Counteracting in Vitro Toxicity of the Ionophoric Mycotoxin Beauvericin - Synthetic Receptors to the Rescue

Article abstract of DOI:10.1021/acs.joc.9b01665

Beauvericin (BEA) and enniatins are toxic ionophoric cyclodepsipeptides that mainly occur in grains. As such, their presence in food commodities poses a concern for public health. To date, despite recent European Food Safety Authority emphasis on the need for more data to evaluate long-term toxicity effects, no suitable affinity reagents are available to detect the presence of BEA and derivatives in food samples. We here report on the synthesis of a small library of artificial receptors with varying cavity sizes and different hydrophobic building blocks. Immobilization of one of the receptors on solid support resulted in a strong retention of beauvericin, thus revealing promising properties as solid-phase extraction material for sample pretreatment. Furthermore, treatment of HepG2 cells with the most promising receptor markedly reduced beauvericin-induced cytotoxicity, hinting toward the possibility of using synthetic receptors as antidotes against ionophoric toxins.

Full text of DOI:10.1021/acs.joc.9b01665

Article
pubs.acs.org/joc
Cite This: J. Org. Chem. XXXX, XXX, XXX−XXX
Counteracting in Vitro Toxicity of the Ionophoric Mycotoxin
BeauvericinSynthetic Receptors to the Rescue
Vincent Ornelis,† Andreja Rajkovic,‡ Marlies Decleer,‡,∥ Benedikt Sas,§ Sarah De Saeger,∥
and Annemieke Madder*,†
†Organic and Biomimetic Chemistry Research Group, Department of Organic and Macromolecular Chemistry, Ghent University,
Krijgslaan 281, 9000 Ghent, Belgium
‡Department of Food Safety and Food Quality, Laboratory of Food Microbiology and Food Preservation and §Department of Food
Safety and Food Quality, Food2Know, Ghent University, Coupure Links 653, 9000 Ghent, Belgium
∥Department of Bioanalysis, Laboratory of Food Analysis, Ghent University, Ottergemsesteenweg 460, 9000 Ghent, Belgium
S
* Supporting Information
ABSTRACT: Beauvericin (BEA) and enniatins are toxic iono-
phoric cyclodepsipeptides that mainly occur in grains. As such, their
presence in food commodities poses a concern for public health. To
date, despite recent European Food Safety Authority emphasis on
the need for more data to evaluate long-term toxicity effects, no
suitable affinity reagents are available to detect the presence of BEA
and derivatives in food samples. We here report on the synthesis of a
small library of artificial receptors with varying cavity sizes and
different hydrophobic building blocks. Immobilization of one of the
receptors on solid support resulted in a strong retention of
beauvericin, thus revealing promising properties as solid-phase
extraction material for sample pretreatment. Furthermore, treatment of HepG2 cells with the most promising receptor markedly
reduced beauvericin-induced cytotoxicity, hinting toward the possibility of using synthetic receptors as antidotes against
ionophoric toxins.
INTRODUCTION
■
When it comes to food-borne illnesses, most attention goes to
the causative microorganisms1 such as the well-known
Salmonella spp., Listeria monocytogenes, Staphylococcus aureus,
Shiga toxin-producing Escherichia coli, or the less familiar
norovirus.2 Yet, an equally important and often forgotten cause
of food poisoning is related to the ingestion of toxins and not
to the microorganisms themselves.
Beauvericin (BEA) is a mycotoxin produced by fungi of the
Fusarium genus, which infect grains such as wheat, barley, and
corn.3 These fungi also produce another kind of cyclo-
depsipeptide, namely, enniatins (ENN),4 which only differ
Figure 1. Chemical structure of receptor R1 with color-coded
indication of the possible interactions with BEA or the most common
ENN subtypes.
slightly in structure from beauvericin (Figure 1) and show
comparable toxicity profiles. Quantification of BEA and ENN
content in 228 Norwegian grain samples collected at the grain
delivery site revealed a high abundance of these toxins (32% of
samples contained BEA, 25% ENN A, 67% ENN A1, 100%
ENN B, and 94% ENN B1). Concentrations up to 7.4 mg/kg
were detected for the sum of the toxins.5 Moreover, infected
grain kernels have been shown to contain as much as 520 mg/
kg beauvericin.6 The observation that ENN B is frequently the
most abundant enniatin subtype found in grain samples has
also been confirmed by others.7 Due to their thermal stability,
these toxins survive most food-processing techniques, which
explains their presence in food commodities.
The primary toxic action of BEA and ENN is related to their
ability to form ion channels and transport NH4 or K+ ions
+
across the cell membrane, resulting in disturbance of the ion
homeostasis and eventually cell death.8 As reported by
Tonshin and co-workers, addition of 3.8 μM BEA causes
membrane depolarization of human neural Paju cells.9
Received: June 21, 2019
© XXXX American Chemical Society
A
DOI: 10.1021/acs.joc.9b01665
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Concentrations in the range of 10 μM are cytotoxic,10 and due
to its lipophilic nature, beauvericin has been shown to
accumulate in the liver and adipose tissue.11 Moreover, passage
through the blood−brain barrier has also been shown.12
This combination of toxic properties with the occurrence in
food commodities poses a considerable risk to human health.
In contrast to a variety of commercial immunoaffinity
cartridges being available for sample preparation in food
analysis for other mycotoxins (such as aflatoxins, deoxyniva-
lenol, zearalenone, or fumonisins) and given the unavailability
of antibodies or other suitable retention molecules, there is
currently a complete lack of suitable tools for detection of BEA
and derivatives in food. Furthermore, no antidote is available
against BEA, ENN, or related ionophores, which is also
attributed to the difficulty in generating antibodies against
these nonimmunogenic fungal metabolites. Physicians can
therefore only rely on general detoxification procedures such as
gastric lavage, plasma exchange, or hemodialysis, to remove the
toxins from the blood circulation once they have been
ingested. The development of a drug that is able to sequester
such ionophoric cyclodepsipeptides is thus urgently needed. In
that regard, artificial receptors targeting these toxins might
provide an answer to the aforementioned problems.
Many synthetic receptors have been developed for ligands
such as anions,13 cations,14 saccharides,15 small peptides, and
even large entities such as fullerene16 or entire proteins.17,18
Yet despite these efforts, some ligands, including, e.g., medium-
sized peptides, have remained unexplored. Indeed, most
synthetic receptors for peptide-related structures reported to
date only show affinity for di- or tripeptides19−21 and the
recognition of structures of intermediate size (between small
peptides and large proteins), such as the cyclodepsipeptides
described in this work, has proven particularly difficult.
Moreover, no prior efforts toward the application of synthetic
receptors in food safety and mitigation of food contamination
risks have been deployed.
In an effort to fill this void, we previously described receptor
R1 (Figure 1)22 which was able to recognize beauvericin with
an affinity of 4.0 × 105 M−1 in CHCl3. The design principle for
receptor R1 was based on two key features of the toxin,
namely, its lipophilic nature and its ability to bind cations such
as NH4+ and K+. The incorporation of steroid side arms in the
receptor allows interaction with the hydrophobic side chains of
the toxin, while the protonated amine functionality connected
to the central part of the carbazole linker mediates the
ionophoric interaction. Due to the presence of the fluorogenic
carbazole moiety in close proximity to the protonated amine,
binding of BEA could be monitored via fluorescence
measurements. However, for receptor R1 to be truly applicable
as an antidote, some caveats still remain. The most prominent
one resides in the rather lengthy synthesis (12 steps), in which
certain steps caused problems during early scale-up attempts.
In this work, we improved on our initial design by
developing receptor R2 (Scheme 1), which can be synthesized
in fewer steps and on a larger scale. Additionally, a series of
receptor analogues featuring varying cavity sizes and different
hydrophobic moieties were prepared to probe the impact of
such changes on the selectivity and affinity for ionophoric
toxins, which themselves have different sizes. To explore
possible applications of the synthetic receptors, immobilization
of the most promising analogue on solid support was
performed. This led to a strong retention of beauvericin,
thus opening the possibility of using this construct for solid-
phase extraction (SPE) in food sample pretreatment.
More importantly, we report on the application of the
synthetic receptor as antidote against beauvericin (Figure 2).
Figure 2. Use of the synthetic receptor to prevent beauvericin-
mediated cell death.
We show that addition of 10 equiv of receptor allows to
significantly reduce beauvericin-induced cytotoxicity in HepG2
(liver carcinoma) cells using multiple cell assays. This is the
first example showing the use of an artificial receptor to
counteract food-borne ionophore toxicity in vitro.
RESULTS AND DISCUSSION
■
Synthesis of Receptor R2. The main hurdles in the
synthetic route toward previously described receptor R1 were
related to the functionalization of the C23 end of the steroid
side arms. We reasoned that receptor R2 with ester
functionalities at the C23 extremities would be more accessible
synthesis-wise without compromising the binding with BEA
(Scheme 1). The synthesis of R2 started with the Fisher
esterification of commercially available lithocholic acid 1,
followed by acylation of the C3−OH using bromoacetyl
bromide. Subsequent second-order nucleophilic substitution
reaction with NaN3 gave us steroid 3 in good yield. The
carbazole fragment 5 was synthesized as described pre-
viously.22 Briefly, alkylation of 2,7-dibromocarbazole 4 was
followed by a Sonogashira reaction23 using 2-methyl-3-butyn-
2-ol. Removal of the 2-hydroxy-isopropyl protecting group
using NaH in toluene yielded compound 5 in satisfactory yield.
Subsequently, both fragments 3 and 5 were coupled via a
copper-catalyzed alkyne−azide cycloaddition (CuAAC) using
Fokin−Sharpless conditions.24 Receptor R2 was obtained after
removal of the tert-butyloxycarbonyl (Boc)-group using 4 M
HCl in dioxane.25
The synthesis of R2 involves only eight steps, which is
considerable less than the 12 steps that were necessary for
receptor R1. Moreover, the modular nature of the synthetic
route depicted in Scheme 1 enables the straightforward
generation of a variety of receptor analogues (vide infra).
Before embarking on that journey, the affinity of receptor R2
for BEA was assessed.
Determination of the Binding Affinity of Receptor R2.
From the many techniques that are available for the
determination of binding constants, fluorescence titration has
proven to be particularly useful. This principle has been used
by us before for receptor R1, in which case binding with BEA
resulted in a pronounced fluorescence increase of the blue
carbazole emission at 373 and 389 nm in CHCl3.
Given the structural similarity between receptor R1 and R2,
we used the same methodology in this work to asses if receptor
B
DOI: 10.1021/acs.joc.9b01665
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Scheme 1. Synthetic Route toward Receptor R2
R2 is still capable of binding BEA with reasonable affinity. It is
worth noting that the choice of excitation wavelength (λex) for
the titration is important. A common mistake in fluorescence
titration is the negligence of the inner filter effect.26 Briefly, if
the added guest (e.g., BEA) absorbs at the excitation or
emission wavelength, a decrease in emission intensity would be
observed, which is unrelated to complex formation. This can
result in misinterpretation of the results. A UV−vis experiment
with BEA showed that this toxin absorbs considerably below
300 nm. However, above this value, the absorption (hence the
inner filter effect) is negligible [Figure S1 in the Supporting
Information (SI)]. This explains why 329 nm (n → π*
transition) was used as λex and not the 250 nm region (π → π*
transition).
As shown in Figure 3, addition of BEA to R2 in CHCl3
resulted in an increase of the carbazole emission, similar to
what has been observed before for R1. An association constant
of 1.1 × 105 M−1 4.5 × 104 [Ka standard deviation (SD)]
was obtained, based on the mean of three independent
titrations. This value was determined by fitting the
fluorescence titration data at 389 nm to a theoretical binding
isotherm for a 1:1 binding model.22
The observed complexation between R2 and beauvericin is
as expected, best explained considering a 1:1 binding
stoichiometry. Indeed, when the titration data were fitted to
a 2:1 or 1:2 host−guest stoichiometry,27,28 no proper fit could
be obtained (data not shown).
To provide further evidence for the binding between BEA
and R2, we explored the possibility of using NMR titration and
isothermal titration calorimetry (ITC). Both techniques
require concentrations in the millimolar range. However,
under such conditions, aggregation of the receptor in CHCl3
was observed, which prevented comparison with the
fluorescence titration data. As fluorescence titration can be
performed at micromolar concentrations (UV−vis experiments
showed that R2 followed the Beer−Lambert law in this
concentration range (data not shown)). Although receptor R2
is soluble in dimethylformamide (DMF)-d7, the added value of
a NMR titration in this solvent is questionable. DMF, being a
so-called dipolar aprotic solvent, will interact strongly with the
ammonium cation of the receptors, which is, at the same time,
vital for binding to the toxin. As such, data obtained in DMF
cannot be compared with the fluorescence titration performed
in CHCl3. Therefore, other methods were considered to
further support the binding of R2 with BEA (vide infra SPE
experiments and cell toxicity assays).
We have previously shown22 that titration of the carbazole
fragment alone, without any steroid side arms, yielded a 1000-
fold lower binding affinity for BEA compared to receptor R2
(Figure S3 in SI). On the other hand, when the Boc-protected
receptor was tested, no change in fluorescence intensity was
observed upon BEA addition (Figure S4 in SI). Taken
together, these results indicate that both the ammonium
functionality and the hydrophobic cavity generated by the
steroid side arms are required for high-affinity binding and
sensing of beauvericin. This observation was also confirmed by
the application experiments reported in this work (SPE
experiments and use of receptor R2 as antidote, vide infra).
Figure 3. Fluorescence titration of 0−5 equiv BEA to 10 μM receptor
R2 in CHCl3 using λex = 329 nm. The inset shows the mean
fluorescence intensity of R2 at 389 nm as a function of added BEA
concentration. The mean and standard errors at each titration point
were calculated based on three independent titrations. Data points
were fitted to a 1:1 stoichiometry (red line). See Figure S2 in the SI
for a larger representation of the inset graph.
C
DOI: 10.1021/acs.joc.9b01665
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
The fourfold lower binding affinity of R2 for BEA compared
to R1, as determined using fluorescence titration, is somewhat
surprising, considering that the only difference between both
receptors lies in the C23 extremity. A possible explanation
might be that the large bromine atoms in R1 induce a
conformational change in the receptor, which is favorable for
the interaction with beauvericin. Nevertheless, the shorter
synthesis of R2 and its scalability outweigh the fourfold lower
affinity. Moreover, certain solid-phase extraction (SPE)
applications (vide infra) do not necessarily require high-
affinity binders because the lower affinity can be compensated
for by using an excess of immobilized receptor.
Variation of Cavity Size and/or Hydrophobic Moiety.
The discovery that receptor R2 has a fourfold lower affinity
than R1 prompted us to investigate if it would be possible to
synthesize analogues of R2 with higher affinity. As shown in
Figure 4, in principle, the cavity sizes of receptors R1 and R2
the secondary alcohol with PCC, yielding the corresponding
ketone 7 in 65% yield (over two steps). The nucleophilic
attack of the bulky sulfur ylide at the equatorial position
resulted in the formation of the corresponding epoxide in a
modest 41% yield,29 which was ring-opened with NaN3.
Building blocks 3 (Scheme 1) and 9 (Scheme 2) are both
based on the lithocholic acid precursor that contains a cis-
decalin ring fusion, responsible for its concave inner surface.
Although this has usually been one of the reasons why this
molecule is considered as scaffold for synthetic receptors, we
were wondering if substitution with a hydrophobic moiety
containing a trans-decalin system would influence the binding
affinity with the ionophoric toxins.
An example of such trans-decalin scaffold is abietic acid,
which is isolated from resin or conifer extracts. Multiple
derivatives with interesting biological activity have been
synthesized starting from this natural product.30 Additionally,
being a chiral building block, it has been employed as ligand
for the Zn-catalyzed enantioselective Friedel−Crafts alkylation
of N-methylindoles31 and as chiral solvating agent for NMR
analysis.32 Yet, its use as a scaffold for synthetic receptors
seems to be limited.33
The synthetic route toward abietic acid-based building block
15 follows a similar path to that described for its litocholic acid
counterpart (Scheme 3). Hence, acylation of commercially
Scheme 3. Synthetic Route toward Abietic Acid-Based
Fragment 15
Figure 4. Visual representation of receptor R1 (orange) and BEA
(green) using visual molecular dynamics, illustrating the large cavity
size. The same applies to R2 (not shown).
are considerably large when taking the volume of BEA into
account. Given our efficient and modular synthesis route for
R2, variation of the cavity size was considered as a way to
increase the affinity for BEA.
Synthesis of an alternative steroid fragment 9 (Scheme 2)
featuring a shorter C3 chain started with the Fisher
esterification of lithocholic acid 1, followed by oxidation of
available dehydroabietylamine 13 with bromoacetyl bromide
was followed by substitution with NaN3, which yielded the
desired azido-derivative 15 in 30% yield (two steps).
Having synthesized three different azido-functionalized
building blocks (3, 9, and 15), coupling to carbazole moiety
5 via a CuAAC reaction was performed (Figure 6). As
described earlier, reaction of steroid fragment 3 with carbazole
5 gave the receptor with the largest cavity size (R2).
Interestingly, when conducting the CuAAC reaction using
2.5 equiv of the alternative steroid fragment 9, mainly the
monosubstituted reaction product was obtained. This differ-
ence in reactivity might be due to the closer proximity of the
azide to the steroid skeleton in compound 9, thereby rendering
the azide more sterically crowded. The obtained monosub-
stituted derivative allows the synthesis of asymmetric receptor
analogues. Hence, coupling of 2.5 equiv 9 followed by a second
CuAAC reaction with fragment 3 gave a receptor with an
intermediate cavity size (R3), while coupling of 9 in excess (8
equiv) yielded receptor R4, which has the smallest cavity size.
Finally, the CuAAC reaction using abietic acid-based fragment
15 yielded receptor R5.
Scheme 2. Synthetic Route toward Steroid Fragment 9
Binding Affinity of the Receptor Analogues for the
Different Ionophores. With the different receptor analogues
at hand, we were keen to investigate how changing the cavity
size (receptors R2−R4) or using the abietic acid scaffold
D
DOI: 10.1021/acs.joc.9b01665
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
(receptor R5) would influence the binding affinity for
ionophoric toxins, which themselves have different sizes.
While our previous endeavors were mainly directed toward
the binding of beauvericin or valinomycin (VAL), in this work,
we decided to expand the number of tested ionophores.
Therefore, enniatin B, the enniatin subtype most frequently
found in contaminated food (vide supra), was also included to
obtain a clearer picture on the relationship between receptor
structure and affinity for a particular toxin. Valinomycin (VAL)
was tested because of its strong analogy to BEA. Both are
These results seem to indicate that the use of a scaffold
containing the trans-decalin system is clearly advantageous
compared to one containing a cis-fused ring (e.g., lithocholic
acid), when higher affinities are required. Finally, none of the
receptors showed appreciable interaction with the larger
ionophore valinomycin, which is in line with our previous
observation for receptor R1.22 For further applications, the
more complicated synthesis of receptors R3 and R4 is not
outweighted by the small increase in affinity compared to
receptor R2. Abietic acid-based receptor R5, on the other
hand, does show a significantly higher affinity, but the loss of
selectivity for the different ionophores makes it less attractive.
Taking all of the above into account, receptor R2 was chosen
for subsequent experiments aimed at exploring its use in
potential applications.
+
cyclodepsipeptides, able to bind NH4 . Yet their size is
significantly different (Figure 5). To determine the binding
Synthesis and Immobilization of the Receptor for
Solid-Phase Extraction (SPE). Currently, analysis of
mycotoxins is generally achieved using liquid chromatog-
raphy−tandem mass spectrometry (LC−MS/MS), through
direct injection of the sample in a diluted fashion without any
pretreatment step. Even though this can be done for simple
food matrices such as pasta, bread, rice, etc., the analysis of
complex food samples is far more challenging. Because of the
complexity of the matrix in the latter, sample pretreatment is
usually required. This is mainly performed using SPE
cartridges that contain a solid support with affinity for the
analyte. After addition of a sample to the cartridge, the analyte
will be retained more effectively on the solid support compared
to matrix components. By using an appropriate eluent (eluent
1 in Figure 7), the interfering matrix can be eluted, while the
analyte remains bound to the sorbent. Subsequently switching
toward a more competitive eluent (eluent 2) allows isolation of
the analyte.
Figure 5. Chemical structure of valinomycin (left) and tabulation of
the molecular weights of the different ionophores (right).
affinity, fluorescence titration was used (Figures S5−S8 in the
SI), as already described above for receptor R2. The resulting
association constants for BEA and ENN B are depicted in
Figure 6 (KBEA and KENN, respectively).
A couple of interesting observations can be made when
looking at the experimental data. Upon decreasing the cavity
size, the binding affinity for BEA increased, while this was not
the case for the smaller toxin ENN B. Rather unexpectedly, the
highest affinities were obtained with the abietic acid-based
receptor R5, which showed an affinity of 1.6 × 106 M−1. This is
a 10-fold increase in binding affinity for BEA compared to
receptor R2. However, the higher affinity came at the cost of
completely losing the selectivity between the individual
ionophores.
For most mycotoxins, highly specific SPE cartridges exist
based on antibodies. However, due to the lack thereof for BEA
and ENN, such cartridges are unavailable for these toxins.
Therefore, only nonselective SPE cartridges have been used,
such as C18. Recently, also a graphitized carbon black sorbent
Figure 6. CuAAC reaction of carbazole 5 with different azido-functionalized hydrophobic scaffolds (3, 9, and 15) gave intermediates 6, 10−12, and
16 (structures not shown). Subsequent Boc deprotection yielded receptors R2−R5. The association constants of each receptor for BEA (KBEA) and
ENN B (KENN) were determined using fluorescence titration (λex = 329 nm) in CHCl3 (mean of three independent titrations). Standard deviation
values for each titration can be found in Table S2 (SI).
E
DOI: 10.1021/acs.joc.9b01665
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
trinitrobenzenesulfonic acid (TNBS) test indicated incomplete
coupling. Therefore, a capping step with Ac2O and Et3N was
performed to block any residual free amine on the resin. The
successful capping was confirmed by the TNBS test.
Next to the immobilization of R6 (Si-R6), acetylated resin
(Si-NHAc) (Scheme 4B) and 3-aminopropyl base resin (Si-
NH2) were evaluated as controls. While the former should only
feature background absorption related to the resin material, the
latter can show nonselective ionophoric interactions with the
toxin.
Figure 7. Illustration of the principle behind SPE.
To assess which of the three resins was most effective at
retaining BEA, each resin was incubated overnight with 3 mL
of a 100 nM BEA solution in hexane. The next day, the solvent
was removed (=rest fraction) and the resin was subjected to a
gradient elution (hexane/EtOAc) (Figure 8). The beauvericin
concentration in each collected fraction was determined via
LC−MS/MS.35
It is to be expected that the stronger the binding between
the resin and BEA, the more polar the eluent must be to
disrupt this interaction. As shown in Figure 8 (bottom left), a
pronounced difference between the resins could be observed.
The toxin is retained far more efficiently on the resin
containing immobilized R6 (pink) compared to the acetylated
(blue) or amino-functionalized resin (yellow). The observed
retention capabilities of the resins increase in the order: Si-
NHAc < Si-NH2 < Si-R6. This corresponds to the expected
interaction strength between the modified solid supports and
the toxin (vide supra).
has been employed for the extraction of ENN and BEA from
human urine and plasma.34
Our goal was to synthesize an analogue of R2 containing
carboxylic acid handles rather than esters (R6) for coupling
onto 3-aminopropyl-functionalized silica (Scheme 4A). This
would potentially endow the sorbent material with affinity for
BEA. Such material could find practical use in the solid-phase
extraction of ionophoric toxins.
Scheme 4. Immobilization of R6 onto Solid Support (A)
and Acetylation of the Resin (B)
More importantly, our immobilized receptor proved to be
superior in retaining BEA compared to Si-NH2, which can be
attributed to a combination of the hydrophobic cavity size with
the ammonium functionality interacting in concert with the
toxin. As such, this SPE experiment further validates the
important contribution of the receptor cavity with respect to
toxin binding and is consistent with the 1000-fold lower
affinity measured for the individual carbazole fragment
compared to the receptor, as mentioned earlier. Moreover, it
is worth noting that BEA is retained poorly on the acetylated
resin, which indicates that hydrophobic interactions alone are
not sufficient for binding the toxin. This is in line with our
observation that when the amine of the receptor is Boc-
protected (hence only hydrophobic interactions are available
for toxin binding), no change in fluorescence intensity is
observed (vide supra). The fact that Si-R6 is able to retain
BEA provides further evidence for the specific interaction
between our synthetic receptor and BEA, in accordance with
the obtained fluorescence titration data (vide supra).
The same fluorescence titration experiments also revealed
that receptor R2 did not bind the larger ionophore
valinomycin. Upon using a 100 nM VAL solution, interest-
ingly, all three resins behave similarly with respect to VAL
retention (Figure 8, bottom right), which is in strong contrast
with the results obtained for BEA. The selectivity observed for
R2 in fluorescence titration is thus maintained upon its
immobilization onto solid support. These experiments
independently confirm the observations made during the
fluorescence titrations regarding the affinity of the receptor for
BEA. At the same time, they constitute a first step toward the
development of a selective SPE cartridge for beauvericin.
Application of R2 in Cell Cultures. Although our
endeavors in developing synthetic receptors for ionophoric
toxins were initially driven with analytical applications in mind
(e.g., their use in solid-phase extraction), it became tempting
The synthesis of R6 was accomplished via a route analogous
to the one described in Scheme 1 (see Experimental Section).
Subsequent immobilization on 3-aminopropyl-functionalized
silica (Si-NH2) using 2-(1H-benzotriazol-1-yl)-1,1,3,3-tetrame-
thyluronium hexafluorophosphate (HBTU) and N,N-diisopro-
pylethylamine (DIPEA) proceeded in 51% yield (Scheme 4A).
Indeed, red coloration of the beads after the 2,4,6-
F
DOI: 10.1021/acs.joc.9b01665
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Figure 8. Illustration of the SPE experimental setup (top). Gradient elution of BEA (bottom left) and VAL (bottom right) using acetylated resin
(blue), 3-aminopropyl-functionalized silica resin (yellow) modified with R6 (pink). The area under the curve of the most abundant BEA or VAL
fragment ion in each collected fraction (LC−MS/MS) is shown on the vertical axis.
to consider if the recognition capabilities of receptor R2
toward BEA could be used in a medicinal/pharmaceutical
setting. By binding beauvericin, R2 might be able to act as
antidote and prevent cell death. Currently, only one procedure
has been reported in the literature, in which the authors use
the antioxidant resveratrol to counteract the toxicity of
beauvericin in vitro.36 However, this method only tackles
one of the downstream effects of BEA toxicity (e.g., reactive
oxygen species (ROS) production) and not the actual
interaction of BEA with host cells. This interaction goes
deeper and beyond ROS production. Work by Jow and co-
workers has shown that addition of 10 μM BEA to human
leukemia cells (CCRF-CEM) induced the release of
cytochrome c from mitochondria, which in turn causes
apoptosis via the caspase-3 pathway.10 Evidence has also
been put forward for the binding of beauvericin to DNA37 and
its ability to inhibit the enzyme acyl-CoA/cholesterol
acyltransferase.38 By sequestering the toxin using a synthetic
receptor, all downstream pathways would be inhibited, not
only ROS production.
As a proof of principle, we investigated if a mixture of BEA
and R2 would be less toxic in vitro compared to BEA alone.
For these experiments, human liver carcinoma (HepG2) cells
were chosen based on earlier reports that BEA accumulates in
the liver when administered to mice.11 Cell viability was
assessed using three complementary assays. The first assay is
the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazoliumbro-
mide (MTT) assay, which is based on the reduction of
MTT by viable cells to the purple dye formazan.39 It is worth
noting that although this assay is often interpreted as a cell
viability assay, it is actually a metabolic activity assay. If a
compound is employed which is nontoxic but inhibits the
reduction of MTT to formazan, the assay would incorrectly be
interpreted as a decrease in viable cell count.
Hence, in this work, we conducted two additional assays to
validate the results obtained with MTT, namely, the neutral
red (NR) and sulforhodamine (SRB) assay. The NR assay is
based on the accumulation of the dye neutral red in lysosomes
of living cells, which is linearly dependent on the number of
viable cells.40 The SRB assay, on the other hand, relies on the
binding of sulforhodamine B to proteins and, as such, is a
measure of protein content (e.g., cell density, unless the
compound under study severely alters protein expression).41
While none of these assays are fail-safe on their own as a
measure of cell viability, using all three of them in concert
improves the reliability of the interpretation.
In a first experiment, different quantities of BEA were
dissolved in DMF, followed by addition of growth medium to
obtain solutions containing 1% DMF and BEA concentrations
ranging from 0.2 to 30 μM. This solvent was chosen due to its
solvating capability toward both the toxin and the receptor.
Control experiments showed that 1% DMF is well tolerated by
the cells while higher concentrations are cytotoxic (data not
shown). The three assays clearly showed signs of toxicity after
24 h incubation of HepG2 cells with beauvericin, starting at 2.5
μM BEA (Figures S10 and S12 in the SI). At 15 μM, complete
cell death was observed, which is consistent with what has been
reported by others on the same cell line.42
Before performing the toxicity assays of receptor R2−
beauvericin mixtures, the toxicity of receptor R2 alone was
investigated. Indeed, if the receptor would be highly toxic on
its own, its use as an antidote would be jeopardized. The
results obtained showed no toxicity of R2 in any of the
deployed assays (MTT, NR, and SRB), even at the highest
dose tested of 200 μM (Figures S14 and S16 the in SI).
G
DOI: 10.1021/acs.joc.9b01665
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Based on the above results, we decided to prepare mixtures
of 5 μM BEA in combination with 1 or 10 equiv R2, while
maintaining the total DMF concentration constant. As shown
in Figure 9, addition of 1 equiv R2 to 5 μM BEA was not
sufficient to significantly reduce the toxicity observed visually
under the light microscope (Figure 9B vs Figure 9C).
Figure 10. Results of the NR assay after 24 h incubation of HepG2
cells at 37 °C with different concentrations of beauvericin alone or in
combination with 1 or 10 equiv receptor R2. The error bars represent
the relative SD value.
Figure 9. Image under the light microscope of HepG2 cells after 24 h
incubation at 37 °C with (A) 1% DMF, (B) 5 μM BEA, (C) 5 μM
BEA + 1 equiv R2, and (D) 5 μM BEA + 10 equiv R2.
excess receptor, the concentration of free toxin decreases,
which in turn reduces the observed toxicity.
Finally, as a control, the Boc-protected derivative of R2 was
also tested, which lacks the ionophoric interaction with the
toxin. In this case, no decrease in cytotoxicity was notable
(Figure S23 in the SI), thereby highlighting the importance of
the ammonium functionality in R2. This is in accordance with
the observation that the Boc-protected receptor does not show
any increase in fluorescence intensity upon addition of
beauvericin (vide supra).
It is important to stress that if receptor R2 would not be able
to bind BEA, a reduction in cytotoxicity should not be
observed. Likewise, during the SPE experiments, the strong
retention of BEA by the immobilized receptor confirms the
binding event. While one might want to confirm the binding
affinity with more classical techniques such as ITC or NMR,
solubility limitations prevented their use (vide supra). Both the
fact that the receptor is able to retain BEA in an SPE retention
experiment and the observation that the receptor reduces
beauvericin-induced cytotoxicity provide clear and independ-
ent evidence of the affinity and selectivity of the receptor for
beauvericin.
However, addition of 10 equiv R2 clearly led to a lower
toxicity (Figure 9D) compared to cells treated with 5 μM BEA
alone (Figure 9B). This was also confirmed with the MTT,
SRB, and NR assay (Figures S18 and S20 in the SI). A similar
observation was made in the experiments using 2.5 μM BEA
(SI).
More interestingly, a mixture of 10 equiv R2 and 15 μM
BEA still showed a cell survival rate of 60% in the NR assay.
This is in strong contrast to those cells that were treated with
15 μM BEA alone, which causes complete cell death (Figure
10). These results demonstrate the remarkable ability of R2 to
protect cells from BEA-mediated cytotoxicity, which was also
confirmed with the SRB assay (Figure S20 in the SI).
It is not entirely surprising that an equimolar mixture of R2
and BEA at 15 μM concentration only has a reduced
protecting effect on the cells, having determined that the
stability constants of the 1:1 complexes are in the range of
105−106 M−1. Indeed, under those conditions, a great extent of
the ionophore is still free in solution. The protecting effect of
R2 increases as the number of equivalents is increased. This
result can be considered a direct consequence of the reduction
of free ionophore in solution. Furthermore, also nonselective
binding of R2 to compounds present in the complex cell
medium (e.g., serum albumin) could play a role. It is worth
noting that during the course of the experiment, fetal bovine
serum was always present. While we could have opted to avoid
the presence of serum during the treatment of the cells with
the receptor−toxin mixtures, we feel that performing the
experiments with this complex additive present is more
representative of the real-life situation. Moreover, considering
the fact that the affinity of receptor R2 for BEA was in the 105
M−1 range, addition of only 1 equiv R2 is not sufficient to bind
all of the toxin in solution. As such, the amount of free toxin is
still high enough to cause cytotoxicity. Upon addition of an
CONCLUSIONS
■
In summary, we here report on an improved synthetic route
toward an artificial receptor with affinity for beauvericin (8
steps compared to 12 steps reported previously). The binding
affinity of the receptor for beauvercin was confirmed by three
independent methods: fluorescence titration, SPE retention
experiments, and cell assays. Additionally, the highly optimized
convergent synthetic route was further exploited in a versatile
way to obtain a series of new receptor analogues with varying
cavity sizes or hydrophobic moieties, all of which were tested
for affinity against three different ionophoric toxins. While
receptor cavity size had a moderate impact on the affinity, the
use of an abietic-based scaffold instead of lithocholic acid
increased the binding affinity considerably, albeit at the cost of
H
DOI: 10.1021/acs.joc.9b01665
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
56.7, 76.8, 167.0, 174.9. HRMS m/z: calcd for C27H47BrNO4 [M +
NH4]+ 528.2683, found 528.2680.
selectivity. The obtained data allow to rationalize and guide
further design efforts.
Methyl-3α-azidoacetoxy-5β-cholan-24-oate (3). Compound 2
(1.3 g, 2.5 mmol) was dissolved in dry DMF, to which NaN3 (963
mg, 14.8 mmol, 6 equiv) was added, followed by stirring at rt for 24 h.
The mixture was poured in 150 mL of H2O and extracted with
EtOAc. The organic phase was concentrated in vacuo, and the crude
was dry-loaded onto 4 g of silica. Purification via flash chromatog-
raphy (eluent 90:10 hexane/EtOAc) yielded compound 3 as a white
powder (987 mg, 84%). 1H NMR (300 MHz, CDCl3) δ: 0.65 (s, 3H),
0.80−2.04 (m, 35H), 2.29 (m, 2H), 3.67 (s, 3H), 3.83 (s, 2H), 4.84
(m, 1H). 13C{1H} NMR (75 MHz, CDCl3) δ: 12.2, 18.4, 21.0, 23.5,
24.4, 26.5, 26.8, 27.2, 28.4, 31.19, 31.25, 32.3, 34.8, 35.1, 35.5, 36.0,
40.3, 40.6, 42.1, 42.9, 50.8, 51.7, 56.2, 56.7, 76.6, 168.0, 175.0. IR
(νmax in cm−1): 2935, 2864, 2106, 1754, 1731. HRMS m/z: calcd for
C27H47N4O4 [M + NH4]+ 491.3592, found 491.3601.
Taking into account the widespread occurrence of
beauvericin in food samples and the fact that a recent
European Food Safety Authority report43 urges the acquisition
of more data to evaluate the chronic toxicity of beauvericin,
improved detection methods for this toxin are urgently needed.
To date, there is a total lack of SPE cartridges that are selective
for beauvericin. This limits the possibility of detecting the toxin
in complex food samples, which require a pretreatment step
with SPE. In this context, we immobilized the most promising
receptor on solid support, which led to a strong retention of
beauvericin in a solid-phase extraction setup. This yielded the
first sorbent material with affinity for beauvericin with the
potential to discriminate between BEA and other ionophoric
cyclodepsipeptides such as valinomycin. As such, the
developed material can be valuable for future application in
the capture and detection of beauvericin in complex food
samples.
Finally, we showed for the first time that this type or artificial
receptor can be used to reduce beauvericin-mediated
cytotoxicity in HepG2 cells, thus paving the way toward its
potential use as an antidote.
Future work will be aimed at shining light on the mechanism
behind this protective effect. In addition, the differential
response of each receptor analogue for the ionophoric toxins
opens up the possibility of using this small receptor library in
the development of a chemical nose sensor44 for the detection
of beauvericin and enniatins.
Methyl (4R)-4-((3R,8S,9R,13S,14R)-3-(2-(4-(9-(2-((tert-
Butoxycarbonyl)amino)ethyl)-7-(1-(2-(((3R,5R,8R,9S,10S,13R,14S)-
17-((R)-5-methoxy-5-oxopentan-2-yl)-10,13-di- methylhexadeca-
hydro-1H-cyclopenta[a]phenanthren-3-yl)oxy)-2-oxoethyl)-1H-
1,2,3-triazol-4-yl)-9H-carbazol-2-yl)-1H-1,2,3-triazol-1-yl)acetoxy)-
10,13-dimethylhexadecahydro-1H-cyclopenta[a]phenanthren-17-
yl)pentanoate (6). Compound 5 (synthesized as reported previously
starting from 2,7-dibromocarbazole 4)22 (24 mg, 0.066 mmol) and
compound 3 (78 mg, 0.166 mmol, 2.5 equiv) were dissolved in 4 mL
of DMF, followed by addition of 6 mL of tBuOH. CuSO4·5H2O (33
mg, 0.132 mmol, 2 equiv) and Na-L-ascorbate (52 mg, 0.264 mmol, 4
equiv) were separately dissolved in 1 mL of H2O each and added to
the reaction mixture. The reaction was flushed under N2 for 15 min
and stirred at 60 °C overnight. The solvent was partially evaporated,
and the residue was poured in a mixture of 80 mL of H2O and 30 mL
of EtOAc. The aqueous phase is extracted two more times with
EtOAc. The collected fractions were evaporated, and the residue was
dry-loaded onto 1 g of silica. Purification via flash chromatography
(eluent 1:1 hexane/EtOAc) yielded compound 6 as a pink powder
EXPERIMENTAL SECTION
■
1
(69 mg, 80%). H NMR (400 MHz, CDCl3) δ: 0.64 (s, 6H), 0.87−
General Information. Chemical shifts are reported in parts per
million (ppm) using tetramethylsilane as an internal standard. All
chemicals were purchased from Sigma-Aldrich with the exception of
beauvericin, which was purchased from Fermentek, and 2,7-
dibromocarbazole, which was purchased from TCI. Column
chromatography was performed on 200 μm silica gel (Davisil,
Grace Davison). Yields are reported for the isolated compounds.
High-resolution mass spectrometry (HRMS) images were recorded
on an Agilent 1100 series high-performance liquid chromatography
connected to an Agilent 6220A time-of-flight high-resolution mass
detector equipped with an electrospray ionization (ESI)/atmospheric-
pressure chemical ionization multimode ionization source. Unless
otherwise stated, ionization was performed in positive ESI mode.
Infrared spectra were recorded on a PerkinElmer 1000 FTIR
spectrometer equipped with a Pike ATR module.
2.00 (m, 78H), 2.28 (m, 4H), 3.63 (s + m, 8H), 4.54 (t, 3J = 5.4 Hz,
2H), 4.72 (t, 3J = 5.2 Hz, 1H), 4.87 (m, 2H), 5.20 (s, 4H), 7.67 (d, 3J
= 8.1 Hz, 2H), 8.02 (s, 4H), 8.09 (d, 3J = 8.1 Hz, 2H). 13C{1H} NMR
(101 MHz, CDCl3) δ: 12.2, 18.5, 21.0, 23.5, 24.4, 26.5, 26.7, 27.2,
28.4, 28.5, 31.18, 31.24, 32.3, 34.8, 35.1, 35.5, 36.0, 40.3, 40.6, 42.1,
42.9, 51.4, 51.7, 56.1, 56.6, 77.3, 106.2, 117.8, 120.9, 121.4, 123.0,
128.5, 141.7, 149.2, 166.0, 175.0. HRMS m/z: calcd for
C77H109N8O10 [M + H]+ 1305.8261, found 1305.8272.
2-(2-(1-(2-(((3R,5R,8R,9S,10S,13R,14S)-17-((R)-5-Methoxy-5-oxo-
pentan-2-yl)-10,13-dimethylhexadecahydro-1H-cyclopenta[a]-
phenanthren-3-yl)oxy)-2-oxoethyl)-1H-1,2,3-triazol-4-yl)-7-(1-(2-
(((3R,8S,9R,13S,14R)-17-((R)-5-methoxy-5-oxopentan-2-yl)-10,13-
dimethylhexadecahydro-1H-cyclopenta[a]phenanthren-3-yl)oxy)-
2-oxoethyl)-1H-1,2,3-triazol-4-yl)-9H-carbazol-9-yl)ethan-1-amo-
nium Chloride (R2). To compound 6 (6.9 mg, 0.005 mmol), 2 mL of
4 M HCl in dioxane was added. The mixture was stirred at rt for 1 h.
Evaporation of the solvent yielded receptor R2 quantitatively as a pink
Chemical Synthesis. Methyl-3α-bromoacetoxy-5β-cholan-24-
oate (2). Lithocholic acid 1 (2.0 g, 5.5 mmol) was dissolved in 30 mL
of MeOH, followed by addition of 0.4 mL of concentrated H2SO4.
The mixture was stirred at room temperature (rt) overnight. The
solvent was removed in vacuo, and the residue was dissolved in
EtOAc. After washing with H2O, the organic phase was evaporated,
yielding the methyl ester as a white solid. The ester was dissolved
without any further purification in 20 mL of dry tetrahydrofuran
(THF), followed by addition of 4 equiv pyridine (1.8 mL, 22.1
mmol). The mixture was cooled to 0 °C, and bromoacetyl bromide
(1.0 mL, 11.0 mmol, 2 equiv) was added, followed by stirring for 2 h
at 0 °C. The mixture was poured in H2O, extracted with
dichloromethane (DCM), and the organic phase was evaporated in
vacuo. The crude was dry-loaded onto 4 g of silica and purified via
flash chromatography (eluent 95:5 hexane/EtOAc) yielding com-
1
powder. H NMR (400 MHz, DMF-d7) δ: 0.67 (s, 6H), 0.84−2.01
(m, 70H), 2.32 (m, 4H), 3.64 (s, 6H), 3.68 (m, 2H), 4.83 (m, 2H),
5.10 (t, 3J = 6.6 Hz, 2H), 5.60 (s, 4H), 7.93 (d, 3J = 7.8 Hz, 2H), 8.29
(d, 3J = 7.8 Hz, 2H), 8.56 (s, 2H), 8.97 (s, 2H), 9.05 (s, 3H). 13C{1H}
NMR (101 MHz, DMF-d7) δ: 12.6, 18.9, 21.7, 23.8, 25.0, 27.2, 27.4,
27.9, 29.0, 32.0, 33.0, 36.8, 39.8, 41.0, 41.3, 42.7, 43.6, 44.6, 51.9,
56.9, 57.3, 61.9, 69.0, 71.9, 72.0, 73.7, 77.1, 107.4, 118.1, 121.7, 123.6,
124.2, 130.2, 142.7, 149.0, 168.1, 175.0. HRMS m/z: calcd for
C72H101N8O8 [M + H]+ 1205.7737, found 1205.7736.
Methyl-3-oxo-5β-cholan-24-oate (7). Lithocholic acid 1 (1 g, 2.65
mmol) was dissolved in 30 mL of MeOH. Concentrated H2SO4 (0.4
mL) was injected, and the mixture was stirred overnight at rt. After
evaporation of the solvent, the residue was dissolved in 40 mL of
EtOAc and washed with H2O. The organic phase was evaporated, and
the crude was used without further purification. The obtained methyl
ester was dissolved in 15 mL of dry THF, followed by addition of 2.9
g of pyridinium chlorochromate (13.25 mmol, 5 equiv). The mixture
was stirred at rt for 24 h. The solvent was removed in vacuo, and the
1
pound 2 as a white solid (1.3 g, 45%). H NMR (400 MHz, CDCl3)
δ: 0.64 (s, 3H), 0.87−2.00 (m, 34H), 2.28 (m, 2H), 3.66 (s, 3H),
3.80 (s, 2H), 4.79 (m, 1H). 13C{1H} NMR (101 MHz, CDCl3) δ:
12.2, 18.5, 21.0, 23.5, 24.4, 26.47, 26.56, 26.60, 27.2, 28.4, 31.19,
31.25, 32.1, 34.8, 35.1, 35.5, 36.0, 40.3, 40.6, 42.1, 42.9, 51.7, 56.2,
I
DOI: 10.1021/acs.joc.9b01665
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
residue was dry-loaded onto 3 g of silica. Purification via flash
chromatography (eluent 90:10 hexane/EtOAc) yielded compound 7
as a white powder (672 mg, 65% in two steps). 1H NMR (400 MHz,
mmol) and compound 9 (72 mg, 0.16 mmol, 8 equiv) were separately
dissolved in 4 mL of DMF each and added to the reaction flask,
followed by 4 mL of tBuOH. CuSO4·5H2O (10 mg, 0.04 mmol, 2
equiv) and Na-L-ascorbate (16 mg, 0.08 mmol, 4 equiv) were
separately dissolved in 1 mL of H2O each and added to the flask. The
mixture was flushed for 15 min with N2 and stirred at 60 °C for 24 h.
The solvent was removed in vacuo, and the crude was purified via
flash chromatography (eluent 70:30 EtOAc/hexane), yielding title
3
CDCl3) δ: 0.67 (s, 3H), 0.91 (d, J = 6.5 Hz, 3H), 1.00 (s, 3H),
1.03−2.41 (m, 28H), 2.68 (t, 3J = 14.3 Hz, 1H), 3.65 (s, 3H).
13C{1H} NMR (101 MHz, CDCl3) δ: 12.2, 18.5, 21.4, 22.9, 24.3,
26.0, 26.8, 28.4, 31.14, 31.2, 35.1, 35.5, 35.7, 37.2, 37.4, 40.2, 40.9,
42.5, 42.9, 44.5, 51.7, 56.1, 56.6, 174.9, 213.6. IR (νmax in cm−1):
2925, 2868, 2850, 1735, 1711. HRMS m/z: calcd for C25H44NO3 [M
+ NH4]+ 406.3316, found 406.3329.
1
compound 11 as a yellow powder (16 mg, 66%). H NMR (700
MHz, CDCl3) δ: 0.66 (s, 6H), 0.79−2.05 (m, 74H), 2.30 (m, 4H),
3.43 (m, 2H), 3.67 (s, 6H), 4.39 (m, 2H), 4.96 (s, 1H), 7.38−8.24
(m, 8H). 13C{1H} NMR (176 MHz, CDCl3) δ: 12.2, 18.4, 21.3, 23.8,
24.4, 26.6, 26.7, 28.4, 28.5, 29.9, 30.0, 31.2, 31.3, 35.1, 35.6, 35.8,
38.0, 40.0, 40.3, 42.9, 51.6, 56.2, 56.7, 71.9, 117.6, 120.8, 123.0, 141.5,
175.0. HRMS m/z: calcd for C75H109N8O8 [M + H]+ 1249.8363,
found 1249.8373.
(3S)-Spiro[24-methyl-5β-cholan-24-oate-3,2′-oxirane] (8). Tri-
methylsulfoxonium iodide (1.1 g, 5.12 mmol, 1.5 equiv) and NaH
(273 mg, 6.82 mmol, 2 equiv) were dissolved in 10 mL of dry
dimethyl sulfoxide (DMSO). This mixture was stirred for 1 h at rt.
Compound 7 (1.3 g, 3.41 mmol) dissolved in 5 mL of dry THF was
added, and the resulting mixture was stirred at rt for 24 h. The
mixture was poured into 100 mL of H2O and extracted with DCM.
The combined organic phases were evaporated in vacuo, and the
crude was dry-loaded onto 3 g of silica. Purification via flash
chromatography (eluent 90:10 hexane/EtOAc) yielded compound 8
2-(2-(1-(((3S,5R,8R,9S,10S,13R,14S)-3-Hydroxy-17-((R)-5-me-
thoxy-5-oxopentan-2-yl)-10,13-dimethylhexadecahydro-1H-
cyclopenta[a]phenanthren-3-yl)methyl)-1H-1,2,3-triazol-4-yl)-7-(1-
(((3S,8S,9R,13S,14R)-3-hydroxy-17-((R)-5-methoxy-5-oxopentan-2-
yl)-10,13-dimethylhexadecahydro-1H-cyclopenta[a]phenanthren-
3-yl)methyl)-1H-1,2,3-triazol-4-yl)-9H-carbazol-9-yl)ethan-1-amo-
nium Chloride (R4). HCl (2 mL, 4 M) in dioxane was added to
compound 11 (6.6 mg) and the mixture was stirred at rt for 1 h.
Evaporation of the solvent in vacuo yielded receptor R4 as a yellow
powder quantitatively. 1H NMR (400 MHz, DMF-d7) δ: 0.66 (s, 6H),
0.82−2.11 (m, 85H), 2.30 (m, 4H), 3.64 (m, 8H), 4.51 (s, 4H), 5.08
1
as a white powder (565 mg, 41%). H NMR (500 MHz, CDCl3) δ:
0.66 (s, 3H), 0.78−2.00 (m, 32H), 2.29 (m, 3H), 2.61 (s, 2H), 3.66
(s, 3H). 13C{1H} NMR (126 MHz, CDCl3) δ: 12.3, 18.5, 21.3, 23.6,
24.4, 26.4, 26.7, 28.33, 28.37, 31.19, 31.24, 33.7, 34.4, 34.8, 35.6, 35.8,
40.2, 40.4, 40.9, 43.0, 51.7, 53.8, 56.2, 56.8, 59.4, 174.9. HRMS m/z:
calcd for C26H43O3 [M + H]+ 403.3207, found 403.3223.
3
3
3
(3S)-Methyl-3-azidomethyl-3-hydroxy-5β-cholan-24-oate (9).
Compound 8 (142 mg, 0.35 mmol) was dissolved in 10 mL of dry
DMF, followed by addition of 229 mg of NaN3 (3.52 mmol, 10
equiv). The mixture was stirred at 80 °C for 2 days. The mixture was
poured into 100 mL of H2O and extracted with EtOAc, followed by
evaporation of the organic phase in vacuo. The crude was purified via
flash chromatography (eluent 50:48:2 DCM/hexane/EtOAc) yielding
(t, J = 5.9 Hz, 2H), 7.93 (d, J = 8.3 Hz, 2H), 8.26 (d, J = 8.7 Hz,
2H), 8.53 (s, 2H), 8.84 (s, 2H), 9.02 (s, 3H). 13C{1H} NMR (101
MHz, DMF-d7) δ: 12.7, 18.9, 22.0, 24.2, 25.0, 27.2, 27.7, 28.9, 31.5,
31.9, 32.2, 36.2, 36.7, 38.7, 39.9, 40.5, 41.1, 41.6, 43.6, 44.6, 51.9,
57.0, 57.4, 61.9, 62.4, 71.7, 72.1, 73.7, 107.3, 118.2, 121.7, 123.6,
124.2, 130.6, 142.7, 148.3, 175.0. ESI-MS m/z: calcd for C70H101N8O6
[M + H]+ 1149.78386, found 1149.6.
1
compound 9 as a white powder (73 mg, 47%). H NMR (400 MHz,
Methyl (4R)-4-((3R,8S,9R,13S,14R)-3-(2-(4-(9-(2-((tert-
Butoxycarbonyl)amino)ethyl)-7-(1-(((3S,5R,8R,9S,10S,13R,14S)-3-
hydroxy-17-((R)-5-methoxy-5-oxopentan-2-yl)-10,13-dimethylhex-
adecahydro-1H-cyclopenta[a]phenanthren-3-yl)methyl)-1H-1,2,3-
triazol-4-yl)-9H-carbazol-2-yl)-1H-1,2,3-triazol-1-yl)acetoxy)-10,13-
dimethylhexadecahydro-1H-cyclopenta[a]phenanthren-17-yl)-
pentanoate (12). Compound 10 (26 mg, 0.032 mmol) and
compound 3 (30 mg, 0.064 mmol, 2 equiv) were dissolved in 2 mL
of DMF, followed by the addition of 10 mL of tBuOH. CuSO4·5H2O
(16 mg, 0.064 mmol, 2 equiv) and Na-L-ascorbate (25 mg, 0.13
mmol, 4 equiv) were separately dissolved in 1 mL of H2O each and
added to the mixture. After 15 min flushing with N2, the mixture was
stirred at 60 °C for 24 h. The solvent was partially evaporated in
vacuo, and the crude was purified via flash chromatography (eluent
70:30 DCM/EtOAc, followed by 60:40 DCM/EtOAc), yielding title
3
CDCl3) δ: 0.64 (s, 3H), 0.90 (d, J = 6.4 Hz, 3H), 0.94−2.00 (m,
32H), 2.28 (m, 2H), 3.24 (s, 2H), 3.66 (s, 3H). 13C{1H} NMR (101
MHz, CDCl3) δ: 12.2, 18.4, 21.2, 23.8, 24.3, 26.4, 26.7, 28.3, 29.8,
31.17, 31.22, 31.33, 35.1, 35.48, 35.52, 35.8, 38.0, 39.9, 40.3, 42.9,
51.7, 56.2, 56.7, 63.1, 72.6, 174.9. IR (νmax in cm−1): 3508, 2935,
2866, 2100, 1715. HRMS m/z: calcd for C26H47N4O3 [M + NH4]+
463.3643, found 463.3641.
Methyl (4R)-4-((3S,5R,8R,9S,10S,13R,14S)-3-((4-(9-(2-((tert-
Butoxycarbonyl)amino)ethyl)-7-ethynyl-9H-carbazol-2-yl)-1H-
1,2,3-triazol-1-yl)methyl)-3-hydroxy-10,13-dimethylhexadecahy-
dro-1H-cyclopenta[a]phenanthren-17-yl)pentanoate (10). Com-
pound 5 (21 mg, 0.06 mmol) and compound 9 (66 mg, 0.15
mmol, 2.5 equiv) were separately dissolved in 4 mL of DMF each and
added to the reaction flask, followed by 20 mL of tBuOH. CuSO4·
5H2O (30 mg, 0.12 mmol, 2 equiv) and Na-L-ascorbate (47 mg, 0.24
mmol, 4 equiv) were separately dissolved in 2 mL of H2O each and
injected into the flask. The mixture was flushed for 15 min with N2,
heated to 60 °C, and stirred at that temperature for 24 h. The solvent
was removed in vacuo. The crude was purified via flash
chromatography (eluent 70:30 DCM/EtOAc), which yielded
1
compound 12 as a yellow powder (21 mg, 52%). H NMR (700
MHz, CDCl3) δ: 0.65 (m, 6H), 0.78−2.04 (m, 76H), 2.29 (m, 4H),
3.62 (m, 8H), 4.42 (m, 4H), 4.88 (m, 1H), 5.26 (m, 3H), 7.40−8.72
(m, 8H). 13C{1H} NMR (176 MHz, CDCl3) δ: 12.3, 15.6, 18.5, 21.1,
21.3, 23.6, 23.8, 24.4, 26.5, 26.6, 26.8, 27.2, 28.4, 29.9, 31.2, 31.3,
31.4, 32.3, 34.8, 35.1, 35.6, 35.8, 35.91, 35.97, 40.0, 40.3, 40.7, 42.1,
42.9, 43.0, 51.7, 56.2, 56.6, 175.2. HRMS m/z: calcd for C76H109N8O9
[M + H]+ 1277.8312, found 1277.8317.
1
compound 10 as a brown powder (19 mg, 41%). H NMR (400
MHz, CDCl3) δ: 0.65 (s, 3H), 0.75−2.01 (m, 45H), 2.29 (m, 2H),
3
3.15 (s, 1H), 3.57 (s, 2H), 3.67 (s, 3H), 4.41 (s, 4H), 4.72 (dd, J =
2-(2-(1-(((3S,5R,8R,9S,10S,13R,14S)-3-Hydroxy-17-((R)-5-me-
thoxy-5-oxopentan-2-yl)-10,13-dimethylhexadecahydro-1H-
cyclopenta[a]phenanthren-3-yl)methyl)-1H-1,2,3-triazol-4-yl)-7-(1-
(2-(((3R,8S,9R,13S,14R)-17-((R)-5-methoxy-5-oxopentan-2-yl)-
10,13-dimethylhexadecahydro-1H-cyclopenta[a]phenanthren-3-
yl)oxy)-2-oxoethyl)-1H-1,2,3-triazol-4-yl)-9H-carbazol-9-yl)ethan-
1-amonium Chloride (R3). HCl (2 mL, 4 M) in dioxane was added
to compound 12 (2.3 mg, 0.0018 mmol) at 0 °C, followed by stirring
for 30 min at 0 °C. The solvent was removed with compressed air
while keeping the solution at 0 °C. Title compound R3 was obtained
53.8 and 6.9 Hz, 2H), 4.86 (m, 1H), 7.34 (m, 1H), 7.51 (m, 2H),
7.93 (m, 2H), 8.08 (s, 1H), 8.29 (m, 1H). 13C{1H} NMR (101 MHz,
CDCl3) δ: 12.3, 18.5, 21.3, 23.8, 24.4, 26.5, 26.6, 28.4, 28.5, 29.9,
31.2, 31.3, 35.1, 35.6, 35.8, 38.0, 40.0, 40.3, 42.9, 51.7, 56.2, 56.6,
70.3, 72.0, 72.5, 74.4, 80.0, 112.9, 117.7, 119.8, 120.5, 121.2, 123.0,
123.2, 123.7, 141.2, 156.2, 166.3, 168.5, 175.1. HRMS m/z: calcd for
C49H66N5O5 [M + H]+ 804.5059, found 804.5062.
Methyl (4R)-4-((3S,8S,9R,13S,14R)-3-((4-(9-(2-((tert-
Butoxycarbonyl)amino)ethyl)-7-(1-(((3S,5R,8R,9S,10S,13R,14S)-3-
hydroxy-17-((R)-5-methoxy-5-oxopentan-2-yl)-10,13-dimethylhex-
adecahydro-1H-cyclopenta[a]phenanthren-3-yl)methyl)-1H-1,2,3-
triazol-4-yl)-9H-carbazol-2-yl)-1H-1,2,3-triazol-1-yl)methyl)-3-hy-
droxy-10,13-dimethylhexadeca-hydro-1H-cyclopenta[a]-
phenanthren-17-yl)pentanoate (11). Compound 10 (16 mg, 0.02
1
quantitatively as a yellow powder (2.2 mg). H NMR (400 MHz,
DMF-d7) δ: 0.66 (s, 3H), 0.68 (s, 3H), 0.83−2.44 (m, 99H), 3.60
3
(m), 4.51 (s, 2H), 4.83 (m, 1H), 5.08 (t, J = 6.5 Hz, 2H), 5.60 (s,
2H), 7.92 (double t, 3J = 8.1 Hz, 2H), 8.28 (d, 3J = 8.1 Hz, 2H), 8.52
(s, 2H), 8.81 (s, 1H), 8.93 (s, 1H), 8.95 (s, 3H). 13C{1H} NMR (101
J
DOI: 10.1021/acs.joc.9b01665
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
MHz, DMF-d7) δ: 12.7, 19.0, 19.2, 21.8, 22.0, 23.5, 23.8, 24.2, 25.0,
27.3, 27.5, 27.9, 29.0, 31.6, 32.0, 33.1, 36.6, 36.9, 38.8, 40.5, 41.1,
41.3, 41.6, 42.8, 43.6, 46.8, 52.0, 57.0, 57.3, 62.0, 62.4, 64.5, 69.0,
69.6, 71.7, 72.0, 74.1, 77.1, 118.2, 121.7, 123.7, 130.7, 142.6, 148.2,
168.1, 175.0. HRMS m/z: calcd for C71H101N8O7 [M + H]+
1177.7788, found 1177.7775.
m/z: calcd for C69H89N10O5 [M + HOAc−OH]− 1137.70174, found
1137.7.
2-(2,7-Bis(1-(2-((((1R,4aS,10aR)-7-isopropyl-1,4a-dimethyl-
1,2,3,4,4a,9,10,10a-octahydrophenanthren-1-yl)methyl)amino)-2-
oxoethyl)-1H-1,2,3-triazol-4-yl)-9H-carbazol-9-yl)ethan-1-amo-
nium Chloride (R5). HCl (2 mL, 4 M) in dioxane was added to
compound 16 (6.1 mg, 0.0056 mmol) and the mixture was stirred at
rt for 1 h. Evaporation of the solvent with compressed air yielded
2-Bromo-N-(((1R,4aS,10aR)-7-isopropyl-1,4a-dimethyl-
1,2,3,4,4a,9,10,10a-octahydrophenanthren-1-yl)methyl)acetamide
(14). Dehydroabietylamine 13 (4.4 g, 9.2 mmol) was dissolved in 20
mL of dry DCM, followed by 3 mL of dry pyridine (36.8 mmol, 4
equiv). The mixture was cooled at 0 °C and bromoacetyl bromide
(1.6 mL, 18.4 mmol, 2 equiv) was slowly added, followed by stirring
at 0 °C for 1 h. The mixture was poured in a solution of 50 mL of
H2O and 30 mL of DCM. Due to poor phase separation, the mixture
was centrifuged to isolate the bottom organic layer. MgSO4 was
added, the mixture was filtered, and the filtrate was evaporated in
vacuo. Purification of the crude with flash chromatography (eluent
50:49:1 DCM/hexane/EtOAc) yielded compound 14 as a white
powder (1.35 g, 36%). On NMR, a small impurity was observed (at
4.07 and 6.62 ppm), which could not be removed during purification.
However, when performing the next reaction, removal of this impurity
could be achieved (see compound 15). 1H NMR (400 MHz, CDCl3)
1
receptor R5 as a yellow powder quantitatively. H NMR (400 MHz,
3
DMF-d7) δ: 0.71−2.39 (m, 137H), 3.03 (q, J = 5.9 Hz, 2H), 3.22
3
(m, 4H), 4.14 (s, 1H), 4.21 (s, 1H), 4.29 (m, 2H), 4.61 (t, J = 7.7
3
Hz, 4H), 5.40 (s, 4H), 6.91 (s, 2H), 7.02 (d, J = 8.3 Hz, 2H), 7.22
(d, 3J = 8.3 Hz, 2H), 7.84 (m, 2H), 8.23 (s, 2H), 8.27 (d, 3J = 8.2 Hz,
3
2H), 8.44 (t, J = 5.9 Hz, 2H), 8.51 (s, 3H), 8.79 (s, 2H). 13C{1H}
NMR (101 MHz, DMF-d7) δ: 19.5, 19.58, 19.79, 24.60, 24.63, 24.67,
26.0, 28.0, 29.3, 34.3, 36.9, 38.4, 38.6, 39.3, 40.3, 44.5, 44.6, 46.0,
50.7, 53.2, 62.0, 64.5, 69.0, 69.5, 72.0, 73.7, 74.2, 107.0, 118.0, 121.7,
123.3, 124.7, 125.3, 127.8, 130.2, 130.8, 136.0, 142.5, 146.4, 148.4,
148.7, 167.2. ESI-MS m/z: calcd for C64H81N10O3 [M + HOAc−
OH]− 1037.64931, found 1037.6.
3α-Bromoacetoxy-5β-cholan-24-oic Acid (17). Lithocholic acid 1
(2 g, 5.3 mmol) was dissolved in 10 mL of dry THF. Pyridine (1.7
mL, 21.3 mmol, 4 equiv) was added and the mixture was cooled at 0
°C, followed by addition of 0.9 mL of bromoacetyl bromide (10.6
mmol, 2 equiv). The mixture was stirred at that temperature for 2 h.
The reaction was quenched by addition of 20 mL of H2O and
extracted with DCM. The organic phase was evaporated in vacuo, and
the residue was dissolved in DCM/MeOH (95:5) and dry-loaded
onto 3 g of silica. Purification via flash chromatography (eluent 95:5
DCM/EtOAc) yielded compound 17 as a white solid (1.6 g, 60%).
1H NMR (500 MHz, CDCl3) δ: 0.65 (s, 3H), 0.83−2.01 (m, 36H),
2.33 (m, 2H), 3.80 (s, 2H), 4.79 (m, 1H). 13C{1H} NMR (126 MHz,
CDCl3) δ: 12.1, 18.3, 20.9, 23.3, 24.2, 26.3, 26.39, 26.42, 27.0, 28.2,
30.78, 30.83, 31.9, 34.6, 34.9, 35.3, 35.8, 40.1, 40.5, 41.9, 42.8, 56.0,
56.5, 76.7, 166.8, 179.3. HRMS m/z: calcd for C26H45BrNO4 [M +
NH4]+ 514.2527, found 514.2535.
3
δ: 0.70−1.98 (m, 46H), 2.30 (d, J = 12.9 Hz, 2H), 2.75−3.33 (m,
7H), 3.91 (m, 2H), 4.07 (m), 6.53 (m, 1H), 6.62 (m), 6.89 (s, 1H),
3
3
7.00 (d, J = 8.2 Hz, 1H), 7.17 (d, J = 8.2 Hz, 1H). 13C{1H} NMR
(101 MHz, CDCl3) δ: 18.75, 18.77, 19.3, 24.3, 25.6, 29.9, 30.6, 33.7,
36.5, 37.7, 38.5, 43.2, 45.9, 47.8, 50.4, 50.8, 51.1, 124.1, 124.5, 127.2,
134.9, 145.9, 147.1, 165.4. HRMS m/z: calcd for C22H36BrN2O [M +
NH4]+ 423.2006, found 423.2022. Note: a peak corresponding to [M
+ NH4]+ + 4 Da was also observed, which is due to a reduced
impurity that was present in the starting material (dehydroabietyl-
amine). Because of the high similarity between both compounds,
separation could not be achieved.
2-Azido-N-(((1R,4aS,10aR)-7-isopropyl-1,4a-dimethyl-
1,2,3,4,4a,9,10,10a-octahydrophenanthren-1-yl)methyl)acetamide
(15). Compound 14 (408 mg, 1 mmol) was dissolved in 15 mL of dry
DMF, followed by the addition of 10 equiv NaN3 (650 mg, 10 mmol)
and stirring at rt for 24 h. The mixture was poured into 80 mL of H2O
and extracted with EtOAc. The organic phase was evaporated, and the
crude was dry-loaded onto 2 g of silica. Purification via flash
chromatography (eluent 70:30 hexane/EtOAc) yielded azide 15 as a
transparent gel (304 mg, 82%). 1H NMR (400 MHz, CDCl3) δ:
0.69−1.98 (m, 50H), 2.30 (d, 3J = 12.9 Hz, 2H), 2.75−3.34 (m, 7H),
3α-Azidoacetoxy-5β-cholan-24-oic Acid (18). Compound 17 (1
g, 2.0 mmol) was dissolved in 15 mL of dry DMF. NaN3 (796 mg, 12
mmol, 6 equiv) was added and the mixture was stirred at rt for 24 h.
Water (60 mL) was added, followed by extraction with EtOAc. The
organic phase was evaporated in vacuo and the crude was purified via
flash chromatography (first 99:1 DCM/EtOAc, then 95:5 DCM/
EtOAc once the product started to elute). Compound 18 was
1
obtained as a white powder (603 mg, 64%). H NMR (400 MHz,
3
CDCl3) δ: 0.65 (s, 3H), 0.92 (m, 6H), 0.98−2.01 (m, 28H), 2.33 (m,
2H), 3.84 (s, 2H), 4.84 (m, 1H). 13C{1H} NMR (101 MHz, CDCl3)
δ: 12.2, 18.4, 21.0, 23.5, 24.4, 26.5, 26.8, 27.2, 28.4, 30.9, 31.1, 32.3,
34.8, 35.1, 35.5, 36.0, 40.3, 40.6, 42.1, 42.9, 50.8, 56.2, 56.6, 76.6,
168.0, 180.1. IR (νmax in cm−1): 3448, 2931, 2866, 2102, 1732, 1714,
1673, 1641. HRMS m/z: calcd for C26H45N4O4 [M + NH4]+
477.3435, found 477.3443.
4.00 (s, 2H), 6.31 (m, 1H), 6.90 (s, 1H), 7.00 (d, J = 8.3 Hz, 1H),
7.17 (d, 3J = 8.2 Hz, 1H). 13C{1H} NMR (101 MHz, CDCl3) δ: 18.7,
18.9, 19.2, 24.2, 25.6, 30.5, 33.6, 36.4, 37.6, 37.7, 38.4, 45.6, 50.0,
52.9, 53.0, 124.1, 124.4, 127.2, 135.0, 146.0, 147.2, 166.7. IR (νmax in
cm−1): 3316, 2925, 2867, 2101, 1655. HRMS m/z: calcd for
C22H33N4O [M + H]+ 369.2649, found 369.2661.
tert-Butyl (2-(2,7-Bis(1-(2-((((1R,4aS,10aR)-7-isopropyl-1,4a-di-
methyl-1,2,3,4,4a,9,10,10a-octahydrophenanthren-1-yl)methyl)-
amino)-2-oxoethyl)-1H-1,2,3-triazol-4-yl)-9H-carbazol-9-yl)ethyl)-
carbamate (16). Compound 5 (20 mg, 0.057 mmol) was dissolved in
2 mL of DMF, followed by addition of 52 mg of compound 15 (0.14
mmol, 2.5 equiv) and 8 mL of tBuOH. CuSO4·5H2O (28 mg, 0.11
mmol, 2 equiv) and Na-L-ascorbate (45 mg, 0.23 mmol, 4 equiv) were
separately dissolved in 1 mL of H2O each and added to the mixture,
followed by flushing for 15 min under N2. The mixture was stirred at
60 °C for 24 h. The solvent was partially removed in vacuo and
poured in 60 mL of H2O, followed by extraction with DCM. The
organic phase was evaporated, and the crude was purified via flash
chromatography (eluent 60:40 EtOAc/DCM), yielding title com-
(4R)-4-((3R,8S,9R,13S,14R)-3-(2-(4-(9-(2-((tert-Butoxycarbonyl)-
amino)ethyl)-7-(1-(2-(((3R,5R,8R,9S,10S,13R,14S)-17-((R)-4-carbox-
ybutan-2-yl)-10,13-dimethylhexadecahydro-1H-cyclopenta[a]-
phenanthren-3-yl)oxy)-2-oxoethyl)-1H-1,2,3-triazol-4-yl)-9H-car-
bazol-2-yl)-1H-1,2,3-triazol-1-yl)acetoxy)-10,13-dimethylhexade-
cahydro-1H-cyclopenta[a]phenanthren-17-yl)pentanoic Acid (R6).
Compound 18 (603 mg, 1.3 mmol, 2.5 equiv) and compound 5 (188
mg, 0.52 mmol) were dissolved in 6 and 2 mL of DMF, respectively,
and added to the reaction flask, followed by the addition of 20 mL of
tBuOH. CuSO4·5H2O (262 mg, 1 mmol, 2 equiv) and Na-L-ascorbate
(415 mg, 2.1 mmol, 4 equiv) were each dissolved in 2 mL of H2O and
added separately to the reaction. After flushing for 15 min under N2,
the mixture was stirred at 60 °C for 20 h. The solvent was removed in
vacuo, and the crude was purified via flash chromatography (first 1:1
hexane/EtOAc + 1% HOAc, then 70:30 EtOAc/hexane + 1%
HOAc). The collected fractions were dissolved in 40 mL of DCM and
washed with H2O to remove the HOAc. Evaporation of the organic
1
pound 16 as a yellow powder (52 mg, 84%). H NMR (400 MHz,
CDCl3) δ: 0.55−2.30 (m, 120H), 2.73−3.34 (m, 17H), 4.28 (s, 2H),
3
4.70 (s, 1H), 5.19 (s, 6H), 6.87 (s, 2H), 6.98 (d, J = 8.6 Hz, 2H),
7.13 (d, 3J = 8.6 Hz, 2H), 7.46 (m, 2H), 8.00 (m, 6H). 13C{1H} NMR
(101 MHz, CDCl3) δ: 18.65, 18.72, 19.3, 24.2, 25.5, 28.6, 28.8, 29.9,
30.4, 33.6, 36.4, 37.6, 37.8, 38.4, 46.0, 50.7, 53.8, 106.3, 117.3, 121.2,
123.1, 124.1, 124.4, 127.1, 134.9, 141.4, 145.9, 147.4, 165.4. ESI-MS
1
phase yielded receptor R6 as a white powder (372 mg, 56%). H
NMR (300 MHz, CDCl3) δ: 0.64 (s, 7H), 0.78−2.19 (m, 112H),
2.31 (m, 4H), 3.64 (m, 2H), 4.53 (m, 2H), 4.87 (m, 2H), 4.97 (m,
K
DOI: 10.1021/acs.joc.9b01665
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
3
1H), 5.23 (s, 4H), 7.65 (d, J = 8.6 Hz, 2H), 8.07 (m, 6H). 13C{1H}
antibiotics (PenStrep, 10 000 units/mL penicillin + 10 000 μg/mL
streptomycin, purchased from Gibco, Thermo Fisher), and 1%
nonessential amino acids and cell medium (Dulbecco’s modified
Eagle’s medium + GlutaMAX, purchased from Gibco, Thermo
Fisher). Cells were incubated in an atmosphere containing 5% CO2.
Preparation of the Well Plates. Once the cells had a suitable
confluence, they were trypsinized and reconstituted in growth
medium. Of this cell suspension, 100 μL was taken and added to
an eppendorf tube containing 100 μL of trypan blue staining solution
NMR (75 MHz, CDCl3) δ: 12.2, 18.4, 21.0, 23.5, 24.3, 26.4, 26.7,
27.2, 28.4, 28.5, 29.9, 31.0, 32.2, 34.7, 35.1, 35.6, 35.9, 39.8, 40.3,
40.6, 42.1, 42.9, 51.7, 56.0, 56.6, 106.4, 117.8, 121.0, 121.7, 123.1,
128.1, 141.7, 149.0, 165.9, 179.2. HRMS m/z: calcd for
C75H105N8O10 [M + H]+ 1277.7948, found 1277.7968.
Immobilization of R6 on Solid Support. Dry DMF (2 mL) was
added to 29 mg of 3-aminopropyl-functionalized silica (loading 1
mmol/g), followed by addition of receptor R6 (33 mg, 0.026 mmol, 1
equiv), HBTU (59 mg, 0.16 mmol, 6 equiv), and DIPEA (0.03 mL,
0.16 mmol, 6 equiv). The mixture was shaken at rt for 24 h. The
solvent was filtered off, and the residue was rinsed with DMF and
DCM, yielding 33 mg of immobilized R6 (51% yield). As expected,
TNBS test showed red coloration, thus indicating the presence of free
amines. Therefore, a capping step was performed as follows: 2 mL of
dry DMF was added to the resin, followed by addition of 0.05 mL of
Ac2O (0.58 mmol, 20 equiv) and 0.08 mL of Et3N (0.58 mmol, 20
equiv). The mixture was shaken for 24 h at rt. The solvent was filtered
off, and the resin was washed with DMF and DCM. Full capping was
confirmed with the TNBS test. Finally, 2 mL of 4 M HCl in dioxane
was added to the resin and the mixture was shaken for 1 h at rt. The
solvent was filtered off, and the residue was washed with DCM. TNBS
test showed intense red coloration of the beads, thus indicating
successful Boc deprotection.
Acetylation of the Solid Support. Dry DMF (2 mL) was added
to 146 mg of 3-aminopropyl-functionalized silica (loading 1 mmol/g),
followed by 0.14 mL of Ac2O (1.46 mmol, 10 equiv) and 0.2 mL of
Et3N (1.46 mmol, 10 equiv). The mixture was shaken at rt for 24 h.
The solvent was filtered off, and the residue was washed with DMF
and DCM, yielding 127 mg of acetylated resin (83%). The TNBS test
was colorless, thus indicating complete acetylation.
SPE Experiments. The resins (3-aminopropyl-functionalized
silica (Si-NH2), Si-NHAc, and Si-R6) were placed in reaction vessels
for peptide synthesis containing a microporous filter. Beauvericin or
valinomycin solution (3 mL, 100 nM) in hexane was added, and the
mixture was shaken overnight (the BEA and VAL solutions were
prepared from 2.8 and 1 mM stock solutions in CHCl3, respectively).
The next day, the solvent was filtered off and kept aside (=rest
fraction). The resin was eluted with 3 mL of the appropriate solvent,
as shown in Figure 8. The collected fractions were evaporated, and the
residue in each sample was reconstituted in 1 mL of MeOH.
Quantification of the BEA or VAL concentration in each sample was
performed via LC−MS/MS, as reported previously,35 by determining
the area under the curve in the total ion count chromatogram of the
most abundant product ion (244.3 Da for BEA and 343.5 Da for
VAL).
Fluorescence Titrations. Fluorescence measurements were
performed on a Cary Eclipse Fluorescence Spectrophotometer
(Varian Australia Pty. Ltd.). The slit width for both excitation and
emission was 5 nm. The temperature was maintained at 25 °C during
the measurements. For receptor R2, the following parameters were
used: photomultiplier tube (PMT) voltage = 420 V, λex = 329 nm, λem
= 340−800 nm. A 10 μM solution of receptor R2 was prepared by
adding 62.4 μL of a 0.4 mM stock solution in CHCl3 to a cuvette
containing 2438 μL of CHCl3. After recording the emission spectrum,
beauvericin was added (from a 1 mM stock solution) in amounts
corresponding to 0.1, 0.2, 0.3, 0.5, 0.7, 1, 1.5, 2, 3, 4, and 5 equiv (in
total), and the fluorescence intensity was measured in between each
addition.
The titrations of the receptor analogues with the different
ionophores were done similar to what has been described for R2
and BEA. All receptors were excited at λex = 329 nm in CHCl3, and
titrations were performed in triplicate. Depending on the receptor,
different concentrations and PMT voltages were used. Fitting of the
titration data was done similar to what has been described for R2,
except the used maximum λem was dependent on the receptor (Table
S1).
(0.4%). A fraction of this mixture was added to a Burker counting
̈
chamber, and the number of cells were counted manually using a light
microscope. Once the cell concentration was known, cell solutions
were prepared in such a way that each well contained 0.2 mL of
medium and 40 000 cells. The outer wells were filled with phosphate-
buffered saline (PBS) only and did not contain any cells. The
resulting plates are then incubated overnight. The next day, the
treatment was performed (vide infra).
General Procedure for Performing the MTT Assay. After the
treatment, 100 μL of medium was removed from each well and 20 μL
of the MTT solution was added (5 mg/mL in PBS, 0.22 μm filter
sterilized). The plate was incubated for 2 h at 37 °C. The liquid was
completely removed from each well, and 200 μL DMSO was added.
The solution was pipetted up and down to homogenize, and the UV−
vis absorbance was measured at 570 nm using a plate reader.
General Procedure for Performing the SRB Assay. After the
treatment, 50 μL of a trichloroacetic acid solution (50% in Milli-Q
water) was added, followed by storage of the plate at 4 °C for 1 h. The
plate was carefully rinsed with tap water multiple times. Once dry, 200
μL SRB staining solution was added (0.4% in 1% acetic acid). After 30
min, the plate was rinsed with 1% acetic acid to remove the unbound
dye (three times), followed by drying. Then, 200 μL of a 10 mM Tris
buffer solution was added and each well was properly homogenized.
The UV−vis absorbance was measured at 490 nm using a plate
reader.
General Procedure for Performing the NR Assay. After the
treatment, each well was washed once with growth medium (without
antibiotics or serum). The NR staining solution was made by adding
0.3 mL of a NR stock solution to 30 mL of growth medium (without
antibiotics or serum). The stock solution was prepared by dissolving
40 mg of neutral red dye in 10 mL of PBS. The solution was first
placed in a water bath at 37 °C for 30 min to assure solubility of the
dye, followed by filtration through a 0.22 μm filter. Then, the liquid in
each well was replaced by 100 μL of this staining solution and the
plates were incubated at 37 °C for 3 h. The cells were washed once
with PBS and dried. Then, 100 μL of desorption solution was added
(1% HOAc, 50% EtOH, and 49% H2O). After 20 min, the UV−vis
absorbance is measured at 540 nm using a plate reader.
Treatment of the Cells. The growth medium was removed from
the plate (which was prepared as described in Preparation of the Well
Plates). The wells were subsequently filled with 200 μL of the
appropriate treatment condition (see preparation below).
Treatment with BEA Alone. Taking the 5 μM BEA treatment
condition as an example, a diluted solution A was made by taking 20
μL of the BEA stock solution (6.378 × 10−3 M, in DMF), followed by
addition of 180 μL of DMF. Then, 47 μL of A was taken and added to
a falcon tube together with 13 μL of DMF. Then, this was added to
5940 μL of growth medium. For the 2.5 μM BEA condition, 24 μL of
A was taken and added to a falcon tube containing 36 μL of DMF.
Then, this was added to 5940 μL of growth medium. For the 1%
DMF condition, 60 μL of DMF was added to a falcon tube containing
5940 μL of growth medium. The other treatment conditions were
prepared in an analogous way. Note that the total DMF volume was
always kept constant, namely, 1%. The plate was incubated for 24 h at
37 °C and stained according to the general MTT staining procedure
described above. The same procedure was followed for the plates
stained with SRB or NR.
Treatment with R2 Alone. Taking the 100 μM solution as an
example, 66 μL from a 1.82 × 10−5 M stock solution of R2 (in DMF)
was added to 174 μL of DMF (=solution B). From this solution, 60
μL was taken and added to 5940 μL of growth medium. For the 50
Cell Experiments: Cell Growth. HepG2 cells were grown in T25
or T75 flasks containing 4 or 12 mL of growth medium, respectively.
The growth medium contained 10% fetal bovine serum, 1%
L
DOI: 10.1021/acs.joc.9b01665
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
μM solution, 120 μL of B was taken and added to 120 μL of DMF.
From this solution, 60 μL was taken and added to 5940 μL of growth
medium. The other treatment solutions were prepared in a similar
fashion, while making sure that the total DMF percentage was always
kept constant (1%). The plate was incubated for 24 h at 37 °C and
stained according to the general MTT staining procedure described
above. The same procedure was followed for the plates stained with
SRB or NR.
Treatment with BEA + R2. Taking the 5 μM BEA + 10 equiv R2
treatment condition as an example, 20 μL of a 6.38 × 10−3 M stock
solution of BEA (in DMF) was added to 80 μL of DMF (=solution
C). From a stock solution of R2 (1.82 × 10−5 M in DMF), 50 μL is
taken and added to 50 μL of DMF (=solution D). Then, 16.5 μL is
taken from this solution and added to a mixture containing 23.5 μL of
solution C and 20 μL of DMF. This solution was subsequently added
to 5940 μL of growth medium. For the 5 μM BEA treatment solution,
23.5 μL of solution C was added to 36.5 μL of DMF and the resulting
mixture was added to 5940 μL of growth medium. While for the 10
equiv R2 treatment solution, 16.5 μL of solution D was added to 43.5
μL DMF and the resulting mixture was added to 5940 μL of growth
medium. The other treatment solutions were prepared in a similar
fashion while making sure that the total DMF concentration always
remained constant (1%). The plate was incubated for 24 h at 37 °C
and stained according to the general MTT staining procedure
described above. The same procedure was followed for the plates
stained with SRB or NR.
(2) Hardstaff, J. L.; Clough, H. E.; Lutje, V.; McIntyre, K. M.; Harris,
J. P.; Garner, P.; O’Brien, S. J. Foodborne and Food- Handler
Norovirus Outbreaks: A Systematic Review. Foodborne Pathog. Dis.
2018, 15, 589−597.
(3) Gupta, S.; Krasnoff, S. B.; Underwood, N. L.; Renwick, J. A. A.;
Roberts, D. W. Isolation of Beauvericin as an Insect Toxin from
Fusarium semitectum and Fusarium moniliforme var. Subglutinans.
Mycopathologia 1991, 115, 185−189.
(4) Shemyakin, M. M.; Ovchinnikov, Y. A.; Kiryushkin, A. A.;
Ivanov, V. T. The Structure and Total Synthesis of Enniatin B.
Tetrahedron Lett. 1963, 4, 885−890.
(5) Uhlig, S.; Torp, M.; Heier, B. T. Beauvericin and Enniatins A,
A1, B and B1 in Norwegian Grain: A Survey. Food Chem. 2006, 94,
193−201.
(6) Ritieni, A.; Moretti, A.; Logrieco, A.; Bottalico, A.; Randazzo, G.;
Monti, S. M.; Ferracane, R.; Fogliano, V. Occurrence of
Fusaproliferin, Fumonisin B1, and Beauvericin in Maize from Italy.
J. Agric. Food Chem. 1997, 45, 4011−4016.
(7) Juan, C.; Manes, J.; Raiola, A.; Ritieni, A. Evaluation of
̃
Beauvericin and Enniatins in Italian Cereal Products and Multicereal
Food by Liquid Chromatography Coupled to Triple Quadrupole
Mass Spectrometry. Food Chem. 2013, 140, 755−762.
(8) Kouri, K.; Lemmens, M.; Lemmens-Gruber, R. Beauvericin-
Induced Channels in Ventricular Myocytes and Liposomes. Biochim.
Biophys. Acta, Biomembr. 2003, 1609, 203−210.
(9) Tonshin, A. A.; Teplova, V. V.; Andersson, M. A.; Salkinoja-
Salonen, M. S. The Fusarium Mycotoxins Enniatins and Beauvericin
Cause Mitochondrial Dysfunction by Affecting the Mitochondrial
Volume Regulation, Oxidative Phosphorylation and Ion Homeostasis.
Toxicology 2010, 276, 49−57.
Treatment with Boc-Protected Receptor 6. The treatment
solutions were prepared in a similar fashion to BEA + R2 using a 1.82
× 10−5 M stock solution of 6 (in DMF). The plate was incubated for
24 h at 37 °C and stained according to the general MTT staining
procedure described above.
(10) Jow, G.-M.; Chou, C.-J.; Chen, B.-F.; Tsai, J.-H. Beauvericin
Induces Cytotoxic Effects in Human Acute Lymphoblastic Leukemia
Cells Through Cytochrome c Release, Caspase 3 Activation: The
Causative Role of Calcium. Cancer Lett. 2004, 216, 165−173.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.joc.9b01665.
■
S
(11) Rodríguez-Carrasco, Y.; Heilos, D.; Richter, L.; Sussmuth, R.
̈
D.; Heffeter, P.; Sulyok, M.; Kenner, L.; Berger, W.; Dornetshuber-
Fleiss, R. Mouse Tissue Distribution and Persistence of the Food-
born Fusariotoxins Enniatin B and Beauvericin. Toxicol. Lett. 2016,
247, 35−44.
Spectroscopic characterization, used equations for the
data fitting, additional fluorescence titration spectra, and
results from cell assays (PDF)
(12) Taevernier, L.; Bracke, N.; Veryser, L.; Wynendaele, E.;
Gevaert, B.; Peremans, K.; De Spiegeleer, B. Blood-Brain Barrier
Transport Kinetics of the Cyclic Depsipeptide Mycotoxins Beauver-
icin and Enniatins. Toxicol. Lett. 2016, 258, 175−184.
(13) Langton, M. J.; Serpell, C. J.; Beer, P. D. Anion Recognition in
Water: Recent Advances from a Supramolecular and Macromolecular
Perspective. Angew. Chem., Int. Ed. 2016, 55, 1974−1987.
(14) Gokel, G. W.; Leevy, W. M.; Weber, M. E. Crown Ethers:
Sensors for Ions and Molecular Scaffolds for Materials and Biological
Models. Chem. Rev. 2004, 104, 2723−2750.
AUTHOR INFORMATION
Corresponding Author
*E-mail: Annemieke.Madder@Ugent.be.
ORCID
Sarah De Saeger: 0000-0002-2160-7253
Annemieke Madder: 0000-0003-0179-7608
Notes
The authors declare the following competing financial
interest(s): The technology described in the manuscript is
part of a pending patent application by V.O., A.R., B.S., S.D.S.
and A.M.
■
(15) Mazik, M. Molecular Recognition of Carbohydrates by Acyclic
Receptors Employing Noncovalent Interactions. Chem. Soc. Rev. 2009,
38, 935−956.
́
(16) Canevet, D.; Perez, E. M.; Martín, N. Wraparound Hosts for
Fullerenes: Tailored Macrocycles and Cages. Angew. Chem., Int. Ed.
2011, 50, 9248−9259.
ACKNOWLEDGMENTS
■
(17) Shinoda, S.; Tsukube, H. Molecular Recognition of
Cytochrome c by Designed Receptors for Generation of In Vivo
and In Vitro Functions. Chem. Sci. 2011, 2, 2301−2305.
(18) Chinai, J. M.; Taylor, A. B.; Ryno, L. M.; Hargreaves, N. D.;
Morris, C. A.; Hart, P. J.; Urbach, A. R. Molecular Recognition of
Insulin by a Synthetic Receptor. J. Am. Chem. Soc. 2011, 133, 8810−
8813.
The authors thank Charlotte Grootaert for her guidance in the
practical aspects of the cell experiments. This research was
funded by the BOF Special Research Fund from Ghent
University, GOA project no. 01G02213. V.O. is grateful to the
Fonds voor Wetenschappelijk Onderzoek for a VLAIO
research grant.
(19) Schmuck, C.; Rupprecht, D.; Wienand, W. Sequence-Depend-
ent Binding of Dipeptides by an Artificial Receptor in Water. Chem. −
Eur. J. 2006, 12, 9186−9195.
(20) Bernard, J.; Wennemers, H. Macrocyclic Diketopiperazine
Receptors: Effect of Macrocyclization on the Binding Properties of
Two-Armed Receptors. Org. Lett. 2007, 9, 4283−4286.
REFERENCES
■
(1) Valderrama, W. B.; Dudley, E. G.; Doores, S.; Cutter, C. N.
Commercially Available Rapid Methods for Detection of Selected
Food-borne Pathogens. Crit. Rev. Food Sci. Nutr. 2016, 56, 1519−
1531.
M
DOI: 10.1021/acs.joc.9b01665
J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
̈
(21) Gersthagen, T.; Hofmann, J.; Klarner, F.-G.; Schmuck, C.;
Schrader, T. Ditopic Arginine-Aspartate Binders Recognize RGD
Loops. Eur. J. Org. Chem. 2013, 2013, 1080−1092.
(39) Sieuwerts, A. M.; Klijn, J. G. M.; Peters, H. A.; Foekens, J. A.
The MTT Tetrazolium Salt Assay Scrutinized: How to Use this Assay
Reliably to Measure Metabolic Activity of Cell Cultures In Vitro for
the Assessment of Growth Characteristics, IC50-Values and Cell
Survival. Eur. J. Clin. Chem. Clin. Biochem. 1995, 33, 813−823.
(40) Repetto, G.; del Peso, A.; Zurita, J. L. Neutral Red Uptake
Assay for the Estimation of Cell Viability/Cytotoxicity. Nat. Protoc.
2008, 3, 1125−1131.
(41) Vichai, V.; Kirtikara, K. Sulforhodamine B Colorimetric Assay
for Cytotoxicity Screening. Nat. Protoc. 2006, 1, 1112−1116.
(42) Juan-García, A.; Ruiz, M.-J.; Font, G.; Manyes, L. Enniatin A1,
Enniatin B1 and Beauvericin on HepG2: Evaluation of Toxic Effects.
Food Chem. Toxicol. 2015, 84, 188−196.
(43) EFSA Panel on Contaminants in the Food Chain (CONTAM).
Scientific Opinion on the Risks to Human and Animal Health Related
to the Presence of Beauvericin and Enniatins in Food and Feed. EFSA
J. 2014, 12, 3802.
(44) Umali, A. P.; Anslyn, E. V. A General Approach to Differential
Sensing Using Synthetic Molecular Receptors. Curr. Opin. Chem. Biol.
2010, 14, 685−692.
(22) Ornelis, V.; Rajkovic, A.; Sas, B.; De Saeger, S.; Madder, A.
Development of a Synthetic Receptor for the Food Toxin
Beauvericin: A Tale of Carbazole and Steroids. Org. Lett. 2018, 20,
6368−6371.
(23) Sonogashira, K.; Tohda, Y.; Hagihara, N. A Convenient
Synthesis of Acetylenes: Catalytic Substitutions of Acetylenic
Hydrogen with Bromoalkenes, Iodoarenes and Bromopyridines.
Tetrahedron Lett. 1975, 16, 4467−4470.
(24) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. A
Stepwise Huisgen Cycloaddition Process: Copper(I)-Catalyzed
Regioselective ‘Ligation’’ of Azides and Terminal Alkynes. Angew.
Chem., Int. Ed. 2002, 41, 2596−2599.
(25) Han, G.; Tamaki, M.; Hruby, V. J. Fast, Efficient and Selective
Deprotection of the Tert-Butoxycarbonyl (Boc) Group Using HCl/
dioxane (4M). J. Pept. Res. 2001, 58, 338−341.
(26) van de Weert, M. Fluorescence Quenching to Study Protein-
ligand Binding: Common Errors. J. Fluoresc. 2010, 20, 625−629.
(27) Ulatowski, F.; Dabrowa, K.; Balakier, T.; Jurczak, J. Recognizing
the Limited Applicability of Job Plots in Studying Host-Guest
Interactions in Supramolecular Chemistry. J. Org. Chem. 2016, 81,
1746−1756.
(28) Hibbert, D. B.; Thordarson, P. The Death of the Job Plot,
Transparency, Open Science and Online Tools, Uncertainty
Estimation Methods and Other Developments in Supramolecular
Chemistry Data Analysis. Chem. Commun. 2016, 52, 12792−12805.
(29) Aher, N. G.; Pore, V. S.; Mishra, N. N.; Shukla, P. K.; Gonnade,
R. G. Design and Synthesis of Bile Acid-Based Amino Sterols as
Antimicrobial Agents. Bioorg. Med. Chem. Lett. 2009, 19, 5411−5414.
́
(30) Gonzalez, M. A. Synthetic Derivatives of Aromatic Abietane
Diterpenoids and their Biological Activities. Eur. J. Med. Chem. 2014,
87, 834−842.
(31) Huang, W.-G.; Wang, H.-S.; Huang, G.-B.; Wu, Y.-M.; Pan, Y.-
M. Enantioselective Friedel-Crafts Alkylation of N-Methylindoles with
Nitroalkenes Catalyzed by Chiral Bifunctional Abietic-Acid-Derived
Thiourea-ZnII Complexes. Eur. J. Org. Chem. 2012, 2012, 5839−5843.
̈
̈
̈
(32) Laaksonen, T.; Heikkinen, S.; Wahala, K. Synthesis and
Applications of Secondary Amine Derivatives of (+)-Dehydroabietyl-
amine in Chiral Molecular Recognition. Org. Biomol. Chem. 2015, 13,
10548−10555.
(33) Fei, B.-L.; Yin, B.; Li, D.-D.; Xu, W.-S.; Lu, Y. Enantiopure
Copper(II) Complex of Natural Product Rosin Derivative: DNA
Binding, DNA Cleavage and Cytotoxicity. J. Biol. Inorg. Chem. 2016,
21, 987−996.
(34) Serrano, A. B.; Capriotti, A. L.; Cavaliere, C.; Piovesana, S.;
̀
Samperi, R.; Ventura, S.; Lagana, A. Development of a Rapid LC-MS/
MS Method for the Determination of Emerging Fusarium Mycotoxins
Enniatins and Beauvericin in Human Biological Fluids. Toxins 2015,
7, 3554−3571.
(35) Decleer, M.; Rajkovic, A.; Sas, B.; Madder, A.; De Saeger, S.
Development and Validation of Ultra-High-Performance Liquid
Chromatography-Tandem Mass Spectrometry Methods for the
Simultaneous Determination of Beauvericin, Enniatins (A, A1, B,
B1) and Cereulide in Maize, Wheat, Pasta and Rice. J. Chromatogr. A
2016, 1472, 35−43.
(36) Mallebrera, B.; Brandolini, V.; Font, G.; Ruiz, M.
Cytoprotective Effect of Resveratrol Diastereomers in CHO-K1
Cells Exposed to Beauvericin. Food Chem. Toxicol. 2015, 80, 319−
327.
́
(37) Pocsfalvi, G.; Di Landa, G.; Ferranti, P.; Ritieni, A.; Randazzo,
G.; Malorni, A. Observation of Non-Covalent Interactions Between
Beauvericin and Oligonucleotides Using Electrospray Ionization Mass
Spectrometry. Rapid Commun. Mass Spectrom. 1997, 11, 265−272.
(38) Tomoda, H.; Huang, X.-H.; Cao, J.; Nishida, H.; Nagao, R.;
Okuda, S.; Tanaka, H.; Omura, S.; Arai, H.; Inoue, K. Inhibition of
Acyl-CoA: Cholesterol Acyltransferase Activity by Cyclodepsipeptide
Antibiotics. J. Antibiot. 1992, 45, 1626−1632.
N
DOI: 10.1021/acs.joc.9b01665
J. Org. Chem. XXXX, XXX, XXX−XXX