Angewandte
Communications
Chemie
of indole, in which one enantiomer of the substrate undergoes
substrates, though perturbation on substrate structure
C3-allylation while the other undergoes N-allylation in
a comparable rate in the presence of a chiral Pd catalyst.
This hypothesis fits into the concept of parallel kinetic
affected the resolution result to some extent [Eq. (2)]. Ethyl
as the R group in place of methyl (substrate 8a) and tuning of
1
the electronic property of the indole core by methoxy
substitution (substrate 8b) still rendered the reaction under
the regime of PKR, albeit the resolution deviated from
perfect parallelity. This implies subtle matches among certain
reaction components to be the key to an ideal PKR.
[19]
resolution (PKR).
In further reaction condition optimization, we found that
organic bases could enhance the performance of the Pd/L1
catalytic system and allowed for the formation of significant
amount of N-attack product (entries 5 and 6). Switching the
Pd source to [Pd(allyl)Cl] was slightly beneficial (entry 7).
2
Systematic optimizations (Table S1) revealed that a diluted
mesitylene solution, addition of molecular sieves, and a low
reaction temperature were the key to success (entry 8): the
reaction could be performed with only 1 mol% of Pd and
1
.5 mol% of L1 to afford (+)-7 and (+)-11 in high enantio-
selectivities and nearly equal yields. The absolute configu-
ration of (+)-7 was determined after having accomplished the
synthesis of arborisidine and later confirmed by single crystal
X-ray diffraction (XRD) analysis, while that of (+)-11 was
also identified by XRD analysis (vide infra). Since both
In order to gain mechanistic insights into this intriguing
PKR, DFT calculation was performed. We noticed that
a number of computational studies on Pd-catalyzed AAA
[
22]
(R,R)- and (S,S)-L1 are commercially available, we reasoned
reactions have been reported,
focused on the Pd/L1 catalytic system.
AAA transition states (TSs) in this reaction, we adopted the
but only a few of them
[
23]
that (À)-7 could be prepared following the same way leading
In modeling the
to the natural (+)-arborisidine.
[
23a]
We believed that this reaction represents a regiodivergent
PKR of (Æ)-8, as supported by the following evidences
established Lloyd-Jones/Norrby model
while taking the
effect of TMG into consideration due to the observed base
effect (Table 1). Surprisingly, our initial calculation on
(
Figure 2): (1) The absolute configurations of (+)-7 and
+
a cationic mechanism involving [(R,R)-L1-Pd(allyl)] inter-
mediates failed to reproduce the observed selectivity (see the
Supporting Information for details). Given that this PKR
proceeded in a nonpolar solvent, we reasoned that a tight ion-
[24]
pair mechanism may operate instead of a cationic mech-
anism. Therefore, a mechanism involving neutral [(R,R)-L1-
+
À
Pd(allyl)] ·OAc intermediates was modeled, in which vari-
ous possible TSs with different geometries and anion binding
modes were surveyed (Figure 3).
Among them, TS-(R)-C/TS-(R)-N and TS-(S)-C/TS-(S)-
N were the most favorable C-/N-attack TSs for (R)- and (S)-
À
substrates, where AcO binds to the amide N-H by hydrogen
bonding and TMG abstracts the indole N-H to assist the
reaction. For the (R)-substrate, TS-(R)-C is 4.8 kcalmol
À1
mentary crystallographic data for this paper. These data are provided
free of charge by the joint Cambridge Crystallographic Data Centre and
favorable than TS-(R)-N; for the (S)-substrate, TS-(S)-N is
À1
2
.0 kcalmol
favorable than TS-(S)-C. This is in good
agreement with the observation that the PKR favors C3-
attack for (R)-8 and N-attack for (S)-8, as well as the high ee
of product (S)-11 owing to the highly selective C-attack in the
(R)-pathway. The unfavorable N-attack in the (R)-pathway is
likely caused by the repulsion of the cyclohexane ring in L1
with the methylene group adjacent to the p-allyl unit and the
Boc group in TS-(R)-N; on the other hand, in TS-(S)-C, the
repulsion of the Boc group and the indole core with the
phenyl rings in L1 might be responsible for the unfavorable C-
attack in the (S)-pathway (see the double-headed red arrows
in Figure 3).
(
(
+)-11 determined by XRD analysis [R for (+)-7 and S for
+)-11 at the aza-quaternary center]; (2) Parallel generation
of the two products repeatedly observed at different con-
versions; and (3) Fitness of the ee values and yields of the two
[
20]
products to the equation ee ·[x] = ee ·[y]. Since proposed in
x
y
1
997 by Vedejs, PKR has been recognized as a rare type of
resolution with strict kinetic requirements, and its application
[
21]
in natural product synthesis was extremely rare.
The
present reaction is the first example of PKR using the Pd-
catalyzed AAA reaction, which set the absolute configura-
tions of two quaternary chiral centers in a single step and
overcame the shortcomings of traditional KR in the intra-
molecular AAA. Gratifyingly, we found that the optimal
reaction conditions were applicable to the PKR of more
Successful preparation of (+)-7 paved the way towards
(À)-arborisidine (Scheme 2). Treatment of (+)-7 with benze-
[25]
nesulfenyl chloride
led to the crude chloride 12 as
diastereomeric mixture (dr= 1.2:1). The following elimina-
tion reaction was highly dependent on the base used, and the
[26]
n-Bu NCl/2,4,6-collidine system was the optimal choice to
4
Angew. Chem. Int. Ed. 2021, 60, 1 – 6
ꢀ 2021 Wiley-VCH GmbH
3
These are not the final page numbers!