Organic Letters
Letter
bearing a para or meta substituent on the aryl group afforded
3b−3h in high yields with 84−92% ee and 3:1 to 13:1 dr. The
substituent at the ortho position furnished 3i in a moderate
yield of 48% with 13:1 dr and 65% ee, possibly due to steric
hindrance upon a Michael step. The 2-naphthyl or furyl group
instead of an aryl ring was also compatible with this
methodology to produce 3j and 3k in 80% yield with high
stereocontrol. Moreover, this protocol could also be applied to
various substituents at the alkyne tether, affording 3l−3q in
76−85% yields with 86−90% ee and 3:1 to 7:1 dr. In addition,
the absolute configuration of the resulting 3 was determined by
the X-ray analysis of 3e.15
Scheme 4. Proposed Reaction Mechanism
We then turned our attention to the scope of substituted 3-
pyrrolyloxindoles 2. As shown in Scheme 3, this catalytic
strategy was compatible with several substituents on the
aromatic ring of oxindole. The electronic properties of the
substituents at the C5 position of oxindole rings showed a
weaker effect on this transformation, and either electron-rich
groups (CH3 and OCH3) or electron-poor groups (Cl and Br)
gave the target cycloadducts 3r−3u with good yields (70−
84%), diastereoselectivity (3:1), and high enantioselectivities
(85−92% ee). Similarly, good results were obtained with the
substituent at the C6 or C7 position of oxindoles (3w and 3y).
We also tested the substituent at the C4 position, which
decreased the reactivity (3v) and enantioselectivity because of
the steric hindrance. Interestingly, a switch of the N substituent
from methyl to benzyl led to a positive effect on the
stereoselectivity and afforded product 3z in 84% yield, 8:1
dr, and 97% ee (vs 3b). Extensive study indicated that the N-
benzyl-substituted 3-pyrrolyloxindoles could deliver the
corresponding products with better stereocontrol (3t vs 3t′,
3u vs 3u′, and 3w−3y vs 3w′−3y′). Further study of the N
protection indicated that the N-allyl substituent also afforded
3A with a comparable result. Variation of the N substituent
with i-Pr, phenyl, or H could also give the corresponding
products 3B−3D in good yield and enantioselectivity.
However, N protection with electron-poor groups, such as
Boc or Cbz, failed to afford the desired products.
Scheme 5. Scale-Up and Further Transformations
we envisoned that the branched nitroenynes could also be
employed as biselectrophiles to trigger the asymmetric cascade
cyclization with 3-pyrrolyloxindoles in the presence of
bifunctional organocatalysts to produce the spiropiperidinyl
oxindole derivatives (Scheme 1c). The challenge of this design
arises from the control of the regioselective and stereoselective
Michael/Friedel−Crafts cascade process, especially the endo-
cyclization without the addition of Lewis acidic metal.
A plausible catalytic reaction model for this cascade
cyclization was proposed. As illustrated in Scheme 4, the
reaction was initiated from the deprotonation of 3-pyrrolylox-
indole 2a by the quinuclidine nitrogen of C8, which resulted in
formal indol-2-ol. Meanwhile 3-nitroenyne 1a was activated by
double H-bonding between the squaramide moiety and the
nitro group of 1a. The formal indol-2-ol then underwent a
Michael addition from the Re-face. The nitroenyne approaches
with its Si-face because of the bulky N-tert-butyl group on C8.
Presumably, the N-benzyl group of 2 shielded the Si-face attack
to afford better enantioselectivity. The subsequent proton
transfer of intermediate I resulted in allenyl II as its resonance
form. After that, the pyrrolyl moiety of 2a was added to
nitroallene II in a 6-endo-trig intramolecular Friedel−Crafts
cyclization. The further isomerization of III gave rise to the
target aza-spirooxindole 3a.
To demonstrate the practicability of the cascade cyclization,
a scale-up synthesis of 3 was conducted on a 2.2 mmol scale,
giving the desired 3z in 76% yield with maintained dr and ee
values. Additionally, the treatment of 3z with n-Bu3SnH
selectively reduced the CC bond and afforded product 4 in
68% yield with excellent stereoselectivity. The absolute
stereochemistry of 4 at the two new chiral centers was
confirmed by NOESY analysis (Scheme 5).
To test the feasibility of our reaction design, we probed a
model reaction between branched nitroenyne 1a and 3-
pyrrolyloxindole 2a under a bifunctional organocatalyst in
DCM. As shown in Table 1, with thioureas C1−C3 as the
catalysts, the reactions gave the desired aza-spirooxindole 3a in
63−97% yields and acceptable dr with poor enantioselectivity
at ambient temperature (Table 1, entries 1−3, respectively).
The screening of other chiral bifunctional organocatalysts
(C4−C6) was then carried out. To our delight, bifunctional
squaramide C6 was shown to be an effective catalyst, delivering
spiropiperidinyl oxindole 3a in 98% yield with 84% ee and
1.9:1 dr (entry 7). The subsequent solvent screening and the
addition of Ag2O as an additive did not improve the
stereoselectivity, albeit with comparable reactivity (entries 8−
15). Decreasing the reaction temperature to 0 °C had no effect
on the enantioselectivity. Finally, we focused our attention to
varying the type of sterically hindering groups on the
squaramide moiety. Gratifyingly, quinine-derived squaramide
C8 (entry 18), bearing a tert-butyl N substituent, resulted in
the best outcome, producing the desired 3a in a good yield of
85% with 3:1 dr and 92% ee.
With the optimized conditions in hand, the scope of various
branched nitroenynes (1) for the cyclization with 2a was first
examined (Scheme 2). The reactions of branched nitroenyne
2275
Org. Lett. 2021, 23, 2273−2278
Ni, Qijian
