Taming Radical Pairs in the Crystalline Solid State: Discovery and Total Synthesis of Psychotriadine

Article abstract of DOI:10.1021/jacs.1c01100

Solid-state photodecarbonylation is an attractive but underutilized methodology to forge hindered C-C bonds in complex molecules. This study discloses the use of this reaction to assemble the vicinal quaternary stereocenter motif present in bis(cyclotryptamine) alkaloids. Our strategy was enabled by experimental and computational investigations of the role of substrate conformation on the success or failure of the solid-state photodecarbonylation reaction. This informed a crystal engineering strategy to optimize the key step of the total synthesis. Ultimately, this endeavor culminated in the successful synthesis of the bis(cyclotryptamine) alkaloid "psychotriadine,"which features the elusive piperidinoindoline framework. Psychotriadine, a previously unknown compound, was identified in the extracts of the flower Psychotria colorata, suggesting it is a naturally occurring metabolite.

Full text of DOI:10.1021/jacs.1c01100

Made available for a limited time for personal research and study only License.
pubs.acs.org/JACS
Article
Taming Radical Pairs in the Crystalline Solid State: Discovery and
Total Synthesis of Psychotriadine
Jordan J. Dotson, Ieva Liepuoniute, J. Logan Bachman, Vince M. Hipwell, Saeed I. Khan, K. N. Houk,*
Neil K. Garg,* and Miguel A. Garcia-Garibay*
Cite This: J. Am. Chem. Soc. 2021, 143, 4043−4054
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Solid-state photodecarbonylation is an attractive
but underutilized methodology to forge hindered C−C bonds in
complex molecules. This study discloses the use of this reaction to
assemble the vicinal quaternary stereocenter motif present in
bis(cyclotryptamine) alkaloids. Our strategy was enabled by
experimental and computational investigations of the role of
substrate conformation on the success or failure of the solid-state
photodecarbonylation reaction. This informed a crystal engineer-
ing strategy to optimize the key step of the total synthesis.
Ultimately, this endeavor culminated in the successful synthesis of
the bis(cyclotryptamine) alkaloid “psychotriadine,” which features
the elusive piperidinoindoline framework. Psychotriadine, a
previously unknown compound, was identified in the extracts of
the flower Psychotria colorata, suggesting it is a naturally occurring metabolite.
With the aforementioned motivations in mind, we sought to
evaluate solid-state photodecarbonylation chemistry in the
context of the bis(cyclotryptamine) alkaloids (Figure 1b). This
class of natural products, arising from common biosynthon 8,
features vicinal quaternary stereocenters and has been popular
among synthetic chemists for many decades.17−22 In addition,
this family of small molecules possesses a rich history,
stemming from the isolation report of calycanthine (4) in
1888 and the subsequent discovery of related23 natural
products (e.g., 5 and 6), representing a total of four unique
constitutional isomers.24 Interestingly, a fifth piperidinoindo-
line isomer 7 was proposed by Robinson in 1954 but has never
been isolated from a natural source.25−27 Movassaghi and
coworkers have synthesized a highly substituted derivative of 7
in the context of communesin alkaloid total synthesis studies.17
Given the widespread interest in these compounds and their
complex structures bearing vicinal quaternary stereocenters, we
deemed the bis(cyclotryptamine) alkaloids an ideal arena to
test the solid-state photodecarbonylation methodology.
INTRODUCTION
■
Photochemistry has become an increasingly powerful tool in
modern organic synthesis.1−4 One particularly interesting class
of photochemical transformations with great potential involves
reactions conducted in the crystalline solid state. Such
transformations are attractive because of the opportunity to
control various selectivities (stereo-, regio-, and chemo-),
potential for scalable green chemistry, and ability to form
strained or congested frameworks.5−7
Our laboratory has been particularly interested in photo-
decarbonylation reactions that occur in the crystalline solid
state, and we have demonstrated their potential for the
stereospecific assembly of vicinal quaternary stereocenters
(Figure 1a).5,8−13 The introduction of vicinal quaternary
stereocenters has remained a long-standing challenge in
organic synthesis.14−16 An attractive approach to this motif,
1, is via direct coupling of two prochiral aliphatic radicals 2.17
In turn, radicals 2 can be generated from the photo-
decarbonylation of hexasubstituted ketone 3. While in solution,
the coupling of prochiral radicals can lead to complex product
mixtures, the corresponding process in the crystalline solid
state can result in high-yielding stereospecific recombination.5
Despite promising investigations into fundamental reactivity,
this solid-state approach has seen little use in natural product
synthesis10,11 and has not been used to construct complex
alkaloids. Moreover, the understanding of how substrate
conformations relate to the success or failure of solid-state
photodecarbonylation reactions has remained underexplored.
Our laboratory recently described the successful application
of solid-state photodecarbonylation chemistry in the total
Received: January 28, 2021
Published: March 8, 2021
https://dx.doi.org/10.1021/jacs.1c01100
J. Am. Chem. Soc. 2021, 143, 4043−4054
© 2021 American Chemical Society
4043

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 2. Retrosynthetic analysis of biosynthon 8.
At the outset of this research pursuit, it was anticipated that
a major challenge in our synthetic strategy would be the
preparation of 9 bearing amines on the ortho position of the
aromatic rings. Introduction of protected amines at this
position early in the synthesis would be ideal, but the steric
crowding imparted by ortho substituents in close proximity to
the quaternary centers would make the assembly of the
corresponding ketone challenging.29 Additionally, it was
envisioned that the increased steric bulk resulting from these
substituents may also hinder C−C bond formation from
radical pair 10. To address these concerns, two classes of
ketone substrates, ortho C−H (11a−c) and ortho C−X
(11d−f) (Figure 2), that would be synthetically accessible
from a common route were targeted. It was hypothesized that
the photodecarbonylation of ketones 11a−c would occur
selectively and the resulting products 9a−c could be
functionalized by directed ortho C−H metalation/amination.
It was predicted that ketones 11d−f could be competent
substrates for photodecarbonylation to access 9d−f, and that
the ortho-Cl, -Br, and -OMe substituents could function as
synthetic handles for a metal-catalyzed amination.30
Figure 1. Solid-state photodecarbonylation to introduce vicinal
quaternary stereocenters and overview of select bis(cyclotryptamine)
alkaloids.
synthesis of a bis(cyclotryptamine) alkaloid.28 Here, we
disclose our full investigation of the key photodecarbonylation
step, including the evaluation of 11 different crystalline
substrates. By analysis of substrate conformation via X-ray
structures and computational modeling, we provide the
physical organic underpinnings that explain the success or
failure of the key photodecarbonylation step. These efforts
demonstrate that the transformation can be optimized by
manipulating the crystalline conformation of the necessary
hexasubstituted ketone substrate. Ultimately, our crystalline
conformation-based approach enabled the formation of the
challenging vicinal quaternary stereocenters present in
biosynthon 8. These efforts culminated in the total synthesis
of a natural product that had previously been overlooked in
plant isolations and features the final, elusive piperidinoindo-
line scaffold of the bis(cyclotryptamine) alkaloids.
Synthesis and Photodecarbonylation of Crystalline
Ketones. Photodecarbonylation substrates 11a−c were
prepared in a straightforward manner from readily available
diesters 14 (Scheme 1).31 Malonic esters 14 were subjected to
a two-step alkylation/reductive cyclization sequence to furnish
pyrrolidinones 15 in good yield. N-methylation of 15 under
basic conditions provided 16 in 60−87% yield. In turn, N-
methylpyrrolidinones 16 were cleanly converted to 17 via a
one-pot protocol involving ester saponification and subsequent
acid-mediated decarboxylation. Finally, treatment of lactams
17 with LiHMDS provided the corresponding enolates, which
were coupled with phosgene (13) to deliver ketones 11a,b in
49−64% yield and with high diastereoselectivity. While 11a
was a high-melting crystalline solid, 11b was a viscous oil at
room temperature (and therefore not a candidate for a solid-
state photodecarbonylation reaction).32 However, after effi-
ciently cleaving the aryl-methyl ethers of 11b using BBr3, we
were delighted to find that phenolic ketone 11c was a
crystalline solid.
RESULTS AND DISCUSSION
■
Retrosynthetic Analysis. Our retrosynthetic analysis of
common biosynthon 8 is depicted in Figure 2. It was
envisioned that 8, or a dehydrated synthetic congener thereof,
would arise from late-stage bis(amination) and reduction of
intermediate 9. In turn, 9, bearing the requisite vicinal
quaternary stereocenters, would be accessed via a key solid-
state photodecarbonylation of hexasubstituted ketone 11 via
the intermediacy of radical pair 10. Although both radical
centers would be stereochemically labile in intermediate 10, it
was anticipated that the rigidity of the crystalline lattice would
affect the transfer of stereochemical information from 11 to 9.
Finally, 11 would be constructed rapidly by the convergent
coupling of two equivalents of lithium enolate 12 with an
equivalent of phosgene (13).
The preparation of ketones 11d−f is shown in Scheme 2.31
The beginning of this synthetic sequence mirrored that for
4044
https://dx.doi.org/10.1021/jacs.1c01100
J. Am. Chem. Soc. 2021, 143, 4043−4054

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Scheme 1. Synthesis of Ketone Substrates 11a−c
Scheme 2. Synthesis of Ketone Substrates 11d−f
Figure 3. Photodecarbonylation of ketones 11 (yields resulting from
decarbonylation of substrates 11c−f determined using qNMR).
respectively.34 Although identifiable side products were not
observed from the photodecarbonylation of 11c,35 the reaction
of 11a also gave rise to products 17a and 19a (Figure 3, top),
presumably via a competing radical pair disproportionation
process.36 Unfortunately, ortho-substituted ketones 11d−f
performed poorly in the desired C−C bond formation.37
Solid-state irradiation of ketones 11d and 11e resulted in 9%
and 8% yields of 9d and 9e, respectively, while irradiation of
ketone 11f gave no identifiable recombination product 9f.38 As
was the case for ketone 11a, the photodecarbonylation of
ketones 11d−f displayed a competitive disproportionation
reaction pathway that was the major reaction outcome for
these substrates.39
Interestingly, we noted a correlation between the relative
selectivity of radical pair 10 undergoing recombination over
disproportionation to the size of the ortho-substituent
(quantified by the A-value).40 Ketones 11a and 11c bearing
only a hydrogen atom (A-value = 0) at the ortho position
underwent recombination to give 9a and 9c, respectively, as
the major products. Ketones 11d−f, bearing ortho-substitu-
ents, displayed lower yields of 9d−f and showed a higher
incidence of disproportionation. Intrigued by these results, we
sought to understand the structural factors that guide the
outcome of the solid-state photodecarbonylation reaction.
On the basis of close inspection of substrate X-ray crystal
structures, we offer the analysis shown in Figure 4 to explain
the sensitivity of the photodecarbonylation reaction to
aromatic substituents. We postulate that radical pair 10 is
prevented from forming a C−C bond due to steric clash
between the ortho-substituent on one radical fragment and the
ketones 11a−c. A two-step annulation of malonic esters 14
and subsequent N-methylation cleanly furnished pyrrolidi-
nones 16. Saponification of the ester in 16, followed by
decarboxylation, provided enolate precursors 17 in 79−90%
yield. Unfortunately, we observed that the reaction of 17d−f
with phosgene (13) did not provide ketone 11 and instead
resulted in substantial nonspecific decomposition (not shown).
To circumvent this issue, we constructed acid chloride 18 in a
two-step sequence from pyrrolidinone 16. This electrophile
was then efficiently coupled with the lithium enolate derived
from 17 to provide ketones 11d−f in good to excellent yields
and with high diastereoselectivity.
Once we had prepared ketones 11a−f, we next turned our
attention to evaluating their reactivities in the key solid-state
photodecarbonylation reaction that forges the vicinal quater-
nary stereocenters present in the bis(cyclotryptamine)
alkaloids (Figure 3).33 In exploring the photodecarbonylation
of these, a wide range of reactivity was observed. Substrates
11a and 11c were irradiated in the solid-state and proceeded
smoothly to provide 9a and 9c in 59% and 35% yield,
4045
https://dx.doi.org/10.1021/jacs.1c01100
J. Am. Chem. Soc. 2021, 143, 4043−4054

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
imentation, we were unable to functionalize the arene and
most attempts either returned starting material or resulted in
substantial nonspecific decomposition.47 Interestingly, we
found that the use of alkyllithium bases resulted in
deprotonation of the N-methyl groups on 9a to give the
corresponding primary carbanion (not shown).48 In order to
circumvent this issue, we sought to construct a derivative of 9a
with cleavable protecting groups on the nitrogen. This would
allow for the introduction of N-substituents that would not
undergo competitive activation.
In pursuit of a derivative of 9a with cleavable N-substituents,
ketone precursor 20 (Figure 5) was prepared.49 To our
Figure 4. Crystalline conformation dictates radical−radical recombi-
nation trajectory.
aromatic ring on the other (Figure 4a). The proposed repulsive
interactions can further be observed in the X-ray crystal
structures of the substrates (Figure 4b and c). For ketones
11d−f, steric congestion is observed in the interaction between
the ortho-substituent on one-half of the molecule and the aryl
ring on the other (interaction depicted with a red arrow).
Notably, such interactions are not present in ketones without
ortho functionality (i.e., 11a and 11c).41 Since the radical
centers in 10 must move closer to one another by
approximately 1.2 Å to generate a C−C bond, these close
interactions in the substrate are likely to become strongly
repulsive at the transition state. It is important to note that
reactions that take place in the crystalline solid state are
governed by the topochemical principle and favor reaction
pathways with minimal atomic motion.42 As such, the
conformational features observed in the X-ray crystal structures
of the substrate are likely to also be present in the reactive
intermediates and transition-state structures.
Figure 5. Conformational change could enhance chemoselectivity of
the photodecarbonylation reaction.
surprise, inspection of the X-ray crystal structure of 20 revealed
a different ketone conformation compared to our previous
observations, leading us to suspend our efforts toward C−H
activation and instead focus on manipulating the conformation
of the ketone substrate to optimize the solid-state photo-
decarbonylation reaction (Figure 5). As noted previously, it is
likely that the solid-state conformations of ketones 11d−f
prevent successful decarbonylative C−C bond formation due
to a growing steric clash in radical pair 10. On the basis of the
striking change in crystalline conformation between 11a and
20 (Figure 5a), we hypothesized that the conformation of
ortho-functionalized ketone 21 could also be altered by
modifying the N-substituents. An alternative conformation of
21 could lead to hypothetical radical pair 22 that does not
feature a steric clash that prevents the formation of 23.
Therefore, we sought to explore whether conformational
manipulation could be used to give a ketone that both includes
ortho- aryl halides or pseudohalides for late-stage metal-
catalyzed amination and the ability to undergo efficient solid-
state photodecarbonylation to forge the desired vicinal
quaternary stereocenters.
Given the efficiency of the solid-state photodecarbonylation
of 11a to give 9a, we pursued a C−H functionalization
approach to install the requisite ortho-C−N bonds in 9a
(Scheme 3).43−45 We hypothesized that the pyrrolidinone
oxygen atoms may function as convenient directing groups for
either transition metal-catalyzed ortho functionalization or
directed lithiation.46 Unfortunately, despite extensive exper-
Scheme 3. Attempted Directed Ortho-Metalation/
Functionalization
In order to prepare ketones 21 bearing modifiable N-
substituents, we conducted the synthetic sequence depicted in
4046
https://dx.doi.org/10.1021/jacs.1c01100
J. Am. Chem. Soc. 2021, 143, 4043−4054

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Scheme 4.50 First, ortho-chloro and ortho-bromo pyrrolidi-
nones 15 were cleanly converted to para-methoxybenzyl
Scheme 4. Synthesis of Ketones 28 and 29
Figure 6. Solid-state photochemistry of ketones 28d and 28e (the R-
groups on imides 28d and 28e were removed from the X-ray
renderings for clarity).
(PMB) protected substrates 24 under basic conditions. At
this point, the synthesis diverged. Esters 24 were saponified
and then thermally decarboxylated to provide 25 in 79−90%
yield. Esters 24 were also converted to acid chlorides 26 by a
two-step saponification/dehydrochlorination protocol. Pyrroli-
dinones 25 were then deprotonated with LiHMDS and the
resultant enolates were coupled with acid chlorides 26 to
furnish PMB-protected amides 27. As was observed in the
construction of 11, formation of 27 occurred with exquisite
diastereoselectivity, presumably due to a highly ordered
transition state mediated by Li+ chelation.49 In efforts to
cleave the PMB protecting groups, ketones 27 were then
treated with ceric ammonium nitrate (CAN), which
unexpectedly furnished a mixture of imide products 28 and
29. Interestingly, ketones 28 and 29 were both high-melting
point crystalline solids. Given this observation, we then sought
to test their efficiency in the solid-state photodecarbonylation
reaction.
The results of the solid-state photodecarbonylation of
symmetric chloride-containing ketone 28d are depicted in
Figure 6a. We were pleased to find that initial efforts toward
the conversion of 28d to 30d proceeded smoothly and with
retention of stereochemistry, as verified by single-crystal X-ray
diffraction. Furthermore, a competing disproportionation
reaction pathway was not observed in the photodecarbonyla-
tion of 28d. This result validated our hypothesis that
exchanging the N-substituents could alter the chemoselectivity
of the photodecarbonylation reaction, presumably due to the
new conformation of 28d relative to ketone 11d. Furthermore,
the efficiency of the reaction demonstrated that solid-state
photodecarbonylation is a competent transformation to forge
the vicinal quaternary stereocenter motif present in the
bis(cyclotryptamine) alkaloids.
Despite initial success, it was found that the solid-state
photodecarbonylation reaction was difficult to reproduce with
new batches of 28d (Figure 6a). While exposure of 28d to
ultraviolet light initially led to full consumption of starting
material within 24−48 h, similar irradiation of new samples of
28d only resulted in recovered starting material. After verifying
the robustness of the light source and ruling out the possibility
of impurities being present in the substrate, we realized that a
new, more stable polymorph of 28d, termed “unreactive
polymorph,” had formed. Single-crystal X-ray diffraction
revealed that this new polymorph adopts a different
conformation than the “reactive polymorph” (see X-ray crystal
structures in Figure 6a). Unfortunately, once the unreactive
polymorph was present, it was exceedingly difficult to prepare
the reactive polymorph, despite extensive screening of
crystallization conditions. These results suggest that the
reactive polymorph was likely formed kinetically, while the
unreactive polymorph was thermodynamically favored.51,52
Trace quantities of the more stable and unreactive polymorph
can act as seed crystals in subsequent recrystallization attempts,
thereby preventing further preparation of the reactive
polymorph.53
Due to the poor reaction reproducibility stemming from the
conformational polymorphism of 28d, we turned our attention
to evaluating the ortho-bromo substrate 28e (Figure 6b). While
two conformationally distinct polymorphs were observed for
28d, we only identified a single polymorph of 28e. The
conformation of 28e observed in the X-ray structure was
almost identical to that of the 28d unreactive polymorph. As
4047
https://dx.doi.org/10.1021/jacs.1c01100
J. Am. Chem. Soc. 2021, 143, 4043−4054

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
was expected, based on the similar conformation, 28e was also
unreactive under solid-state irradiation.
the nearest C−C bond of the aromatic π-system (see bonds
highlighted in blue in the microelectron diffraction (microED)
structures) are 66° and 68° for 29d, and 70° and 72° for
29e.55,56 These values, which were similar to those seen in the
reactive polymorph of 28d, presumably provide enough
hyperconjugative stabilization to the breaking C−C bonds to
facilitate decarbonylation. Fortuitously, we have only observed
reactive polymorphs of 29d and 29e.
Computational Analysis of C−C Bond Formation. In
order to better understand the role of substrate conformation
in reaction selectivity, we conducted a computational study of
the radical pairs 10a and 10c−f (Figures 8 and 9). At the
The dramatically different behavior of 28d (reactive
polymorph) relative to 28d (unreactive polymorph) and 28e
is likely controlled by substrate conformation in the crystalline
solid state (Figure 6c). Solid-state photodecarbonylation
requires stabilization of the breaking C−C sigma bonds by
neighboring π-systems. The extent of these hyperconjugative
interactions in substrate 28d (both reactive and unreactive
polymorphs) and 28e can be correlated to the dihedral angle
between the breaking C−C σ bond and the nearest C−C bond
of the aromatic π-system. A dihedral angle of 90° is ideal,
allowing for maximum orbital overlap (see bonds highlighted
in blue). Alternatively, if the dihedral angle is 0°, the C−C σ-
bond and π-system will be orthogonal, resulting in no
electronic stabilization (see bonds highlighted in red). In
considering substrate 28d (unreactive polymorph) and 28e,
the relevant dihedral angles are 82° and 18° in the former and
85° and 20° in the latter. The smaller of the two dihedral
angles for each substrate, 18° and 20°, presumably leads to
negligible orbital overlap and failed bond homolysis. The
relevant dihedral angles in 28d (reactive polymorph) are 64°
and 68°, which we surmise provide sufficient orbital overlap to
facilitate decarbonylation.
Having explored the reactivity of symmetric ketones 28 in
solid-state photochemistry, we next investigated nonsym-
metrical ketones 29 (Figure 7). To our delight, both 29d
Figure 8. QM/MM hybrid approach to model radical−radical
recombination in the crystalline solid state. “Ball-and-stick”
representation was used for the atoms treated by the DFT method
(ωB97X-D/6-311G(d,p) level of theory) and a “wireframe”
representation for the atoms in the low-level layer (UFF).
Figure 7. Solid-state photochemistry of ketones 29d and 29e (the R-
groups on imides 29d and 29e were removed from the X-ray
renderings for clarity). Crystal structures obtained by both X-ray
diffraction and microcrystal electron diffraction (microED).
and 29e underwent productive photodecarbonylation. Upon
exposure to UV light and subsequent pyrrolidinone depro-
tection, 29d and 29e efficiently provided 31d and 31e,
respectively. Unlike substrate 28d, the solid-state trans-
formations of 29d and 29e were reproducible even when
performed on >300 mg scale.54 These results were consistent
with our understanding of the role of solid-state conformation
and photochemical lability of the ketone carbonyl. The
relevant dihedral angles between breaking C−C bonds and
Figure 9. Computational investigation of radical pair 10.
4048
https://dx.doi.org/10.1021/jacs.1c01100
J. Am. Chem. Soc. 2021, 143, 4043−4054

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
outset, we recognized that effective computational simulation
of the crystalline lattice would be critical in achieving
meaningful results. While gas-phase and simple solvation
models would be unlikely to accurately simulate the crystalline
environment, full density functional theory (DFT) treatment
of the crystalline lattice would be untenable due to prohibitive
computational cost. Therefore, the ONIOM57 approach within
the hybrid QM/MM58 method was employed, in order to
estimate the intermolecular interactions within the crystal
lattice environment.59 The central part consisted of a single
radical pair computed with the unrestricted open-shell
dispersion-corrected ωB97X-D/6-311G(d,p) level of theory.60
The surrounding shell of molecules, extracted from the
experimentally determined crystal structure, was computed
with the electronically embedded UFF force field.61
Using this methodology, we investigated the recombination
barriers for radical pairs 10 in the solid state (Figure 9).
Radical pairs 10 were generated computationally by removing
the carbonyls from ketones 11 and then performing an energy
minimization calculation. Interestingly, the success of the
ground-state optimization for radical fragments was critically
dependent on embedding them in their respective crystal
cavities. In the gas phase, the radical pair would spontaneously
recombine to form the C−C bond. However, when conforma-
tionally restrained by the crystal cavity, the radical pair was an
energetic minimum and existed as a discrete species.
Transition states for the recombination reaction were then
identified for each of the radical pairs in their respective
lattices. Radical pairs 10a and 10c, which lack ortho
substitution, displayed the lowest barriers of 4.8 and 1.7
kcal/mol, respectively. Recombination barriers for ortho-
functionalized radical pairs 10d−f were uniformly higher,
with values between 6.1 and 6.8 kcal/mol. Given the successful
calculation of transition states for C−C bond formation from
radical pairs 10, we were additionally interested in probing the
competing disproportionation process. The results of these
calculations suggest that hydrogen atom tunneling may be
responsible for the observed disproportionation (for further
discussion, see SI).
After determining the recombination barriers for radical
pairs 10, we turned our attention to radical pair 22d (arising
from decarbonylation of ketone 28d) (Figure 10). For this
calculation, we used the coordinates derived from the reactive
polymorph of 28d. While radical pairs 10 all displayed small
but non-negligible barriers to C−C bond formation, the
corresponding process for radical pair 22d was barrierless. In
fact, all attempts at energy minimization of 22d resulted in
radical recombination to form the C−C bond. This result was
consistent with experimental results described previously,
which demonstrated that photodecarbonylation of ketone
28d (reactive polymorph) led to efficient C−C bond
formation (Figure 6a).
Figure 10. Computational investigation of radical pair 22d.
reaction, realized by the presence of an activation barrier.
Comparatively, the reactive polymorph of ketone 28d likely
undergoes a barrierless recombination due to a favorable
ground state conformation. This is reflected both by the
barrierless recombination observed in computational analysis
and by the high-yielding formation of photoproduct 30d.
Additionally, the magnitude of the barrier to radical
recombination for radical pairs derived from ketones 11 is
related to the presence of an ortho-substituent. The computa-
tional results clearly show a higher activation energy for
recombination when an ortho-substituent is present (6.1−6.8
kcal/mol) and a diminished barrier for radical pairs that lack
ortho substitution (1.7−4.8 kcal/mol).62
Summary of the Optimization and Mechanistic
Analysis of the Solid-State Photodecarbonylation
Reaction. The optimization and analysis en route to the
bis(cyclotryptamine) alkaloids gave insight into the role of
substrate conformation on the success or failure of the solid-
state photodecarbonylation reaction. It was found that the
performance of the transformation was highly sensitive to
remote substituent effects. While the reaction was low-yielding
for N-methyl ketones 11d−f, the solid-state photodecarbony-
lation reaction of N-acyl ketones 28d and 29d,e gave up to
61% yield of the desired vicinal quaternary stereocenter-
containing product. We hypothesize that changes to the solid-
state conformation of the substrate underpin these dramatic
substituent effects on photodecarbonylation. This hypothesis
was further supported computationally through study of the
C−C bond forming events for radical pairs 10a, 10c−f, and
22d. Additionally, a key stereoelectronic relationship between
the conformation of the substrate and the photolability of the
ketone carbonyl was identified. This led to dramatically
different reactivity of two polymorphic forms of 28d, the
failed reaction of 28e, and efficient decarbonylative C−C bond
formation from 29d,e. Ultimately, this optimization employed
crystal engineering of the solid-state conformation to improve
On the basis of the computational results depicted in Figures
9 and 10 and the experimental results previously discussed, we
were able to draw a number of conclusions about the key
photodecarbonylation reaction. Conformational restriction can
sometimes lead to activation barriers for recombination. In the
photodecarbonylation reaction of ketones 11a and 11c−f, the
computationally generated radical pairs 10a and 10c−f were
not energetic minima unless they were confined by the
crystalline lattice. This is consistent with our hypothesis that in
the solid state, unfavorable ground-state substrate conforma-
tions can lead to repulsive interactions in the course of the
4049
https://dx.doi.org/10.1021/jacs.1c01100
J. Am. Chem. Soc. 2021, 143, 4043−4054

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
the key step of our synthetic endeavor. To our knowledge, this
is the first use of a crystal engineering-based optimization in a
natural product total synthesis campaign.
Total Synthesis of a Bis(Cyclotryptamine) Alkaloid
with the Piperidinoindoline Framework. Having opti-
mized the solid-state photochemical methodology to assemble
the daunting vicinal quaternary stereocenter motif, we were
poised to elaborate 31 to access members of the bis-
(cyclotryptamine) alkaloid family. We first focused our efforts
on synthetic elaboration of 31d (Scheme 5). It was found that
effect amination of the aryl bromides, we found that a
modification of Ma’s copper-catalyzed azidation conditions
yielded the desired double azidation product, 33.64 We then
subjected 33 to LiAlH4 to enact a global reduction in an effort
to access an equivalent of biosynthon 8, and thereby the
bis(cyclotryptamine) framework. To our surprise, under
strongly reducing conditions, 33 underwent a skeletal
rearrangement to provide piperidinoindoline 36.65 While
there are several mechanistic possibilities to explain this
reaction, one plausible pathway involves double azide
reduction to give 34, followed by Lewis acid-mediated
transamidation to generate bis(oxindole) 35. Finally, reductive
cyclocondensation of 35 resulted in formation of piperidi-
noindoline 36. Although single-crystal X-ray diffraction
ultimately confirmed the structure, we were initially unsure if
the correct structural assignment of 36 was the piperidinoindo-
line depicted. As shown in Scheme 7, it was plausible that
intermediate 34 could have undergone double cyclodehydra-
tion to provide dehydrobhesine (37), followed by mono-
amidine reduction to give bhesine (38).66
Scheme 5. Failed Cross-Coupling Attempts Using
Bis(Arylchloride) 9d
a
Scheme 7. Possible Competing Pathway to Give 37 or 38
treatment of 31d with sodium hydride and iodomethane
furnished N-methylamide 9d in 85% yield. Next, we attempted
to install the requisite nitrogen substituents on the aromatic
rings through a metal-catalyzed double amination reaction.
Although robust methods for amination of hindered aryl
chlorides have been reported,63 attempts at amination of 9d
were uniformly unsuccessful. We surmise that the dense steric
encumbrance of the vicinal quaternary stereocenters and the
potential for the amide carbonyls in 9d to chelate catalysts in a
bidentate fashion is responsible for the poor reactivity.
Unabated by the difficulty of functionalizing the C−Cl bond
of 9d, we turned our efforts to the potentially more reactive
ortho-bromo variant, 9e (Scheme 6). Efficient N-methylation
of 31e was realized through a base-mediated alkylation to
provide 9e. After subjecting 9e to a variety of conditions to
a
Not observed.
Before we ultimately obtained the single-crystal X-ray
structure of 36, we sought to confirm the structural identity
of 36 by oxidizing the aminal to the corresponding amidine
(Scheme 8). To accomplish this transformation, 36 was
converted to bis(amidine) 39 under Ley−Griffith oxidation
conditions.67 Initially surmising that we may have generated
dehydrobhesine (37), we compared our synthetic material to
an authentic sample of 37 extracted from the leaves of the
Psychotria colorata flower.68 Upon NMR spectroscopic analysis,
it was observed that the natural sample contained a 7:1 mixture
Scheme 6. Cross-Coupling and Assembly of
“Dihydropsychotriadine” 36
Scheme 8. Total Synthesis of “Psychotriadine” 39
4050
https://dx.doi.org/10.1021/jacs.1c01100
J. Am. Chem. Soc. 2021, 143, 4043−4054

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
of 37 and a second, unknown isomeric alkaloid. Surprisingly,
the spectrum of our synthetic material 39 matched that of the
unknown alkaloid present in the natural material. On the basis
of this analysis and the unambiguous crystallographic
characterization of 36, we propose that our synthetic material
bears the depicted piperidinoindoline scaffold. Furthermore, its
presence in the extracts of Psychotria colorata suggests it is a
naturally occurring alkaloid that we have termed “psycho-
triadine.”
United States; orcid.org/0000-0002-8387-5261;
Email: mgg@chem.ucla.edu
Neil K. Garg − Department of Chemistry and Biochemistry,
University of California, Los Angeles, California 90095,
United States; orcid.org/0000-0002-7793-2629;
Email: neilgarg@chem.ucla.edu
Miguel A. Garcia-Garibay − Department of Chemistry and
Biochemistry, University of California, Los Angeles,
California 90095, United States; orcid.org/0000-0002-
6268-1943; Email: houk@chem.ucla.edu
CONCLUSIONS
■
Authors
We completed the first total synthesis of a naturally occurring
bis(cyclotryptamine) alkaloid featuring the elusive piperidi-
noindoline scaffold and confirmed its natural occurrence in the
Psychotria colorata flower. These studies demonstrate that all
five of the structurally distinct bis(cyclotryptamine) ring
systems originally proposed by R. B. Woodward and Robert
Robinson are, in fact, seen in natural products. Essential to the
success of this endeavor was the use of the solid-state
photodecarbonylation reaction to stereoselectively forge the
daunting vicinal quaternary stereocenter motif and a fortuitous
reductive rearrangement of bis(azide) 33 to give piperidinoin-
doline 36.
Jordan J. Dotson − Department of Chemistry and
Biochemistry, University of California, Los Angeles,
California 90095, United States
Ieva Liepuoniute − Department of Chemistry and
Biochemistry, University of California, Los Angeles,
California 90095, United States; orcid.org/0000-0003-
2965-8642
J. Logan Bachman − Department of Chemistry and
Biochemistry, University of California, Los Angeles,
California 90095, United States
Vince M. Hipwell − Department of Chemistry and
Biochemistry, University of California, Los Angeles,
California 90095, United States
Psychotriadine (39) marks the most complex, synthetically
challenging natural product accessed by solid-state photo-
chemistry to date. In deploying the key photodecarbonylation
reaction, two primary hurdles were faced: poor chemo-
selectivity for recombination and a lack of substrate reactivity.
Both of these challenges were overcome by manipulation of
the substrate conformation. To our knowledge, this is the first
use of crystal engineering to optimize a reaction in a total
synthesis endeavor.69 While these efforts showcase the power
of crystal engineering as a tool to optimize solid-state reactions,
the unpredictability of crystallization remains a major
challenge. As is reflected in this work, optimization of the
crystalline conformation of a substrate currently relies on trial-
and-error based experimentation. Although the use of easily
modifiable motifs expedites this process, de novo prediction of
crystal conformation represents the ideal solution. As such, the
advent of computational prediction of crystal structures would
offer a rapid means of predictive control of reactivity and
selectivity in solid-state organic reactions.70
Saeed I. Khan − Department of Chemistry and Biochemistry,
University of California, Los Angeles, California 90095,
United States
Complete contact information is available at:
https://pubs.acs.org/10.1021/jacs.1c01100
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The authors thank the National Science Foundation (CHE-
1855342 for M.G.G.) and the Trueblood Family (for N.K.G.)
for financial support. J.J.D. acknowledges the UCLA Graduate
Division for a Dissertation Year Fellowship. We thank
Professor Luisella Verotta (University of Milan, Italy) for
helpful discussions and for providing authentic samples of 37
and 38. We also thank Chih-Te Zee and Jose A. Rodriguez for
their generous help solving the microED structures of 29d and
29e. These studies were supported by shared instrumentation
grants from the NSF (CHE-1048804) and the National Center
for Research Resources (S10RR025631).
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/jacs.1c01100.
■
sı
Detailed experimental procedures and compound
characterization data (PDF)
REFERENCES
■
(1) For other reviews of photochemical reactions being used in
organic synthesis, also see references 2−4. Prier, C. K.; Rankic, D. A.;
MacMillan, D. W. C. Visible Light Photoredox Catalysis with
Transition Metal Complexes: Applications in Organic Synthesis.
Chem. Rev. 2013, 113, 5322−5363.
(2) Kärkäs, M. D.; Porco, J. A.; Stephenson, C. R. J. Photochemical
Approaches to Complex Chemotypes: Applications in Natural
Product Synthesis. Chem. Rev. 2016, 116, 9683−9747.
(3) Bach, T.; Hahn, J. Photochemical Reactions as Key Steps in
Natural Product Synthesis. Angew. Chem., Int. Ed. 2011, 50, 1000−
1045.
Accession Codes
CCDC 2006460−2006463 and 2006465−2006469 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.
AUTHOR INFORMATION
Corresponding Authors
K. N. Houk − Department of Chemistry and Biochemistry,
University of California, Los Angeles, California 90095,
■
(4) Nicholls, T. P.; Leonori, D.; Bissember, A. C. Applications of
Visible Light Photoredox Catalysis to the Synthesis of Natural
Products and Related Compounds. Nat. Prod. Rep. 2016, 33, 1248−
1254.
4051
https://dx.doi.org/10.1021/jacs.1c01100
J. Am. Chem. Soc. 2021, 143, 4043−4054

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(5) For other reviews of photochemical reactions conducted in the
solid state, also see references 6 and 7. Dotson, J. J.; Perez-Estrada, S.;
Garcia-Garibay, M. A. Taming Radical Pairs in Nanocrystalline
Ketones: Photochemical Synthesis of Compounds with Vicinal
Stereogenic All-Carbon Quaternary Centers. J. Am. Chem. Soc.
2018, 140, 8359−8371.
Perophoramidine, from the Philippine As-cidian Perophoranamei. J.
Org. Chem. 2002, 67, 7124−7126.
(24) Eccles, R. G. Calycanthine. Druggists’ Circular and Chemical
Gazette 1888, 32, 65.
(25) Robinson, R.; Teuber, H. J. Reactions with Nitrosodisulfonate.
IV. Calycanthine and Calycanthidine. Chem. Ind. (London) 1954,
783−784.
(26) Manske, R. H. The Alkaloids of Calycanthacaea. Alkaloids
1965, 8, 581−589.
(6) Ramamurthy, V.; Venkatesan, K. Photochemical Reactions of
Organic Crystals. Chem. Rev. 1987, 87, 433−481.
(7) Ramamurthy, V.; Sivaguru, J. Supramolecular Photochemistry as
a Potential Synthetic Tool: Photocycloaddition. Chem. Rev. 2016,
116, 9914−9993.
(27) While this structure has never been isolated from a natural
source, a structure with this ring system (albeit the meso isomer) was
reported synthetically. For further details, see the following reference.
Hall, E. S.; McCapra, F.; Scott, A. I. Biogenetic-Type Synthesis of the
Calycanthaceous Alkaloids. Tetrahedron 1967, 23, 4131−4141.
(28) Dotson, J. J.; Bachman, J. L.; Garcia-Garibay, M. A.; Garg, N. K.
Discovery and Total Synthesis of a Bis(Cyclotryptamine) Alkaloid
Bearing the Elusive Piperidinoindoline Scaffold. J. Am. Chem. Soc.
2020, 142, 11685−11690.
(8) Hernandez-Linares, M. G.; Guerrero-Luna, G.; Perez-Estrada, S.;
Ellison, M.; Ortin, M.-M.; Garcia-Garibay, M. A. Large-Scale Green
Chemical Synthesis of Adjacent Quaternary Chiral Centers by
Continuous Flow Photodecarbonylation of Aqueous Suspensions of
Nanocrystalline Ketones. J. Am. Chem. Soc. 2015, 137, 1679−168.
(9) Mortko, C. J.; Garcia-Garibay, M. A. Green Chemistry Strategies
Using Crystal-to-Crystal Photoreactions: Stereoselective Synthesis
and Decarbonylation of trans-α,α’-Dialkenoylcyclohexanones. J. Am.
Chem. Soc. 2005, 127, 7994−7995.
(29) When a ketone bearing an ortho-amine was synthesized,
spontaneous condensation of the amine onto the ketone to form the
corresponding spirocyclic aminal was observed.
(10) Ng, D.; Yang, Z.; Garcia-Garibay, M. A. Total Synthesis of
( )-Herbertenolide by Stereospecific Formation of Vicinal Quater-
nary Centers in a Crystalline Ketone. Org. Lett. 2004, 6, 645−647.
(11) Natarajan, A.; Ng, D.; Yang, Z.; Garcia-Garibay, M. A. Parallel
Synthesis of (+)- and (−)-)-Cuparenone by Radical Combination in
Crystalline Solids. Angew. Chem., Int. Ed. 2007, 46, 6485−6487.
(12) Yang, Z.; Photodecarbonylation of crystalline ketones and its
application to the synthesis of tochuinyl acetate, Ph.D. Thesis. University
of California, Los Angeles, 2003. Available from ProQuest
Dissertations & Theses Global database. (Accession Order No.
AAT 3133088).
(30) Although direct amination of the methoxy-substituted arene
may be challenging, 9f could also be converted to the corresponding
aryl triflate, which should be more reactive toward cross-coupling.
(31) The common synthetic route used to access of ketones 11 was
adapted from studies previously reported by our group. For details see
the following reference: Resendiz, M. J. E.; Natarajan, A.; Garcia-
Garibay, M. A. Diastereoselective Synthesis and Spin-Dependent
Photodecarbonylation of Di(3-Phenyl-2-Pyrrolidinon-3-yl)Ketones:
Synthesis of Nonadjacent and Adjacent Stereogenic Quaternary
Centers. Chem. Commun. 2008, 193−195.
(32) Dotson, J. J.; Garg, N. K.; Garcia-Garibay, M. A. Evaluation of
the Photodecarbonylation of Crystalline Ketones for the Installation
of Reverse Prenyl Groups on the Pyrrolidinoindoline Scaffold.
Tetrahedron [Online early access]. DOI: 10.1016/j.tet.2020.131181.
Published Online: Apr 10, 2020. https://www.sciencedirect.com/
science/article/abs/pii/S0040402020303161?via%3Dihub.
(33) It should be noted that all of the ketones that are disclosed in
this study crystallized as racemic mixtures and their photo-
decarbonylation resulted in racemic products. For discussion of the
solid-state photodecarbonylation being used to prepare enantioen-
riched products, see reference 2.
(13) Chang, T. Y.; Dotson, J. J.; Garcia-Garibay, M. A. Org. Lett.
2020, 22, 8855−8859.
(14) Peterson, E. A.; Overman, L. E. Contiguous Stereogenic
Quaternary Carbons: A Daunting Challenge in Natural Products
Synthesis. Proc. Natl. Acad. Sci. U. S. A. 2004, 101, 11943−11948.
(15) Long, R.; Huang, J.; Gong, J.; Yang, Z. Direct Construction of
Vicinal All-Carbon Quaternary Stereocenters in Natural Product
Synthesis. Nat. Prod. Rep. 2015, 32, 1584−1601.
(16) Liu, Y.; Han, S.; Liu, W.; Stoltz, B. M. Catalytic
Enantioselective Construction of Quaternary Stereocenters: Assembly
of Key Building Blocks for the Synthesis of Biologically Active
Molecules. Acc. Chem. Res. 2015, 48, 740−751.
(17) Pompeo, M. M.; Cheah, J. H.; Movassaghi, M. Total Synthesis
and Anti-Cancer Activity of All Known Communesin Alkaloids and
Related Derivatives. J. Am. Chem. Soc. 2019, 141, 14411−14420.
(18) For other reviews on the isolation and synthesis of
bis(cyclotryptamine) alkaloids, also see references 18−21. May, J.
A.; Stoltz, B. The Structural and Synthetic Implications of the
Biosynthesis of the Calycanthaceous Alkaloids, the Communesins,
and Nomofungin. Tetrahedron 2006, 62, 5262−5271.
(19) Schmidt, M. A.; Movassaghi, M. New Strategies for the
Synthesis of Hexahydropyrroloindole Alkaloids Inspired by Bio-
synthetic Hypotheses. Synlett 2008, 2008, 313−324.
(20) Steven, A.; Overman, L. E. Total Synthesis of Complex
Cyclotryptamine Alkaloids: Stereocontrolled Construction of Quater-
nary Carbon Stereocenters. Angew. Chem., Int. Ed. 2007, 46, 5488−
5508.
(21) Trost, B. M.; Osipov, M. Recent Advances on the Total
Synthesis of Communesin Alkaloids and Perophoramidine. Chem. -
Eur. J. 2015, 21, 16318−16343.
(22) Xu, J.-B.; Cheng, K.-J. Studies on the Alkaloids of the
Calycanthaceae and their Syntheses. Molecules 2015, 20, 6715−6738.
(23) The fifth scaffold, not depicted, has only been observed in a
higher oxidation state and possesses opposite relative stereochemistry
compared to 4-7. For this scaffold, see Verbitski, S. M.; Mayne, C. L.;
Davis, R. A.; Concepcion, G. P.; Ireland, C. M. Isolation, Structure
Determination, and Biological Activity of a Novel Alkaloid,
(34) Note that starting material (11a) remained after irradiation.
The yield based on remaining starting material is 54% (9a).
(35) The loss of mass balance was attributed to unidentified,
insoluble decomposition products.
(36) The yield of disproportionation was determined based on the
yield of 17a.
(37) It was noted that a significant amount of mass was lost in the
photodecarbonylation reactions that were low yielding (reactions of
ketones 11d, 11e, and 11f). In these cases, we observed a significant
quantity of insoluble solid material following irradiation. We were
unable to identify the chemical identity of this unknown solid but
believe it to be the primary source of lost mass balance.
(38) Note that starting material (11d) remained after irradiation.
The yield based on remaining starting material is 17% (9d) and 24%
(17d).
(39) Even substrates that exclusively undergo disproportionation
upon solution-phase irradiation show full suppression of the
disproportionation pathway in the crystalline solid state. For more
details, see Choe, T.; Khan, S. I.; Garcia-Garibay, M. A. Combination
vs. Disproportionation in Dialkyl Biradicals. Selectivity Reversal in a
Crystalline Solid. Photochem. Photobiol. Sci. 2006, 5, 449−451.
(40) Anslyn, E. V.; Dougherty, D. A. Modern Physical Organic
Chemistry; University Science Books: Sausalito, CA, 2006; pp 104.
(41) Consistent with our experimental results, this repulsive
interaction results in diminished recombination yields for ketones
with larger ortho moieties.
4052
https://dx.doi.org/10.1021/jacs.1c01100
J. Am. Chem. Soc. 2021, 143, 4043−4054

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(42) The topochemical principle states that reactions in the
crystalline solid state will favor pathways that feature a minimum
amount of atomic or molecular movement. For further discussion of
the topochemical principle, see reference 2c and the following book:
Schmidt, G. M. J.; et al. in Solid State Photochemistry (Ed.: Ginsburg,
D.) Verlag Chemie: Weinheim, 1976; pp, 1−280.
(43) We were also interested in a directed C-H amination of 9c;
however, we recognized that the meta-hydroxy moieties would need
to be reductively cleaved to provide the bis(pyrrolidinoindoline)
alkaloids. Functionalization of 9a would lead to a more ideal synthesis
and was therefore prioritized in these studies. For functionalization of
C(sp2)-H bonds ortho to an amide directing group, also see
references 44,45. Qiu, F.-C.; Yang, W.-C.; Chang, Y.-Z.; Guan, B.-T.
Palladium-Catalyzed ortho-Halogenation of Tertiary Benzamides.
Asian J. Org. Chem. 2017, 6, 1361−136.
reference 56 Rodriguez, J. A.; Eisenberg, D. S.; Gonen, T. Taking the
measure of MicroED. Curr. Opin. Struct. Biol. 2017, 46, 79−86.
(56) Jones, C. G.; Martynowycz, M. W.; Hattne, J.; Fulton, T. J.;
Stoltz, B. M.; Rodriguez, J. A.; Nelson, H. M.; Gonen, T. The
CryoEM Method MicroED as a Powerful Tool for Small Molecule
Structure Determination. ACS Cent. Sci. 2018, 4, 1587−1592.
(57) Dapprich, S.; Komáromi, I.; Byun, K. S.; Morokuma, K.; Frisch,
M. J. A New ONIOM Implementation in Gaussian98. Part I. The
Calculation of Energies, Gradients, Vibrational Frequencies and
Electric Field Derivatives. J. Mol. Struct.: THEOCHEM 1999, 461, 1−
21.
(58) Vreven, T.; Byun, K. S.; Komáromi, I.; Dapprich, S.;
Montgomery, J. A.; Morokuma, K.; Frisch, M. J. Combining Quantum
Mechanics Methods with Molecular Mechanics Methods in ONIOM.
J. Chem. Theory Comput. 2006, 2, 815−826.
(59) This type of QM/MM strategy to model a reaction in the
crystalline solid state was reported in the following reference: Keating,
A. E.; Shin, S. H.; Houk, K. N.; Garcia-Garibay, M. A. Combining
Quantum Mechanical Reaction Pathways with Force Field Lattice
Interactions to Model a Solid-State Phototransformation. J. Am.
Chem. Soc. 1997, 119, 1474−1475.
(44) Wykypiel, W.; Lohmann, J.-J.; Seebach, D. Lithiation in W-
Position to the N-Atom of Triphenylacetamides from Cyclic
Secondary Amines. Rearrangements of Metalated Triphenylaceta-
mides by 1,3-Shift of Carbamoyl Groups. Helv. Chim. Acta 1981, 64,
1337−1346.
(45) Yeung, C. S.; Zhao, X.; Borduas, N.; Dong, V. M. Pd-Catalyzed
ortho-Arylation of Phenylacetamides, Benzamides, and Anilides with
Simple Arenes Using Sodium Persulfate. Chem. Sci. 2010, 1, 331−336.
(46) For a review of directed aromatic lithiation of tertiary amides,
see: Beak, P.; Snieckus, V. Directed Lithiation of Aromatic Tertiary
Amides: An Evolving Synthetic Methodology for Polysubstituted
Aromatics. Acc. Chem. Res. 1982, 15, 306−312.
(47) We did not investigate electrophilic or radical mediated
amination of 9a, because such transformations would likely suffer
from issues of poor site-selectivity (i.e., ortho vs para functionaliza-
tion). This could conceivably be mitigated with removable blocking
groups installed at the 4-position of the phenyl rings, however, such
strategies were not pursued.
(60) Chai, J. D.; Head-Gordon, M. Long-Range Corrected Hybrid
Density Functionals with Damped Atom-Atom Dispersion Correc-
tions. Phys. Chem. Chem. Phys. 2008, 10, 6615−6620.
(61) We favor the employment of a molecular mechanics method to
describe the low-level layer (over semiempirical methods or quantum
chemical calculations with very small basis sets) since the MM force
field by definition includes a description of dispersion effects between
all atoms in the MM region and between QM and MM regions.
Indeed, this hybrid approach has been shown to be accurate in similar
computational studies of solid-state transformations of stilbene where
crystalline environment calculations were included in the low-level
layer (see reference 59).
(48) Formation of the carbanion was determined by quenching with
D2O followed by NMR spectroscopic analysis.
(62) We do not consider the ordering of barriers for 10d-f to be
meaningful as they fall within a 0.7 kcal/mol range and are therefore
within computational error of one another.
(63) For a review of catalytic amination of aryl halides, see Heravi,
M. M.; Kheilkordi, Z.; Zadsirjan, V.; Heydari, M.; Malmir, M.
Buchwald-Hartwig Reaction: An Overview. J. Organomet. Chem. 2018,
861, 17−104.
(64) Unfortunately, despite extensive efforts, 33 could not be fully
purified and was, therefore, carried forward directly in the subsequent
step.
(65) Note that the double azidation reaction to convert 9e to 33 was
challenging to enact and resulted in < 35% yield (based on 1H NMR
analysis of the semipure material). As such, the reductive cyclization
to convert 33 into 36 likely took place in > 75% yield.
(66) The observed 1H and 13C NMR and spectra of 36 were not
identical to the published spectra of bhesine (38). We hypothesized
that we may have either produced an aminal isomer of 38 or subtle
changes in solvent acidity may have shifted the 1H and 13C
resonances. For the first isolation of 38, see Balayer, A.; Sévenet, T.;
Schaller, H.; Haudi, A. H. A.; Chiaroni, A.; Riche, C.; Païs, M.
Dihydroquinoline-type Alkaloids from Bhesa Paniculata, Celastraceae.
Nat. Prod. Lett. 1993, 2, 61−67.
(49) Compound 20 was easily prepared in one step from a ketone
previously reported by our group. For details, see reference 31.
(50) Since we anticipated that aryl halides would be ideal latent
electrophiles for metal-catalyzed amination, we pursued the synthesis
of chloro- and bromoketones 27d and 27e.
(51) The occurrence of a new, more stable polymorph that
precludes formation of the desired polymorph is well-known and is
often referred to as the “disappearing polymorph problem.” For
further discussion of this topic, also see reference 52 Cruz-Cabeza, A.
J.; Bernstein, J. Conformational Polymorphism. Chem. Rev. 2014, 114,
2170−2191.
(52) Snider, D. A.; Addicks, W.; Owens, W. Polymorphism in
Generic Drug Product Development. Adv. Drug Delivery Rev. 2004,
56, 391−395.
(53) While highly irreproducible, several successful crystallizations to
provide the reactive polymorph were accomplished when they were
conducted in a different lab room, using new glassware, and new lab
coat to avoid contamination of the unreactive polymorph. In these
instances, small quantities of the reactive polymorph were used to
seed crystallization.
(54) Larger scale transformations were not attempted. However, we
surmise that the inherent scalability of suspension photochemistry
would permit increased scale up without loss of reaction performance.
See the following references for the use of crystalline suspensions to
scale up the solid-state photodecarbonylation reaction. (For examples,
see reference 8 and Veerman, M.; Resendiz, M. J. E.; Garcia-Garibay,
M. A. Large-Scale Photochemical Reactions of Nanocrystalline
Suspensions: A Promising Green Chemistry Method. Org. Lett.
2006, 8, 2615−2617.
(55) The structures of 29d and 29e were elucidated using both
single-crystal X-ray diffraction as well as microED. Both methods
resulted in nearly identical solutions for both structures. For a side-by-
side comparison, see Figure S1 in the Supporting Information. For
further discussion of microED-based structural determination, also see
(67) Oxidation conditions adapted from: Higuchi, K.; Sato, Y.;
Tsuchimochi, M.; Sugiura, K.; Hatori, M.; Kawasaki, T. First Total
Synthesis of Hinckdentine A. Org. Lett. 2009, 11, 197−199.
(68) Verotta, L.; Pilati, T.; Tató, M.; Elisabetsky, E.; Amaor, T. A.;
Nunes, D. S. Pyrrolidinoindoline Alkaloids from Psychotria colorata. J.
Nat. Prod. 1998, 61, 392−396.
(69) We anticipate that the rapid structural determination enabled
by microED technology will streamline the use of crystal engineering-
based reaction optimization in future studies. Only two structures
(29d and 29e) were obtained using microED in this report because
the requisite technological capabilities were not available to us until
the end of this study.
(70) Computational prediction of crystal structures has been
recognized as an unsolved problem in the field of computational
4053
https://dx.doi.org/10.1021/jacs.1c01100
J. Am. Chem. Soc. 2021, 143, 4043−4054

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
chemistry. For more discussion, see the following review: Houk, K.
N.; Liu, F. Holy Grails for Computational Organic Chemistry and
Biochemistry. Acc. Chem. Res. 2017, 50, 539−543.
4054
https://dx.doi.org/10.1021/jacs.1c01100
J. Am. Chem. Soc. 2021, 143, 4043−4054