Organic Letters
Letter
π-orbital of aromatic compounds can also interact with a σ-
hole.12 Even if the nucleophile NuH has neither an n-orbital
nor an acidic proton, the conformation of the pyrrole N-
tethered allylic alcohol, (E)-2a, would be fixed in a similar way
to the OH case to furnish the vinyl-substituted tetrahy-
droindolidine, (S)-3a.
Table 1 shows the results of a screening and optimization
(entries 1−8) and the scope and limitation (entries 9−19).13
Since all of the substrates in Table 1 possess an E
configuration, the prefix “(E)-” was truncated. Under the
influence of 1 mol % of (R)-1, 2a was quantitatively converted
to the desired 1−2/2-allylation product, (S)-3a, with a 99:1 er
in DMA at 50 °C for 3 h (Table 1, entry 1). The reliability of
the result was confirmed on a 10-g scale (Table 1, entry 2).
The catalyst loading could be reduced to 0.1 mol %, although
48 h was required to complete the reaction (Table 1, entry 3).
An increase in temperature slightly decreased the enantiose-
lectivity and generated a small amount of a di-allylated side-
product (Table 1, entry 4). In t-BuOH, the catalyst
performance was decreased in comparison to DMA (Table
1, entry 5). With other solvents, such as tetrahydrofuran
(THF), CH2Cl2, acetone, and toluene, little reaction occurred
(Table 1, entry 6). The Cl atom of the Cl-Naph-PyCOOH
ligand could be replaced with Br with little change in the
enantioselectivity, while the selectivity slightly decreased to a
97:3 ratio with I-Naph-PyCOOH (Table 1, entries 7 and 8).
With a diastereomeric (Z)-2a, an endo-trig cyclization at C(1)
of allyl moiety was predominant to generate an 8-membered-
cyclic product in 81% yield.13 Methyl or ethyl substituents
could be introduced at the β or γ position of the allylic alcohol
moiety (Table 1, entries 9−12). With an alkyl group at the γ
carbon, a quaternary carbon stereogenic center could be
controlled (Table 1, entries 11 and 12). Methoxycarbonylation
at C(3) of (S)-3e gave a key synthetic intermediate for ent-
rhazinilam.14 Introduction of an ethyl, phenyl, methylthio, or
benzoyl group at C(2) of 2a afforded the corresponding C(3)-
substituted tetrahydroindolidines, 3f−3i, with er values of
93:7−98:2 (Table 1, entries 13−16), although the electron-
withdrawing COC6H5 substituent significantly decreased the
reactivity (Table 1, entry 16). Cyclization using 2j with a C(3)-
methyl substituent afforded 3j with a 9:1 C(2)/C(5)
regioselectivity and in a 78% total yield (Table 1, entry 17).
Elongation of the methylene tether from three to four carbons
gave a pyrrole-fused vinylazepane 3k with er = 97:3, whereas
shortening from three carbons to two carbons led to no
production of the desired dihydropyrrolidine 3l (Table 1,
entries 18 and 19), in contrast to the corresponding OH case.8
Replacement of the pyrrole of 2a with indole resulted in no
reaction. The general trend of enantioselectivity observed with
these E substrates was the same as that observed in the (R)-1-
catalyzed dehydrative cyclization of NuH (Figure 2a).8−10 The
same trend of enantioselectivity suggests that a similar
mechanism to the OH case11 is operating.
Figure 1. Catalytic asymmetric intramolecular allylations of pyrroles
patterned by the pyrrole atom positions of condensation and
allylation; e.g. “2−3/3-allyl type” denotes C(2)−C(3)-ring con-
densation and C(3)-allylation. X is a leaving group.
Figure 2. (a) CpRu(II)/(R)-1-catalyzed dehydrative enantioselective
cyclization of allylic alcohols with a protic nucleophile (NuH) and the
σ-allyl intermediate proposed by the mechanistic study.11 The catalyst
was generated in situ using (R)-Cl-Naph-PyCOOAll (All =
CH2CHCH2). (b) Standard reaction for investigating the
possibility of the dehydrative 1−2/2-type pyrrole allylative cyclization.
with OH as the terminal NuH strongly supported that the
major S enantiomer should be produced via a σ-allyl Ru
intermediate.11 The nucleophilic OH and the allylic OH are
trapped by the σ-hole12 of the Cl atom and the H+ of the
COOH group in (R,SRu)-1, respectively. The double activation
of the substrate facilitates oxidative addition of Ru(II) to
generate the macrocyclic σ-allyl intermediate, in which the OH
oxygen atom quickly attacks the C(3) carbon in a trans-annular
SN2′ manner to give the exo-trig-cyclized product.
Figure 3a shows the D-labeling experiments performed to
confirm the mechanism similarity to the OH case:11
A
88.5:11.5 mixture of (S)-2a-d and (R)-2a-d was converted
under the same conditions as those of entry 1 in Table 1 to a
86.5:11.2:0.33:1.97 mixture of (S)-(Z)-COCF3-4a-d (SZ), (S)-
(E)-COCF3-4a-d (SE), (R)-(Z)-COCF3-4a-d (RZ), and (R)-
(E)-COCF3-4a-d (RE) (S:R = 97.7:2.3; Z:E = 86.7:13.3) after
acylation of 3a-d, followed by chiral HPLC separation of the
enantiomers.13 The most tangible pathway would be one in
which the 88.5 part and 11.5 part are transformed mainly to
The new mechanism encouraged us to try the dehydrative
1−2/2-type pyrrole allylative cyclization (Figure 2b), because a
B
Org. Lett. XXXX, XXX, XXX−XXX