Dramatic Effect of γ-Heteroatom Dienolate Substituents on Counterion Assisted Asymmetric Anionic Amino-Cope Reaction Cascades

Article abstract of DOI:10.1021/jacs.1c00745

We report a dramatic effect on product outcomes of the lithium ion enabled amino-Cope-like anionic asymmetric cascade when different γ-dienolate heteroatom substituents are employed. For dienolates with azide, thiomethyl, and trifluoromethylthiol substituents, a Mannich/amino-Cope/cyclization cascade ensues to form chiral cyclohexenone products with two new stereocenters in an anti-relationship. For fluoride-substituted nucleophiles, a Mannich/amino-Cope cascade proceeds to afford chiral acyclic products with two new stereocenters in a syn-relationship. Bromide- and chloride-substituted nucleophiles appear to proceed via the same pathway as the fluoride albeit with the added twist of a 3-exo-trig cyclization to yield chiral cyclopropane products with three stereocenters. When this same class of nucleophiles is substituted with a γ-nitro group, the Mannich-initiated cascade is now diverted to a β-lactam product instead of the amino-Cope pathway. These anionic asymmetric cascades are solvent- and counterion-dependent, with a lithium counterion being essential in combination with etheral solvents such as MTBE and CPME. By altering the geometry of the imine double bond from E to Z, the configurations at the R1 and X stereocenters are flipped. Mechanistic, computational, substituent, and counterion studies suggest that these cascades proceed via a common Mannich-product intermediate, which then proceeds via either a chair (X = N3, SMe, or SCF3) or boat-like (X = F, Cl, or Br) transition state to afford amino-Cope-like products or β-lactam in the case of X = NO2.

Full text of DOI:10.1021/jacs.1c00745

pubs.acs.org/JACS
Article
Dramatic Effect of γ‑Heteroatom Dienolate Substituents on
Counterion Assisted Asymmetric Anionic Amino-Cope Reaction
Cascades
Pradipta Das, Michael D. Delost, Munaum H. Qureshi, Jianhua Bao, Jason S. Fell, Kendall N. Houk,*
and Jon T. Njardarson*
Cite This: J. Am. Chem. Soc. 2021, 143, 5793−5804
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: We report a dramatic effect on product outcomes of the lithium ion
enabled amino-Cope-like anionic asymmetric cascade when different γ-dienolate
heteroatom substituents are employed. For dienolates with azide, thiomethyl, and
trifluoromethylthiol substituents, a Mannich/amino-Cope/cyclization cascade ensues to
form chiral cyclohexenone products with two new stereocenters in an anti-relationship.
For fluoride-substituted nucleophiles, a Mannich/amino-Cope cascade proceeds to afford
chiral acyclic products with two new stereocenters in a syn-relationship. Bromide- and
chloride-substituted nucleophiles appear to proceed via the same pathway as the fluoride
albeit with the added twist of a 3-exo-trig cyclization to yield chiral cyclopropane products
with three stereocenters. When this same class of nucleophiles is substituted with a γ-
nitro group, the Mannich-initiated cascade is now diverted to a β-lactam product instead
of the amino-Cope pathway. These anionic asymmetric cascades are solvent- and
counterion-dependent, with a lithium counterion being essential in combination with
etheral solvents such as MTBE and CPME. By altering the geometry of the imine double bond from E to Z, the configurations at the
R1 and X stereocenters are flipped. Mechanistic, computational, substituent, and counterion studies suggest that these cascades
proceed via a common Mannich-product intermediate, which then proceeds via either a chair (X = N3, SMe, or SCF3) or boat-like
(X = F, Cl, or Br) transition state to afford amino-Cope-like products or β-lactam in the case of X = NO2.
substituted with electron withdrawing groups such as esters
and amides in the 4-position succumbs to the unwanted retro-
aldol pathway.8 In the context of realizing asymmetric
rearrangement goals with such substrates, Schneider,9−11
Nakai,12 Black,13 and Davies14 have demonstrated that
protection of the troublesome 3-hydroxyl with a silyl or
benzoate group allows the oxy-Cope rearrangement to proceed
albeit at very high temperatures in addition to two steps
(protection/deprotection) being added to the overall synthetic
sequence.
In the present study, we report on our investigations
employing dienolates substituted with a heteroatom at the γ-
position. We postulated that, in addition to further defining the
scope and behavior of these new Cope substrates, the electron
donating and withdrawing properties of the heteroatom X-
substituents would provide invaluable insights into the
INTRODUCTION
■
We recently reported a new asymmetric lithium-ion-mediated
amino-Cope-type rearrangement for assembling chiral cyclic
and acyclic products from simple building blocks.1 This work
built upon mostly overlooked earlier explorations,2−6 whose
aim was to establish an anion-accelerated amino-Cope version
of Evan’s seminal anion-accelerated oxy-Cope rearrangement.7
These studies, as well as our own, revealed that the anionic
amino-Cope version proceeds via a nonconcerted reaction
pathway wherein the anionic Mannich adduct first fragments
(retro-Mannich) followed by a recombination affording formal
amino-Cope ([3,3]) or [1,3]-rearrangement products. Chal-
lenges associated with controlling the recombination step while
ensuring that all retro-Mannich fragmentations proceeded
productively are most likely the reasons these contributions did
not become widely recognized. We postulated that these
challenges could be addressed by in situ chelation of the two
fragments with a metal counterion. We demonstrated the
feasibility of this hypothesis with the productive marriage
between a sulfinamide nitrogen anion, dienolate, and a lithium
counterion resulting in formation of amino-Cope products
with high selectivity and broad substrate scope. Our
observations are particularly noteworthy when contrasted
with the corresponding oxy-anion version, which when
Received: January 20, 2021
Published: April 12, 2021
https://doi.org/10.1021/jacs.1c00745
J. Am. Chem. Soc. 2021, 143, 5793−5804
© 2021 American Chemical Society
5793

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
mechanism of the anionic cascade (Mannich/Cope vs Michael
addition). The γ-position was chosen over the α-position, as it
was postulated that (a) it would result in a less sterically
congested Mannich addition step, (b) its inductive effects on
the dienolate and in turn the strength of the counterion
coordination would be expected to be on par with the α-
position, and (c) the product would contain an additional high
value second sp3-stereocenter whose stereochemistry would
afford significant clues on the transition state. Based on our
earlier insights and efforts, our expectations with respect to the
mechanism and anticipated stereochemical outcomes as a
result of introducing γ-dienolate substituents are presented in
Figure 1. Following a reversible syn-selective Mannich addition
Hydrogen (X = H). Shown in Scheme 1 are the amino-
Cope results for 3-phenyl methyl butenoate (1) using both
Scheme 1. Anion-Accelerated Amino-Cope Rearrangement
Results for 3-Phenyl Methyl Butenoate
enantiomers (2 and 5) of the sulfonamide auxiliary for imine
electrophiles with E-double bonds. In both cases, the anionic
Mannich-amino-Cope-cyclization cascades afford the expected
chiral cyclohexenone products (3 and 6) in excellent yields as
single stereoisomers but opposite stereochemistry. Upon
oxidation of the product auxiliaries to their corresponding t-
butyl sulfonamide (Bus) groups, 4 and 7 are formed, which
display identical NMRs but opposite optical rotations,
suggesting that this anion-accelerated amino-Cope rearrange-
ment can be used to access either of the enantiomeric series (4
and 7, see Supporting Information Figure S1). The dienolate
of 1 also successfully engages an imine with a Z-double bond
(8), albeit in a bit lower yields, to deliver 9 which following
sulfonamide oxidation is shown to be identical in all respects to
7.
Figure 1. Proposed anionic amino-Cope studies.
With the viability of the new 3-aryl-substituted nucleophile
established, we turned our attention to investigating 3-aryl
methyl butenoates containing heteroatom substituents in the γ-
position.
step between an E-conjugated Ellman-imine and a dienolate,
resulting from deprotonation of γ-heteroatom-substituted
propionates, a lithium guided anion-accelerated amino-Cope
rearrangement would then proceed followed by a rapid
enamide cyclization to form a cyclohexenone product with
two new stereocenters.15 We predicted, based on our earlier
studies, that the two new stereocenters would be anti to each
other (R, S configured at R1 and X, respectively, for an S-
sulfinamide) in the expected product. Furthermore, it would
then be expected that the three other products shown in the
highlighted box could be specifically obtained by employing
either E- or Z-conjugated imines in combination with the
appropriately configured sulfinamide chiral auxiliary (R or S).
Azide (X = N3). Given the important role nitrogen atoms
play as essential components of the majority of approved
pharmaceuticals16,17 and countless natural product structures,
we initiated our heteroatom substituent studies by identifying a
suitable nitrogen atom functional group. It quickly became
evident to us that azides represented an excellent nitrogen
atom representative for our studies.18 Azides are not only one
of the smallest nitrogen atom substituents as well as a versatile
functional group handle for further manipulations, but their
unique structure helps avoid any unwanted nucleophilic or
basic behavior of many typical nitrogen atom substituents.
Furthermore, for our purposes, the γ-substituted azide 3-aryl
butenoates that we required for our studies would be easy to
synthesize and stable as a result of electronic deactivation by
enoate and a relatively high carbon atom count.
We were delighted to learn that when azide-containing
enoate 10 was treated with lithium bis(silyltrimethyl)amide
(LiHMDS) in the presence of chiral sulfiname imine 11 the
expected anionic amino-Cope cascade ensued to afford a chiral
cyclohexenone product 12 in very good yield as a single
stereoisomer. For this cascade, methyl t-butyl ether (MTBE)
was shown to afford the product in higher yield compared to
tetrahydrofuran (THF). Proving the stereochemistry at the
RESULTS AND DISCUSSION
■
For our debut amino-Cope studies, we utilized the esters of
tiglic acid and 3,3-dimethyl acrylic acid and their trifluor-
omethyl variants as dienolate nucleophiles. For the present
study, we opted to push the limits further in terms of the size
of the nucleophile by employing 3-aryl methyl butenoates,
which we postulated would provide the added benefits of
providing electronic tuning by altering the nature of the 4-aryl
substituent while affording more diverse and potentially higher
value products.
5794
https://doi.org/10.1021/jacs.1c00745
J. Am. Chem. Soc. 2021, 143, 5793−5804

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
newly formed stereocenter proved to be challenging using
NMR, as coupling constants between the C−H’s of the
adjacent stereocenters were not deemed reliable. This
assignment was confidently confirmed following treatment of
the sulfinamide group with an oxidant (m-CPBA) to deliver a
crystalline sulfonamide product, thereby allowing single crystal
X-ray analysis and confirmation of an anti-relationship between
the two sp3-stereocenters. Absolute configuration was assumed
based on crystal structures lacking the azide group obtained in
our earlier studies. Crystal structure reveals that the azide and
the adjacent aryl group both adopt axial positions, thus
accounting for the surprising coupling constants observed in
the proton NMR spectra.
Scheme 2. γ-Azide-Substituted Butenoates Are Successful
Nucleophile Partners for the Anionic Amino-Cope Cascade
We have evaluated this new Mannich-amino-Cope-cycliza-
tion cascade for a series of imines and dienolates (Table 1).
Scheme 3. Z-Imines Engage Azide Nucleophile
Table 1. Amino-Cope Cascade Scope with Azide
Nucleophiles
with our optimized amino-Cope cascade reaction conditions
followed by oxidation of the auxiliary to a t-butyl sulfonamide
(Bus) group. These studies revealed that our structural
assignment of 32 was indeed correct and that the minor
product (16) was the other anti-product matching the oxidized
product originating from an amino-Cope cascade originating
from E-imine 2.
We next turned our attention to evaluating the compatibility
of sulfur substituents for the anionic cascade. Following the
four key elements of life (C, H, O, and N), which account for
the structural bulk and architectural frameworks of US FDA
approved pharmaceuticals, it is sulfur (S) which reigns
supreme as the fifth most commonly occurring element in
drugs.19
Thiomethyl (X = SCH3). We were pleased to learn that,
when we employed dienolate nucleophiles substituted with a γ-
thiomethyl group (36), the same novel Mannich-amino-Cope-
cyclization cascade we had observed for azides ensued (Table
2). Yields are very good for aromatic substituents (R and Ar),
and as with other series, the yield drops a bit when alkyl (R)
substituents are used. Using cyclopentyl methyl ether (CPME)
as the solvent, we observed improved yields compared to THF
and MTBE. Unlike the azide series, the thiomethyl substituent
affords two amino-Cope products, which we have named cyclic
(C, cyclization following the amino-Cope step) and acyclic
(A). The structures of these products do not only differ in their
enoate double bond geometry, with a Z-enoate for the cyclic
Yields are generally very good with higher yields observed for
imines with aromatic compared to alkyl substituents, which is
consistent with observations we made in our first report.
Interestingly, for alkyl substrates 22 (n-butyl) and 23
(cyclohexyl), an acyclic amino-Cope product is obtained
with an E-configured double bond and syn-configuration
between the newly formed stereocenters.
We next evaluated if a Z-imine would be compatible with
our asymmetric cascade (Scheme 3). Treatment of Z-imine 8
with azide nucleophile 10 revealed that a Z-imine is indeed
compatible, affording two cyclized products 32 and 16 in a 3:1
ratio and good yield, albeit lower yield than the corresponding
E-imine. We predicted that cyclohexenone 32, with opposite
stereochemistry of the newly formed stereocenters to that
obtained from E-imines, but leaving no stones unturned, we
decided to confirm this assignment unambiguously. This was
accomplished by treating E-imines 5 and 2, which only differ in
the absolute stereochemistry of the sulfinamide chiral auxiliary,
5795
https://doi.org/10.1021/jacs.1c00745
J. Am. Chem. Soc. 2021, 143, 5793−5804

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Table 2. Amino-Cope Cascade with Thiomethyl
Nucleophiles
Scheme 4. Z-Imine Reacts with Thiomethyl Nucleophile
This powerful feature of these new asymmetric anionic
cascades enables access to either anti-configuration of the
product from a single chiral auxiliary by simple choice of E- or
Z-imine reaction partners.
Following the success of the thiomethyl nucelophiles, we
next decided to further probe the impact of electronics of the
sulfur substituent by replacing the methyl group with a
trifluoromethyl group, which would minimally alter the size of
the nucleophile while significantly impacting dienolate electron
density. It is worth mentioning that the trifluoromethyl thiol
group (SCF3) possesses attractive physicochemical properties
such as high lipophilicity (Hansch parameter Π = 1.44).20 As
such, this moiety continues to receive interest in disciplines
such as medicinal chemistry and agrochemistry.21 Therefore,
novel trifluoromethylthiolation strategies, in particular asym-
metric, continue to be of interest.
Trifluoromethylthiol (X = SCF3). The results of our
fluorine inductive effect studies with trifluoromethylthiol-
substituted nucleophile 48 are reported in Table 3. First, this
Table 3. Trifluoromethylthiol Nucleophiles
product and an E-enoate for the acyclic product, but the newly
formed stereocenters have a syn-relationship in case of acyclic
product compared to the anti-relationship for the cyclic
product. In all cases, the cyclic product is the major product,
and overwhelmingly so for most substrates. We were able to
convert one of the products (37) into imine enol ether product
46 under mild O-alkylation reaction conditions. Most
gratifyingly, enol ether 46 was crystalline, which allowed us
to secure a single X-ray crystal structure of it and therefore to
unambiguously determine the relative and absolute stereo-
chemistry of the cyclic product series and confirm that its
structure matched our observation for the azide nucleophiles.
It is noteworthy to mention that the starting double bond
geometry of the enoate does not seem to influence the
outcome as azide nucleophile 10 (Z-double bond) and
thiomethyl nucleophile 36 (E-double bond) both result in
cyclic products with the same relative and absolute
configuration.
We next evaluated the compatibility of Z-imines as partners
in the anionic cascade with a thiomethyl nucleophile (Scheme
4). Reaction between Z-imine 8 and thiomethyl-substituted
dienolate shown under our standard reaction conditions
afforded two cyclized products 47 and 37 in a 2.5:1 ratio,
albeit in lower yield than the corresponding E-imine. This
result is similar to what we observed with the azide substitution
(Scheme 3, top part). Substrate 47 also has the anti-
stereochemistry at the aryl- and thiomethyl stereochemistry
but opposite configuration to those originating from E-imines.
nucleophile is indeed compatible with the Mannich-amino-
Cope-cyclization cascade affording expected cyclized products
in fair to good yields and as single stereoisomers with the anti-
relationship between newly formed stereocenters as observed
for azides and thiomethyl-substituted nucleophiles. Interest-
ingly, and in a slight contrast to the thiomethyl nucleophiles, all
of the products with an aromatic imine substituent proceeded
to cyclize with no observed acyclic products. The exception to
5796
https://doi.org/10.1021/jacs.1c00745
J. Am. Chem. Soc. 2021, 143, 5793−5804

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
this was when alkyl substituents (cyclohexyl and cyclopropyl,
55 and 56) were employed, where we observed the acyclic
product as the major component for the cyclohexyl substituent
or as the sole product in the case of cyclopropyl.
Table 4. Amino-Cope Cascade with Fluorinated
Nucleophile
It is fair to say that fluorine has taken the pharmaceutical22
and agrochemical industries by storm, as is evident from the
large number and continued increase of approved small
molecules containing fluorine atoms in the last two decades.
This is not entirely surprising given fluorine’s small size (not
much larger than hydrogen) while packing a punch with
respect to electronic impact, strong C−F bond, and a
multitude of other attractive attributes.23,24 We decided to
next explore the behavior and compatibility of fluorine-
substituted nucleophiles.
Fluoro (X = F). When we subjected imine 57 to the
dienolate nucleophile resulting from deprotonation of γ-fluoro-
substituted enoate 58, a new product (59, Scheme 5) was
Scheme 5. Fluorine-Substituted Nucleophile Participates in
the Amino-Cope Cascade
we conducted,25 wherein it was revealed that as of recently
there were more approved drugs containing a chlorine atom
than fluorinated ones, with only a handful being decorated
with either a bromine or iodine atom.
Chloro and Bromo (X = Cl and Br). We proceeded to
evaluate the nucleophilic addition behavior of the dienolate
originating from γ-chloro enoate 69 (Scheme 6). Treatment of
formed. Initial NMR analysis of compound 59 suggested an
acyclic product similar to what we had observed in the
thiomethyl series (Table 2). Fortunately, we were able to grow
crystals of this new product and in turn secure a single X-ray
crystal structure analysis, which revealed not only that the
product was acyclic with an E-enoate double bond but that the
newly formed stereocenters had a syn-relationship (not anti as
in the azide, thiomethyl, and trifluoromethylthio series). More
importantly, with the chiral auxiliary still in place, this crystal
structure allowed us to unequivocally confirm both the
absolute and relative stereochemistry of 59.
Scheme 6. Chlorine-Substituted Nucleophile Participates in
the Amino-Cope Cascade, Forming Cyclopropane Products
We have proceeded to further evaluate how other imine
electrophiles behave when treated with fluorine-substituted
nucleophile 58 (Table 4). The imines we have evaluated to
date all proceeded to undergo a Mannich-amino-Cope cascade
and afford acyclic products with a syn-stereochemical relation-
ship between the fluorine and imine substituents. Yields are
mostly very good with MTBE being shown to be the most
suitable ethereal solvent. The highly fluorinated product 67,
which originated from a cascade employing a trifluoromethyl-
substituted imine, is particularly noteworthy.
Not surprisingly, given the unexpected outcome that the
fluorinated nucleophile exhibited compared to the other
heteroatom substituents explored, we decided to explore
other halogen substituents to learn if they followed the same
reaction pathway as fluorine. In this regard, we were most
intrigued in exploring the behavior of chlorinated nucleophiles.
This was inspired by a recent pharmaceutical structural analysis
5797
https://doi.org/10.1021/jacs.1c00745
J. Am. Chem. Soc. 2021, 143, 5793−5804

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
69 in the presence of imine 68 and MTBE as solvent resulted
in a rapid anionic cascade, which produced a more complex
proton NMR spectra than we had expected, including the
presence of two methyl ester groups as well as a chlorine atom
attached to a double bond. Close inspection of the NMR
spectra soon revealed the presence of a cyclopropane group,
resulting from intramolecular 3-exo-tet cyclization of the in situ-
formed enamide resulting from initial Mannich-amino-Cope
(or Michael addition) steps. Remarkably, the cyclopropane
imine could not be isolated, as second addition of a dienolate
nucleophile proved to be faster than addition to the starting
imine, thus affording 70 as the product. This outcome is quite
a contrast with the fluorine nucleophile addition results,
wherein no second addition of a dienolate nucleophile was
observed. With such a complex product, containing five
contiguous sp3-stereocenters as well as two double bond
stereoisomers, we were not confident relying on NMR for
satisfactory structural analysis. Fortunately, after much
experimentation, we were finally successful in growing crystals
of 70 that allowed us to secure a single X-ray crystal structure
and thus enabling definite confirmation of its structure as well
as its absolute and relative stereochemical assignments. Most
importantly, working backward based on the stereochemistry
of the cyclopropane, we were able to deduce that the
chlorinated dienolate nucleophile followed exactly in the
footsteps of the fluorinated nucleophile to afford an acyclic
product with a syn-relationship between the two new
stereocenters as well as an E-enoate double bond. We have
evaluated this amazing Mannich-amino-Cope-cyclization-Man-
nich anionic cascade that also proceeded for additional imines,
all of which afforded similar products (71−73) in very good
yields. We were also able to isolate the mono addition imine
product (see Supporting Information Scheme S3) for the R =
4-CF3C6H4 substituent.
electronics and size impacted imine additions and ensuing
anionic cascades.
Nitro (X = NO2). We chose a nitro group as a great γ-
substituent that would strongly influence the electronics of the
dienolate nucleophile, while also delivering a versatile nitrogen
atom substituent for post-cascade applications. Toward that
end, we synthesized nitro-substituted enoate 76 and subjected
it under our standard reaction conditions to imine 11. The
reaction proceeded to form two products in a 5.5:1 ratio,
which NMR revealed still contained three olefinic protons and
absence of a methyl ester signal (Scheme 8). Further 1D and
Scheme 8. Nitro-Substituted Nucleophile Diverts the Path
toward β-Lactam Formation
2D NMR, especially NOE analyses, soon uncovered that these
new products were indeed diastereomeric divinyl β-lactams
(77 and 78) differing only in their syn-relationship of the vinyl
substituents.
The complete syn-selectivity and moderate diastereoselec-
tivity (5.5:1) are strikingly similar to what we observed in a
recent vinylogous aza-Darzens vinyl aziridine study26 wherein
it was established that a competing retro-Mannich pathway
accompanied by imine isomerization was responsible for low
diastereoselectivity. Close inspection of Hammett and
modified Swain−Lupton constants27 reveals a much higher
σp value for a nitro group compared to that of F, Cl, Br, N3,
SMe, and SCF3 may help explain why a strong electron
withdrawing group could suppress the amino-Cope pathway
from opening the door for β-lactam formation. We hypothesize
that not only the double bond geometry of the nitro olefin
adds in concert a steric burden on the amino-Cope transition
state but the presence of such a strong electron withdrawing
group also increases the electrophilicity of the adjacent double
bond carbon which destabilizes the amino-Cope transition
state and leads to a β-lactam pathway. Alternatively, it is also
possible that β-lactam formation occurs via the intermediacy of
a ketene intermediate.
To help shed additional light on how the Mannich addition
step and resulting amino-Cope or β-lactam cascades are
impacted by γ-nucleophile substituents, we decided to explore
more electron donating substituents represented by methyl
and methoxy groups.
Methyl (X = CH3). Deprotonation of methyl-substituted
enoate 79b in the presence of imine 57 only afforded Mannich
product 80 with the R = CF3 group (Scheme 9). Importantly,
this Mannich product 80 does not undergo either an amino-
Cope reaction or a β-lactam formation but instead readily
undergoes an isomerization to the most-substituted Mannich
product when resubjected to base (LiHMDS). However, when
With chlorine nucleophile results firmly established, we
turned our attention to evaluating the addition and anionic
cascade behavior of γ-bromo enoate 74 (Scheme 7). The
Scheme 7. Bromine-Substituted Nucleophile Is Also
Competent for the Amino-Cope Cascade
dienolate of this enoate progressed to follow the exact same
multistep anionic cascade that the chloride nucleophile had
established. In the bromo case, there was a small deviation in
outcome as two diastereomeric products were isolated (75),
differing only in their stereochemistry at the ester sp3-
stereocenter. This could be a result of the less selective
second addition step or post-epimerization following addition.
We also evaluated an iodo-substituted nucleophile and
learned it did afford the expected cyclopropane product albeit
in very low yield (less than 10%), which we tentatively
attribute to its larger atomic size compared to the other
halogens. With the halogens behaving differently to azide,
thiomethyl, and trifluoromethylthiol groups, we decided to
explore how extremes in terms of γ-nucleophile substituent
5798
https://doi.org/10.1021/jacs.1c00745
J. Am. Chem. Soc. 2021, 143, 5793−5804

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
suppressing the dissociations that plagued earlier investiga-
tions. This interesting reaction outcome has allowed us to
confirm that both Mannich products undergo a retro-Mannich
reaction to form lithium dienolate and sulfinylimine but only
the matched lithium dienolate recombines with sulfinylimine
to form the desired cylcohexenone product as the only
diastereomer.
Scheme 9. Methyl-Substituted Nucleophile Undergoes
Mannich/Isomerization
To further understand the aforementioned lithium chelation
effects and the significance of the retro-Mannich pathway, the
isolated Mannich adduct 86 was subjected to a competition
experiment with excess 4-Br-phenyl imine 11 and LiHMDS. In
this case, both the chiral cyclohexanones 15 and 12 were
formed in the ratio 1.5:1 (Scheme 12). Taken together, these
R = Ph (79a), a Mannich adduct was noticed, but over time, it
underwent a retro-Mannich reaction.
Trifluoroethoxy (X = OCH2CF3). We next evaluated the
dienolate originating from deprotonation of trifluoroethoxy (X
= OCH2CF3)-substituted enoate 81 (Scheme 10). When
Scheme 11. Counterion Effects
Scheme 10. Dienolate of the Trifluoroethoxy Substituent
Adds Unselectively to Imine
reacted with imine 11, a slow reaction ensued, which seemed
to mostly form a Mannich adduct but over time produced a
new product (82) as a nonselective 3:1:0.5 mixture of three
diastereomers. This mixture could have arisen from either a
Mannich-amino-Cope cascade or a direct Michael addition.
We were curious to learn more about this structure, which we
pursued through a series of degradation experiments. These
experiments revealed that the imine was the same in all
structures and the auxiliary had not been compromised,
leading us to postulate that the newly formed methoxy
stereocenter and the ester double bond geometry were most
likely responsible for this mixture.
results suggest that the favored mechanism is indeed a stepwise
dissociative one. Upon treatment with excess base, adduct 86
Similarly, when we employed a methoxy (X = OCH3)-
substituted nucleophile (not shown, see Supporting Informa-
tion Scheme S5), we observed the same product as in the
trifluoroethoxy case but in lower yield (32%) and as a mixture
of three diastereomers in the ratio of 2:1:1.
Scheme 12. Mannich Crossover Experiment
MECHANISTIC INVESTIGATIONS
■
We were curious to learn how the change in counterion would
affect these γ-substituted asymmetric amino-Cope reaction
cascades. We performed counterion studies with isolated
Mannich adduct 86 (generated by quenching the reaction at
−78 °C) and resubjected it to alternative bases (LiHMDS,
NaHMDS, and KHMDS). It is important to note that the
initial Mannich reaction proceeded to form a mixture of two
products in a 1.5:1 ratio with E and Z at the ester sp3-
stereocenter. Our preliminary investigations have revealed that
the Li counterion was critical for the success of the
rearrangement while Na and K were detrimental. This suggests
that the Li counterions are playing a key role, most likely by
5799
https://doi.org/10.1021/jacs.1c00745
J. Am. Chem. Soc. 2021, 143, 5793−5804

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
a
Scheme 13. Free Energy Diagram for the γ-Substituted Asymmetric Amino-Cope Reaction Cascades
a
S = solvent, i.e., THF. * means when X = F (fluorine), the reaction stopped at acyclic imine with a Z double bond at the ester moiety. X = N3,
SMe, SCF3, F, Cl, and Br unless otherwise stated. The DFT-calculated energies for each species for each substrate are tabulated in Table 5.
a
Table 5. DFT-Calculated Energies for Each Species of the Rearrangement Mechanism for Each Substrate
Cope-
Int
anti-
TS2
syn-
TS2
anti-Cope-
Prod
syn-Cope-
Prod
3-Exo-Tet-
TS
Enamide-
TS
3-Exo-Tet-
Prod
Enamide-
Int
Enamide-
Prod
reactant
TS1
When X = N3, SMe, SCF3
when X = N3
when X = SMe
when X = SCF3
8.0
9.6
8.6
0.0
1.5
0.2
5.9
11.3
8.6
13.6
18.5
15.9
−10.9
−5.3
−6.6
−8.4
−3.5
−6.0
15.9
31.6
14.4
2.9
9.2
8.3
−3.8
13.9
−10.6
−2.6
3.1
2.3
26.3
29.2
31.6
When X = F, Cl, Br
when X = F
when X = Cl, Br
Energies are given in kcal mol−1.
8.9
7.8
6.8
2.3
10.6
12.7
9.1
7.7
−11.6
−9.2
−12.1
−10.6
26.4
7.0
5.5
6.6
11.0
−7.2
−4.5
−4.8
25.3
24.9
a
undergoes a retro-Mannich reaction, which is then followed by
a Li-chelation assisted Mannich recombination of the lithium
dienolate with both sulfinylimines. These fragments recombine
in the desired chair transition state, resulting in the formation
of anti-Cope products 15 and 12 as single diastereomers. In
other words, the lithium chelation assisted amino-Cope step
serves as the gatekeeper that allows the amino-Cope reaction
to proceed (matched case) or to be rejected (mismatched
case) via a retro-Mannich step (see Supporting Information
Figures S4−S6).
and substrate are detailed in Table 5. The model truncates the
tert-butyl group of the sulfonamide to a methyl, and a lithium
counterion with two explicit THF solvent molecules are
coordinating to the anionic structure.28 (See the Supporting
Information for full details on the computational methods.)
As we have previously discussed, the favored mechanism is a
stepwise dissociative mechanism.1 The first step TS1 88 in the
rearrangement mechanism is the cleavage of the C3−C4 to
form a lithium dienolate and sulfinylimine that coordinate to
the lithium counterion Cope-Int 89. These fragments
recombine in a chair transition state (TS2) to form the new
carbon−carbon bond between C1−C6, which we calculate to
have a slightly higher energy than TS1. These fragments
recombine to form either the syn- or anti-Cope product (92
and 93). The formation of the Cope products is calculated to
be exergonic for both product isomers for all substitutions.
Subsequently, after the formation of the Cope product, each
substrate continues to rearrange either through (1) an
COMPUTATIONAL RESULTS AND DISCUSSION
■
Density functional theory (DFT) with the M06-2X functional
and 6-311+G(d,p) basis set was employed to shed further light
on the mechanism of the new amino-Cope rearrangement
cascades (Scheme 13). The DFT-calculated energies for each
species of the rearrangement mechanism for each intermediate
5800
https://doi.org/10.1021/jacs.1c00745
J. Am. Chem. Soc. 2021, 143, 5793−5804

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
intramolecular SN2-like transition state between C2−C6−X 3-
exo-tet-TS 94 to form a cyclopropane ring 3-exo-tet-Prod 96 or
(2) C2−C(ester) Enamide-TS 95 to form a six-membered ring
intermediate Enamide-Int 97 that quickly releases a methoxy
anion to generate cyclohexanone products.
though the energy of forming the Enamide-Int 97 is less
favorable than that of the 3-exo-tet-Prod 96.
We next investigate the full rearrangement for the fluoro-
substrate. This substrate forms the syn-Cope-Prod 92 and does
not further cascade into either the 3-Exo-Tet-Prod or enamide
product (Enamide-Int). The calculated energy for the syn-TS2
is 1.5 kcal mol−1 lower in energy than the respective anti-TS2
(Figure 3), and the calculated energy for the syn-Cope-Prod is
The relative configuration of the neutral sulfinylimine has a
lasting impact on what stereoisomeric products are available.
The sulfinylimine fragment within the intermediate complex
Cope-Int 89 contains two single bonds that can potentially
rotate around to generate sets of unique conformers. One set
can be either s-cis or s-trans around the internal single bond,
and one set can position the lone pairs of the N and S of the
sulfonamide can either be syn or anti (see Supporting
Information Figure S7). Our modeling suggests that the s-
trans conformer is more stable by 2.8 kcal mol−1, and just as in
our previous study,2 the lone pairs of the N and S atoms are
more stable in the anti-conformation by 2.5 kcal mol−1.29 This
fragment is prevented from complete free rotation due to steric
clashes with the methyl group of the sulfonamide and the
surrounding THF molecules and methyl ester group on the
enolate fragment. These preferences combined with limited
rotational freedom of the sulfinylimine to completely rotate
within the intermediate complex ensure that the re-face of this
fragment is the only face open for the recombination step (TS2
90 and 91).
Figure 3. DFT-calculated structures of γ-F-substituents.
We first investigated the full rearrangement mechanism for
the N3/SMe/SCF3 substrates. These substrates preferentially
form the anti-Cope product 93 and further cascade to form the
enamide species. The anti- and syn-TS2 for the N3-substrate is
shown in Figure 2. The anti-TS2 is 7.7 kcal mol−1 lower in
0.5 kcal mol−1 lower in energy than the respective anti-Cope-
Prod. Both the syn- and anti-TS2 arrange to form a chair-like
structure, with the fluorine atom preferentially being in the
axial (F-Syn-TS2) position due to favorable hyperconjugation
from the C1−H s-bond.
But, on the other hand, formation of the 3-Exo-Tet-Prod 96
is highly endergonic due to the large enthalpy of formation of
fluoride. The formation of the enamide is prevented due to the
unfavorable E-arrangement of the C4 ester and C5 methyl
groups along the newly formed C4−C5 π-bond (Scheme 14).
To form the enamide intermediate 100, this region of the
substrate must rotate inward toward the imine bond, which
would place the methyl group inside the forming cyclohexane
ring. However, for the N3/SMe/SCF3 substrates, these groups
are in a Z-arrangement, which rotates the methyl group out of
the forming ring 101. If these groups were in a Z-arrangement,
the formation of the enamide product would be facile.
The rearrangement mechanism for the chloro-substrate Cl-
React follows a similar path to that of the fluoro-substrate in
that the syn-Cope product 92 is observed, however following
the formation of the Cope-product the substrate reorganizes to
form the 3-Exo-Tet-Prod. The syn- and anti-TS2 arrange to
form chair-like structures, with the chlorine atom preferentially
being in the axial (Scheme 13) position due to favorable
hyperconjugation from the C1−H s-bond. The Cl-syn-Cope-
Prod is prevented to form the enamide intermediate due to the
same steric restrictions of the C4 ester and C5 methyl groups,
similar to that of the fluoro-substrate. However, the barrier to
form the 3-exo-tet is nearly isoenergetic to Cl-syn-TS2, and the
formation of a chloride is much more thermodynamically
favorable to that of the Cl-Enamide-Pro and methoxy anion
(Table 5).
Figure 2. DFT-calculated structures of γ-N3-substituents.
energy than the syn-TS2, and the anti-Cope-Prod is 0.6−2.5
kcal mol−1 lower in energy than the syn-Cope-Prod. Each of
these substrates has the capacity to form the 3-exo-tet and
enamide products, yet the enamide product is exclusively
observed. We find that the barrier to form the enamide is much
lower in energy than the barrier to form the 3-exo-tet TS by
nearly 9−22 kcal mol−1, respectively, due to the ring strain
associated with the formation of a cyclopropane ring and poor
leaving group formation. At cryogenic temperatures, the
formation of the kinetic product (enamide) is favored even
In summary, calculated structures, supported by exper-
imental results, lead us to the following conclusions: (1) The
stepwise dissociation−recombination mechanism was found to
be favored. The stereochemistry of the chiral auxiliary
5801
https://doi.org/10.1021/jacs.1c00745
J. Am. Chem. Soc. 2021, 143, 5793−5804

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Scheme 14. γ-N3-Substituent Promotes Enamide
Cyclization
Scheme 15. Examples of Transformations for a
Representative Azide-Substituted Amino-Cope Product
determines the stereochemistry of the Cope product by
influencing the orientation of the lithium dienolate and
sulfinylimine fragments in the recombination step. (2) For
the N3/SMe/SCF3 substrates, the anti-Cope product for-
mation is preferred over syn with an azide group in the
equatorial position to avoid steric clashes with the methyl
group. Because of this equatorial placement of the azide group,
a favorable Z-arrangement between the C4 ester and C5 methyl
groups can be achieved, which further cascade to form the
enamide species. (3) In the case of halogens (X= F, Cl, and
Br), the syn-Cope product formation is preferred over anti with
the halogen group in the axial position due to favorable
hyperconjugation from the C1−H s-bond. This axial preference
of the halogen groups further prevents the enamide formation
because it requires an unfavorable inward rotation around the
C5 methyl group toward the imine bond.
tively, heating of 108 in toluene forms a nitrene, which inserts
into a C−H bond of the neighboring bromo-aryl group to
deliver carbazole 110.31
CONCLUSIONS
■
In summary, we have further established and expanded upon
the great synthetic promise of our new lithium-enabled
asymmetric amino-Cope rearrangement cascade. By exploring
dienolates substituted with a range of γ-substituents, we have
uncovered most intriguing selectivity outcomes, wherein γ-
azide, thiomethyl, and trifluoromethylthio substituents proceed
via a chair amino-Cope transition state to form chiral
cyclohexenone products with two new stereocenters with an
anti-relationship. Intriguingly and unexpectedly, the corre-
sponding γ-halo (F, Cl, and Br) dienolate substituents also
undergo the proposed lithium-enabled asymmetric amino-
Cope cascade but form products with a syn-relationship
between the two new stereocenters as well as an E-enoate. In
the case of chlorides and bromides, the intermediate lithium
enamide undergoes a remarkably facile 3-exo-tet cyclization to
form cyclopropane products. Further explorations on the
extremes of the size and electronics of γ-substituents revealed
that electron donating groups such as methyl, methoxy, and
trifluoroethoxy proceed to only undergo the initial Mannich
step or undergo unselective conjugate additions. Interestingly
and of great interest is the diverted pathway a nitro-substituted
APPLICATION EXAMPLES
■
Highlighted in Scheme 15 are interesting reaction trans-
formation examples for azide-substituted cyclohexenone
amino-Cope product 12. Treatment of 12 with electrophilic
halogen reagents such as N-chlorosuccinimide (NCS) and
selectfluor affords chloro- and fluoro-substituted products 104
and 105, respectively, as single diastereomers. The Staudinger
reaction is realized with trimethyl phosphite,30 which following
in situ hydrolysis yields amino-phosphate 106. Selective O-
allylation can be achieved using allyl iodide and cesium
carbonate as base (107). Dehydrogenation to form aryl azide
108 takes place in the presence of 2,3-dichloro-5,6-dicyano-p-
benzoquinone (DDQ). Aryl azide 108 readily undergoes a
Staudinger reaction to yield aminophosphate 109. Alterna-
5802
https://doi.org/10.1021/jacs.1c00745
J. Am. Chem. Soc. 2021, 143, 5793−5804

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
enoate undergoes to yield valuable β-lactam products. It is
evident that exciting adventures and discoveries await to be
unraveled for this new asymmetric amino-Cope reaction
platform.
REFERENCES
■
(1) Chogii, I.; Das, P.; Fell, J. S.; Scott, K. A.; Crawford, M. N.;
Houk, K. N.; Njardarson, J. T. New class of anion-accelerated amino-
Cope rearrangements as gateway to diverse chiral structures. J. Am.
Chem. Soc. 2017, 139, 13141−13146.
(2) Spules, T. J.; Galpin, J. D.; Macdonald, D. Charge-accelerated
Cope rearrangements of 3-amino-1,5-dienes. Tetrahedron Lett. 1993,
34, 247−250.
ASSOCIATED CONTENT
■
sı
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/jacs.1c00745.
(3) Allin, S. M.; Button, M. A. C. The first example of asymmetric
induction in an anionic amino-Cope rearrangement. Tetrahedron Lett.
1998, 39, 3345−3348.
Experimental procedures, characterization data, and
spectroscopic data for all new compounds and computa-
tional methods (PDF)
(4) Yoo, H. Y.; Houk, K. N.; Lee, J. K.; Scialdone, M. A.; Meyers, A.
I. New paradigm for anionic heteroatom Cope rearrangements. J. Am.
Chem. Soc. 1998, 120, 205−206.
(5) Dobson, H. K.; LeBlanc, R.; Perrier, H.; Stephenson, C.; Welch,
T. R.; Macdonald, D. [1,3] and [3,3] Rearrangements of 3-amino-1,5-
hexadienes: Solvent effect on the regioselectivity. Tetrahedron Lett.
1999, 40, 3119−3122.
Accession Codes
CCDC 2017739−2017740 and 2039911−2039912 contain
the supplementary crystallographic data for this paper. These
data can be obtained free of charge via www.ccdc.cam.ac.uk/
data_request/cif, or by emailing data_request@ccdc.cam.ac.
uk, or by contacting The Cambridge Crystallographic Data
Centre, 12 Union Road, Cambridge CB2 1EZ, UK; fax: +44
1223 336033.
(6) Allin, S. M.; Button, M. A. C. Evidence for alternative
mechanisms in the amino-Cope rearrangement. Tetrahedron Lett.
1999, 40, 3801−3802.
(7) Evans, D. A.; Golob, A. M. [3,3]-Sigmatropic rearrangements of
1,5-diene alkoxides. The powerful accelerating effects of the alkoxide
substituent. J. Am. Chem. Soc. 1975, 97, 4765−4766.
(8) The Seebach group has proposed a rare example which did not
succumb to the retro-aldol pathway: Seebach, D.; Misslitz, U.;
Uhlmann, P. The dienolate of (R)-2-tert-butyl-6-methyl-1,3-dioxin-
4(2H)-one. A chiral acetoacetic ester 4-reagent. Angew. Chem., Int. Ed.
Engl. 1989, 28, 472−473.
(9) Schneider, C.; Rehfeuter, M. Stereoselective Cope rearrange-
ments of syn- and anti-aldol products. Synlett 1996, 1996, 212−214.
(10) Schneider, C. The silyloxy-Cope rearrangement of syn-aldol
products: Evolution of a powerful synthetic strategy. Synlett 2001,
2001, 1079−1091.
(11) Weise, C. F.; Immel, S.; Richter, F.; Schneider, C. The oxy-
Cope rearrangement of aldol products. A combined experimental and
theoretical study. Eur. J. Org. Chem. 2012, 2012, 1520−1529.
(12) Tomooka, K.; Nagasawa, W.; Wei, S.; Nakai, T. Chiral
dienolate chemistry in remote asymmetric induction: The asymmetric
aldol/oxy-Cope strategy for asymmetric synthesis of γ,δ-dichiral α,β-
unsaturated acid derivatives. Tetrahedron Lett. 1996, 37, 8899−8900.
(13) Black, W. C.; Giroux, A.; Greidanus, G. Silyloxy-Cope
rearrangement of chiral aldol adducts. Tetrahedron Lett. 1996, 37,
4471−4474.
AUTHOR INFORMATION
■
Corresponding Authors
Kendall N. Houk − Department of Chemistry and
Biochemistry, University of California, Los Angeles,
California 90095, United States; orcid.org/0000-0002-
8387-5261; Email: houk@chem.ucla.edu
Jon T. Njardarson − Department of Chemistry and
Biochemistry, University of Arizona, Tucson, Arizona 85721,
United States; orcid.org/0000-0003-2268-1479;
Email: njardars@email.arizona.edu
Authors
Pradipta Das − Department of Chemistry and Biochemistry,
University of Arizona, Tucson, Arizona 85721, United States
Michael D. Delost − Department of Chemistry and
Biochemistry, University of Arizona, Tucson, Arizona 85721,
United States
(14) Davies, H. M. L.; Beckwith, R. E. J. Catalytic asymmetric
reactions for organic synthesis: The combined C-H activation/
silyloxy-Cope rearrangement. J. Org. Chem. 2004, 69, 9241−9247.
(15) It is worth noting that this same amino-Cope enamide
intermediate could also be envisioned to arise from an asymmetric
macrocyclic Michael addition, wherein the lithium counterion of the
dienolate interacts with the imine nitrogen to enable an asymmetric
union. This second lithium chelate scenario would be expected to be
much weaker, and the fact that Mannich intermediates can be isolated
suggests it is less likely.
(16) Vitaku, E.; Smith, D. T.; Njardarson, J. T. Analysis of the
structural diversity, substitution patterns and frequency of nitrogen
heterocycles among US FDA approved pharmaceuticals. J. Med. Chem.
2014, 57, 10257−10274.
(17) Das, P.; Delost, M. D.; Qureshi, M. H.; Smith, D. T.;
Njardarson, J. T. A survey of the structures of US FDA approved
combination drugs. J. Med. Chem. 2019, 62, 4265−4311.
(18) Brase, S.; Banert, K. Organic Azides: Syntheses and applications;
John Wiley & Sons Ltd.: West Sussex, U.K., 2010.
(19) Scott, K. A.; Njardarson, J. T. Analysis of US FDA Drugs
Containing Sulfur Atoms. Top. Curr. Chem. 2019, 376, 1−34.
(20) Leroux, F.; Jeschke, P.; Schlosser, M. α-Fluorinated Ethers,
Thioethers, and Amines: Anomerically Biased Species. Chem. Rev.
2005, 105, 827−856.
Munaum H. Qureshi − Department of Chemistry and
Biochemistry, University of Arizona, Tucson, Arizona 85721,
United States
Jianhua Bao − Department of Chemistry and Biochemistry,
University of Arizona, Tucson, Arizona 85721, United States
Jason S. Fell − Department of Chemistry and Biochemistry,
University of California, Los Angeles, California 90095,
United States; orcid.org/0000-0001-6680-2936
Complete contact information is available at:
https://pubs.acs.org/10.1021/jacs.1c00745
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We would like to thank the National Science Foundation
(CHE-1855708) for financial support of this research and for
supporting the acquisition and installation of a new NMR
spectrometer (MRI-grant CHE-1920234). Special thanks and
gratitude to Dr. Andrei Astachkine for helping with data
collection and analysis for all reported crystal structures.
5803
https://doi.org/10.1021/jacs.1c00745
J. Am. Chem. Soc. 2021, 143, 5793−5804

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(21) Toulgoat, F.; Alazet, S.; Billard, T. Direct Trifluoromethylth-
iolation Reactions: The “Renaissance” of an Old Concept. Eur. J. Org.
Chem. 2014, 2014, 2415−2428.
(22) Ilardi, E. A.; Vitaku, E.; Njardarson, J. T. Data-mining for sulfur
and fluorine: An evaluation of pharmaceuticals to reveal opportunities
for drug design and discovery. J. Med. Chem. 2014, 57, 2832−2842.
(23) Gillis, E. P.; Eastman, K. J.; Hill, M. D.; Donnelly, D. J.;
Meanwell, N. A. Application of fluorine in medicinal chemistry. J.
Med. Chem. 2015, 58, 8315−8359.
(24) Meanwell, N. A. Fluorine and fluorinated motifs in the design
and applications of bioisosteres for drug design. J. Med. Chem. 2018,
61, 5822−5880.
(25) Smith, B. R.; Eastman, C. M.; Njardarson, J. T. Beyond C, H, O
and N! Analysis of the elemental composition of US FDA approved
drug architectures. J. Med. Chem. 2014, 57, 9764−9773.
(26) Chogii, I.; Das, P.; Delost, M. D.; Crawford, M. D.; Njardarson,
J. T. Asymmetric Vinylogous Aza-Darzens Approach to Vinyl
Aziridines. Org. Lett. 2018, 20, 4942−4945.
(27) Hansch, C.; Leo, A.; Taft, R. W. A survey of Hammett
substituent constants and resonance and field parameters. Chem. Rev.
1991, 91, 165−195.
(28) Essafi, S.; Tomasi, S.; Aggarwal, V. K.; Harvey, J. N.
Homologation of boronic esters with organolithium compounds: A
computational assessment of mechanism. J. Org. Chem. 2014, 79,
12148−12158.
(29) Zhang, H.; Wu, W.; Ahmed, B. M.; Mezei, G.; Mo, Y.
Adjacentlone pair (ALP) effect: A computational approach for its
origin. Chem. - Eur. J. 2016, 22, 7415−7421.
(30) Das, P.; Njardarson, J. T. Synthesis of 1,2,3,6-Tetrahydropyr-
idines via aminophosphate enabled anionic cascade and acid catalyzed
approaches. Org. Lett. 2015, 17, 4030−4033.
(31) Smith, P. A. S.; Brown, B. B. The Synthesis of Heterocyclic
Compounds from Aryl Azides. I. Bromo and Nitro Carbazoles. J. Am.
Chem. Soc. 1951, 73, 2435−2437.
5804
https://doi.org/10.1021/jacs.1c00745
J. Am. Chem. Soc. 2021, 143, 5793−5804