Full Paper
We wished to exploit the benefits of alkyne hydroacylation
in a route to complex pyrroles; our proposed retrosynthesis is
shown in Scheme 1, and involves the hydroacylative union of
aldehydes and propargylic amines to deliver g-amino enone
products.[11] Dehydrative cyclisation of these enones would de-
liver the targeted pyrroles. Employing appropriately substitut-
ed aldehydes and alkynes would allow the direct synthesis of
highly substituted ring systems. In this article we demonstrate
how recent advances in catalyst design have allowed us to de-
velop efficient syntheses of both pyrroles and dihydropyrroles
using highly selective intermolecular alkyne and alkene hydro-
acylation reactions.
(Table 1). Our concerns over regioselectivity were soon proved
correct, with reactions employing a catalyst incorporating the
ligand DPEphos[14c] delivering the hydroacylation adducts as
a 2:1 mixture of linear and branched enones, respectively
(entry 1). We have recently shown that small-bite-angle, meth-
ylene-bridged diphosphines generate efficient and selective
catalysts for a variety of hydroacylation reactions.[22] Unfortu-
nately, in the present study, both bis(diphenylphosphino)me-
thane (dppm) and bis(dicyclohexylphosphinomethane) (dcpm)
ligands offered only modest improvement (entries 2–4). How-
ever, it was the electronic parameters of the PNP ligand
system, specifically the ligand (Cy2)P2NMe, that offered a signifi-
cant increase in the level of regiocontrol (entry 5).[23] To achieve
higher conversions we turned our attention to evaluating the
effect of concentration and temperature on the reactions. In-
creasing the concentration to 1.0m with respect to the alde-
hyde, allowed complete conversion for both the PCP and PNP
systems (entry 4 and 6). When assessing the effect of tempera-
ture, the general trend indicated that decreased temperatures
accomplished greater linear selectivity, without any loss in con-
version under ambient conditions (entry 8). Reaction at 08C
provided a more selective transformation but required a slightly
longer reaction time (entry 9). We also evaluated Cbz- and Ts-
protected propargylic amines in order to determine whether
the nature of the amine-protecting group played a significant
role in the selectivity of the reactions. In both cases the PNP-
derived catalyst outperformed the corresponding dcpm
system, and for the Cbz substrate was able to provide the de-
sired linear adduct with reasonable, but diminished with re-
spect to Boc-derivative, selectivity (entries 10 and 11). The N-Ts-
substrate was poorly selective with either catalyst (entries 12
and 13). For convenience we decided to undertake subsequent
reactions using the room temperature conditions shown in
entry 8, and to employ N-Boc-protected amines.
Results and Discussion
One of the main limitations of intermolecular hydroacylation
reactions employing RhI-catalysts is a competing reductive de-
carbonylation pathway.[6] The majority of approaches to over-
come this issue invoke some form of substrate chelation, be it
from the aldehyde,[12–16] or alkene or alkyne.[17] Although there
are now a number of successful non-chelation intermolecular
hydroacylation methods,[18,19] the most general and selective
processes invariably rely on chelation control and have allowed
the development of reactions that operate under mild condi-
tions, employ low loadings of catalyst and achieve high levels
of selectivity.[20] Accordingly, we selected chelating aldehydes
as our preferred substrates. The presence of coordinating func-
tionality on the alkene or alkyne coupling partner has also
been shown to have a significant impact on the course of sev-
eral hydroacylation reactions, and has been employed to facili-
tate both reactivity and selectivity.[17] For example, both Dong
and Suemune employed alkenes bearing coordinating func-
tionality to achieve branched-selective reactions,[21] required to
develop enantioselective processes. We were therefore con-
scious that our proposed alkyne substrates—propargylic
amines—may well be involved in coordination to the metal
catalyst and that this could effect the regioselectivity of the
proposed reactions. Poor regiocontrol would have significant
consequences for our designed pyrrole synthesis, as only the
linear hydroacylation adducts are capable of undergoing a pyr-
role-forming cyclisation (Scheme 2).
We next investigated whether this intermolecular hydroacy-
lation could maintain high levels of selectivity when varying
the propargylic amine substrates (Table 2). Pleasingly, introduc-
ing substituents a to the nitrogen atom displayed enhanced
selectivity for the linear product (3b–d), demonstrating that
further improvements in regiocontrol could be achieved by
substrate design. A range of aldehydes, including substituted
aryl (3 f), alkyl (3g), alkenyl (3h), heteroaromatic variants (3i),
and the use of a nitrogen-chelating atom (3j), were also intro-
duced without incident. We then expanded the process to in-
clude internal alkyne substrates, which following cyclisation
would allow the direct formation of tetra-substituted pyrroles.
The hydroacylation of an internal alkyne proceeded efficiently,
although an alternative ligand was required (3e); the PNP(Cy)-
based catalyst did not deliver the necessary regiocontrol (2:1,
linear: branched), however, when using dppm, a PCP-based
ligand system, acceptable levels of linear-selectivity were ach-
ieved (8:1). This brief study also established that it was possi-
ble to employ a catalyst loading of only 1 mol% (3d), although
for pragmatic reasons the majority of reactions were per-
formed using a 5 mol% loading.
We began our investigation by studying the coupling of o-
SMe-benzaldehyde 1a and N-Boc propargylic amine 2a
With a selective hydroacylation reaction in hand, we next
studied the cyclisation to generate the desired pyrroles. g-
Scheme 2. Linear versus branched selectivity in alkyne hydroacylation.
Chem. Eur. J. 2016, 22, 7879 – 7884
7880 ꢀ 2016 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim