Organic Letters
Letter
a
greatly, providing product 3an in trace amount. The apparent
electron effect of a substituent in the 3-position was also
observed in substrate 1ao, in which isoquinolones 3ao was
obtained in 22% yield. Moreover, 4-methoxy-2-methylbenza-
mide 1ap, 2,4-dimethylbenzamide 1aq, and 2-naphthamide 1ar
were compatible in this transformation, and moderate yields of
corresponding products 3ap, 3aq, and 3ar were obtained.
Beyond the N-methoxy group, a variety of N-protecting
groups were investigated and electronic and steric effects were
observed (Scheme 3). For example, substrate 1ba containing
Scheme 4. Substrates Scope of Epoxides
a
Scheme 3. Substrates Scope of N-Protecting Groups
a
b
Isolated yields. Yields of corresponding isochromans.
extensive pharmacological activities, leading to 3cn in 58%
yield.
The practicability of this protocol was also demonstrated by
the total syntheses of rupreschstyril,21 siamine,22 and cassiarin
A;23 which, however, were previously synthesized in slightly
lower overall yields and many steps. Removal of the methoxy
group and subsequent deprotection of the MOM group in
isoquinolin-1-one 3af′ provided rupreschstyril in 83% yield.
Similarly, siamine (2), another alkaloid of Cassia siamea, was
obtained in 76% yield from isoquinolin-1-one 3ca. As for
cassiarins A (3), treatment of 3ca with sodium hydride at 120
°C removed the methoxy group, followed by treatment with N-
phenyl-bis(trifluoromethanesulfonimide) providing triflate 4 in
73% yield in one pot, which subsequently coupled with the in
situ generated propyne to provide alkyne 5 in 87% yield. Then,
demethylation of 5 with BBr3 and spontaneous 6-endo-dig
cyclization gave cassiarin A (3) in a total of 4 steps and 37%
overall yield from 1aa (Scheme 5).
a
b
Isolated yields. Corresponding isochromans were not observed.
an N-ethoxy group afforded a better yield compared with the
N-methoxy substituent by increasing the nucleophilicity of the
nitrogen atom. The reaction of linear N-butoxy substrate 1bd
offered the desired isoquinolone in 81% yield, while substrate
1bb bearing an isopropoxy gave 3bb in 56% yield and tert-
butoxy substituted substrate did not provide the desired
product. It is pleasing that the reaction of N-isobutoxy
substituted substrate 1bc afforded 3bc in 92% yield, and this
N-protecting group was also employed in 4-methoxy/methyl
benzamides 1bf and 1bg, and synthetically useful yields were
obtained. When the N-protecting group was benzyloxy,
isoquinolone 3bh was obtained in 47% yield; however, N-
phenoxy substrate 1bi did not offer the desired product.
Finally, the scope of epoxides was extensively tested
(Scheme 4). Alkylation with simple propylene oxide and
other 2-alkyloxiranes gave N-methoxyisoquinolones 3ca−3cc
in 80−85% isolated yields. Actually, phenyl and protected
hydroxyl groups (3 cd−3ch, 3cm) gave inferior yields (61−
75%) compared with 2-alkyloxiranes, presumably due to the
additional coordinating oxygen atom or π-electrons of the
phenyl of oxiranes might inhibit the β-hydrogen elimination
process for coordinating to the palladium catalyst. Surprisingly,
2-ylmethanol oxirane was also tolerant in this reaction,
affording 3ci in 46% yield. Further exploration demonstrated
that α-ester and vinyl moieties (1cj, 1ck) were not compatible
with this protocol. Besides, isoindoline-1,3-dione worked well
in this transformation, and a moderate yield of corresponding
product 3cl was obtained.
To gain insights into the mechanism, coupling of 1al with 2a
in the absence of 4 Å molecular sieve and TEA has been
conducted. As a result, isochroman 3as, 3-hydroxy-3,4-
dihydroisoquinolin-1-one 3at, and isoquinolone 3al were
obtained in 24%, 13%, and 21% yields, respectively. Notably,
3at could fully convert to isoquinolone 3al in the presence of 4
Å molecular sieve at 100 °C in HFIP. When the reaction was
performed under an O2 or Ar atmosphere, 3al was obtained
respectively in 69% and 14% yields, indicating the critical role
of oxygen in this reaction. According to our preliminarily
mechanistic experiments and the literature precedent,24
a
plausible mechanism is proposed (Scheme 6). Cyclometalation
of benzamide with palladium(II) generates a palladacyclic
intermediate A,25 which reacted with epoxide via an SN2
nucleophilic ring-opening process generating Pd-alkoxide
species B.17 Intermediate B undergoes β-hydrogen elimination
to form the ketone carbonyl product D and Pd-hydride
species,20a,26 followed by the nucleophilic addition of nitrogen
Moreover, this methodology was successfully applied to the
late stage modification of estrone, a compound possessing
865
Org. Lett. 2021, 23, 863−868