic, we evaluated carbamate 11b and observed the more
efficient formation of 12b in 33% yield (Table 1, entry 2); the
structure of 12b was confirmed by single-crystal X-ray diffraction.
Increasing the directing group Lewis basicity further was
beneficial, and urea 11c afforded 12c in 42% yield (Table 1,
entry 3). Importantly, this directing group also conferred a
significant rate enhancement (24 vs 48 h for 12a,b) and,
ultimately, this allowed lower reaction temperatures (see
below).19 However, the more Lewis basic methoxy-urea 11d
was inefficient, and only traces of 12d were formed (Table 1,
entry 4). Throughout these initial studies, we found that cationic
Rh(I) sources (e.g., [Rh(cod)2]BF4) were less effective due to
competitive formation of enamine byproducts (cf. Scheme
B
dx.doi.org/10.1021/ja401936c | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
2A).15 Further studies involving [Rh(cod)Cl]2 and 11c identified
P(3,5-(CF3)2C6H3)3 as an effective ligand and PhCN as a
suitable solvent. Under these conditions, and using 7.5 mol% Rh
catalyst, full conversion of 11c occurs at 130 °C to afford 12c in
72% isolated yield (Table 1, entry 6).20 Notably, carbonyl-based
directing groups are critical to the process; other classes of N-
protecting group (e.g., R = SO2Ar) did not afford appreciable
amounts of product. These results highlight the key role of the N-
directing group and show that, in these cases, the alkyne moiety is
not effective at orchestrating catalysis.
Table 3. Substitution of the Alkyne Tether
Our prototypical catalytic process is tolerant of a wide range of
electron-neutral and electron-rich alkynes (Table 2). Sterically
distinct alkyl- and aryl-substituted systems 13a−e are tolerated,
and the target heterocycles 14a−e were obtained in good yield.
Electron-deficient alkynes also participate but provide the
products with diminished efficiency. For example, cyclization
of 13f, which possesses an electron-deficient aryl-substituted
alkyne, provided 14f in 30% yield. The lower yield here possibly
reflects decreased alkyne coordination at the stage of the
metallacyclic intermediate 2 (see Scheme 1A). Terminal alkynes
(R1 = H) are not tolerated using this first-generation protocol,
and only traces (≤15%) of product are observed.21 These results
demonstrate the current scope and limitations of the process
with respect to the alkyne.
substituent control product selection by influencing the site of
oxidative addition (bond a vs b). Carbonylative cyclization of 17a
(R1 = Me) generated 18a (20:1 dr),22 which is derived from
metal insertion into the less hindered bond a. However, this
product was accompanied by appreciable levels of the alternate
regioisomer 19a, which is derived from insertion into the more
hindered bond b (5:2 18a:19a). In the case of 17b (R1 = n-Bu),
the more pronounced steric demands of the butyl moiety result
in higher regioselectivity for the formation of 18b (5:1 18b:19b;
18b >20:1 dr).22
Table 2. Scope of the Alkyne Component
Scheme 3. Substitution of the Aminocyclopropane
In summary, we present protecting group directed carbon-
ylative C−C bond activation of aminocyclopropanes as the basis
of a flexible entry to N-heterobicyclic systems. The present work
demonstrates proof-of-principle for the underlying catalysis
platform and outlines the preliminary scope of an approach that
potentially enables entries to a wide range of chiral scaffolds.
Future work will concentrate on expanding the methodologies
available, including the provision of asymmetric variants.23 We
will also target the development of enhanced catalyst systems
that are tolerant of a suite of synthetically useful directing groups.
Heterocyclic products of greater stereochemical complexity
are generated by increasing the substitution of the starting
material. For example, elaboration of the alkyne tether (e.g.,
15a−c) provides diastereoselective access to adducts 16a−c
(Table 3).22 Here, substitution on the tether results in shorter
reaction times compared to the examples shown in Table 2 (36−
60 vs 72−132 h). This is presumably reflective of a conforma-
tionally enhanced rate of alkyne insertion into the metalla-
cyclopentanone. To probe the origins of diastereoselectivity, the
two diastereomers of 16c were separated and independently re-
subjected to the catalysis conditions. No equilibration between
these diastereomers was observed, which supports diastereose-
lection during cyclization rather than by epimerization of the
C(7a) stereocenter of the product.
ASSOCIATED CONTENT
* Supporting Information
Experimental details, characterization data, and crystallographic
data (CIF). This material is available free of charge via the
Internet at http://pubs.acs.org.
■
S
AUTHOR INFORMATION
Corresponding Author
john.bower@bris.ac.uk
Notes
■
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
Dr. Mairi Haddow (University of Bristol) is thanked for X-ray
crystallographic analysis of 12b and 14c. Dr. Craig Butts
■
Substitution of the cyclopropane moiety is also possible
(Scheme 3). In these cases, the steric demands of the R1
C
dx.doi.org/10.1021/ja401936c | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
(10) For a recent review on C−C bond activation of cyclobutane
derivatives, see: (a) Seiser, T.; Saget, T.; Tran, D. N.; Cramer, N. Angew.
Chem., Int. Ed. 2011, 50, 7740. For a review detailing transition metal
insertion into cyclopropanes, see: (b) Jennings, P. W.; Johnson, L. L.
Chem. Rev. 1994, 94, 2241. Selected reviews on metal insertion into C−
C bonds: (c) Crabtree, R. H. Chem. Rev. 1985, 85, 245. (d) Murakami,
M.; Ito, Y. In Activation of Unreactive Bonds and Organic Synthesis; Murai,
S., Ed.; Springer: Berlin, 1999; p 97. (e) Rybtchinski, B.; Milstein, D.
Angew. Chem., Int. Ed. 1999, 38, 870. (f) Murakami, M.; Matsuda, T.
Chem. Commun. 2011, 47, 1100. (g) Aïssa, C. Synthesis 2011, 21, 3389.
(11) For a review covering asymmetric alkene cyclopropanation with
diazo compounds, see: Pellissier, H. Tetrahedron 2008, 64, 7041.
(12) Previous strategies for controlling metallacyclopentanone
formation rely upon either the inherent electronic or steric bias of the
substrate or the use of a π-unsaturated directing group (e.g., alkene or
alkyne) (see refs 3, 5, and 7). In the case of aminocyclopropanes, the
requirement to protect nitrogen engenders a significant steric
impediment for metal insertion into the more hindered cyclopropane
C−C bond. The use of the π-unsaturate to direct metallacycle formation
would necessarily limit the scope of the strategy outlined in Scheme 1A.
(13) Bart, S. C.; Chirik, P. J. J. Am. Chem. Soc. 2003, 125, 886. Here,
Pd-, Ir-, and Pt-based systems were also effective.
(14) For a review on substrate-directed reactions, see: (a) Hoveyda, A.
H.; Evans, D. A.; Fu, G. C. Chem. Rev. 1993, 93, 1307. For a recent
review on removable or catalytic directing groups, see: (b) Rousseau, G.;
Breit, B. Angew. Chem., Int. Ed. 2011, 50, 2450.
(15) For [Rh(cod)2]BF4, significant conversion to 6 was observed even
at 60 °C; this indicates that oxidative addition is reasonably facile. The
faster rate of formation of enamine vs [Rh(cod)Cl]2 may be due to an
additional vacant coordination site facilitating β-hydride elimination.
Longer reaction times provide higher conversions of 4 to 5.
(16) Cationic Rh(I) systems are particularly effective for alkene
isomerization: Tanaka, K. In Comprehensive Organometallic Chemistry
III; Crabtree, R. H., Mingos, D. M. P., Ojima, I., Eds.; Elsevier: Oxford,
U.K., 2007; Vol. 10, p 71.
(University of Bristol) is thanked for assistance with NMR
analysis. Mr. Antony Burton (University of Bristol) is thanked for
contributions during preliminary studies. M.H.S. thanks the
Bristol Chemical Synthesis Doctoral Training Centre, funded by
the EPSRC (EP/G036764/1), and Syngenta for the provision of
a Ph.D. studentship. J.F.B. is indebted to the Royal Society for the
provision of a University Research Fellowship.
REFERENCES
■
(1) For insightful discussions, see: (a) Lovering, F.; Bikker, J.;
Humblet, C. J. Med. Chem. 2009, 52, 6752. (b) Walters, W. P.; Green, J.;
Weiss, J. R.; Murcko, M. A. J. Med. Chem. 2011, 54, 6405. (c) Nadin, A.;
Hattotuwagama, C.; Churcher, I. Angew. Chem., Int. Ed. 2012, 51, 1114.
(2) Atom-economic processes are a key ideal of green chemistry, see:
(a) Trost, B. M. Science 1991, 254, 1471. (b) Trost, B. M. Angew. Chem.,
Int. Ed. Engl. 1995, 34, 259.
(3) (a) Narasaka, K.; Koga, Y. Chem. Lett. 1999, 28, 705. This process
is postulated to proceed via a rhodacyclopentanone intermediate. For a
related process that employs cyclopropenes, see: (b) Li, C.; Zhang, H.;
Feng, J.; Zhang, Y.; Wang, J. Org. Lett. 2010, 12, 3082.
(4) (a) Jiao, L.; Lin, M.; Zhuo, L.-G.; Yu, Z.-X. Org. Lett. 2010, 12,
2528. This process requires vinyl-substituted cyclopropanes and does
not proceed via a metallacyclopentanone. Alkenylidenecyclopropanes
undergo carbonylative (3+2+1) cycloadditions where the alkenylidene
moiety is incorporated into the carbocyclic ring of the product: (b) Lu,
B.-L.; Wei, Y.; Shi, M. Organometallics 2012, 31, 4601. (c) Mazumder, S.;
Shang, D.; Negru, D. E.; Baik, M.-H.; Evans, P. A. J. Am. Chem. Soc. 2012,
134, 20569.
(5) Leading references  Rh-catalyzed cyclobutanone C−C reduction
and decarbonylation: (a) Murakami, M.; Amii, H.; Ito, Y. Nature 1994,
370, 540. (b) Murakami, M.; Amii, H.; Shigeto, K.; Ito, Y. J. Am. Chem.
Soc. 1996, 118, 8285. These studies indicate that, under appropriate
conditions, carbonylative rhodacyclopentanone formation is reversible.
Rh-catalyzed intramolecular alkene insertions into cyclobutanones:
(c) Murakami, M.; Itahashi, T.; Ito, Y. J. Am. Chem. Soc. 2002, 124,
13976. (d) Xu, T.; Dong, G. Angew. Chem., Int. Ed. 2012, 51, 7567. In
both cases, the alkene may direct insertion. For an asymmetric variant,
see: (e) Xu, T.; Ko, H. M.; Savage, N. A.; Dong, G. J. Am. Chem. Soc.
2012, 134, 20005. Directed Pd-catalyzed insertions into cyclo-
butanones: (f) Ishida, N.; Ikemoto, W.; Murakami, M. Org. Lett.
2012, 14, 3230.
(6) Related processes that do not proceed via metallacyclopentanones
 Ni-catalyzed alkyne and alkene insertion into cyclobutanones:
(a) Murakami, M.; Ashida, S.; Matsuda, T. J. Am. Chem. Soc. 2005, 127,
6932. (b) Liu, L.; Ishida, N.; Murakami, M. Angew. Chem., Int. Ed. 2012,
51, 2485. Rh-catalyzed ring expansion of vinyl cyclobutanones:
(c) Wender, P. A.; Correa, A. G.; Sato, Y.; Sun, R. J. Am. Chem. Soc. 2000,
122, 7815.
(17) The procedure for the formation of 10 was adapted from earlier
work by McQuillin (see ref 9b). Attempts to isolate complexes related to
10 by exposure of 9 to phosphine-ligated Rh(I) systems under a CO
atmosphere have so far been unsuccessful.
(18) To the best of our knowledge this is the first X-ray structure of a
dimeric rhodacyclopentanone and the first X-ray structure of a
metallacyclopentanone formed by carbonylation of a cyclopropane.
For an X-ray structure of a bimetallic rhodacyclopentanone derived from
insertion of an allene into a carbonyl-ligated Rh−Ru complex, see:
Chokshi, A.; Rowsell, B. D.; Trepanier, S. J.; Ferguson, M. J.; Cowie, M.
Organometallics 2004, 23, 4759.
(19) 11a,b reached approximately 80−90% conversion after 48 h. Full
conversion of 11c occurred within 24 h.
(20) Under optimized conditions, elevated pressures of CO are
detrimental to reaction efficiency. The use of Na2SO4 as desiccant was
found to have a small but reproducible benefit. Carbamate 11b affords
enone 12b in 20% yield under the conditions outlined in Table 2. In the
absence of CO, enamine formation is observed (cf. Scheme 2A). PPh3 is
not an effective ligand for these carbonylative cyclizations, even though
it is suitable for the processes in Scheme 2A.
(21) Inefficiencies here may be due to competing Rh-vinylidene
formation. For a review, see: Bruce, M. I. Chem. Rev. 1991, 91, 197.
(22) The relative stereochemistry of the major diastereomers was
assigned using NOE experiments (see the Supporting Information).
The lack of stereocontrol in the formation of 19a,b may reflect the
lability of the R1-substituted stereocenter.
(7) Leading references  Rh-catalyzed carbonylation of cyclopropane
to form cyclobutanone: (a) Hidai, M.; Orisaku, M.; Uchida, Y. Chem.
Lett. 1980, 9, 753. Rh-catalyzed carbonylation of spiropentanes to form
cyclopentenones: (b) Matsuda, T.; Tsuboi, T.; Murakami, M. J. Am.
Chem. Soc. 2007, 129, 12596. Co-catalyzed carbonylation of
alkylidenecyclopropanes: (c) Kurahashi, T.; de Meijere, A. Angew.
Chem., Int. Ed. 2005, 44, 7881. Cyclopropylamine undergoes Rh-
catalyzed carbonylation to γ-lactams: (d) Iqbal, A. F. M. Tetrahedron
Lett. 1971, 37, 3381. Polarized cyclopropanes undergo regioselective
oxidative addition, for example: (e) Tamaki, T.; Ohashi, M.; Ogoshi, S.
Angew. Chem., Int. Ed. 2011, 50, 12067.
(8) Vinylcyclopropanes readily undergo Rh-catalyzed ring expansion,
but carbonylative variants do not proceed via metallacyclopentanones:
Liu, P.; Sirois, L. E.; Cheong, P. H.-Y.; Yu, Z.-X.; Hartung, I. V.; Rieck,
H.; Wender, P. A.; Houk, K. N. J. Am. Chem. Soc. 2010, 132, 10127.
(9) Metallacyclopentanones have been characterized  Rhodium:
(a) Wilkinson, G.; Roundhill, D. N.; Lawson, D. N. J. Chem. Soc. (A)
1968, 845. (b) McQuillin, F. J.; Powell, K. C. Dalton Trans. 1972, 2129.
Cobalt: (c) Theopold, K. H.; Bergman, R. G. J. Am. Chem. Soc. 1980,
102, 5694.
(23) At the present stage of reaction development, elongation of the
alkyne linker is not tolerated (e.g., quinoline-based scaffolds are not
accessible). Processes that involve replacement of the alkyne component
with an alkene are currently being optimized, and these studies will be
reported in due course.
D
dx.doi.org/10.1021/ja401936c | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX