Evolution
of a Synthetic
Strategy
toward
the Syntheses
of Bis-
tetrahydroisoquinoline
Alkaloids
Aurapat
Ngamnithiporn,
Eric R. Welin,
Gerit
Pototschnig,
and Brian
M. Stoltz
*
Cite This:
Acc. Chem.
Res.
2024,
57, 1870−1884
Read
Online
ACCESS
Metrics
& More
Article
Recommendations
CONSPECTUS:
The
bis-tetrahydroisoquinoline
(bis-THIQ)
natural
products
represent
a
medicinally
important
class
of
isoquinoline
alkaloids
that
exhibit
broad
biological
activities
with
particularly
potent
antitumor
properties,
as
exemplified
by
the
two
U.S.
FDA
approved
molecules
trabectidin
and
lurbinectedin.
Accordingly,
other
members
within
the
bis-THIQ
family
have
emerged
as
prime
targets
for
synthetic
chemists,
aiming
to
innovate
an
orthogonal
chemical
production
of
these
compounds.
With
the
ability
of
these
complementary
strategies
to
reliably
and
predictably
manipulate
molecular
structures
with
atomic
precision,
this
should
allow
the
preparation
of
synthetic
derivatives
not
existing
in
nature
as
new
drug
leads
in
the
development
of
novel
medicines
with
desired
biological
functions.
Beyond
the
biological
perspective,
bis-THIQ
natural
products
also
possess
intricate
and
unique
structures,
serving
as
a source
of
intellectual
stimulation
for
synthetic
organic
chemists.
Within
our
laboratory,
we
have
developed
an
integrated
program
that
combines
reaction
development
and
target-directed
synthesis,
leveraging
the
architecturally
complex
molecular
framework
of
bis-
THIQ
natural
products
as
a driving
force
for
the
advancement
of
novel
reaction
methodologies.
In
this
Account,
we
unveil
our
synthetic
efforts
in
a comprehensive
story,
describing
how
our
synthetic
strategy
toward
bis-THIQ
natural
products,
specifically
jorunnamycin
A
and
jorumycin,
has
evolved
over
the
course
of
our
studies
through
our
key
transformations
comprising
(a)
the
direct
functionalization
of
isoquinoline
N
-oxide
to
prepare
the
bis-isoquinoline
(bis-IQ)
intermediate,
(b)
the
diastereoselective
and
enantioselective
isoquinoline
hydrogenation
to
forge
the
pentacyclic
skeleton
of
the
natural
product,
and
(c)
the
late-stage
oxygenation
chemistry
to
adjust
the
oxidation
states
of
the
A-
and
E-rings.
First,
we
detail
our
plan
in
utilizing
the
aryne
annulation
strategy
to
prepare
isoquinoline
fragments
for
the
bis-THIQ
molecules.
Faced
with
unpromising
results
in
the
direct
C
−
H
functionalization
of
isoquinoline
N
-oxide,
we
lay
out
in
this
Account
our
rationale
behind
the
design
of
each
isoquinoline
coupling
partner
to
overcome
these
challenges.
Additionally,
we
reveal
the
inspiration
for
our
hydrogenation
system,
the
setup
of
our
pseudo-
high-throughput
screening,
and
the
extension
of
the
developed
hydrogenation
protocols
to
other
simplified
isoquinolines.
In
the
context
of
non-natural
bis-THIQ
molecules,
we
have
successfully
adapted
this
tandem
coupling/hydrogenation
approach
in
the
preparation
of
perfluorinated
bis-THIQs,
representing
the
first
set
of
electron-deficient
non-natural
analogues.
Finally,
we
include
our
unsuccessful
late-stage
oxygenation
attempts
prior
to
the
discovery
of
the
Pd-catalyzed
C
−
O
cross-coupling
reaction.
With
this
full
disclosure
of
the
chemistry
developed
for
the
syntheses
of
bis-THIQs,
we
hope
our
orthogonal
synthetic
tactics
will
provide
useful
information
and
serve
as
an
inspiration
for
the
future
development
of
bis-THIQ
pharmaceuticals.
■
KEY REFERENCES
•
Welin,
E.
R.;
Ngamnithiporn,
A.;
Klatte,
M.;
Lapointe,
G.;
Pototschnig,
G.
M.;
McDermott,
M.
S.
J.;
Conklin,
D.;
Gilmore,
C.
D.;
Tadross,
P.
M.;
Haley,
C.
K.;
Negoro,
K.;
Glibstrup,
E.;
Gru
̈
nanger,
C.
U.;
Allan,
K.
M.;
Virgil,
S.
C.;
Received:
May
2,
2024
Revised:
May
30,
2024
Accepted:
May
31,
2024
Published:
June
14,
2024
Article
pubs.acs.org/accounts
© 2024
The Authors.
Published
by
American
Chemical
Society
1870
https://doi.org/10.1021/acs.accounts.4c00262
Acc. Chem.
Res.
2024,
57, 1870
−
1884
This article is licensed under CC-BY-NC-ND 4.0
Slamon,
D.
J.;
Stoltz,
B.
M.
Concise
Total
Syntheses
of
(
−
)-Jorunnamycin
A
and
(
−
)-Jorumycin
Enabled
by
Asymmetric
Catalysis.
Science
2019
,
363
,
270
−
275.
1
This
study
discloses,
for
the
first
time,
a
summary
of
the
synthetic
strategy
toward
jorumycin,
a bis-tetrahydroiso-
quinoline
alkaloid,
via
the
key
sequential
cross-coupling/
hydrogenation
reactions.
•
Gilmore,
C.
D.;
Allan,
K.
M.;
Stoltz,
B.
M.
Orthogonal
Synthesis
of
Indolines
and
Isoquinolines
via
Aryne
Annulation.
J. Am.
Chem.
Soc
.
2008
,
130
,
1558
−
1559.
2
This
report
unveils
the
reactivity
of
benzyne
with
enamides
in
the
preparation
of
isoquinoline
derivatives.
•
Allan,
K.
M.;
Hong,
B.
D.;
Stoltz,
B.
M.
Expedient
Synthesis
of
3-Hydroxyisoquinolines
and
2-Hydroxy-1,4-
Naphthoquinones
via
One-Pot
Aryne
Acyl-Alkylation/
Condensation.
Org.
Biomol.
Chem
.
2009
,
7
,
4960
−
4964.
3
This
work
details
the
development
of
a
one-pot
procedure,
featuring
an
acyl-alkylation
of
arynes
followed
by
a condensation
with
ammonia
for
the
synthesis
of
3-
hydroxyisoquinolines.
•
Kim,
A.
N.;
Ngamnithiporn,
A.;
Welin,
E.
R.;
Daiger,
M.
T.;
Gru
̈
nanger,
C.
U.;
Bartberger,
M.
D.;
Virgil,
S.
C.;
Stoltz,
B.
M.
Iridium-Catalyzed
Enantioselective
and
Diastereoselective
Hydrogenation
of
1,3-Disubstituted
Isoquinolines.
ACS
Catal
.
2020
,
10
,
3241
−
3248.
4
An
extension
of
the
developed
enantioselective
and
diaster-
eoselective
hydrogenation
protocol
in
a
simplified
1,3-
disubstituted
isoquinoline
system.
•
Kim,
A.
N.;
Ngamnithiporn,
A.;
Bartberger,
M.
D.;
Stoltz,
B.
M.
Iridium-Catalyzed
Asymmetric
Trans
-Selective
Hydrogenation
of
1,3-Disubstituted
Isoquinolines.
Chem.
Sci
.
2022
,
13
,
3227
−
3232.
This
study
presents
another
extension
of
the
developed
hydrogenation
technology.
It
identifies
crucial
reaction
parameters
that
could
enable
trans
-selective
hydrogenation
of
1,3-
disubstituted
isoquinolines.
■
INTRODUCTION
The
tetrahydroisoquinoline
(THIQ)
alkaloid
family
of
natural
products
has
garnered
interest
from
the
synthetic
community
over
the
past
few
decades
due
to
their
complex
molecular
architectures
and
their
remarkable
biological
activities.
6
Of
this
family,
the
bis-THIQ
scaffold
represents
an
important
motif
that
demonstrates
therapeutic
potential
as
antitumor
antibiotics.
7
Exemplified
by
lurbinectedin
(Zepzelca,
1
),
the
most
recent
bis-
THIQ
alkaloid
granted
approval
by
the
U.S.
FDA
in
2020
for
the
treatment
of
metastatic
small-cell
lung
cancer,
and
the
notable
Et-743
(Yondelis,
2
),
approved
for
the
treatment
of
soft-tissue
sarcomas
in
2004,
this
class
of
natural
products
presents
a
vibrant
research
area,
to
both
the
synthetic
and
medicinal
chemistry
communities.
Despite
the
intensive
study
in
the
field
of
bis-THIQ
synthesis,
the
majority
of
synthetic
efforts
on
these
molecules
still
rely
on
the
use
of
the
biomimetic
approach
8
that
exploits
electrophilic
aromatic
substitution
(EAS)
chemistry,
such
as
Pictet
−
Spengler
or
Bischler
−
Napieralski
reactions.
This
chemistry
is
typically
employed
in
either
a
pseudosymmetrical
9
or
a
linear
retrosynthetic
disconnection
strategy
10
(Scheme
1).
In
the
former
approach,
the
A-
and
E-rings
are
first
forged
to
the
central
piperazine
C-ring,
followed
by
the
successive
Pictet
−
Spengler
cyclizations
to
generate
the
B-
and
D-rings.
In
a similar
fashion,
the
latter
joins
the
THIQ
AB-ring
to
an
E-ring
containing
fragment
via
a
condensation
(or
vice
versa,
coupling
of
the
THIQ
DE-ring
with
an
A-ring
containing
molecule).
A
subsequent
Pictet
−
Spengler
reaction
then
produces
the
[3.3.1]-diazabicycle
of
the
bis-THIQ
pentacyclic
core.
While
these
traditional
approaches
have
proven
to
be
effective,
lack
of
regioselectivity
is observed
in
some
cases,
11
and
the
requirement
of
electron-rich
substrates
to
undergo
cyclization
limits
the
possibility
of
producing
certain
non-natural
analogues,
including
electron-deficient
derivatives.
This
limitation
inspired
us
to
develop
an
orthogonal
synthetic
strategy
toward
bis-THIQ
alkaloids
that
represents
a complete
break
from
previous
studies.
Prior
to
our
synthetic
endeavors
toward
bis-THIQ
alkaloids,
our
laboratory
had
completed
the
total
syntheses
of
mono-
THIQs
(
−
)-lemonomycin
(
3
)
12
and
(
−
)-quinocarcin
(
4
).
13
By
utilizing
an
aryne
annulation/reduction
strategy,
the
THIQ
core
of
4
was
assembled,
directly
resulting
in
the
most
concise
total
synthesis
to
date
(Scheme
2).
Drawing
from
our
experience
in
this
area,
we
were
inspired
to
also
use
arynes
as
reactive
intermediates
to
prepare
bis-THIQ
molecules.
To
this
end,
we
became
intrigued
by
a more-recently
isolated
bis-THIQ
natural
product,
(
−
)-jorumycin
(
5
,
Figure
1).
Jorumycin
(
5
)
was
first
isolated
in
2000
from
the
mantle
and
mucus
of
the
Pacific
nudibranch
Jorunna
funebris
.
14
Preliminary
oncological
studies
have
shown
that
this
natural
product
exhibits
potent
anticancer
activity,
with
subnanomolar
potency
against
multiple
cancer
cell
lines,
such
as
P388
mouse
lymphoma,
A549
human
lung
carcinoma,
HT29
human
colon
carcinoma,
and
MEL28
human
melanoma
cells.
In
addition,
the
molecule
is also
reported
to
inhibit
the
growth
of
various
Gram-positive
bacteria
at
a low
concentration.
Scheme
1. Conventional
Approaches
to the Bis-THIQ
Core
Scheme
2. Retrosynthetic
Analysis
of Quinocarcin
(4)
Accounts
of Chemical
Research
pubs.acs.org/accounts
Article
https://doi.org/10.1021/acs.accounts.4c00262
Acc. Chem.
Res.
2024,
57, 1870
−
1884
1871
Although
jorumycin’s
anticancer
mechanism
of
action
has
not
been
specifically
studied,
we
believe
that
it
is
likely
similar
to
those
of
saframycins,
ranieramycins,
and
Et-743.
The
hemi-
aminal
functionality
in
jorumycin’s
C-ring
serves
as
a
latent
iminium
ion,
which
binds
covalently
to
DNA.
15
Furthermore,
previous
study
has
also
shown
that
double-strand
DNA
cleavage
was
observed
under
aerobic
conditions
where
superoxide
and
hydroxyl
radicals
were
detected.
16
Therefore,
its
broad
bioactivity
can
be
attributed
to
an
aerobic
oxidative
mechanism
as
well.
As
part
of
our
ongoing
research
interest
in
the
syntheses
of
THIQ
natural
products,
we
initiated
and
completed
our
synthetic
study
toward
jorumycin
(
5
)
via
a two-part
strategy,
consisting
of
a
cross-coupling
reaction
for
the
convergent
construction
of
the
requisite
carbon-based
skeleton
and
an
enantioselective
hydrogenation
to
install
all
of
the
key
stereochemistry.
1
While
a
summary
of
the
synthetic
strategy
was
previously
disclosed
in
a separate
report,
in
this
Account,
we
aim
to
provide
a comprehensive
and
detailed
story
of
how
our
strategy
to
access
bis-THIQ
natural
products
has
evolved
over
the
course
of
our
studies.
Specifically,
the
design
of
each
isoquinoline
coupling
partner
for
the
direct
functionalization
reaction,
additional
optimization
results
for
the
enantioselective
hydrogenation,
and
our
attempts
to
perform
late-stage
oxygen-
ation
via
other
methodologies
are
included.
Figure 1.
Structures
of
tetrahydroisoquinoline
(THIQ)
natural
products
and
pharmaceuticals.
Scheme
3. Retrosynthetic
Analysis
for the Total Syntheses
of (
−
)-Jorumycin
and Jorunnamycin
A
Accounts
of Chemical
Research
pubs.acs.org/accounts
Article
https://doi.org/10.1021/acs.accounts.4c00262
Acc. Chem.
Res.
2024,
57, 1870
−
1884
1872
■
RETROSYNTHETIC
ANALYSIS
Since
its
discovery
in
2000,
jorumycin
has
been
the
subject
of
a
number
of
semisynthetic
and
total
synthetic
endeavors.
17
All
total
syntheses
utilized
conventional
approaches,
as
outlined
in
Scheme
1.
In
2005,
Williams
and
co-workers
elected
to
use
a
linear
approach,
where
they
employed
a chiral
glycine
derivative
as
a
chiral
template
to
prepare
the
enantioenriched
building
block
for
both
the
eastern
and
western
halves
of
the
molecule.
17a
Zhu
and
Wu
later
disclosed
in
2009
the
exploitation
of
serine-
derived
aziridines
as
lynchpins
to
form
the
bis-THIQ
intermediate
with
AB-
and
DE-rings.
17b
Similar
to
William’s
and
Zhu’s
approaches,
Liu
17c
and
Chen
17d
and
co-workers
utilized
a series
of
Pictet
−
Spengler
reactions
to
prepare
the
bis-
THIQ
pentacyclic
scaffold.
Additionally,
an
alternative
chemo-
enzymatic
total
synthesis
of
jorumycin
was
achieved
with
the
use
of
engineered
enzymes
that
promote
Pictet
−
Spengler
cycliza-
tions.
17e
Although
these
are
powerful
and
efficient
approaches,
we
envisioned
that
an
orthogonal
route
that
is
not
reliant
on
electrophilic
aromatic
substitution
(EAS)
chemistry
(i.e.,
the
Pictet
−
Spengler
reaction)
would
offer
the
potential
for
innovation
and
allow
the
general
access
to
other
bis-THIQ
natural
products
and
previously
inaccessible
non-natural
analogues.
Our
first
generation
of
retrosynthetic
analysis
is
shown
in
Scheme
3.
First,
we
envisioned
accessing
jorumycin
from
jorunnamycin
A
(
6
),
another
bis-THIQ
molecule
that
contains
an
amino
nitrile
moiety
on
the
central
C-ring
instead
of
the
amino
alcohol.
Taken
from
the
intrinsic
symmetry
of
6
,
we
proposed
that
late-stage
oxygenation
of
the
highlighted
oxygens
could
greatly
simplify
the
preparation
of
arene
starting
materials.
To
access
the
key
pentacyclic
intermediate
7
,
we
planned
to
perform
an
enantioselective
hydrogenation
of
bis-isoquinoline
8
,
which
could
be
assembled
via
a cross-coupling
reaction
of
isoquinolines
9
and
10
.
Finally,
the
aryne
annulation
method-
ologies
developed
in
our
laboratory
2,3
could
facilitate
the
preparation
of
both
the
eastern
(
9
)
and
western
(
10
)
isoquinolines
via
a single
common
aryne
intermediate
11
.
■
SYNTHESIS:
FIRST
GENERATION
In
the
forward
sense,
we
first
prepared
ortho-silyl
triflate
12
,
which
serves
as
the
precursor
to
aryne
11
.
This
was
performed
using
a
three-step
sequence,
consisting
of
(1)
direct
C
−
H
borylation/oxidation/carbamate
formation
cascade,
(2)
direc-
ted
ortho-lithiation/silylation,
and
(3)
carbamate
cleavage/
triflation.
1
With
12
in
hand,
we
were
pleased
to
observe
that
regioselective
nucleophilic
addition
with
this
substrate
was
achieved
(Scheme
4).
This
could
be
attributed
to
the
methoxy
substituent
adjacent
to
the
C
−
C
triple
bond
in
aryne
intermediate
11
,
which
polarizes
and
electronically
differentiates
the
triple
bond.
18
When
enamide
13
or
cesium
enolate
of
methyl
acetoacetate
(
14
)
was
utilized
as
a nucleophile,
aryne
insertion/
annulation
cascade
proceeded
in
the
same
pot
to
directly
provide
isoquinolines
9
or
10
,
respectively.
Of
note,
we
found
the
N
-formyl
group
in
13
to
have
a negative
impact
on
the
scalability
and
efficiency
of
the
aryne
annulation
reaction.
19
Nonetheless,
we
were
able
to
obtain
enough
material
for
a
preliminary
investigation
of
the
next
step
by
performing
many
small-scale
reactions
in
parallel.
The
next
stage
of
our
synthesis
involves
the
cross-coupling
reaction
to
access
the
carbon
skeleton
of
the
bis-THIQ
core
(Scheme
5).
We
elected
to
utilize
the
direct
C
−
H
functionaliza-
tion
of
heterocyclic
N
-oxides
to
avoid
the
need
to
prepare
2-
pyridyl-type
organometallic
nucleophiles,
as
these
are
notorious
for
their
inherent
instability
and
rapid
protodemetalation.
20
As
a
result,
we
advanced
isoquinoline
9
to
N
-oxide
15
and
isoquinoline
10
to
triflate
16
.
21
Upon
subjecting
15
and
16
to
the
Pd-catalyzed
cross-coupling
reaction
developed
by
Fagnou
and
co-workers,
22
we
isolated
cross-coupled
product
17
along
with
isoquinoline
9
,
which
is
the
reduced
form
of
15
(Scheme
Scheme
4. Syntheses
of Isoquinolines
9 and 10 via Aryne Annulation
Strategy
Scheme
5. Initial
N
-Oxide
C
−
H Functionalization
Accounts
of Chemical
Research
pubs.acs.org/accounts
Article
https://doi.org/10.1021/acs.accounts.4c00262
Acc. Chem.
Res.
2024,
57, 1870
−
1884
1873
5a).
Attempts
to
further
improve
the
yield
of
17
were
unsuccessful.
Over
the
course
of
our
optimization,
we
found
that
the
high
Pd
loadings
(50
mol
%)
and
very
high
temperature
(180
°
C)
were
necessary
to
enable
the
product
formation.
It
is
also
worth
noting
that
the
excess
N
-oxide
starting
material
(
15
)
was
fully
reduced
to
isoquinoline
9
.
Having
encountered
this
difficulty,
we
turned
to
a mechanistic
study
of
the
N
-oxide
C
−
H
functionalization
reported
by
Hartwig
and
co-workers.
23
Since
this
report
suggests
C
−
H
activation
as
a rate-determining
step,
we
analyzed
our
transition
state
and
found
that
there
is likely
a significant
steric
interaction
in
our
system
(Scheme
5b).
Drawing
from
this
hypothesis,
we
redesigned
the
N
-oxide
coupling
partner.
By
translocating
the
hindered
methoxy
substituent
to
the
C5
position
(cf.
15
vs
18
),
we
believed
this
would
alleviate
the
clash
and
improve
the
efficiency
of
the
cross-coupling
reaction,
yet
not
interfere
with
our
synthetic
plan
downstream.
■
SYNTHESIS:
REVISED
STRATEGY
We
synthesized
the
new
N
-oxide
(
18
)
via
a 3-step
protocol
from
aryl
bromide
19
(Scheme
6).
This
involves
Sonogashira
coupling
with
an
in situ
oxime
formation
to
produce
20
,
Ag-
catalyzed
6-
endo
-dig
cyclization
of
20
to
generate
isoquinoline
N
-oxide
21
,
and
silyl
deprotection/Jones
oxidation/Fisher
esterification
to
provide
18
.
1
This
sequence
was
found
to
be
highly
scalable
with
little
optimization
required,
allowing
us
to
prepare
>10
g of
intermediate
21
in
a single
pass.
With
18
in
hand,
we
subjected
the
compound
together
with
isoquinoline
triflate
16
to
the
Fagnou
cross-coupling
conditions
(Scheme
7,
top).
Unfortunately,
we
still
could
not
significantly
improve
the
yield
of
bis-IQ
22
even
with
high
catalyst
loading.
Additionally,
the
excess
N
-oxide
starting
material
was
completely
reduced
at
the
end
of
the
reaction.
From
this
observation,
we
posited
that
the
reduction
of
N
-oxide
out-
competed
the
desired
coupling,
and
reducing
the
electrophilicity
of
the
N
-oxide
might
be
beneficial
to
slow
down
this
process.
Therefore,
we
investigated
the
possibility
of
utilizing
inter-
mediate
21
,
which
contains
an
electron-donating
silyloxymethyl
functionality,
as
an
alternative
coupling
partner
(Scheme
7,
bottom).
By
exploiting
isoquinoline
N
-oxide
21
as
a coupling
partner,
we
were
pleased
to
observe
that
the
coupling
proceeded
smoothly,
providing
bis-IQ
23
in
an
excellent
yield.
Moreover,
upon
further
optimization,
we
found
that
the
catalyst
loading
Scheme
6. Synthesis
of Isoquinoline
N
-Oxide
18
Scheme
7. Final
N
-Oxide
C
−
H Functionalization
Conditions
Accounts
of Chemical
Research
pubs.acs.org/accounts
Article
https://doi.org/10.1021/acs.accounts.4c00262
Acc. Chem.
Res.
2024,
57, 1870
−
1884
1874
Scheme
8. Probing
the Possibility
of
N
-Oxide
C3-Arylation
Scheme
9. Double
Boekelheide
Rearrangement
Scheme
10. Elaboration
to Hydrogenation
Precursor
32
Accounts
of Chemical
Research
pubs.acs.org/accounts
Article
https://doi.org/10.1021/acs.accounts.4c00262
Acc. Chem.
Res.
2024,
57, 1870
−
1884
1875
could
be
reduced
to
20
mol
%
and
the
excess
N
-oxide
could
be
quantitatively
recovered
at
the
end
of
the
reaction.
Additional
investigation
to
probe
the
possibility
of
swapping
the
nucleophilic/electrophilic
components
was
conducted,
but
the
reaction
favored
the
competitive
C(sp
3
)-arylation
and
produced
the
undesired
bis-IQ
26
as
the
major
product
(Scheme
8).
This
reactivity
has
been
previously
reported
by
Fagnou
and
co-
workers
24
and
thus
not
pursued
further
due
to
the
unpromising
levels
of
regioselectivity.
In
order
to
access
the
hydrogenation
precursor,
we
still
needed
to
adjust
the
oxidation
levels
in
bis-IQ
23
.
We
specifically
envisioned
exploiting
a
double
Boekelheide
rearrangement
as
a
key
transformation
to
transmute
the
N
-
oxides
to
the
benzylic
acetates
(Scheme
9).
With
this
plan,
we
first
performed
an
N
-oxidation
of
23
via
Sharpless’
protocol,
25
utilizing
catalytic
methyl
trioxorhenium,
to
deliver
the
desired
bis-
N
-oxide
(
27
)
in
an
excellent
98%
yield.
Unfortunately,
the
reaction
profile
of
the
subsequent
double
Boekelheide
rearrangement
was
complicated,
and
we
only
observed
34%
optimized
yield
of
triacetoxy
intermediate
28
that
contains
a
unique
acylal
functionality
that
implies
the
TBS
deprotection
is
occurring
under
the
reaction
conditions.
Hydrolysis
of
28
was
performed
in
the
same
pot,
affording
aldehyde
29
in
quantitative
yield.
During
our
optimization
of
the
double
Boekelheide
rearrange-
ment,
we
found
that
the
rearrangement
of
the
B-ring
azine
was
facile
while
that
of
the
D-ring
was
sluggish,
likely
due
to
the
buildup
of
A
1,3
-like
strain
in
the
transition
state
of
the
rearrangement.
High
temperature
(130
°
C)
as
shown
in
Scheme
9
was
required
as
a
consequence
of
the
slow
D-ring
rearrangement.
In
fact,
when
the
reaction
was
performed
at
room
temperature,
we
observed
a much
cleaner
reaction
profile,
affording
the
monorearranged
product
(
30
)
exclusively.
As
a
result,
we
investigated
an
alternative
pathway
based
on
this
observed
reactivity
and
found
that
differentially
protected
diol
30
could
be
obtained.
Importantly,
the
N
-oxidation/
N
-oxide
rearrangement
could
be
conducted
in
one-pot
by
simply
adding
acetic
anhydride
upon
completion
of
the
oxidation,
affording
30
in
62%
yield
(Scheme
10).
We
successfully
advanced
this
intermediate
to
hydrogenation
precursor
32
through
two
additional
steps,
which
include
(1)
the
N
−
O
bond
cleavage
and
(2)
the
N
-oxyl-mediated
oxidation.
Additional
investiga-
tions
to
preinstall
the
hydroxyl
group
appended
on
the
B-ring
prior
to
the
cross-coupling
were
unsuccessful.
26
■
ENANTIOSELECTIVE
HYDROGENATION
We
planned
to
prepare
the
pentacyclic
scaffold
of
bis-THIQ
molecules
by
utilizing
sequential
stereoselective
hydrogenation.
Through
our
proposed
stereochemical
model
(Scheme
11),
we
expected
a
self-reinforcing
diastereoselective
hydrogenation.
Specifically,
we
envisioned
obtaining
the
all-syn
hydrogenated
product
regardless
of
the
order
of
the
reduction
sequence.
Since
there
are
two
discrete
isoquinoline
units,
the
initial
reduction
could
occur
on
either
the
B-
or
D-rings.
If
the
B-ring
is
reduced
first,
the
newly
formed
stereocenters
should
impart
concavity
on
the
system
and
promote
the
subsequent
D-ring
reduction
from
the
convex
face
to
produce
bis-THIQ
35
.
Alternatively,
when
the
D-ring
is
reduced
first,
we
also
expected
the
same
product
(
35
)
since
the
first
reduction
would
still
generate
similar
conformation
conducive
to
reduction
with
the
preferred
diastereoselectivity.
With
this
stereochemical
rationale,
we
next
surveyed
the
literature
for
enantioselective
hydrogenation
of
isoquinolines,
and
we
were
surprised
by
the
lack
of
precedents.
As
noted
earlier
by
Zhou
and
co-workers
in
their
comprehensive
review,
27
isoquinolines
are
among
the
most
challenging
heterocycles
for
asymmetric
hydrogenation
reactions.
This
could
be
attributed
to
the
highly
basic
piperidine-type
products
which
can
lead
to
catalyst
poisoning/deactivation.
Nonetheless,
at
the
time
when
we
initiated
our
hydrogenation
studies,
there
were
four
available
systems
for
isoquinolines
asymmetric
hydrogenation,
28
of
which
only
one
showed
compatibility
with
1,3-disubstitution
pat-
terns.
28c
By
employing
preformed
dinuclear-halide-bridged
iridium(III)
complexes,
Mashima
and
co-workers
achieved
asymmetric
hydrogenation
of
1,3-disubstituted
isoquinolinium
salts.
We,
however,
believed
that
this
catalytic
system
might
not
be
suitable
for
our
optimization
study
because
(a)
there
is
no
example
that
shows
tolerance
of
additional
Lewis
basic
functionalities,
such
as
alcohols
or
heteroaryl-substituted
isoquinolines,
and
(b)
the
requirement
to
individually
prepare
each
chiral
catalyst
complex
can
limit
our
optimization
throughput.
Accordingly,
we
turned
our
attention
to
another
Ir-system
developed
by
scientists
at
Ciba-Geigy
(now
Syngenta)
for
enantioselective
ether-directed
imine
hydrogenation
(Scheme
12).
29
In
this
case,
the
chiral
catalyst
can
be
generated
in situ
,
allowing
us
to
explore
a wide
variety
of
chiral
ligand
scaffolds
via
a
“pseudo-high-throughput”
approach
(Scheme
13a).
Of
63
chiral
ligands
30
from
the
bisphosphine
and
P,N-
ligand
families,
we
found
that
approximately
one-third
provided
at
least
a 20%
yield
of
partially
hydrogenated
mono-THIQ
33
,
all
Scheme
11. Self-Reinforcing
Diastereoselectivity
in the Proposed
Hydrogenation
Reaction
Accounts
of Chemical
Research
pubs.acs.org/accounts
Article
https://doi.org/10.1021/acs.accounts.4c00262
Acc. Chem.
Res.
2024,
57, 1870
−
1884
1876