Total Synthesis and Computational Investigations of Sesquiterpene-Tropolones Ameliorate Stereochemical Inconsistencies and Resolve an Ambiguous Biosynthetic Relationship

Article abstract of DOI:10.1021/jacs.1c02150

The sesquiterpene-tropolones belong to a distinctive structural class of meroterpene natural products with impressive biological activities, including anticancer, antifungal, antimalarial, and antibacterial. In this article, we describe a concise, modular

Full text of DOI:10.1021/jacs.1c02150

pubs.acs.org/JACS
Article
Total Synthesis and Computational Investigations of Sesquiterpene-
Tropolones Ameliorate Stereochemical Inconsistencies and Resolve
an Ambiguous Biosynthetic Relationship
Christopher Y. Bemis,∥ Chad N. Ungarean,∥ Alexander S. Shved, Cooper S. Jamieson, Taehwan Hwang,
Ken S. Lee, K. N. Houk, and David Sarlah*
Cite This: J. Am. Chem. Soc. 2021, 143, 6006−6017
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: The sesquiterpene-tropolones belong to a distinctive
structural class of meroterpene natural products with impressive
biological activities, including anticancer, antifungal, antimalarial, and
antibacterial. In this article, we describe a concise, modular, and
cycloaddition-based approach to a series of sesquiterpene mono- and
bistropolones, including (−)-epolone B, (+)-isoepolone B, ( )-dehy-
droxypycnidione, and (−)-10-epi-pycnidione. Alongside the develop-
ment of a general strategy to access this unique family of metabolites
were computational modeling studies that justified the diastereose-
lectivity observed during key cycloadditions. Ultimately, these studies
prompted stereochemical reassignments of the pycnidione subclass
and shed additional light on the biosynthesis of these remarkable
natural products.
both pycnidione (1′) and its deoxygenated equivalent
dehydroxypycnidione (2′) are bistropolones, while epolone
A (3′) and epolone B (4′) bear a phenol and (E)-trisubstituted
olefin in lieu of a second tropolone, respectively. The
diastereomeric eupenifeldin series maintains an identical
trend among compounds but is differentiated from the former
by the syn-relationship at the eastern dihydropyran seam (i.e.,
5−7) and a (Z)-trisubstituted olefin in the monotropolone
neosetophomone B (8).27 Importantly, the relative stereo-
chemistry of both flagship members, pycnidione (1′) and
eupenifeldin (5), was determined by single crystal X-ray
diffraction, their absolute stereochemistry was deduced using
circular dichroism studies,1,24−27 and the absolute stereo-
chemistry of other members within each series was assigned by
analogy to 1′ and 5.
INTRODUCTION
■
The sesquiterpene-tropolones are a unique subset of
meroterpenes, with pycnidione (1′, Figure 1a) serving as the
first representative member to be isolated from cultures of
Phoma sp. (MF 4728) in 1993.1 With other tropolone-
containing natural products2−6 and compounds7−14 boasting
impressive bioactivities, pycnidione follows suit by demonstrat-
ing a range of cellular functions including inhibition of
stromelysin (MMP3) and the topoisomerases,1 induction of
erythropoietin (EPO) gene expression,15 HIF-1α protein
accumulation15 and nuclear translocation,15 and HIF-1 DNA
binding and reporter gene transactivation.15 Pycnidione is
strongly cytotoxic to human cancer cell lines (e.g., IC50 = 3.5
nM, HCT-116)16 and also possesses effective antifungal,17,18
antimalarial,19−21 antibacterial,17 and antithelmintic proper-
ties,22 all in the micromolar range. Structurally homologous
compounds dehydroxypycnidione (2′),17 epolone A (3′), and
epolone B (4′) were later isolated alongside pycnidione (1′),23
and over time, a parallel collection was assembled upon the
isolation of additional congeners.24−27 While detailed
structure−activity relationships have yet to be compiled, the
entirety of the class maintains similar biological activity, with
bistropolones generally exhibiting the highest potency and
methylation of this motif rendering the compounds inactive.16
All sesquiterpene-tropolones share a common 11-membered
humulene-derived core displaying either pendent tropolones or
phenols, but subtle features of and variations between
congeners should be highlighted. Within the pycnidione series,
Despite the value of their collective biological effects, no
member of the sesquiterpene-tropolones has ever succumbed
to total synthesis, with only the Baldwin group reporting
several elegant model studies28−33 scrutinizing a biosynthetic
hypothesis (Figure 1b) insinuated in the epolone B (4′)
Received: February 25, 2021
Published: April 7, 2021
https://doi.org/10.1021/jacs.1c02150
J. Am. Chem. Soc. 2021, 143, 6006−6017
© 2021 American Chemical Society
6006

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 1. (a) Originally reported structures of representative members of the sesquiterpene-tropolones belonging to the pycnidione and
eupenifeldin series. (b) Biosynthetic proposal for pycnidione series. (c) Prior key synthetic studies. (d) Biosynthetic studies of epolone B (4′).
Figure 2. Retrosynthetic analysis of pycnidione (1′), dehydroxypycnidione (2′), and epolone B (4′).
isolation report.23 It was postulated that tropolone 9 could
serve as a viable precursor to the reactive tropolone ortho-
quinone methide (o-QM) 10, providing sesquiterpene-
tropolones through modular [4 + 2] inverse electron demand
hetero-Diels−Alder (HDA) reactions with hydroxyhumulene
11. Thus, HDA-based union of tropolone o-QM 10 and the
trisubstituted olefin of macrocycle 11 was envisioned to deliver
epolone B (4′). In an identical fashion, a second HDA was
anticipated to effectively transform 4′ to the corresponding
biscycloadduct pycnidione (1′). Guided by their hypothesis,
the Baldwin group conducted a proof-of-concept study (Figure
1c) by generating transient tropolone o-QM 10 through
pyrolysis of 14 in a neat solution of α-humulene (12),
observing formation of dehydroxyepolone B (15′) as a sole
product. However, due to limited supply of o-QM precursor
14, requiring 20 steps to prepare from commercial furan 13, no
further investigations involving second cycloaddition were
attempted.
While this experimental evidence supported the proposed
biosynthetic pathway, several key questions persisted, most
importantly if hydroxyhumulene 11 was a competent partner
for the cycloaddition. Recently, in vitro biosynthetic and
6007
https://doi.org/10.1021/jacs.1c02150
J. Am. Chem. Soc. 2021, 143, 6006−6017

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 3. Preparation of HDA cycloaddition partners. (a) Synthesis of tropolones 19 and 20. (b) Synthesis of 10-hydroxy-α-humulene (11).
computational studies by Hu and Houk successfully identified
a Diels−Alderase enzyme, EupfF, responsible for the selective
formation of neosetophomone B (8).34 Moreover, these
studies revealed the HDA reaction of 9 and 11 occurred
spontaneously in the absence of EupfF (Figure 1d), generating
an equimolar amount of epolone B (4′) and its diastereomer
isoepolone B (16′). However, since isomeric 16′ was not
previously found from producing organisms, the role of EupfF
further obscures the HDA hypothesis when compared with the
isolation data. It should be clarified that 11 is not a native
substrate of EupfF; therefore, a separate enzyme may be
operable for the selective formation of epolone B and further
transformation to pycnidione. Importantly, no biscycloadducts
(i.e., 1′, 2′, 5, and 6) were detected in either case, absolving
EupfF of the responsibility for this biosynthetic step. However,
the Cox group has recently seen success in detecting formation
of dehydroxypycnidione (2′) through incorporation of the
PycR1 Diels−Alderase in an artificial biosynthetic pathway.35
Finally, by refocusing on phenol-containing natural products 3′
and 7, the same group identified enzymes responsible for a
unique ring contraction of a single tropolone moiety within the
bistropolone eupenifeldin (5), forming noreupenifeldin (7).35
Described below is our synthetic campaign toward the
sesquiterpene-tropolones in the pycnidione series. We
accomplished a concise and scalable preparation of several
tropolone o-QM precursors and hydroxyhumulene. Utilizing
these as cycloaddends, we conducted systematic experimental
and computational investigations of their corresponding HDA
reactions, resulting in the synthesis of several sesquiterpene-
tropolones, and clarifying the origin of the peculiar
diastereoselectivities manifest in their formation. In contrast
to preceding assumptions, this research revealed that a second
cycloaddition transforming epolone B to pycnidione is not a
viable synthetic pathway. Moreover, we unambiguously
demonstrated that the absolute stereochemical configuration
of all members within the pycnidione series had previously
been misassigned. Together, these studies provide a rapid and
selective access to highly substituted tropolones, sesquiter-
penes, and their HDA adducts, as well as foundation for their
future (bio)synthetic considerations.
RESULTS AND DISCUSSION
■
Synthetic Planning. The enticing modularity and brevity
apparent in the biosynthetic hypothesis inspired the bio-
6008
https://doi.org/10.1021/jacs.1c02150
J. Am. Chem. Soc. 2021, 143, 6006−6017

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
mimicry in our synthetic blueprint toward the sesquiterpene-
tropolones. While appraising the competency of hydroxylated
humulene 11, we expected this investigation of the HDA
reaction to provide important insight into the biosynthesis of
these unique natural products (Figure 2). Moreover, since
issues with low material throughput had halted previous
synthetic investigations (i.e., 14, Figure 1c), we were adamant
that the route to all cycloaddends be practical and scalable.
By favoring this modular strategy over a linear approach, the
main synthetic challenge was simplified to the preparation of
two cycloaddition partners, 10-hydroxyhumulene (11) and
tropolone o-QMs (such as 17 or 18), the latter envisioned to
be formed in situ through pyrolytic cleavage of acetal precursor
19 or 20. Inspired by de Mayo’s renowned tactic,36−38 the [2 +
2] photocycloaddition of 25 and 26 followed by base-induced
fragmentation of 27 to γ-tropolone 28,39,40 we decided to
explore a variation of this process that could lead to suitably
functionalized α-tropolones (21 → 19 or 20). Cyclobutane 21,
which contains all desired functionality and a preinstalled
halide to relay the proper tropolone oxidation state, could be
traced back to intermediate 22 through an intramolecular
enone-olefin [2 + 2] cycloaddition reaction.41,42
A tantalizing, single-step allylic oxidation of α-humulene
(12) presented a direct route to 11; however, the inherent
challenges of regioselective chemical oxidation were antici-
pated and further supported by studies performed in our
laboratories as well as others.43 Aware of the recent work from
the Shenvi group disclosing the hydrogen atom transfer
(HAT) mediated retrocycloisomerization of (−)-caryophyl-
lene oxide (29) to humulene oxide 30,44 we envisioned an
application of this methodology that might provide an
alternative route toward the target molecule 11. This
advantageous disconnection would grant rapid access to the
desired compound, since the appropriately hydroxylated
caryophyllene oxide derivative 23 could be swiftly discon-
nected via α-oxidation and olefination, revealing readily
available (−)-kobusone (24) as a suitable chiral starting
material.45,46
Synthesis of Precursors 19 and 20. The synthesis of o-
QM precursors 19 and 20 commenced with O-alkylation of
1,3-cyclopentanedione (31) with the chloromethyl ether of
allylic alcohol 32, formed in situ with paraformaldehyde and
chlorotrimethylsilane,47 furnishing vinyl chloride 22 in 89%
yield (Figure 3a). Exposure of this compound to UV
irradiation promoted an intramolecular [2 + 2] cycloaddition,
which delivered tricycle 33 in 74% yield as an inconsequential
mixture of diastereomers (5:1). Adjustment of the oxidation
state (ketone 33 → 1,2-dicarbonyl 21) for the key
fragmentation step proved troublesome. All classical α-
oxidation conditions led to the decomposition of starting
material, owing to the sensitivity of the tertiary chloride at the
β-position. Gratifyingly, a recently reported one-pot α-
iodination/Kornblum oxidation, which was used for the
oxidation of cyclohexanones into catechols,48 proved suitable
for this task and delivered the desired dicarbonyl in the form of
hydroxyl enone 21 in 52% yield. Having embedded all
necessary functionality into the tricyclic structure 21, we
aimed to induce a de Mayo fragmentation. However, all
attempts to fragment the cyclobutane ring through exposure of
21 to standard Brønsted acid or acid/base media were met
with failure. Decomposition of 21 was pervasive, despite an
extensive screen of reaction conditions, and in some instances
we were able to detect unexpected tropolone 34, likely formed
via acetal deprotection, de Mayo fragmentation, and retro-aldol
reaction.
Realizing that standard fragmentation conditions were not
viable for preparation of desired tropolone 19, we strove to
identify milder conditions that would allow for retention of the
1,3-dioxane ring. The Bach group recently reported a Lewis
acid mediated fragmentation (Figure 3a, inset), wherein the
oxocarbenium 37, generated from fragmentation of cyclo-
butane 36 by boron trifluoride diethyl etherate, was trapped by
allyl- or hydrosilanes (37 → 38).49 In our case, precursor 21
bears an α-proton in place of the quaternary center in 36/37;
thus, we postulated that, in the absence of external
nucleophiles, oxocarbenium 39 could undergo deprotonation
at this position, perhaps by an eliminated halide anion,
restoring the 1,3-dioxane structure and bringing the seven-
membered ring into a full conjugation. Fortuitously, treating
21 with excess boron trifluoride diethyl etherate induced the
de Mayo type fragmentation as planned, providing tropolone
19 in 59% yield on a gram scale. To obtain unambiguous proof
of the structure, the tropolone was readily crystallized from
acetone and analyzed by single-crystal X-ray diffraction. While
the retention of boron difluoride was unexpected after basic
workup, it was not entirely surprising, as BF2-tropolone
complexes have been documented.50 Compound 19 was
remarkably bench-stable and was easily purified by silica gel
chromatography or recrystallization.
While the multifaceted capability of the BF2-group to mask
the tropolone and potentially enhance the reactivity of the o-
QM in the planned HDA reaction is of note, we also explored
more traditional protecting groups, such as methoxytropolone
variant 20. Accordingly, the deprotection of tropolone
difluoroborate 19 and subsequent methylation of free
tropolone could be accomplished; however, this approach
was plagued with low yields and delivered 20 in an inseparable
mixture of constitutional isomers, one of which was inactive in
the HDA reaction (see Supporting Information (SI) for
detail). The hydroxyenone 21 presented a convenient
opportunity to differentiate the two oxygens, and we were
keen to explore alternative fragmentations of premethylated
precursor 35 that could circumvent the troublesome
methylation step. Intermediate 35 was readily obtained from
21 with dimethylsulfate; however, treatment with a variety of
Lewis acids to effect its fragmentation as before provided only
complex mixtures. At this point, alternative rearrangement
pathways were explored, with attention given to the tertiary
halide as a potential handle for initiating the fragmentation.
Indeed, we screened silver salts and found that silver
tetrafluoroborate produced 20 as the sole tropolone product
in 78% yield, whose structure was also confirmed by single
crystal X-ray diffraction analysis. However, our excitement
about this new and improved synthesis was quickly quelled by
the finding that the fragmentation was impractically slow on
scales above 200 mg, and considering the requirement of
stoichiometric amounts of expensive silver salt, this route was
not selected for further scale-up campaigns. Nevertheless, these
findings fomented a deeper exploration of the fragmentation
behavior of 35 and ultimately led to the discovery that
irradiation of benzene solutions with UV light (254 nm)
accomplished the same conversion to 20 in 66% yield on a
multigram scale. This unprecedented and direct photochemical
fragmentation serves as an advantageous step with the ability
to selectively generate the desired constitutional isomer of
6009
https://doi.org/10.1021/jacs.1c02150
J. Am. Chem. Soc. 2021, 143, 6006−6017

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 4. (a) HDA cycloaddition studies involving α-humulene (12) and synthesis of dehydroxypycnidione (2′). (b) Optimized geometry of 2′
and (c and d) scatter plots of predicted 13C NMR shifts of 2′ as compared to both observed diastereoisomers. Insets correspond to error
distributions. Diagonal dashed line: y = x.
methyl tropolone 20. Notably, photosensitizer additives had no
effect, and this transformation could not be induced thermally.
Synthesis of 10-Hydroxy-α-humulene (11). With
tropolone o-QM precursors 19 and 20 in hand, we turned
our focus toward the construction of 10-hydroxy-α-humulene
(11, Figure 3b). While the synthesis of racemic 11 has been
accomplished in 12 steps,51 we decided to pursue the design
and execution of a more streamlined, scalable, and
enantioselective approach. Accordingly, the synthesis began
with readily available kobusone (24), which was treated with
potassium hexamethyldisilazane and Davis’ oxaziridine,52
furnishing hydroxyketone 40 as a single diastereoisomer
(87% yield), the structure of which was confirmed by single-
crystal X-ray diffraction analysis. Silyl protection of the newly
formed α-hydroxy motif and Wittig olefination delivered key
isomerization precursor 23 in 76% overall yield. Using
conditions reported by Shenvi,44 HAT-induced retrocycloiso-
merization of 23 proceeded smoothly to afford epoxide 41 in
95% yield, which was deprotected to alcohol 42 and verified
via X-ray crystallography. The use of an epoxide to disguise the
trisubstituted alkene throughout the sequence was paramount
to its success, as the unmasked motif admitted unwanted side
reactions under the HAT conditions. The final two steps,
stereospecific Re-catalyzed deoxygenation53 of this epoxide
and silyl deprotection of the appended hydroxyl group,
provided the desired macrocycle 11 in 80% overall yield, or
in 73% yield if combined into a one-pot process. Importantly,
while the spectroscopic data for 11 were consistent with the
literature, our synthetic sample had an opposite optical
Dehydroxypycnidione HDA Cycloaddition Studies. As
a prelude to the biomimetic HDA reactions of 19/20 and 11
that we envisaged would deliver the pycnidione family of
sesquiterpene-tropolones, we first profiled the cycloaddition
behavior of 19/20 in the simpler α-humulene system
demonstrated by Baldwin (Figure 4a).30 During initial
experiments, the performance of tropolone o-QM precursor
19 or 20 was evaluated by individually heating them in the
presence of α-humulene (12). Gratifyingly, this study revealed
that both 19 and 20 were competent substates for HDA
reaction under microwave irradiation with 12, generating
monocycloadducts 43 (51% from 19) and 44 (82% from 20)
as single constitutional isomers. Both sesquiterpene-tropolone
products were successfully deprotected under basic conditions
(K2CO3/MeOH, 99% yield from 43; or NaOH, MeOH/H2O,
80% yield from 44) to give racemic dehydroxyepolone B (15′).
Though a large excess (3 equiv) of tropolone o-QM precursors
19 or 20 was used to obtained practical yields, the potential
bisadducts were detected only in trace amounts by LC-MS
analysis of the crude reaction mixtures. These results indicated
that pyrolytically generated o-QM 17 or 18 readily undergoes
the first cycloaddition with humulene; however, we speculated
that their rapid decomposition prevented further union with
the remaining trisubstituted olefin. Moreover, during these
experiments we observed pronounced decomposition of
tropolone difluoroborates (i.e., 19 and 43), which was
attributed to the high temperatures required for this cyclo-
addition to occur. Therefore, to enforce the second cyclo-
addition and to avoid thermolytic degradations, the methylated
monocycloadduct 44 was taken further with the corresponding
o-QM precursor 20 added slowly as a solution over the course
of several hours. We were pleased to find that such an
approach resulted in the formation of biscycloadduct 45, albeit
with a poor yield and diastereoselectivity (13% yield and 1:1
d.r.). Separation of this diastereomeric mixture, followed by
base-induced deprotection, delivered ( )-dehydroxy-
pycnidione (2′).
rotation compared to the material that was obtained and
23
used in tropolone-sesquiterpene biosynthetic studies {[α]D
=
−21.1 (c = 0.04, CH3OH); lit.,34 [α]D = +17.5 (c = 0.04,
CH3OH)}. Finally, the described fragmentation approach to
construct such a macrocycle compared favorably to alternative
routes that were initially explored, involving classical ring-
forming strategies used in both the synthesis of α-humulene
(12) and other macrocycles,54−60 such as Nozaki−Hiyama−
Kishi and McMurry couplings, as well as ring-closing
metathesis (see SI for details).
20
Computed 13C NMR Spectra of Dehydroxypycnidione
(2′). We were fortunate to be provided with 1H NMR spectral
data of dehydroxypycnidione (2′) from the isolation group,17
6010
https://doi.org/10.1021/jacs.1c02150
J. Am. Chem. Soc. 2021, 143, 6006−6017

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 5. (a) HDA cycloaddition studies involving (−)-10-hydroxyhumulene (11) and synthesis of (+)-isoepolone B (ent-16) and (−)-epolone B
(ent-4). (b) Computed transition state (TS) energies for the reaction of 11 and 20 to form 46 and 47. (c) Computed TS energies for the reaction
of 20 and 46. (d) Computed TS energies for the reaction of 20 and 47. Calculations with ωB97X-D-CPCM(Mesitylene)/def2-QZVPP//ωB97X-
D/6-31G(d).
which matched with the structure that we had synthesized.
However, considering that compound 2′ was obtained as a
mixture of diastereoisomers with high structural and
spectroscopic homogeneity, additional characterization data,
specifically the 13C NMR, were crucially needed for additional
structural confirmation. Thus, in lieu of the missing and
unavailable data,61 we took initiative to validate the structure of
this compound by performing density functional theory (DFT)
calculations to predict the 13C NMR chemical shifts of 2′.62
After a short screening of functionals, we turned our attention
to PBE0/pcSseg-2//ωB97X-D3/def2-TZVP(−f), as imple-
mented in ORCA 4.2.1.63−65 To our delight, on a set of
small molecules, used for empirical determination of correction
factors,66 this method displayed minor errors (MAE, mean
average error = 1.2 ppm, MAX, maximum absolute error = 3.4
ppm), thus emerging as compatible with the structural
elucidation of our molecules of interest. For dehydroxypycni-
dione (2′), we found satisfactory correspondence of computed
model of 2′ to one of the isomers isolated in our cycloaddition.
The computed spectrum of the reported compound (Figure
4b) was compared with the experimental spectra of
diastereomers synthesized in our studies. As can be seen
from the graphs (Figure 4c and 4d), despite overall close
correspondence (MAE = 2.2 ppm, MAX = 5.9 ppm) of the
computed NMR model to the experimental spectrum, one can
notice that the computational model is especially good for
comparison of the aliphatic 13C resonances (MAE = 1.4 ppm,
MAX = 3.3 ppm), presumably due to their lower dependency
6011
https://doi.org/10.1021/jacs.1c02150
J. Am. Chem. Soc. 2021, 143, 6006−6017

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
on explicit solvation effects, hydrogen exchange, and/or
hydrogen bonding, as compared to the tropolone units. We
were able to match the distribution of resonances of aliphatic
13C atoms within the error of the model and conclude the
diastereomer of the naturally occurring compound reported
matched one of the diastereomers we had obtained.
within the macrocycle. The other alternative, which was to
overcome the intrinsic diastereoselectivity of the second HDA
reaction of 47, also proved ineffective. The use of different
hydroxy protecting groups as well as manipulation of the
reaction conditions did not provide the desired stereoisomer
that could lead to the target molecule (see SI for details).
Computational Studies of HDA Cycloadditions. Faced
with the remarkable yet undesired diastereoselectivity of the
second cycloaddition of 20 with both monocycloadducts 46
and 47, we sought to gain greater understanding of the factors
that affect the diastereoselectivity of these transformations
(Figure 5b−d). We performed DFT calculations on the in situ
generated tropolone o-QM 18 with 11 to elucidate the
observed selectivities at the ωB97X-D/def2-QZVPP//ωB97X-
D/6-31G(d) level of theory with a CPCM(mesitylene) model
for solvation.69−73 In order to form the S,S and R,R
diastereomeric products 46 and 47, the (−)-hydroxyhumulene
(11) must react at both faces of the dienophile. This indicates
that 11 can readily access a “ring flipped” conformation, similar
to the well-studied humulene system.74,75 The lowest energy
transition states TS-1 and TS-2 that lead to monoadducts are
both exo, defined as the anti-relationship between the
heterodiene and the methyl of the trisubstituted alkene (Figure
5b). The reaction barriers of TS-1 and TS-2 are low and
primarily entropic (ΔG‡ = 19−21 kcal·mol−1), with enthalpic
barriers ranging from ΔH‡ = 5−7 kcal·mol−1. These reactions
will readily occur within minutes after generation of o-QM 18.
The difference in ΔG‡ between TS-1 and TS-2 (ΔΔG‡) is
+0.6 kcal·mol−1 and indicates that 46:47 should form in a 2.8:1
ratio. When corrected for the elevated temperature (473 K),
the product ratio shifts to 1.9:1, and furthermore, the
experimental product ratio is well within the error of the
calculation. We thoroughly evaluated popular DFT methods
for the reaction and present them in the Supporting
Information (see Table S1).
Adding an additional equivalent of tropolone o-QM 18 to
react with monoadduct 46 yields a single bistropolone-
sesquiterpene adduct 51. The diastereomeric adduct (not
shown) is predicted not to be observed kinetically. The
transition states TS-3 and TS-4 have a large ΔΔG‡ of 4.7 kcal·
mol−1 (Figure 5c). We hypothesized that this noticeable
difference in Gibbs free energy could be partially attributed to
the hydrogen bond between the o-QM’s oxygen and the
hydroxyl of 46, which is present in TS-3 (1.84 Å) but absent in
TS-4 (3.01 Å). Similarly, the cycloadditions of 18 with 47 are
predicted to yield a single diastereomer 49. Of note, the
stereochemistry of the second addition is the same in products
51 and 49. The transition states TS-5 and TS-6 lead to
bisadducts 49 and methylated pycnidione (1′), respectively
(Figure 5d). The ΔΔG‡ is 9.1 kcal·mol−1, which indicates that
formation of pycnidione by thermal HDA is not possible and a
catalyst or additives must be employed. The geometries of TS-
5 and TS-6 are strikingly different, despite having similar bond-
forming distances and hydrogen bonding interactions. We
hypothesize that the difference in Gibbs free energy arises from
the energies of the conformations required to achieve the
respective transition states, suggesting that distortion may
control the reactivity. To investigate this, we stripped the
monoadduct from TS-5 and TS-6, optimized the geometry,
and found that indeed the difference between the monoadduct
conformations is 2.9 kcal·mol−1, indicating that the diaster-
eoselectivity is due in part to the energy of reactant
conformations. In summary, these calculations supported the
Epolone B and Pycnidione HDA Cycloadditions
Studies. With a practical HDA protocol established and the
α-humulene series completed, we finally homed in on the
critical cycloadditions with 10-hydroxyhumulene (11, Figure
5a). Once again, we tested both tropolone o-QM precursors 19
and 20 and quickly realized that BF2-tropolone 19 was not an
effective partner for the cycloaddition with 11 due to
noticeable decomposition, attributed to incompatibility of the
allylic alcohol with 19. On the other hand, the application of
methylated tropolone o-QM precursor 20 led to the desired
HDA reaction, furnishing monocycloadducts 46 and 47 in 89%
combined yield and with 1.4:1 diastereoselectivity. After
separation of diastereoisomers, we were able to verify the
configuration of 47 via X-ray analysis of its para-bromophe-
nylcarbamate derivative 48. Ultimately, base-induced depro-
tection of 46 and 47 delivered (+)-isoepolone B (ent-16) and
(−)-epolone B (ent-4). Most of the characterization data (1H
and 13C NMR, MS data) of ent-16 and ent-4 were consistent
with those reported in the literature, but there were
discrepancies in their optical rotations. The synthetic sample
of (−)-epolone B (ent-4) had essentially identical but opposite
23
optical rotation to that of the natural substance {[α]D
=
−84.0 (c = 0.35, CHCl3); lit.,23 [α]D = +85.0 (c = 0.33,
CH2Cl2)}; thus, the synthetic compound was assigned as the
unnatural enantiomer. In contrast, the synthetic sample of
(+)-isoepolone B (ent-16) did not match the reported value
23
34
{[α]D23= +65.5 (c = 1.0, CHCl3); lit., [α]D = +101.3 (c =
0.08, CHCl3)}; however, considering it was formed and
isolated with (+)-epolone B (4) from the nonenzymatic
reaction with ent-11, we assumed that our synthetic sample ent-
16 is the unnatural enantiomer as well.67
20
With monocycloadducts 46 and 47 in hand, the stage was
set for the second cycloaddition. To our surprise, when 46 and
47 were separately taken forward for a second HDA reaction
with 20, they each produced only a single diastereomer
corresponding to 51 and 49 in 24% and 26% yield,
respectively. This outcome was in stark contrast to the
humulene series (44 → 45), which gave a diastereomeric
mixture, highlighting the importance of the neighboring
hydroxy group on the pronounced selectivity of the second
HDA. Further deprotection of 51 and 49 led to bistropolones
52 and 50, neither of which matched the reported spectral data
of pycnidione (1′).68
This predicament left us with two options, both of which
were investigated extensively: (1) correction of the stereo-
chemistry at C10 of (−)-10-epi-pycnidione (52) to arrive at
(−)-pycnidione or (2) override the inherent diastereoselectiv-
ity of the cycloaddition of 47 to provide (+)-pycnidione.
However, efforts to arrive at either enantiomer of natural
product were met with substantial challenges on all fronts.
First, attempting to adjust the alcohol configuration on 51 or
52 through oxidation/reduction sequences with a variety of
hydrides gave back only the initial 51 or 52, and single-electron
carbonyl reductions suffered from decomposition. Despite our
exhaustive search for a suitable protocol, 51 and 52 were
completely inert under all conditions for alcohol inversion,
presumably because the necessary site of attack is buried deep
6012
https://doi.org/10.1021/jacs.1c02150
J. Am. Chem. Soc. 2021, 143, 6006−6017

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 6. (a) Structural reinvestigation of authentic sample of (+)-pycnidione (1). (b) TD-DFT/ ωB97X/def2-TZVP electronic transitions,
plotted with UV/vis intensities (green lines). (c) TD-DFT transitions plotted with ECD intensities.
Figure 7. Updated (bio)synthetic pathways of pycnidione series.
experimental outcome that the first HDA forms a mixture of
diastereoisomers, but the second cycloadditions are highly
stereoselective.
absolute stereochemical determination using the anomalous
scattering technique was complicated, as the heaviest scattering
atoms on ordered molecule were also oxygen atoms.
Commonly employed parameters (Flack, Hooft, Parson; see
SI, Table S7) were all inconclusive of the correct absolute
stereochemistry assignment. Application of Bayesian statistical
analysis on the Bijvoet pair intensities allowed us to make such
an assignment with high confidence.76 The three-sided
probability for incorrect absolute stereochemistry was
determined to be 10−4, and the probability of racemic
twinning, to be 10−34. Thus, given the X-ray anomalous
scattering data, our best hypothesis is that the correct absolute
stereochemical model belongs in the P61 (# 169), and not in
its enantiomorphic P65 (# 170) space group.
In order to corroborate our conclusions about the absolute
stereochemistry from the single crystal XRD experiment, we
decided to calculate the electronic circular dichroism
transitions and compare them to the originally reported data
(Figure 6b and 6c). We obtained TD-DFT transition energies
with single-point ωB97X/def2-TZVP calculation with the
CPCM(Ethanol) implicit solvation model on unoptimized
geometry from our XRD experiment (see SI for details).
According to the original report, positive and negative peaks
should be observed at 260 and 238 nm, respectively, which
perfectly matches the predicted spectrum for a molecular
Revision of Absolute Stereochemistry of Pycnidione.
Based on stereochemical reassignment of epolone B (from 4′
to 4), we postulated that the absolute stereochemistry of all
natural products within the pycnidione series might be
misassigned, including the flagship member pycnidione (1′).
To set the record straight, an authentic sample of pycnidione
was kindly provided by Merck Research Laboratories, identical
to the material analyzed in the original isolation report.1 The
natural product was crystallized from acetonitrile, furnishing a
sample suitable for X-ray analysis (Figure 6a). Upon
refinement of the complete data set, several inconsistencies
with the previously reported data were found. All unit cell
metrics matched well (see SI, Table S7), and therefore we are
highly confident in having examined the same crystal form as
previously reported. Well-ordered acetonitrile molecules were
observed to occupy one of the solvent accessible voids, which
were not found in the previous report. The other solvent
accessible void was occupied with highly disordered water
molecules. The overall composition of crystal was then
established as C35H43NO7·CH3CN·9H2O, which differs from
previously reported composition of C35H43NO7·4H2O. Due to
the substantial presence of disordered oxygen atoms, the
6013
https://doi.org/10.1021/jacs.1c02150
J. Am. Chem. Soc. 2021, 143, 6006−6017

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
geometry which is enantiomeric to that proposed in the
literature. Performing an analogous computation on the X-ray
geometry from the literature, we arrived at the spectrum which
was, expectedly so, inverted vertically and hence displays a
complete mismatch with the literature report. It should be
noted that the original absolute stereochemistry proposal was
made based on the ECD interpretation, which the present
computational data do not corroborate. Having examined both
the anomalous scattering data and the ECD data, we have high
confidence that the absolute stereochemistry of pycnidione was
incorrectly determined.
construction of other tropolone-containing natural products.
Likewise, the requisite hydroxyhumulene was prepared in a
rapid and practical fashion from chiral pool starting material.
The robustness of the chosen routes permitted assembly of all
key building blocks on a gram to decagram scale and enabled
comprehensive investigations of their union through HDA
chemistry. Biomimetic cycloaddition studies produced several
natural products, including (−)-epolone B (ent-4), (+)-iso-
epolone B (ent-16), and ( )-dehydroxypycnidione (2).
Furthermore, two new sesquiterpene bistropolones 8,9-
epi,epi-(+)-pycnidione (50) and 10-epi-(−)-pycnidione (52)
were formed in highly selective HDA reactions from ent-4 and
ent-16. Computational modeling effectively explained the
origins of selectivities observed in the HDA reactions. These
studies uncover stereochemical inconsistencies and conse-
quently merited structural reinvestigation of natural pycnidione
(1), resulting in stereochemical reassignments of all natural
products within the pycnidione series.
These endeavors also secured additional insight into the
biogenesis of these metabolites. The previously assumed
biomimetic pathway from epolone B (4) to pycnidione (1)
is not viable through spontaneous HDA-assembling mecha-
nisms, in contrast to nonenzymatic cycloaddition leading to
epolone B (4). Further biosynthetic studies will be required to
uncover the potential role of enzymatic involvement in the
second HDA, or in any alternative pathways leading to
formation of 1. Likewise, additional synthetic studies are
warranted to reach pycnidione and to explore the chemical
biology and medicinal chemistry of these unique metabolites.
The HDA stratagem described above may be adapted to afford
various analogues of sesquiterpene-tropolones and can also be
useful for the synthesis of other natural products, including
members of the eupenifeldin series.
Alternative Biosynthetic Considerations. The synthetic
studies described above, supported by the computational data
and stereochemical reassignments, shine additional light on the
biogenesis of the sesquiterpene-tropolones (Figure 7). Thus,
α-humulene (12) undergoes oxidation to (+)-(10R)-hydrox-
yhumulene (ent-11) followed by HDA reaction with tropolone
o-QM 10 to give epolone B (4) and isoepolone B (16). While
these transformations were already reported during the
eupenifeldin biosynthetic studies by Hu and co-workers
(Figure 1d),34 our asymmetric total synthesis (11 → 46 +
47, Figure 5a) led to reassignment of absolute stereochemistry
of all cycloaddition components (ent-11, 4, and 16). On the
other hand, although there are no biosynthetic studies of a
second HDA in native producers, our synthetic campaign
unveiled the formation of sesquiterpenes-bistropolones (51
and 49, Figure 5a) relevant to the proposed biosynthesis.
Notably, the second HDA with epolone B (4) revealed
remarkable macrocyclic stereocontrol, with peripheral attack of
the o-QM occurring at the diastereoface opposite to that which
leads to pycnidione (1). These observations, supported by
DFT computational data, provide strong evidence that
nonenzymatic conversion of 4 to 1 is not feasible. Therefore,
we speculate two different possibilities for the formation of
pycnidione (1). First, separate enzymes may exist that assist in
the second cycloaddition process (4 → 1), as also proposed by
Hu. Alternatively, the formation of pycnidione (1) could be
achieved through a late-stage enzymatic C−H oxidation of
dehydroxypycnidione (2). That in turn could be formed from
α-humulene (12) and double HDA, involving dehydroxyepo-
lone B (15) as an intermediate, as both have been readily
formed during our synthetic studies. We are quick to note,
however, that this is not in line with literature precedent, as 15
has never been isolated and 2 has only once been coisolated
from mixtures containing 1, and therefore our proposal is
merely speculative. Finally, since the second HDA between 10
and 4 or 16 is spontaneous and synthetically feasible, as
demonstrated with the preparation of bistropolones 50 and 52,
it is possible that these compounds are in fact bona fide
sesquiterpene-bistropolone natural products still awaiting
discovery within natural sources (as ent-50 and ent-52).
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/jacs.1c02150.
■
sı
Detailed experimental procedures, spectroscopic data,
computational data, and 1H and 13C NMR spectra
(PDF)
Accession Codes
CCDC 2050423−2050428 contain the supplementary crys-
tallographic 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 Author
■
CONCLUSION
■
David Sarlah − Roger Adams Laboratory, Department of
Chemistry, University of Illinois, Urbana, Illinois 61801,
United States; Cancer Center at Illinois, University of Illinois,
Urbana, Illinois 61801, United States; Email: sarlah@
illinois.edu
Concise, modular, and efficient total syntheses of sesquiter-
pene-tropolones have been developed, which unravel many
chemical and biosynthetic characteristics of this structurally
unique class of macrocyclic meroterpenoids. Within the
described routes to key HDA partners hydroxyhumulene
(11) and o-QM precursors (19 and 20), radical-based
assembly and fragmentation of intermediates served as a
common theme. Of note is a novel, photochemical-based
method to generate tropolones that is complementary to de
Mayo fragmentation and may find application in the selective
Authors
Christopher Y. Bemis − Roger Adams Laboratory,
Department of Chemistry, University of Illinois, Urbana,
Illinois 61801, United States; Cancer Center at Illinois,
University of Illinois, Urbana, Illinois 61801, United States
6014
https://doi.org/10.1021/jacs.1c02150
J. Am. Chem. Soc. 2021, 143, 6006−6017

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(4) Angawi, R. F.; Swenson, D. C.; Gloer, J. B.; Wicklow, D. T.
Malettinin A: A New Antifungal Tropolone from an Unidentified
Fungal Colonist of Hypoxylon Stromata (NRRL 29110). Tetrahedron
Lett. 2003, 44, 7593−7596.
(5) Silber, J.; Ohlendorf, B.; Labes, A.; Wenzel-Storjohann, A.;
Nather, C.; Imhoff, J. F. Malettinin E, an Antibacterial and Anti
Fungal Tropolone Produced by a Marine Cladosporium strain. Front.
Mar. Sci. 2014, 35, 1−6.
(6) Angawi, R. F.; Swenson, D. C.; Gloer, J. B.; Wicklow, D. T.
Malettinins B-D: New Polyketide Metabolites from an Unidentified
Fungal Colonist of Hypoxylon stromata (NRRL 29110). J. Nat. Prod.
2005, 68, 212−216.
(7) For examples of synthetic tropolone products and relevant
studies, see refs 7−14. Yamato, M.; Hashigaki, K.; Kokubu, N.;
Tsuruo, T.; Tashiro, T. Synthesis and Antitumor Activity of
Tropolone Derivatives. J. Med. Chem. 1984, 27, 1749−1753.
(8) Yamato, M.; Hashigaki, K.; Ishikawa, S.; Kokubu, N.; Inoue, Y.;
Tashiro, T. Synthesis and Antitumor Activity of Tropolone
Derivatives. J. Med. Chem. 1985, 28, 1026−1031.
(9) Yamato, M.; Hashigaki, K.; Kokubu, N.; Tashiro, T.; Tsuruo, T.
Synthesis and Antitumor Activity of Tropolone Derivatives. J. Med.
Chem. 1986, 29, 1202−1205.
(10) Yamato, M.; Hashigaki, K.; Sakai, J.; Kawasaki, Y.; Tsukagoshi,
S.; Tashiro, T. Synthesis and Antitumor Activity of Tropolone
Derivatives. J. Med. Chem. 1987, 30, 117−120.
Chad N. Ungarean − Roger Adams Laboratory, Department
of Chemistry, University of Illinois, Urbana, Illinois 61801,
United States; Cancer Center at Illinois, University of Illinois,
Urbana, Illinois 61801, United States
Alexander S. Shved − Roger Adams Laboratory, Department
of Chemistry, University of Illinois, Urbana, Illinois 61801,
United States; Cancer Center at Illinois, University of Illinois,
Urbana, Illinois 61801, United States; orcid.org/0000-
0001-5979-179X
Cooper S. Jamieson − Department of Chemistry and
Biochemistry, University of California, Los Angeles,
California 90095, United States
Taehwan Hwang − Roger Adams Laboratory, Department of
Chemistry, University of Illinois, Urbana, Illinois 61801,
United States; Cancer Center at Illinois, University of Illinois,
Urbana, Illinois 61801, United States
Ken S. Lee − Roger Adams Laboratory, Department of
Chemistry, University of Illinois, Urbana, Illinois 61801,
United States; Cancer Center at Illinois, University of Illinois,
Urbana, Illinois 61801, United States
K. N. Houk − Department of Chemistry and Biochemistry,
University of California, Los Angeles, California 90095,
United States; orcid.org/0000-0002-8387-5261
(11) Yamato, M.; Hashigaki, K.; Sakai, J.; Takeuchi, Y.; Tsukagoshi,
S.; Tashiro, T.; Tsuruo, T. Synthesis and Antitumor Activity of
Tropolone Derivatives. J. Med. Chem. 1987, 30, 1245−1248.
(12) Yamato, M.; Hashigaki, K.; Yasumoto, Y.; Sakai, J.; Luduena, R.
F.; Banerjee, A.; Tsukagoshi, S.; Tashiro, T.; Tsuruo, T. Synthesis and
Antitumor Activity of Tropolone Derivatives. Structure-Activity
Relationships of Antitumor-Active Tropolone and 8-Hydroxyquino-
line Derivatives. J. Med. Chem. 1987, 30, 1897−1900.
Complete contact information is available at:
https://pubs.acs.org/10.1021/jacs.1c02150
Author Contributions
∥C.Y.B. and C.N.U. contributed equally.
Notes
The authors declare no competing financial interest.
(13) Yamato, M.; Ando, J.; Sakaki, K.; Hashigaki, K.; Wataya, Y.;
Tsukagoshi, S.; Tashiro, T.; Tsuruo, T. Synthesis and Antitumor
Activity of Tropolone Derivatives. Bistropolones Containing Con-
necting Methylene Chains. J. Med. Chem. 1992, 35, 267−273.
(14) Yamato, M.; Hirota, Y.; Yoshida, S.; Tanaka, S.; Morita, T.;
Sakai, J.; Hashigaki, K.; Hayatsu, H.; Wataya, Y. Imbalance of
Deoxyribonucleoside Triphosphates and DNA Double-strand Breaks
in Mouse Mammary Tumor FM3A Cells Treated in vitro with an
Antineoplastic Tropolone Derivative. Jpn. J. Cancer Res. 1992, 83,
661−668.
(15) Hsiao, C.-J.; Hsiao, S.-H.; Chen, W.-L.; Guh, J.-H.; Hsiao, G.;
Chan, Y.-J.; Lee, T.-H.; Chung, C.-L. Pycnidione, a Fungus-Derived
Agent, Induces Cell Cycle Arrest and Apopotosis in A549 Human
Lung Cancer Cells. Chem.-Biol. Interact. 2012, 197, 23−30.
(16) Kaneko, M.; Matsuda, D.; Ohtawa, M.; Fukuda, T.; Nagamitsu,
T.; Yamori, T.; Tomoda, H. Potentiation of Bleomycin in Jurkat Cells
by Fungal Pycnidione. Biol. Pharm. Bull. 2012, 35, 18−28.
(17) For antifungal properties, see refs 17 and 18. Overy, D. P.;
Berrue, F.; Correa, H.; Hanif, N.; Hay, K.; Lanteigne, M.; Mquilian,
K.; Duffy, S.; Boland, P.; Jagannathan, R.; Carr, G. S.; Vansteeland,
M.; Kerr, R. G. Sea foam as a Source of Fungal Inoculum for the
Isolation of Biologically Active Natural Products. Mycology 2014, 5,
130−144.
(18) Kerr, R. G.; Overy, D. P.; Berué, F. Anti-Dandruff Composition
Comprising Pycnidione and Epolone. PCT Int. Appl. WO2016/
198848, 2016.
(19) For antimalarial properties, see refs 19−21. Höller, U.; Wright,
A. D.; Matthée, G. F.; Konig, G. M.; Draeger, S.; Aust, H.-J.; Schulz,
B. Fungi from marine sponges: diversity, biological activity and
secondary metabolites. Mycol. Res. 2000, 104, 1354−1365.
(20) Wright, A. D.; Lang-Unnasch, N. Potential Antimalarial Lead
Structures from Fungi of Marine Origin. Planta Med. 2005, 71, 964−
966.
ACKNOWLEDGMENTS
■
Financial support for this work was provided by the University
of Illinois and National Institute of General Medical Sciences,
National Institute of Health (GM 124480 to K.N.H.). Bristol
Myers Squibb, Amgen, Eli Lilly, and FMC are also acknowl-
edged for unrestricted research support. The Bruker 500-MHz
NMR spectrometer was obtained with the financial support of
the Roy J. Carver Charitable Trust, Muscatine, Iowa, USA. We
also thank Dr. D. Olson and Dr. L. Zhu for NMR
spectroscopic assistance, Dr. D. L. Gray and Dr. T. Woods
for X-ray crystallographic analysis assistance, and F. Sun for
mass spectrometric assistance. We thank Dr. Patrick S. Fier
(Merck Research Laboratories) for providing us with an
authentic sample of pycnidione and Dr. Russel G. Kerr
(University of Prince Edward Island) for sharing their NMR
data of dehydroxypycnidione. Finally, we would like to thank
Matthew Bock and Yingzhe Yang for assistance with this work
and Christopher J. Huck for critical proofreading of this paper.
We dedicate this work in memory of Sir Jack Baldwin.
REFERENCES
■
(1) Harris, G. H.; Hoogsteen, K.; Silverman, K. C.; Raghoobar, S. L.;
Bills, G. F.; Lingham, R. B.; Smith, J. L.; Dougherty, H. W.; Cascales,
C.; Pelaez, F. Isolation and Structure Determination of Pycnidione, a
Novel Bistropolone Stromelysin Inhibitor from a Phoma sp.
Tetrahedron 1993, 49, 2139−2144.
(2) For examples of related natural products, see refs 2−6. Ito, T.
Fusariocin C, a New Cytotoxic Substance Produced by Fusarium-
Moniliforme. Agric. Biol. Chem. 1979, 43, 1237−1242.
(3) Park, S. Y.; Choi, H.; Hwang, H.; Kang, H.; Rho, J.-R.
Gukulenins A and B, Cytotoxic Tetraterpenoids from the Marine
Sponge Phorbas gukulensis. J. Nat. Prod. 2010, 73, 734−737.
(21) Pittayakhajonwut, P.; Theerasilp, M.; Kongsaeree, P.; Rungrod,
A.; Tanticharoen, M.; Thebtaranonth, Y. Pughiinin A, a Sesquiterpene
6015
https://doi.org/10.1021/jacs.1c02150
J. Am. Chem. Soc. 2021, 143, 6006−6017

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
from the Fungus Kionochaeta pughii BCC 2878. Planta Med. 2002, 68,
1017−1019.
(22) Ayers, S.; Zink, D. L.; Powell, J. S.; Brown, C. M.; Grund, A.;
Bills, G. F.; Platas, G.; Thompson, D.; Singh, S. B. Noreupenifeldin, a
Tropolone from an Unidentified Ascomycete. J. Nat. Prod. 2008, 71,
457−459.
(23) Cai, P.; Smith, D.; Cunningham, B.; Brown-Shimer, S.; Katz, B.;
Pearce, C.; Venables, D.; Houck, D. Epolones: Novel Sesquiterpene-
Tropolones from Fungus OS-F69284 That Induce Erythropoietin in
Human Cells. J. Nat. Prod. 1998, 61, 791−795.
(24) For eupenifeldin isolation, see refs 24 and 25. Mayerl, F.; Gao,
Q.; Huang, S.; Klohr, S. E.; Matson, J. A.; Gustavson, D. R.; Pirnik, D.
M.; Berry, R. L.; Fairchild, C.; Rose, W. C. Eupenifeldin, A Novel
Cytotoxic Bistropolone From Eupenicillium brefeldianum. J. Antibiot.
1993, 46, 1082−1088.
(25) Gao, Q.; Mayerl, F. Eupenifeldin. Acta Crystallogr., Sect. C:
Cryst. Struct. Commun. 1994, 50, 2033−2036.
(26) For noreupenifeldin isolation, see: Ayers, S.; Zink, D. L.;
Powell, J. S.; Brown, C. M.; Grund, A.; Bills, G. F.; Platas, G.;
Thompson, D.; Singh, S. B. Noreupenifeldin, a Tropolone from an
Unidentified Ascomycete. J. Nat. Prod. 2008, 71, 457−459.
(27) For dehydroxyeupenifeldin and neosetophomone B isolations,
see: El-Elimat, T.; Raja, H. A.; Ayers, S.; Kurina, S. J.; Burdette, J. E.;
Mattes, Z.; Sabatelle, R.; Bacon, J. W.; Colby, A. H.; Grinstaff, M. W.;
Pearce, C. J.; Oberlies, N. H. Meroterpenoids from Neosetophoma sp.:
A Dioxa[4.3.3]propellane Ring System, Potent Cytotoxicity, and
Prolific Expression. Org. Lett. 2019, 21, 529−534.
(28) For previous model studies, see refs 28−33. Baldwin, J. E.;
Mayweg, A. V. W.; Neumann, K.; Pritchard, G. J. Studies toward the
Biomimetic Synthesis of Tropolone Natural Products via Hetero
Diels−Alder Reaction. Org. Lett. 1999, 1, 1933−1935.
(29) Adlington, R. M.; Baldwin, J. E.; Pritchard, G. J.; Williams, A. J.;
Watkin, D. J. A Biomimetic Synthesis of Lucidene. Org. Lett. 1999, 1,
1937−1939.
(30) Adlington, R. M.; Baldwin, J. E.; Pritchard, G. J.; Mayweg, A. V.
W.; Pritchard, G. J. Biomimetic Cycloaddition Approach to
Tropolone Natural Products via a Tropolone Ortho-quinone
Methide. Org. Lett. 2002, 4, 3009−3011.
(31) Baldwin, J. E.; Mayweg, A. V. W.; Pritchard, G. J.; Adlington, R.
M. Expedient Synthesis of a Highly Substituted Tropolone via 3-
oxidopyrylium [5 + 2] Cycloaddition Reaction. Tetrahedron Lett.
2003, 44, 4543−4545.
(32) Rodriguez, R.; Adlington, R. M.; Moses, J. E.; Cowley, A.;
Baldwin, J. E. A New and Efficient Method for o-Quinone Methide
Intermediate Generation: Application to the Biomimetic Synthesis of
( )-Alboatrin. Org. Lett. 2004, 6, 3617−3619.
(33) Rodriguez, R.; Moses, J. E.; Adlington, R. M.; Baldwin, J. E. A
New and Efficient Method for o-Quinone Methide Intermediate
Generation: Application to the Biomimetic Synthesis of Benzopyran
Derived Natural Products ( )-Lucidene and ( )-Alboatrin. Org.
Biomol. Chem. 2005, 3, 3488−3495.
(34) Chen, Q.; Gao, J.; Jamieson, C.; Liu, J.; Ohashi, M.; Bai, J.; Yan,
D.; Liu, B.; Che, Y.; Wang, Y.; Houk, K. N.; Hu, Y. Enzymatic
Intermolecular Hetero-Diels−Alder Reaction in the Biosynthesis of
Tropolonic Sesquiterpenes. J. Am. Chem. Soc. 2019, 141, 14052−
14056.
(35) Schotte, C.; Li, L.; Wibberg, D.; Kalinowski, J.; Cox, R. J.
Synthetic Biology Driven Biosynthesis of Unnatural Tropolone
Sesquiterpenoids. Angew. Chem., Int. Ed. 2020, 59, 23870−23878.
(36) For reviews, see refs 36−38. De Mayo, P. Photochemical
Syntheses. 37. Enone Photoannelation. Acc. Chem. Res. 1971, 4, 41−
47.
(37) Oppolzer, W. The Intramolecular [2 + 2] Photoaddition/
Cyclobutane-Fragmentation Sequence in Organic Synthesis. Acc.
Chem. Res. 1982, 15, 135−141.
(38) Winkler, J. D.; Bowen, C. Mazur.; Liotta, Fina. [2 + 2]
Photocycloaddition/Fragmentation Strategies for the Synthesis of
Natural and Unnatural Products. Chem. Rev. 1995, 95, 2003−2020.
(39) Lange, G. L.; de Mayo, P. Photochemical Synthesis:
Stipitotanic Acid. Chem. Commun. 1967, 704.
(40) Challand, B. D.; Hikino, H.; Kornis, G.; Lange, G.; de Mayo, P.
Photochemical Cycloaddition. Some Applications of the Use of
Enolized β−Diketones. J. Org. Chem. 1969, 34, 794−806.
(41) For reviews, see refs 41 and 42. Crimmins, M. T.; Reinhold, T.
L. Enone Olefin [2 + 2] Photochemical Cycloadditions. Org. React.
1993, 44, 297.
(42) Poplata, S.; Tröster, A.; Zou, Y.-Q.; Bach, T. Recent Advances
in the Synthesis of Cyclobutanes by Olefin [2 + 2] Photo-
cycloaddition Reactions. Chem. Rev. 2016, 116, 9748−9815.
(43) Zigon, N.; Hoshino, M.; Yoshioka, S.; Inokuma, Y.; Fujita, M.
Where is the Oxygen? Structural Analysis of α-Humulene Oxidation
Products by the Crystalline Sponge Method. Angew. Chem., Int. Ed.
2015, 54, 9033−9037.
(44) Crossley, S. W. M.; Barabé, F.; Shenvi, R. A. Simple,
Chemoselective, Catalytic Olefin Isomerization. J. Am. Chem. Soc.
2014, 136, 16788−16791.
(45) Kobusone can be readily synthesized from commercially
available caryophyllene oxide in a single step: Bume, D. D.; Pitts, C.
R.; Ghorbani, F.; Harry, S. A.; Capilato, J. N.; Siegler, M. A.; Lectka,
T. Ketones as Directing Groups in Photocatalytic sp3 C-H
Fluorination. Chem. Sci. 2017, 8, 6918−6923.
(46) For a recent review on chiral pool starting materials for terpene
synthesis, see: Brill, Z. G.; Condakes, M. L.; Ting, C. P.; Maimone, T.
J. Navigating the Chiral Pool in the Total Synthesis of Complex
Terpene Natural Products. Chem. Rev. 2017, 117, 11753−11795.
(47) Bairamov, G. I. Synthesis of New Nitrogen-Containing
Compounds on the Basis of 2,6-Dichloro-5-oxo-2-hexene and Their
Study. Azerb. Khim. Zh. 2008, 3, 174−180.
(48) Liang, Y.-F.; Li, X.; Wang, X.; Zou, M.; Tang, C.; Liang, Y.;
Song, S.; Jiao, N. Conversion of Simple Cyclohexanones into
Catechols. J. Am. Chem. Soc. 2016, 138, 12271−12277.
(49) Brimioulle, R.; Bach, T. [2 + 2] Photocycloaddition of 3-
Alkenyloxy-2-cycloalkenones: Enantioselective Lewis Acid Catalysis
and Ring Expansions. Angew. Chem., Int. Ed. 2014, 53, 12921−12924.
(50) Sato, M.; Ebine, S.; Tsunetsugu, J. Small Ring-annelated Non-
benzenoid Aromatic Compounds: 1,2-Dihydrocyclobuta[e]tropolone.
J. Chem. Soc., Chem. Commun. 1978, 215.
(51) Takahashi, T.; Kitamura, K.; Tsuji, J. Synthesis of New
Humulene Derivatives; (2E,6E,9E)- Cycloundecatrienones, by Intra-
molecular Alkylation of Protected Cyanohydrin. A Route to
Humulene. Tetrahedron Lett. 1983, 24, 4695−4698.
(52) Davis, F. A.; Vishwakarma, L. C.; Billmers, J. G.; Finn, J.
Synthesis of.Alpha.-Hydroxycarbonyl Compounds (Acyloins): Direct
Oxidation of Enolates Using 2-Sulfonyloxaziridines. J. Org. Chem.
1984, 49, 3241−3243.
(53) Nakagiri, T.; Murai, M.; Takai, K. Stereospecific Deoxygenation
of Aliphatic Epoxides to Alkenes under Rhenium Catalysis. Org. Lett.
2015, 17, 3346−3349.
(54) For classic strategies for the synthesis of α-humulene, see refs
54−60. Corey, E. J.; Hamanaka, E. Total Synthesis of Humulene. J.
Am. Chem. Soc. 1967, 89, 2758−2759.
(55) Vig, O. P.; Ram, B.; Atwal, K. S.; Bari, S. S. Terpenoids. 125.
New Synthesis of Humulene (2,6,6,9-tetramethylcycloundeca-1,4,8-
triene). Indian J. Chem. 1976, 14B, 855−857.
(56) Kitagawa, Y.; Itoh, A.; Hashimoto, S.; Yamamoto, H.; Nozaki,
H. Total Synthesis of Humulene. A Stereoselective Approach. J. Am.
Chem. Soc. 1977, 99, 3864−3867.
(57) McMurry, J. E.; Matz, J. R. Stereospecific Synthesis of
Humulene by Titanium-Induced Dicarbonyl Coupling. Tetrahedron
Lett. 1982, 23, 2723−2724.
(58) Miyaura, N.; Suginome, H.; Suzuki, A. New Stereo- and
Regiospecific Synthesis of Humulene by Means of the Palladium-
Catalyzed Cyclization of Haloalkenylboranes. Tetrahedron Lett. 1984,
25, 761−764.
(59) Corey, E. J.; Daigneault, S.; Dixon, B. R. A Biomimetic
Chemical Synthesis of Humulene from Farnesol. Tetrahedron Lett.
1993, 34, 3675−3678.
6016
https://doi.org/10.1021/jacs.1c02150
J. Am. Chem. Soc. 2021, 143, 6006−6017

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(60) Hu, T.; Corey, E. J. Short Syntheses of ( )-δ-Araneosene and
Humulene Utilizing a Combination of Four-Component Assembly
and Palladium-Mediated Cyclization. Org. Lett. 2002, 4, 2441−2443.
(61) Though dehydroxypycnidione was recently reported to be
isolated through combinatorial biosynthesis (ref 35), the reported
spectroscopic data for this natural product did not match our data as
well as data provided to us by the initial isolation group.
(62) Boyko, Y. D.; Huck, C. J.; Ning, S.; Shved, A. S.; Yang, C.; Chu,
T.; Tonogai, E. J.; Hergenrother, P. J.; Sarlah, D. Synthetic Studies on
Selective, Proapoptotic Isomalabaricane Triterpenoids Aided by
Computational Techniques. J. Am. Chem. Soc. 2021, 143, 2138−2155.
(63) Neese, F. Software Update: The ORCA Program System,
Version 4.0. Wiley Interdiscip. Rev.: Comput. Mol. Sci. 2018, 8, e1327.
(64) Neese, F. The ORCA Program System. Wiley Interdiscip. Rev.:
Comput. Mol. Sci. 2012, 2, 73−78.
(65) Jensen, F. Segmented Contracted Basis Sets Optimized for
Nuclear Magnetic Shielding. J. Chem. Theory Comput. 2015, 11, 132−
138.
(66) Lodewyk, M. W.; Siebert, M. R.; Tantillo, D. J. Computational
Prediction of 1 H and 13 C Chemical Shifts: A Useful Tool for
Natural Product, Mechanistic, and Synthetic Organic Chemistry.
Chem. Rev. 2012, 112, 1839−1862.
(67) Based on this data, the naturally obtained isoepolone B (16)
should have the negative optical rotation value.
(68) Oxidation of 51 with DMP delivered the ketone that
spectroscopically (1H and 13C NMR) matched with oxidized and
methylated natural pycnidione.
(69) 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.
(70) Weigend, F.; Ahlrichs, R. Balanced Basis Sets of Split Valence,
Triple Zeta Valence and Quadruple Zeta Valence Quality for H to Rn:
Design and Assessment of Accuracy. Phys. Chem. Chem. Phys. 2005, 7,
3297−3305.
(71) Hehre, W. J.; Ditchfield, R.; Pople, J. A. SelfConsistent
Molecular Orbital Methods. XII. Further Extensions of Gaussian
Type Basis Sets for Use in Molecular Orbital Studies of Organic
Molecules. J. Chem. Phys. 1972, 56, 2257−2261.
(72) Hariharan, P. C.; Pople, J. A. The Influence of Polarization
Functions on Molecular Orbital Hydrogenation Energies. Theor.
Chim. Acta 1973, 28, 213−222.
(73) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A Consistent and
Accurate Ab Initio Parametrization of Density Functional Dispersion
Correction (DFT-D) for the 94 Elements H-Pu. J. Chem. Phys. 2010,
132, 154104.
(74) Shirahama, H.; Osawa, E.; Matsumoto, T. Conformational
Studies on Humulene by Means of Empirical Force Field
Calculations. Role of Stable Conformers of Humulene in Biosynthetic
and Chemical Reactions. J. Am. Chem. Soc. 1980, 102, 3208−3213.
(75) Neuenschwander, U.; Czarniecki, B.; Hermans, I. Origin of
Regioselectivity in α-Humulene Functionalization. J. Org. Chem. 2012,
77, 2865−2869.
(76) Hooft, R. W. W.; Straver, L. H.; Spek, A. L. Determination of
Absolute Structure Using Bayesian Statistics on Bijvoet Differences. J.
Appl. Crystallogr. 2008, 41, 96−103.
6017
https://doi.org/10.1021/jacs.1c02150
J. Am. Chem. Soc. 2021, 143, 6006−6017