Palladium catalyzed cyclizations of oxime esters with 1,2-disubstituted alkenes: Synthesis of dihydropyrroles

Article abstract of DOI:10.1021/ol4023112

Pd-catalyzed cyclizations of oxime esters with 1,2-dialkylated alkenes provide an entry to chiral dihydropyrroles. Substrate and catalyst controlled strategies for selective product formation (vs alternative pyrroles) are outlined.

Full text of DOI:10.1021/ol4023112

ORGANIC
LETTERS
2013
Vol. 15, No. 17
4616–4619
Palladium Catalyzed Cyclizations of
Oxime Esters with 1,2-Disubstituted
Alkenes: Synthesis of Dihydropyrroles
Nicholas J. Race and John F. Bower*
School of Chemistry, University of Bristol, Bristol, BS8 1TS, U.K.
john.bower@bris.ac.uk
Received August 13, 2013
ABSTRACT
Pd-catalyzed cyclizations of oxime esters with 1,2-dialkylated alkenes provide an entry to chiral dihydropyrroles. Substrate and catalyst controlled
strategies for selective product formation (vs alternative pyrroles) are outlined.
Robust and diversity oriented catalysis platforms are vital
to drug discovery. Since the early 1990s, drug candidate
library synthesis has relied heavily upon Pd(0)-catalyzed
cross-couplings.1 These processes convert readily available
starting materials to a wide range of new products via a
mechanistically common aryl-Pd(II) intermediate. This
unified retrosynthetic blueprint, coupled with the opera-
tional simplicity and wide substrate scope offered by these
methods, accounts for their popularity in medicinal chem-
istry. While particularly suited to the formation of new two-
dimensional structures (e.g., biaryls), these processes are
less applicable to the generation of chiral “3D” molecules.
Consequently, over the past 20 years the average “chiral
complexity” of drug candidate libraries has decreased
significantly.2À4 There is a growing realization that small
molecule drug discovery programs can be enhanced by
increasing the “chirality content” of compound libraries.3,4
This, in turn, requires the development of catalysis plat-
forms that not only retain the operational simplicity of
Pd(0)-catalyzed cross-couplings but also provide direct
access to privileged chiral scaffolds.
To address the issues outlined above, we have examined
the possibility of exploiting palladated imines 1 as catalytic
intermediates for accessing chiral N-heterocyclic systems
(Scheme 1a).5,6 Intermediates 1 are generated by oxidative
addition of Pd(0) into the NÀO bond of oxime esters and
undergo Heck-like cyclization with pendant alkenes.7,8
In general, this leads to heteroaromatic products (new
C(sp2)ÀN bond formed).8c,d However, in our previous studies,
following 5-exo imino-palladation, the nature of the alkene
acceptor controlled the direction of β-hydride elimination to
generate chiral products 2a,b (new C(sp3)ÀN bond formed).
The ligand requirements for these processes contrast those
of the conventional Heck reaction, and we have found that
electron deficient triarylphosphines are effective.5,6
Related cyclizations involving 1,2-dialkylated alkenes
are known to generate preferentially pyrrole products and
donot provide access to chiral targets. Narasaka has shown
that cyclization of 3, using PPh3 as the ligand, generated
pyrrole 4 with 8:1 selectivity over dihydropyrrole 5
(Scheme 1b).8a,b,9 Selective product generation requires
strategies to control the direction of β-hydride elimination
(5) Faulkner, A.; Bower, J. F. Angew. Chem., Int. Ed. 2012, 51, 1675.
(6) Faulkner, A.; Scott, J. S.; Bower, J. F. Chem. Commun. 2013, 49,
1521.
(1) (a) Carey, J. S.; Laffan, D.; Thomson, C.; Williams, M. T. Org.
Biomol. Chem. 2006, 4, 2337. (b) Roughley, S. D.; Jordan, A. M. J. Med.
Chem. 2011, 54, 3451.
(2) “Compound chirality” correlates to the percentage of C(sp3)
centers in a compound: Walters, W. P.; Green, J.; Weiss, J. R.; Murcko,
M. A. J. Med. Chem. 2011, 54, 6405.
(7) Tan, Y.; Hartwig, J. F. J. Am. Chem. Soc. 2010, 132, 3676.
(8) (a) Tsutsui, H.; Narasaka, K. Chem. Lett. 1999, 28, 45. (b)
Tsutsui, H.; Kitamura, M.; Narasaka, K. Bull. Chem. Soc. Jpn. 2002,
75, 1451. For reviews, see: (c) Kitamura, M.; Narasaka, K. Chem. Rec.
2002, 2, 268. (d) Narasaka, K.; Kitamura, M. Eur. J. Org. Chem. 2005,
4505. Pentafluorobenzoyl oxime esters are usually employed for these
processes because they are stable to Beckmann rearrangement.
(9) As far as we are aware, this is the only example where cyclization
involving a linear 1,2-dialkylated alkene generates any dihydropyrrole.
(3) There is a pressing demand for the development of efficient
methodologies that target low molecular weight (200À350 Da), 3D
(sp3-rich) scaffolds: Nadin, A.; Hattotuwagama, C.; Churcher, I.
Angew. Chem., Int. Ed. 2012, 51, 1114.
(4) Lovering, F.; Bikker, J.; Humblet, C. J. Med. Chem. 2009, 52,
6752.
r
10.1021/ol4023112
Published on Web 08/27/2013
2013 American Chemical Society

(Ha vs Hb) at the stage of alkyl-Pd(II) intermediate 6
(Scheme 1c). Elimination via Ha accesses directly chiral
target 7, whereas elimination via Hb affords an enamine
intermediate that undergoes isomerization to the undesired
pyrrole 8.8,10 In this study we outline (a) substrate and (b)
catalyst controlled strategies that achieve selectivity for chiral
products 7. This provides a new class of cyclization and
enhances further the scope of this catalysis platform for chiral
scaffold generation. The selectivity issues addressed here did
not apply to our earlier work where the constraints of the alkene
acceptor predetermined β-hydride elimination directionality.5,6
did not result in significant isomerization to 11a.12 These
observations are indicative of kinetic product selection.
In our previous studies, we delineated the scope of this type
of process with respect to substituents on the oxime.5,6 The
present class of cyclization demonstrates similar tolerances,
and alkyl and cyclopropyl substituted systems 9b/c under-
went smooth cyclization to afford 10b and 10c respectively.13
In both cases, the phenyl group enforced good selectivity for
the chiral targets over the corresponding pyrroles 11b/c.
Scheme 2. Scope of the R1 and R2 Groups
Scheme 1
b
a Isolated yield of the major E-alkene isomer. Total cyclization
yields and selectivities were determined by 1H NMR of the crude
reaction mixture (vs trimethoxybenzene as a standard).
The success of these cyclizations mirrors our previous
observations that electron deficient Pd-systems enhance
reaction efficiency.5,6 This is likely due to beneficial effects
upon the properties of the key imino-Pd(II) intermediate 1.
Presumably, an electron deficient Pd(II)-center enhances
σ-donation from the imine moiety which increases the
NÀPd bond strength and decreases N-lone pair basicity.
This then suppresses competitive protodepalladation to the
corresponding NH imine. Additionally, a more electrophilic
Pd-center may increase the rate of imino-palladation via
increased coordination of the pendant alkene.14 By way of
comparison, use of PPh3 as the ligand for the cyclization of 9a
provided a 12:1 mixture of 10a:11a but in only 45% overall
yield. Here, significant quantities of the corresponding ketone
We initially explored scenarios where R2 = aryl because
secondary orbital interactions (π(aryl)fσ*(CÀH)) should
weaken the CÀHa bond of 6 and increase its propensity
to β-hydride elimination.11 In the event, 9a (R2 = Ph)
underwent efficient and selective cyclization, when treated
with a P(3,5-(CF3)2C6H3)3 ligated Pd(0)-system, to pro-
vide a 17:1 mixture of 10a:11a in 78% overall yield, as
1
determined by H NMR analysis of the crude mixture
(Scheme 2). 10a was formed as a 12:1 mixture of geometric
isomers, and chromatographic purification afforded read-
ily the major E-isomer in 64% yield. The product ratio of
10a:11a was constant over the course of the cyclization
process, and resubmission of 10a to the reaction conditions
(10) Isomerization may be promoted either by acid generated during
the aza-Heck process or by Pd(II)-hydride.
(13) Aldoxime esters are challenging substrates due to competitive
Beckmann rearrangement to the corresponding nitrile. Studies to
address this issue are ongoing.
(14) Electron-deficient ligand systems may also accelerate migratory
insertion of the alkene. For studies on alkene insertion into Pd-NPh2
bonds, see: (a) Hanley, P. S.; Hartwig, J. F. J. Am. Chem. Soc. 2011, 133,
15661. See also: (b) Neukom, J. D.; Perch, N. S.; Wolfe, J. P. Organo-
metallics 2011, 30, 1269 and references cited therein.
(11) For studies that delineate factors responsible for β-hydride
elimination selectivity from alkyl-Pd(II) intermediates, see: (a) Werner,
E. W.; Sigman, M. S. J. Am. Chem. Soc. 2011, 133, 9692. (b) Werner,
E. W.; Sigman, M. S. J. Am. Chem. Soc. 2010, 132, 13981.
(12) Product ratios have been determined by 1H NMR at various
time points during the cyclization of 9a.
Org. Lett., Vol. 15, No. 17, 2013
4617

were isolated, and our studies indicate that this forms by
hydrolysis of the NH imine protodepalladation product.15
We have undertaken an assessment of the scope of the
aryl moiety to determine its effects upon product selection
(Scheme 2).16 A range of electron-rich and -poor aryl
groups gave good selectivity for the target dihydropyrroles
10dÀi (10:1 to >20:1 vs 11dÀi).17 In all cases, the major
E-alkene isomer was isolated easily in moderate to good
yield (45À64%). For 10i, the crude reaction mixture was
not readily analyzed by 1H NMR and so in situ selectivities
have not been determined.
some alkene isomerization to the corresponding styrenyl
derivative was observed (6:1 regioisomer ratio).19 As before,
product ratios were largely constant over the time frame of
the reaction, which supports kinetic product selection.
Scheme 4
Scheme 3. Cyclizations Generating 1,2-Dialkylated Alkenes
In the cyclizations outlined in Scheme 3, the R2 group
seemingly provides no significant activation of CÀHa
(vs CÀHb) for β-hydride elimination (see Scheme 1c).
We have carried out an assay of ligand effects (PPh3 vs
P(4-CF3C6H4)3 vs P(3,5-(CF3)2C6H3)3) on the cyclization
of 9k to uncover the origins of product selection in these
cases (Scheme 4a). As expected, overall cyclization effi-
ciency improves as the ligand becomes more electron
deficient. However, both PPh3 and P(4-CF3C6H4)3 af-
forded predominantly the undesired pyrrole 11k.20 These
data indicate that P(3,5-(CF3)2C6H3)3 is able to enforce
selective β-hydride elimination because of the steric effects
of the meta-CF3 substituents (i.e., larger cone angle vs
P(4-CF3C6H4)3).21 It is generally accepted that β-hydride
elimination proceeds through a planar 4-centered transi-
tion state.22 This requires alkyl-Pd(II) intermediate 12 to
adopt transiently a high energy conformation where angle
x deviates from ideality (Scheme 4b). We suggest that
selective β-hydride elimination is facilitated by the inter-
play of the sterically demanding ligand system (PR3 =
P(3,5-(CF3)2C6H3)3) with the bulky dihydropyrrole moi-
ety (vs the smaller propyl group) (conformation I vs II). In
conformation I, steric repulsion between these two groups
is alleviated by progressing to TS-I en route to the desired
dihydropyrrole10k. Inconformation II, the smaller propyl
group suffers less severe steric interactions with the ligand,
and so progression to TS-II, and ultimately 11k, is
a Isolated yield of the major E-alkene isomer; only E-alkene products
were observed. Total cyclization yields and selectivities were deter-
b
mined by 1H NMR of the crude reaction mixture (vs trimethoxybenzene
as a standard). Isolated as a 6:1 mixture of alkene regioisomers.
c
d Isolated as a 12:1 mixture of alkene regioisomers.
In the cyclizations presented in Scheme 2, product
selectivity is dictated by the strategic installation of an aryl
group although the ligand may also play a role (vide infra).
Consequently, in cases where CÀHa is not benzylic (see
Scheme 1c), it was unclear whether cyclizations could be
selective for chiral products. Pleasingly, using the same
reaction conditions as outlined in Scheme 2, cyclization
of 9jÀl afforded dihydropyrroles 10jÀl with between
3:1 and 5:1 selectivity over the undesired pyrroles 11jÀl
(Scheme 3). Here, the lower product selectivities are
offset by the more efficient cyclizations (73À87% total
cyclization yield), and so the targets could still be isolated
in synthetically useful yields (58À60%).18 In the case of 10j,
(15) In the absence of palladium, pentafluorobenzoyl oxime esters
are stable to the reaction conditions (see ref 5).
(16) For the synthesis of oxime esters employed here see the Support-
ing Information. Oxime ester geometry does not affect the efficiency of
cyclization, and facile interconversion likely occurs at the stage of the
imino-Pd(II) intermediate (see ref 8).
(17) A strong trend between the electronics of the aryl moiety and
product selectivity is not evident. However, the importance of the aryl
group is underscored by comparing the cyclization results of 9a and 9k
using PPh3 as the ligand (12:1 vs 1:1.6 selectivity; see text and Scheme 4).
(18) We have also explored the cyclization of compound 3. Under our
optimized conditions we have been able to obtain a 62% overall yield
and 1:3 ratio of 4:5. However, control experiments indicate that dihy-
dropyrrole 5 is not stable to the reaction conditions and so yields and
selectivities are variable in this particular case.
(19) The regioisomer ratio of 10j is constant over the course of the
reaction suggesting that isomerization occurs prior to dissociation of
PdÀH. For related observations, see: Mei, T.-S.; Werner, E. W.;
Burckle, A. J.; Sigman, M. S. J. Am. Chem. Soc. 2013, 135, 6830.
(20) In all cases, the ratio of 10k:11k was largely constant over the
time frame of the reaction, thereby supporting kinetic product selection.
(21) The cone angle for P(3,5-(CF3)2C6H3)3 has been determined as
160° vs 155° for PPh3 and P(4-CF3C6H4)3: Howell, J. A. S.; Fey, N.;
Lovatt, J. D.; Yates, P. C.; McArdle, P.; Cunningham, D.; Sadeh, E.;
Gottlieb, H. E.; Goldschmidt, Z.; Hursthouse, M. B.; Light, M. E.
J. Chem. Soc., Dalton Trans. 1999, 3015.
(22) Koga, N.; Obara, S.; Kitaura, K.; Morokuma, K. J. Am. Chem.
Soc. 1985, 107, 7109.
4618
Org. Lett., Vol. 15, No. 17, 2013

Scheme 5. Synthetic Manipulations of Cyclization Product 10a
Scheme 6
kinetically less favorable.23 The steric and electronic proper-
ties of P(3,5-(CF3)2C6H3)3 are able therefore to effect both
efficient cyclization and selective chiral product formation.
The dihydropyrrole products described here are well suited
to further manipulations, and we have demonstrated this
using adduct 10a (Scheme 5). Mild hydrogenation of the
alkene moiety afforded 13 in good yield. Longer hydrogena-
tion times resulted in the reduction of both the alkene and
imine to generate pyrrolidine 14 (10:1 dr). Reduction of solely
the imine moiety was achieved using DIBAL-H, and, after
N-tosylation, pyrrolidine 15 was isolated in greater than 20:1
diastereoselectivity. The diastereofacial bias associated with
10a makes it well suited to other imine 1,2-addition processes.
Exposure to allyl magnesium chloride afforded 16, which
contains a challenging quaternary amino-stereocenter, in
excellent diastereoselectivity. Access to more complex hetero-
cyclic scaffolds is also possible. For example, diastereoselec-
tive imine reduction of 10a, followed by N-alkylation with
4-bromobutene, enabled an efficient ring closing metathesis
reaction to generate indolizidine bicycle 17.
We have also explored cyclizations of more heavily
substituted systems. Cyclizations of 9m and 9n proceeded
smoothly to generate targets 10m and 10n in good yield
(Scheme 6a). Here, the more bulky dihydropyrrole
moiety (vs 9a to 10a) enforces perfect selectivity for β-
hydride elimination and no traces of pyrrole products were
observed (cf. Scheme 4b). Enhancement of the diastereo-
mer ratios of 10m and 10n by epimerization of the C-4
stereocenter was investigated without success. Under equil-
ibrating conditions (e.g., HCl/MeOH or NaOMe/MeOH)
no diastereoenrichment was evident. Alternative strategies,
based upon the formation of an aza-enolate and kinetic
reprotonation, were complicated by the relatively high
acidity of the C-2 proton. Substitution at C-3 of the starting
material shows more promise for stereoselective synthesis
(Scheme 6b). In the cyclizations of 9o and 9p the size of the
C-3 substituent controls the diastereoselectivity associated
with dihydropyrroles 10o and 10p. The trans-isomer was
favored in both cases, but the larger phenyl group provided
greater levels of diastereocontrol (5:1 dr for R = Ph vs 2:1
dr for R = Me). Once again, these cyclizations were
chemically efficient and exhibited high selectivity for dihy-
dropyrroles 10o and 10p over the corresponding pyrroles
(>20:1 selectivity). Diastereoselective reduction of the
imine moiety (DIBAL-H) of the major diastereomer of
10p provided, after N-tosylation, trisubstituted pyrrolidine
18 (characterized by X-ray diffraction) (Scheme 6c).
The present study outlines, for the first time, substrate
and, perhaps more importantly, catalyst controlled strate-
gies to reverse the established product selectivity for
aza-Heck cyclizations of oxime esters with 1,2-dialkylated
alkenes. The net result is a direct and flexible entry to chiral
dihydropyrrole targets that are then well suited to further
elaboration. In combination with earlier work, we have
now demonstrated that a simple and commercial catalyst
system enables efficient access to a wide range of chiral
pyrrolidine derivatives. These studies will feed into the
design of effective chiral ligands and the development of
complexity generating cascade reactions.
Acknowledgment. EPSRC (EP/J007455/1) and the
Bristol Chemical Synthesis CDT (studentship to N.J.R.),
funded by the EPSRC (EP/G036764/1), are thanked for
support. Nell Townsend and Dr. Mairi Haddow
(University of Bristol) are thanked for X-ray analysis
of 10d and 18. J.F.B. thanks the Royal Society for a
University Research Fellowship.
Supporting Information Available. Experimental pro-
cedures and characterization data for all products. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(23) We favor a cationic aza-Heck pathway as depicted in Scheme 4b
because electron deficient leaving groups are required for efficient
cyclization. The corresponding benzoyl oxime esters are active for
oxidative addition (see: Nishimura, T.; Nishiguchi, Y.; Maeda, Y.;
Uemura, S. J. Org. Chem. 2004, 69, 5342) but provide traces of
cyclization product under the conditions described here.
The authors declare no competing financial interest.
Org. Lett., Vol. 15, No. 17, 2013
4619