Acyl Derivatives of Eudesmanolides To Boost their Bioactivity: An Explanation of Behavior in the Cell Membrane Using a Molecular Dynamics Approach

Article abstract of DOI:10.1002/cmdc.202000783

Semisynthetic analogs of natural products provide an important approach to obtain safer and more active drugs and they can also have enhanced physicochemical properties such as persistence, cross-membrane processes and bioactivity. Acyl derivatives of different natural product families, from sesquiterpene lactones to benzoxazinoids, have been synthesized and tested in our laboratories. These compounds were evaluated against tumoral and nontumoral cell lines to identify selective derivatives with a reduced negative impact upon application. The mode of action of these compounds was analyzed by anti-caspase-3 assays and molecular dynamics simulations with cell membrane re-creation were also carried out. Aryl derivatives of eudesmanolide stand out from the other compounds and are better than current anticancer drugs such as etoposide in terms of selectivity and activity. Computational studies provide evidence that lipophilicity plays a key role and the 4-fluorobenzoyl derivative can pass easily through the cell membrane.

Full text of DOI:10.1002/cmdc.202000783

Accepted Article
Title: Acyl Derivatives of Eudesmanolides to Boost their Bioactivity. An
Explanation of Behavior in the Cell Membrane Using a Molecular
Dynamics Approach.
Authors: Francisco J.R. Mejías, Alexandra G. Durán, Jesús G. Zorrilla,
Rosa M. Varela, José M.G. Molinillo, Manuel M. Valdivia, and
Francisco A. Macias
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemMedChem 10.1002/cmdc.202000783
Link to VoR: https://doi.org/10.1002/cmdc.202000783
01/2020

10.1002/cmdc.202000783
ChemMedChem
FULL PAPER
Acyl Derivatives of Eudesmanolides to Boost their Bioactivity. An Explanation of Behavior in
the Cell Membrane Using a Molecular Dynamics Approach.
Francisco J.R. Mejías[a], Dr. Alexandra G. Durán[a], Jesús G. Zorrilla[a], Prof. Rosa M. Varela[a], Prof.
José M.G. Molinillo[a], Prof. Manuel M. Valdivia[b], Prof. Francisco A. Macías*[a]
[a]
F.J.R. Mejías, Dr. A.G. Durán, J.G. Zorrilla, Prof. R.M. Varela, Prof. J.M.G. Molinillo and Prof. F.A. Macías
Allelopathy Group, Department of Organic Chemistry, Institute of Biomolecules (INBIO), Campus de Excelencia Internacional (ceiA3), School of Science,
University of Cádiz, C/República Saharaui, 7, 11510, Puerto Real, Cádiz, Spain.
E-mail: famacias@uca.es
[b]
Prof. M.M. Valdivia
Department of Biomedicine, Biotechnology and Public Health, Institute of Biomolecules (INBIO), School of Science, University of Cádiz, C/República
Saharaui, 7, 11510, Puerto Real, Cádiz, Spain.
Supporting information for this article is given via a link at the end of the document.
Abstract: Semi-synthetic analogs of natural products provide an
important approach to obtain safer and more active drugs and they
also have enhanced physicochemical properties such as persistence,
cross membrane processes and bioactivity. Acyl derivatives of
different natural product families, from sesquiterpene lactones to
simple derivatizations – such as hydroxylation – were key to
enhancing the activity.[7] In spite of the abundant publications on
these natural products, simple ester derivatives obtained from
hydroxyl-substituted precursors have not been developed to
evaluate their activity by modulation of the physicochemical
benzoxazinoids, have been synthesized and tested in our laboratories. properties. Furthermore, aromatic derivatives mono-halogenated,
These compounds were evaluated against tumoral and non-tumoral
cell lines to identify selective derivatives with reduced negative impact
upon application. The mode of action of these compounds was
analyzed by anti-caspase-3 assays and molecular dynamics
simulations with cell membrane recreation were also carried out. Aryl
derivatives of eudesmanolide stand out from the other compounds
and are better than current anti-cancer drugs, such as etoposide, in
terms of selectivity and activity. Computational studies provide
evidence that lipophilicity plays a key role and the 4-fluorobenzoyl
derivative can pass easily through the cell membrane.
and specifically mono-fluorinated seems to be essential in new
strategies to stop cancer progression. This is the case of
quinazoline 4-fluorophenyl derivatives, which inhibits Hsp90, and
also the case of the 4ʹ-fluoro-2ʹ-hydroxy-chalcone derivatives
which show anti-inflammatory effects.[8–10] Thus, synthesis of
reynosin (2) derivatives focusing in halogenated compound could
be a relevant starting point.
Natural products with other types of skeletons are also relevant in
medicinal chemistry. For example, sesquiterpene lactones with a
guaiane-type skeleton, e.g., cynaropicrin (12) and its derivatives,
have shown anti-proliferation of leukocyte cancer cell lines.[11] In
the case of benzoxazinoids, like APO (14) and its mimics (16 and
18), application against prostate cancer not only showed
favorable results in cell lines,[12] but even in human assays where
a benzoxazinoid-rich diet was applied.[13] Diterpenes such as
gummiferolic acid (20) also have broad applications in medicine,
specifically to promote proliferation of neural precursor cells when
tested in neurosphere assays.[14]
Introduction
Natural product research has produced the majority of medicinal
drugs that are successfully tested in phase II clinical trials.
Nevertheless, just 20% of these natural product leads are taken
through unchanged to the final drug product.[1] Thus, small
structural modifications to generate derivative banks could
produce new and effective drugs with improved bioactivity profiles,
as well as better absorption, distribution, metabolism and
excretion properties.
In the work described here, costunolide was isolated from the
roots of Saussurea costus and reynosin was synthesized to obtain
acyl derivatives by reaction with the hydroxyl functional group. A
structure-activity relationship study was performed to evaluate the
effect of bioavailability of the reynosin derivatives, and the IC50
values were obtained for cytotoxicity on HeLa cells. After selection
of the best derivative, caspase-3 assays were carried out to
analyze the mode of action, and an HEK cell bioassay was carried
out to ascertain the selectivity of the compounds against tumoral
targets. Simulation by molecular dynamics in the cell membrane
allowed the stability and ease of membrane crossing processes
to be analyzed. In order to understand the relevance of the best
acyl fragment, different esters of other natural products and
mimics with different skeletons or families were synthesized to
evaluate their bioactivity.
In this respect, hydroxylated costunolide (1) derivatives can be
highlighted as a source of potent drugs with several types of
bioactivity. Reynosin (2) and santamarine (3) (Figure 1) have
shown strong bactericidal activity against Mycobacterium
tuberculosis, with a minimal inhibitory concentration of 64
g/mL.[2] In addition, reynosin also suppresses ROS production
and reduces COX-2 levels via the NF-KB pathway.[3,4] In vivo
studies have recently been carried out on this eudesmanolide,
with hepatoprotective effects observed in mouse liver upon
exposure to thioacetamide in order to generate apoptosis.[5]
Furthermore, reynosin (2) seems to be relevant in neural
protection against Parkinson’s diseases.[6] Semi-synthetic
analogs of this relevant natural product seem to be important to
enhance the inhibition of the tumor necrosis factor-alpha, and
1
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000783
ChemMedChem
FULL PAPER
Figure 1. Isolated and synthesized compounds employed in this study.
2
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000783
ChemMedChem
FULL PAPER
H-3' signals caused a loss of resolution in the signals for H-2 and
H-5 due to their large integral, which broadened the baseline and
made it difficult to measure some of the coupling constants.
Nevertheless, this behavior was more pronounced for compound
11, for which the H-3' to H-7' signals fully overlapped those of H-
2 and H-5.
Results and Discussion
Synthesis of reynosin derivatives
The starting material 1 (Figure 2) was isolated in high yield from
roots of Saussurea costus, as described in the Experimental
Section. This compound was transformed into the target
eudesmanolides by a cyclization reaction in the presence of the
epoxidation agent m-CPBA followed by the addition of a catalytic
amount of p-TsOH in order to open the epoxy ring. Compounds 2
and 3 were generated in 31% and 51% yield, respectively.
The synthesis of acyl derivatives (Figure 2) was carried out at
room temperature in pyridine and this approach gave high yields
of compounds 4–11. A molar excess of acyl chloride is usually
employed in this kind of reaction to avoid yield loss due to side
reactions to give carboxylic acids. Extraction with 1.0 M NaOH
was employed to remove acid derivatives formed by hydrolysis of
the acyl chlorides. The synthesized aryl derivatives enabled an
analysis of the influence of the inclusion of halogen(s), the
position of the halogen atom and the number of halogens in the
reynosin derivatives. The study carried out on the alkyl derivatives
(10–11) showed the influence that the lipophilicity of reynosin by
comparison with the aryl compounds.
It is known that fluorine atoms show NMR activity due to the more
abundant isotope 19F, which has a nuclear spin of ½. This impacts
not only the 1H NMR spectra but 13C NMR spectra and it causes
significant multiplicities. When one fluoro-substituent is present
on the aromatic ring, the ortho-carbon signal appears as a doublet
(J = 21–23 Hz) The meta-carbon signal is also doublet, but its
coupling constant is limited to the range 7.5–10.0 Hz. The carbon
in the para-position has the smallest coupling constant, with a
value in the range 2.8–4.0 Hz. In contrast, the carbon that is
bonded to the fluoro-substituent has the largest constant (250.0–
255.0 Hz). In the case of multiple fluorine substitution, e.g., the
2,4,6-trifluoro derivative (9), the coupling constant exceeds the
ranges mentioned above due to high electronegativity of each
fluorine nucleus. The carbon directly linked to the fluorine atom
has a doublet multiplicity with J >258 Hz and the ortho, meta and
para coupling constants showed 5 Hz in comparison with the
mono-fluorinated derivative.
In the aliphatic derivatives, i.e., butyryl (10) and octanoyl (11), the
13C spectra were simplified due to the absence of 19F.
Nevertheless, the large numbers of protons with similar chemical
1
environments meant that the H spectra between 2.00 and 0.50
ppm contained a large number of signals with high integral values.
The H-2'a and H-2'b signals were shifted the most due to their
proximity to the carbonyl group (-carbon). In these two cases
signals were observed at 2.29 and 2.30 ppm, respectively, with
coupling constants in the range 7.2–7.3 Hz. However, the
coupling constants for the adjacent protons were in the range 1.7–
1.2 ppm. These higher field signals were overlapped for the
octanoyl derivative (11).
Bioactivity studies
The pharmacological potential of reynosin has been described in
previous publications, including anticancer properties.[15,16] One of
the main purposes of the present work was to synthesize reynosin
analogs in an effort to improve the bioavailability and biological
properties. The ultimate aim was to design new and potent drugs.
In the first step, seven acyl derivatives were obtained by modifying
the hydroxyl group on the reynosin backbone (5–11). The
cytotoxicity vales of these compounds were assessed on cervical
carcinoma cells (HeLa) at 100 µM for 24 hours (Figure 3).
Figure 2. Synthesis of hydroxylated eudesmanolides reynosin and
santamarine. Synthesis of ester derivatives of reynosin with aryl fragments
and aliphatic fragments.
In the 1H NMR spectra the H-1 signal for the eudesmane skeleton
was shifted the most after acylation. The aromatic ring at the three
position still had a marked influence and this increased the proton
shift from 4.80 (butyryl) or 4.79 ppm (octanoyl) to 5.04–5.07 ppm
(aromatic derivatives). In the case of compound 10, the H-2' and
3
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000783
ChemMedChem
FULL PAPER
Figure 3. Cell viability by dye exclusion against cervical carcinoma cells
(HeLa). 1-O-Acyl reynosin derivatives were evaluated at 100 µM after 24
h of treatment. Data are expressed as mean ± SD (n = 3).
It has been demonstrated in several studies that lipophilicity is an
important factor for biological effects.[17,18] Thus, it hs been noted
that the presence, or an increasing number, of hydroxyl groups in
the sesquiterpene lactone backbone results in lower
antimycobacterial, fungicidal and cytotoxic activities.[19,20] On the
other hand, the α-methylene-γ-lactone moiety is considered to be
an essential structural requirement for significant cytotoxic
activity.[21] As a consequence, all of the derivatives synthesized in
this work contain this framework. The results show that
modification of the hydroxyl group on the reynosin skeleton leads
to a significant improvement in the activity (Figure 3). Derivatives
with an aromatic group (4–9) were more active than those with a
linear chain (10–11). In terms of activity it is worth highlighting
compounds 5 and 9.
Figure 4. Cell viability by dye exclusion against cervical carcinoma cells
(HeLa). 4-Fluorobenzoyl derivatives were evaluated at 100 µM for 24 h.
Experiments were performed in triplicate and data are expressed as mean
± SD.
Results observed to every aromatic derivative of reynosin (2) are
statistically similar, highlighting mono-fluorinated and tri-
fluorinated derivatives. According to literature, mono-halogenated
derivatives are broadly applied in biological studies and present
better results due to lower persistence and better absorption and
distribution in vivo.[8–10,23]In view of the results outlined above,
three promising derivatives (4, 5 and 17) were selected, on the
basis of their high activity levels, to perform a cell viability study at
various concentrations (from 0.1 to 100 µM) in tumoral (HeLa) and
non-tumoral (HEK-293, human embryonic kidney) cell lines for 24
h to study their selectivity. The data were compared to those of
the clinically used anticancer drug etoposide (Figure 5).
In view of the findings outlined above, the 4-fluorobenzoyl
framework was selected to modify other mimics and natural
compounds. Two additional sesquiterpene lactones (santamarine
(3) and cynaropicrin (12)), 2-amino-3H-phenoxazin-3-one (APO
(14)), two benzoxazinoid mimics (16 and 18) and a diterpene
(guiferolic acid (20)) were selected for the next study. Moreover,
in cases where hydrolysis of the ester bond occurred, the activity
observed may be due to 4-fluorobenzoic acid and this compound
was also tested.
Table 1. Lipophilicity values for each compound. CLog P
values calculated with ChemDraw Professional® v. 17.1.
The results of the cytotoxicity study are shown in Figure 4. The
most potent cytotoxic effects were observed when the 4-
fluorobenzoyl unit was introduced onto the santamarine skeleton
(4). In this case a cell viability value below 5% obtained, thus
highlighting the relevance of the eudesmane skeleton.
Nevertheless, modification of cynaropicrin (12) did not lead to an
improvement in the anti-proliferative activity against the tumor
cells. This finding could be due to the increased lipophilicity (Table
1), with a CLog P value of 5.83, in respect of the established
‘Lipinski’s rule of five’ for pharmaceutical candidates.[22] In the
case of benzoxazinoids and their mimics, the introduction of this
framework into the APO (14) core also caused a loss of cytotoxic
activity (derivative 15). This result indicates that the molecular site
of action had been modified and the amino group is essential to
retain the biological activity. On the other hand, modification of
benzoxazinoid mimic 16 gave the new derivative 17, which had a
four-fold higher anti-proliferative activity. However, this variation
was not observed for benzoxazinoid mimic 18, as the level of
activity of its derivative (19) increased only slightly. Similar
behavior was also observed on modifying the gummiferolic acid
(20) core to afford derivative 21. In both of these cases the
pharmacophore was not affected. Overall, the modification of
amino groups to form amide derivatives led to a decrease in the
activity, while ester formation on the hydroxyl groups resulted in
an increase in cytotoxic activity. It is also worth highlighting that
4-fluorobenzoic acid did not display any activity. Thus, the activity
observed was due solely to the chemical compound tested.
Compound
Reynosin (2)
Santamarine (3)
CLog P
1.183
1.183
4.201
4.201
4.058
4.771
4.201
3.607
3.187
5.303
0.046
5.827
1.136
2.970
3.019
7.223
4-Fluorobenzoylsantamarine (4)
4-Fluorobenzoylreynosin (5)
Benzoylreynosin (6)
4-Chlorobenzoylreynosin (7)
3-Fluorobenzoylreynosin (8)
2,4,6-Trifluorobenzoylreynosin (9)
Butyrylreynosin (10)
Octanoylreynosin (11)
Cynaropicrin (12)
Bis(4-fluorobenzoyl)cynaropicrin (13)
APO (14)
4-Fluorobenzoyl-N-APO (15)
DiS-OH (16)
Disulfanediylbis(2,1-phenylene) bis(4-
fluorobenzoate) (17)
DiS-NH2 (18)
N,N'-(Disulfanediylbis(2,1-phenylene))bis(4-
fluorobenzamide) (19)
Gummiferolic acid (20)
4-Fluorobenzoic gummiferolic anhydride (21)
2.736
5.062
7.180
9.304
The results show that the sesquiterpene lactone derivatives
tested (4 and 5) were more active than the benzoxazinoid mimic
and etoposide against HeLa cells. IC50 values of 2.85 (5, R2 =
0.9682), 14.53 (4, R2 = 0.9783), 69.04 (17, R2 = 0.9356) and 59.26
(etoposide, R2 = 0.9765) were obtained, respectively, for HeLa
cells. Furthermore, it is worth noting that compound 5 showed
certain specificity between the two cell lines tested. Cell viability
values of less than 50% were noted at 1 and 5 µM against tumoral
cells, while values of 89% were obtained for the same
concentrations against non-tumoral cells. In order to corroborate
these findings, the selectivity index (SI) was calculated. This index
4
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000783
ChemMedChem
FULL PAPER
(expressed as IC50 ratio in HEK-293 versus HeLa cells) allows a
determination of the selectivity of a compound on the cell lines
tested, with a high selectivity corresponding to an SI value above
3.[24] A high selectivity value (SI = 5.63) was obtained for
compound 5 whereas lower selectivities were determined for
derivatives 4 and 17 (SI ≤ 3). This new derivative could be a
promising lead compound since it is synthesized in only two steps
from costunolide (multigram-scale production), which in turn is
of a series of caspases that mediate the cell destruction.
Caspase-3 is considered to be the most important of the
executioner caspases and it is activated by any of the initiator
caspases. Activation results in DNA fragmentation and PARP
(Poly (ADP-Ribose) Polymerase) cleavage, which in turn triggers
a caspase-dependent apoptotic signaling cascade.[26,27] In this
study, immunofluorescence analysis was used to detect caspase-
3. The results reveal that the expression intensity was higher after
readily available (100 mg of costunolide/1 g of extract). In contrast, treatment with compound 5 in comparison to the negative control
etoposide
podophyllotoxin extracted from Sinopodophyllum hexandrum
T.S.Ying (Himalayan mayapple) on milligram scale.
is
produced
semi-synthetically
from
(−)-
(0.1% DMSO) (Figure 6). DNA fragmentation could also be
observed (indicated with arrows).
a
Furthermore, new etoposide production strategies are being
developed due to etoposide shortages in the past decade and as
the species from which it is obtained is considered to be
endangered.[25]
Molecular dynamics simulation
In order to delve further into the lipophilicity and bioactivity results,
a molecular dynamics simulation was carried out on the cellular
membrane.
A
total
of
128
molecules
of
dipalmitoylphosphatidylcholine (DPPC) were employed to
generate the bilayer. The cubic box simulation space was then
solvated and Na+ and Cl– ionic species were added to simulate
the ionic strength of a real cell.
Three markedly different lipophilic derivatives were used in this
study to elucidate more varied behavior: reynosin (2),
octanoylreynosin (11) and 4-fluorobenzoylreynosin (5).
Specifically, 4-fluororeynosin was the most active compound, so
this study allowed us to pinpoint the behavior that could explain
the bioactivity results. These compounds were placed in the
center-of-mass of the box, in the middle of bilayer, parallel to
DPPC molecules, and the performance was analyzed during 10
ns to ascertain whether the compounds could move easily
through the phospholipid, or if they are stabilized in one phase of
the micellar simulation.
Figure 5. Cell viability by dye exclusion of derivatives 4, 5, 17 and the
positive control (etoposide) at various concentrations against: A) cervical
carcinoma (HeLa) and B) human embryonic kidney (HEK-293) cells, both
for 24 h. Experiments were performed in triplicate and data are expressed
as mean ± SD.
It can be seen in Figure S1 how reynosin (2) is displaced from the
mass-center to the interface between the DPPC molecules and
water above and below the bilayer. This finding is consistent with
the behavior expected for its lipophilicity value (CLogP = 1.183).
Nevertheless, this compound could be retained when the
phosphatidyl and choline groups of the DPPC were solvated by
water, as shown in Table 2. Reynosin (2) presents the strongest
interaction with DPPC and also with water molecules, both of
which support its amphiphilic behavior as expressed in Figure
S1E, where at the end of the simulation reynosin (2) has the same
spatial layout as DPPC. After 10 ns of simulation it can be
observed how the lactone ring of the molecule becomes oriented
towards the aqueous phase, with the dipole moment pointing
towards the water. This situation is in agreement with that
represented in Figure 8.
The RMSD plot (Figure 7A) of this compound shows significant
fluctuations due to the dipole orientation, but it is stable at around
1.5 Å during the 10 ns simulation. These low variations are
indicative of the high stability of reynosin (2) in the position
represented in Figure 7A. This means that this compound will
encounter problems in the membrane‒crossing processes and it
will be retained at the interface.
Figure 6. (A) DNA fluorescence staining by Hoechst (B)
immunofluorescence staining results of Caspase-3 and (C) merged in
Lane 3, after treatment with negative control (DMSO 0.1%), positive
control (etoposide) and 4-fluorobenzoate reynosin derivative (5) at 100 µM
for 20 h on HeLa cells.
In the case of compound 5, programed suicide death by apoptosis
was assessed. Apoptosis can be intrinsically activated by the
release of cytochrome C from mitochondria or extrinsically
activated by death receptors. Both pathways lead to the activation
In the case of octanoylreynosin (11) the dipole moment changes
markedly with respect to reynosin. Instead of being oriented
towards the eudemasne skeleton, the carbonyl group provides
5
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000783
ChemMedChem
FULL PAPER
sufficient polarity to move the dipole moment towards the six-
membered ring in the sesquiterpene. Notwithstanding, it can be
seen from Figure 7C and Figure S2 that the molecule changes its
conformation to orient the long saturated chain parallel to the
DPPC molecules and the eudesmane fragment is in the same
spatial layout as reynosin (2).
snapshots observed in Figure S2 do not show any significant
movements of the molecules and they remain in the middle of the
bilayer in all cases. The diffusion coefficient of octanoylreynosin
(11) is the lowest of all of the tested compounds and this is due to
the high molecular volume according to the Stokes–Einstein
equation (Equation 1), with a similar value to the DPPC molecule,
which hinders the movement by diffusion. All of the results
outlined above, i.e., RMSD plot, energy values and diffusion
coefficient, lead us to believe that this compound will not easily
pass through the membrane and, as a consequence, it has a
lower bioactivity than the other two compounds simulated.
Reynosin (2)
A
2.5
2
kB · T
D =
1.5
1
6 · π · η · r
Equation 1. Stokes–Einstein relation for the diffusion coefficient, where
the constant r is related to the radius of the molecule, which in turn is
directly proportional to molecular volume. (kB is the Boltzmann constant, T
is the temperature and  is the dynamic viscosity).
0.5
0
Table 2. Summary of molecular dynamics simulations data acquired
4-Fluorobenzoyl-Reynosin (5)
B
for each derivative at 298 K.
3
2.5
2
Reynosin
Octanoyl-
Fluoro-
System/Values
(2)
Reynosin (11) Reynosin (5)
DPPC-Lig Energy
(Kcal/mol)
–26.4976
–9.33472
–9.5189
–6.7482
–2.88359
–3.04115
Water-Lig Energy
(Kcal/mol)
1.5
1
DPPC Diffusion Coef.
(cm2/s) (10–5)
0.0163 ±
0.0051
0.0084 ±
0.0033
0.0074 ±
0.0016
Lig Diffusion Coef.
(cm2/s) (10–5)
0.0312 ±
0.0675
0.0269 ±
0.0051
0.1115 ±
0.0640
0.5
0
Octanoyl-Reynosin (11)
C
2.5
2
1.5
1
0.5
0
0
2
4
6
8
10
Time / ns
Figure 7. RMSD plot of reynosin (2) (A), 4-fluorobenzoyl-reynosin (5) (B)
and octanoyl-reynosin (11) (C) in the simulated membrane system.
The octanoyl ester derivative 11 has a high lipophilicity value
(CLogP = 5.303), which is more similar to that of DPPC (CLogP =
4.004) than reynosin (2). However, according to the data in Table
1 compound 11 has a lower binding energy. This fact can be
explained by the lower energetic interaction between the non-
polar groups than interactions such as hydrogen bonds
established by polar groups.
The RMSD plot in Figure 7C showed a small but continuous
fluctuation from the start to the end of the MD simulation.
Nevertheless, the RMSD did not exceed 1.5 Å. In addition, the
Figure 9. Snapshot of 4-fluorobenzoyl-reynosin (5) after 10 ns of
molecular dynamics in the simulated membrane system.
6
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000783
ChemMedChem
FULL PAPER
4-Fluorobenzoylreynosin (5) shows the highest interaction energy
with DPPC and water molecules. In addition, compound 5 has the
highest diffusion coefficient, which could indicate an easy
passage through the membrane. The diffusion coefficient is quite
remarkable as it is three- and four-fold higher than the values for
simulation are the most relevant in comparison with others, as
shown by the RMSD plot in Figure 7B. Changes from 0 to 2.5 Å
in a steady rise are also observed in the snapshots (Figure 9),
which means that the compound is not retained to any extent and
it can move more easily than other derivatives from the aqueous
reynosin (2) and octanoylreynosin (11), respectively. Furthermore, phase through the DPPC molecules.
the calculated diffusion coefficient is comparable to that of
pyridine in water (2.8·10–6 cm2/s).[28]
It can be observed from Figure S3 that the fluoro-substituted
fragment, which is the most polar unit, is oriented towards the
water phase and the dipole moment points towards the
phosphatidyl and choline groups.
This derivative has the best lipophilicity value to be considered as
a good drug candidate, according to Lipinski’s rule (CLogP =
4.201), as a compromise between water solubility and interaction
with fatty derivatives (DPPC). Furthermore, fluctuations along the
Reynosin
Dipole moment: 6.2652 Debye
Fluor-Reynosin
Dipole moment: 6.3349 Debye
Octanoyl-Reynosin
Dipole moment: 6.2082 Debye
Figure 8. Calculated dipole moments for three derivatives selected for MD simulations employing HF/6-31G.
occupancy with the DPPC. This finding
supports the RMSD plot and snapshots
shown in Figure 7A, which indicates a very
stable arrangement of the compound in the
membrane and high energetic interaction
with the polar fragment of the DPPC in the
interphase.
4-Fluororeynosin ‒ Solvent
Octanoylreynosin ‒ Solvent
2
1.5
1
2
1.5
1
0.5
0
0.5
0
0
1000
2000
3000
4000
5000
0
1000 2000 3000 4000 5000
Neither octanoylreynosin (11) nor 4-
fluorobenzoylreynosin (5) form hydrogen
bonds with DPPC molecules and these
compounds show a similar number of
hydrogen bonds with the solvent (Table S15
and S16). Nevertheless, octanoylreynosin
(11) only forms H-bonds in the early steps of
the molecular dynamics simulation (before 4
ns). The structural similarity with DPPC
molecules means that 11 is located in the
Reynosin ‒ Solvent
Reynosin ‒ DPPC
2
1.5
1
2
1.5
1
0.5
0
0.5
0
0
1000
2000
Frames
3000
4000
5000
0
1000 2000 3000 4000 5000
Frames
middle of the membrane and it establishes other secondary
interactions with phosphatidyl derivatives (DPPC). In the case of
4-fluorobenzoylreynosin (5), the solvent donor hydrogen bond
interaction has the lowest occupancy, with values below 0.4%.
However, these bonds are present in all stages of the MD. This
means that the molecule switches throughout the simulation and
bonds with different solvent molecules, as shown in Figure 7B,
and this facilitates membrane crossing.
Figure 10. Analysis of hydrogen bonds for each bioactive molecule with
the solvent and DPPC at every step of the molecular dynamics simulation.
All molecules interact with the solvent, but only reynosin (2) forms H-bonds
with DPPC.
A hydrogen bond analysis fixed at a distance of 3.0 Å and 20
degrees as the cut-off angle showed that reynosin (2) is the only
compound that is able to establish this intermolecular interaction
with DPPC molecules due to its hydroxyl group. Only two
hydrogen bonds were established, with DPPC79 and DPPC75
(12.84% and 1.49% of occupancy, respectively), but these had a
long lifetime (Figure 10) and high occupancy throughout the 10 ns
of the simulation. The interaction of reynosin (2) with the solvent
was also the most abundant but this had a shorter lifetime than
the H-bonds with DPPC. The occupancies of the atoms involved
did not exceed 2.5% (Table S14) – in contrast with data for
Conclusion
Natural hydroxylated derivatives of costunolide remain relevant
for future drug development. These compounds have previously
had important applications against different tumors, but ester
derivatives have never been tested. Eight ester-eudesmane
derivatives have been synthesized in quantitative yield in a one-
7
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000783
ChemMedChem
FULL PAPER
Cell viability assays. Tumoral (HeLa) and non-tumoral cells
(HEK-293) were used for the experiments. 1.5 × 105 cells per mL
were seeded in 6-well plates (VWR, Germany) in complete
medium. Compounds were dissolved in DMSO (0.1% v/v) at the
appropriate concentration for 24 h. Control cultures, including
cells treated either with 0.1% DMSO or etoposide, were also
included in each experiment. Cell viability was evaluated by the
Trypan blue dye exclusion assay. Trypan blue solution was mixed
1:1 with a sample of control or treated cells. After incubation for 2
min, a fraction of blue-stained cells was assessed using an
Automated TC20 Cell Counter (Bio-Rad). IC50 values were
determined with GraphPad (Prism software v. 5.00). All
experiments were repeated at least three times with similar results.
The results are expressed as the mean ± S.D.
pot synthesis from a starting material that can be obtained on
multi-gram scale from plant material.
Ester-eudesmane derivatives showed strong cytotoxicity against
HeLa cells, with values of less than 20% cell viability observed in
most cases after 24 hours of treatment. The derivatives with an
aromatic framework were more active than compounds with a
linear ester chain. In view of these findings, other mimics and
natural compounds were modified by introducing
a
4-
fluorobenzoyl core. However, the boost in bioactivity shown by
reynosin esters does not have a counterpart in other natural
products. Overall, the modification of amino groups to yield amide
derivatives led to a decrease in the activity, while the best results
were obtained by modifying the hydroxyl groups to form ester
derivatives. It is worth highlighting the potential cytotoxic effects
of compounds 4, 5 and 17. These findings encouraged us to
perform a cell viability study with these promising compounds at
various concentrations against tumoral (HeLa) and non-tumoral
(HEK-293) cells. It is worth noting the high selectivity of derivative
5 between the two cell lines tested, with a selectivity index above
three. Immunofluorescence staining analysis showed that this
derivative induces apoptosis and DNA fragmentation on HeLa
cells.
Immunofluorescence staining. HeLa cells were grown on 11
mm × 22 mm coverslips (Thomas Scientific, USA) until 70% of
confluence and then incubated for 20 h with 4-fluorobenzoate
reynosin derivative (4) or etoposide (positive control) at 100 µM in
culture medium as described above. Caspase-3 activity assay
was performed according to the manufacturer’s protocol.[29] In
brief, cover slips with HeLa cells were fixed in methanol for 10 min
at –20 ºC, later washed in PBS, and incubated with the primary
rabbit active caspase-3 antibody at 1:500 dilution in PBS for 45
min at 37 ºC. Thereafter, cells were washed in PBS in triplicate
for 5 min each and incubated with the secondary antibody
fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit IgG
(dilution 1:100) in PBS for 45 min at 37 ºC. Finally, cells were
washed three times with PBS, and mounted in PBS–glycerol
containing Hoechst at 1 µg/ml. The expression of Caspase-3 in
cells was observed under a Zeiss Axiophot fluorescence
microscope (Carl Zeiss, West Germany) equipped with a Zeiss
Axiocam 503 camera.
In an effort to elucidate the origin of the bioactivity, molecular
dynamics simulations were carried out with a cell membrane
construction and the cross-membrane processes was linked with
the cell viability results. These simulations showed how 4-
fluorobenzoylreynosin (5) is not retained by either phosphatidyl
derivatives (DPPC) that constitute the membrane or solvent
molecules, so it can move through the system more easily than
reynosin (2) or other derivatives. This compound has the highest
diffusion coefficient and establishes hydrogen bonds with a short
lifetime, which could explain theoretically the level of bioactivity.
Isolation of costunolide (1). Saussurea costus CH2Cl2 root
extract (50.0 g) was employed for the isolation of 2.5 g of
costunolide. 6 L of hexane/ethyl acetate 5% was used to purify
the compound by column chromatography.
Synthesis of reynosin (2) and santamarine (3). The method of
Zorrilla et al. with modifications was employed.[30] Costunolide
(100 mg, 4.30·10–4 moles) was dissolved in CH2Cl2 (6 mL). m-
CPBA (89 mg, 5.16·10–4 moles, 1.2 eq.) was added to the above
solution and the reaction mixture was stirred at room temperature
for 24 h. A catalytic amount of p-TsOH was added and the
reaction mixture was stirred for a further 2 h. The mixture was
extracted three times with 1.0 M NaOH and the combined organic
phase was dried with anhydrous Na2SO4. Purification by CC was
carried out with a Hexane:AcOEt gradient (80:20–50:50) to give
santamarine (2) (58.8 mg, 58.8%) and reynosin (3) (30.5 mg,
30.5%).
Experimental Section
Chemicals. For the cytotoxicity bioassays, Dulbecco's Modified
Eagle's Medium (DMEM) was supplied by Lonza (Verviers,
Belgium), premixed phosphate buffer saline solution (PBS, 10X)
was supplied by Roche (Steinheim, Germany), fetal bovine serum,
penicillin/streptomycin, l-glutamine, sodium pyruvate, trypsin and
minimum essential medium non-essential amino acids (MEM
NEAA) were purchased from Gibco (Paisley, UK). Anti-active
caspase 3 rabbit polyclonal antibody, fluorescein isothiocyanate
(FITC)-conjugated goat anti-rabbit IgG, etoposide and trypan blue
solution (0.4%) were purchased from Sigma Aldrich (Steinheim,
Germany). Bisbenzimidine H 33342 fluorochrome (Hoechst) was
obtained from Calbiochem. 4-Fluorobenzoyl chloride, benzoyl
chloride, 3-fluorobenzoyl chloride, 4-chlorobenzoyl chloride,
fluorobenzoic acid, butyryl chloride, octanoyl chloride, 2-
aminothiophenol, 2-hydroxythiophenol, meta-chloroperbenzoic
acid were supplied by Sigma-Aldrich®. 2,4,6-Trifluorobenzoyl
chloride and p-TsOH were supplied by Alfa-Aesar®.
Synthesis of 4-fluorobenzoylsantamarine (4). Santamarine (2)
(30 mg, 1.21·10-4 moles) was dissolved in dry pyridine (2 mL)
under an inert atmosphere. Two equivalents of 4-fluorobenzoyl
chloride (2.42·10–4 moles, 28.6 L) were added to the flask and
the reaction mixture was stirred for 12 h at room temperature. The
mixture was extracted three times with saturated aqueous CuSO4
and the organic phase was washed three times with 0.1 M NaOH
solution. Chromatographic separation was carried out with a
gradient from 10% to 40% (Hexane:AcOEt) to give the product in
85% yield. NMR data see Table S1, Figure S4 and S5. Calculated
m/z for [C22H23FO4]Na+ 393.1473, obtained 393.1498. IR cm–1):
2935.5, 2854.3, 1771.0, 1716.9, 1603.1, 1507.8, 1271.2, 1237.4,
Cell lines and cell cultures. HeLa (human cervix carcinoma) and
HEK-293 (human embryonic kidney 293) cells were cultured as
monolayers in DMEM (GIBCO) supplemented with 10% fetal
bovine serum, 5% glutamine, 5% non-essential amino acids, 5%
penicillin-streptomycin and 5% sodium pyruvate. Cells were
maintained in a HERA Cell 150i (Thermo Scientific) incubator at
37 ºC, 5% CO2 and 95% humidity.
1114.9, 1090.3, 982.8, 964.6, 854.4, 767.3. UV (CH3CN), max
:
244 nm.
8
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000783
ChemMedChem
FULL PAPER
General synthesis of acyl derivatives of reynosin (5–11).
Reynosin (3) (30 mg, 1.21·10–4 moles) was dissolved in dry
pyridine (2 mL) under an inert atmosphere. Two equivalents of the
acyl chloride (Table 2) were added to the flask and the reaction
mixture was stirred for 12 h at room temperature. The mixture was
extracted three times with saturated aqueous CuSO4 and the
organic phase was washed three times with 0.1 M NaOH solution.
Chromatographic separation was carried out with a gradient from
10% to 40% (Hexane:AcOEt) to give the desired compounds in
the yields shown in Table 3.
1368.7, 1257.0, 1227.8, 1179.7, 1149.2, 1092.6, 1050.6, 986.5,
966.0. UV (CH3CN), max: 194 nm.
Octanoylreynosin (11). NMR data see Table S8, Figure S18 and
S19. Calculated m/z for [C23H34O4]Na+ 397.2349, obtained
397.2330. IR (cm–1): 2930.1, 2856.7, 1774.0, 1733.8, 1457.8,
1412.8, 1377.7, 1257.8, 1148.5, 1051.3, 966.2. UV (CH3CN),

max: 196 nm.
Isolation of cynaropicrin (12). Cynara cardunculus var.
scolymus AcOEt leaf extract (94.1 g) was employed to isolate
0.24 g of cynaropicrin. The isolation procedure was similar to that
described in the literature.[31]
Table 3. Experimental data in synthesis of acyl derivatives
Compound
Volume (L) Yield (%)
4-fluorobenzoyl chloride (5)
benzoyl chloride (6)
28.6
28.1
31.0
29.0
88
99
85
86
76
87
80
Synthesis
of
bis(4-fluorobenzoyl)cynaropicrin
(13).
Cynaropicrin (12) (30 mg, 8.66·10–5 moles) was dissolved in dry
pyridine (2 mL) under an inert atmosphere. Three equivalents of
4-fluorobenzoyl chloride (2.60·10–4 moles, 30.7 L) were added
to the flask and the reaction mixture was stirred for 12 h at room
temperature. The mixture was extracted three times with
saturated aqueous CuSO4 and the organic phase was washed
three times with 0.1 M NaOH solution. Chromatographic
separation was carried out with a gradient from 0% to 30%
(Hexane:Acetone) and HPLC was subsequently employed to
obtain the product at 7.4 min (35% Hexane:Acetone) (45% yield)
using a Merck-Hitachi system (Tokyo, Japan) with a refractive
index detector (Elite LaChrom L-2490). A semipreparative
LiChrospher 10 μm 250-10 Si 60 (Merck) column was employed
with a flow rate of 3 mL/min. NMR data see Table S9, Figure S20
and S21. Calculated m/z for [C33H28F2O8]Na+ 613.1644, obtained
613.1644. IR cm–1): 2927.7, 1770.2, 1722.3, 1603.2, 1508.2,
1270.3, 1241.1, 1153.6, 1113.7, 854.7, 768.1. UV (CH3CN), max
201 nm.
4-chlorobenzoyl chloride (7)
3-fluorobenzoyl chloride (8)
2,4,6-trifluorobenzoyl chloride (9) 31.4
butyryl chloride (10)
octanoyl chloride (11)
25.1
41.3
4-Fluorobenzoylreynosin (5). NMR data see Table S2, Figure S6
and S7. Calculated m/z for [C22H23FO4]Na+ 393.1473, obtained
393.1492. IR cm–1): 2933.0, 2851.6, 1771.9, 1716.8, 1682.6,
1605.0, 1508.4, 1271.8, 1240.9, 1153.1, 1114.0, 1091.3, 854.7,
768.8. UV (CH3CN), max: 248 nm.
Benzoylreynosin (6). NMR data see Table S3, Figure S8 and S9.
Calculated m/z for [C22H24O4]Na+ 375.1567, obtained 375.1576.
IR (cm–1): 2938.9, 2856.1, 1771.7, 1715.3, 1451.4, 1272.3,
1112.6, 965.8, 712.8. UV (CH3CN), max: 224 and 196 nm.
Synthesis of 2-amino-3H-phenoxazin-3-one (APO) (14). 2-
Aminophenol (1 g, 9.16·10–3 moles) was dispersed in a mixture
50:50 H2O/MeOH and stirred during 24 hours while an oxygen
flow was passed through at room temperature. A dark precipitate
was obtained and this was filtered off. This solid was washed with
cold water and recrystallized in CHCl3 to give APO (99% yield).
4-Chlorobenzoylreynosin (7). NMR data see Table S4, Figure
S10 and S11. Calculated m/z for [C22H23ClO4]Na+ 409.1177,
obtained 409.1180. IR (cm–1): 2933.2, 2854.6, 1771.7, 1717.2,
1594.3, 1271.2, 1116.5, 1103.9, 1032.8, 760.0. UV (CH3CN),
Synthesis of 4-fluorobenzoyl-N-APO (15). APO (14) (30 mg,
1.59·10-4 moles) was dissolved in dry pyridine (2 mL) under an
inert atmosphere. 1.5 equivalents of 4-fluorobenzoyl chloride
(2.39·10–4 moles, 28.2 L) were added to the flask and the
reaction mixture was stirred for 12 h at room temperature. The
mixture was extracted three times with saturated aqueous CuSO4
and the organic phase was washed three times with 0.1 M NaOH
solution. Chromatographic separation was carried out with a
gradient from 0% to 20% (Hexane:AcOEt) to give the target
compound in 69% yield. NMR data see Table S10, Figure S22
and S23. Calculated m/z for [C19H11F2N2O3]H+ 335.0826,
obtained 335.0846. IR (cm–1): 3355.2, 2921.9, 2851.6, 1739.0,
1692.4, 1618.1, 1529.8, 1502.1, 1350.0, 1240.3, 1174.9, 880.7,
847.9, 755.1. UV (CH3CN), max: 255 and 410 nm.

max: 239 and 197 nm.
3-Fluorobenzoylreynosin (8). NMR data see Table S5, Figure S12
and S13. Calculated m/z for [C22H23FO4]Na+ 393.1473, obtained
393.1483. IR (cm–1): 2928.4, 2854.6, 1773.3, 1719.0, 1592.3,
1446.1, 1290,1. 1270.5, 1203.8, 1093.2, 966.1, 755.6. UV
(CH3CN), max: 278, 225 and 192 nm.
2,4,6-Trifluorobenzoylreynosin (9). NMR data see Table S6,
Figure S14 and S15. Calculated m/z for [C22H21F3O4]Na+
429.1284, obtained 429.1299. IR (cm–1): 2938.8, 2850.0, 1772.5,
1730.0, 1641.8, 1618.4, 1444.9, 1272.4, 1130.0, 1048.3, 852.4.
UV (CH3CN), max: 221 and 192 nm.
Synthesis of 2,2'-disulfanediyldiphenol (DiS-OH) (16). This
compound was synthesized by the method of Oliveira et al. with
modifications.[32] 2-Hydroxythiophenol (100 mg, 0.82 L, 7.93·10–
4 moles) was dissolved in acetone (35 mL). A solution of KIO3
(1.36 g, 8 eq. 6.59·10-3 moles) in water (100 mL) was added and
Butyrylreynosin (10). NMR data see Table S7, Figure S16 and
S17. Calculated m/z for [C19H26O4]Na+ 341.1723, obtained
341.1731. IR (cm–1): 2936.5, 2874.8, 1771.5, 1732.7, 1458.1,
9
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000783
ChemMedChem
FULL PAPER
the mixture was stirred for 24 hours at room temperature. The
mixture was extracted five times with AcOEt and the combined
organic phases were dried (Na2SO4). No further purification was
required to give DiS-OH (195 mg, 98% yield).
Molecular
dynamics
simulations.
Dipalmitoylphosphatidylcholine (DPPC) was chosen as the lipid
molecule to simulate the biological membrane, with TIP3P as a
model for water molecules. GROMACS (2019.6 version) was
employed with the OPLS-AA force-field, which is optimal for
liquids, solvated and pseudo-liquid simulations.[33] The ligand was
generated with GaussView (6.0.16 version) and its topology and
parameters were obtained by employing the SwissParam server
(www.swissparam.ch).[34] 128 molecules of DPPC were added in
a cubic box with at least 1 nm from the edges of the box and 2 nm
distance between periodic simulated membrane images, to fulfill
the minimum image convention. Sodium and chloride ionic
species were added to simulate the physiological conditions while
keeping the system neutral. The ligand employed in the simulation
was placed on the center-of-mass of the bilayer, parallel to DPPC
molecules. The system was minimized until the force was
<1000.0 kJ·mol–1·nm–1 employing a short-range of electrostatic
and van der Waals cut-off of 1.2 Å. The system was then
equilibrated at constant volume-temperature (NVT) for 0.1 ns with
an increment of 2 fs per step. Berendsend thermostat modification
(V-rescale) was employed and the temperature was 298 K to
avoid the phase transition of the DPPC (315 K). An isobaric-
isothermal (NPT) equilibration was performed for 1 ns with
increments of 2 fs per step, employing the Nosé–Hoover
thermostat. The increment of time concerning the NVT was due
to heterogeneity of the system because of slow lipid diffusion. The
fully equilibrated system was submitted to molecular dynamics
simulation for 10 ns with 2 fs per step. In this case, the Nosé–
Hoover thermostat was used for temperature control and the
canonical Parrinello–Rahman thermostat for pressure control.
Finally, a correction of the trajectory was carried out by DPPC
bilayer recentering within the cubic box.
Synthesis
of
disulfanediylbis(2,1-phenylene)
bis(4-
fluorobenzoate) (17). DiS-OH (16) (100 mg, 4.00·10–4 moles)
was dissolved in dry pyridine (500 L). Three equivalents of 4-
fluorobenzoyl chloride (1.20·10–4 moles, 141.6 L) were added
and the reaction mixture was stirred for 24 h at room temperature.
The mixture was directly separated by column chromatography
with a gradient from 5% to 30% (Hexane:AcOEt) to give the target
compound in 89% yield. NMR data see Table S11, Figure S24
and S25. Calculated m/z for [C26H16F2O4S2]Na+ 517.0350,
obtained 517.0380. IR (cm–1):1742.2, 1733.9, 1601.3, 1504.6,
1466.6, 1252.6, 1240.8, 1198.7, 1155.9, 1064.8, 1014.4, 856.7,
761.2, 749.8, 682.6, 625.6, 575.3. UV (CH3CN), max: 199 nm.
Synthesis of 2,2'-disulfanediyldianiline (DiS-NH2) (18). This
compound was synthesized by the method of Oliveira et al. with
modifications.[32] 2-Aminothiophenol (1 g, 0.83 mL, 7.99·10–3
moles) was dissolved in a 1:1 mixture of H2O:ethanol (100 mL).
The solution was stirred for 18 h with airflow until a solid
precipitated. The solid was filtered off, washed with cold water and
dried in a vacuum oven overnight to give the target compound in
80% yield.
Synthesis of N,N'-(disulfanediylbis(2,1-phenylene))bis(4-
fluorobenzamide) (19). DiS-NH2 (18) (100 mg, 4.03·10–4 moles)
was dissolved in dry pyridine (500 L). Three equivalents of 4-
fluorobenzoyl chloride (1.21·10–4 moles, 142.8 L) were added
and the reaction mixture was stirred for 24 h at room temperature.
The mixture was separated by column chromatography
employing a gradient from 5% to 30% (Hexane:AcOEt) to give the
target compound in 81% yield. NMR data see Table S12, Figure
S26 and S27. Calculated m/z for [C26H18F2N2O2S2]H+ 493.0851,
obtained 493.0865. IR (cm–1): 3377.5 1770.4, 1676.5, 1601.7,
1576.7, 1525.2, 1500.5, 1431.2, 1311.4, 1295.2, 1233.8, 1162.6,
849.1, 757.0, 749.2, 575.6. UV (CH3CN), max: 264 and 319 nm.
Acknowledgements
This research was supported by the “Ministerio de Economꢀa,
Industria y Competitividad” (MINEICO), Spain, Project AGL2017-
88-083-R. F.J.R.M. thanks the University of Cꢁdiz for predoctoral
support under grant 2018-009/PU/EPIF-FPI-CT/CP. The authors
declare no competing financial interest.
Isolation of gummiferolic acid (20). Margotia gummifera CH2Cl2
root extract (592.5 g) was employed to isolate 17.58 g of
gummiferolic acid. The extract was concentrated and crystallized
from hexane/CH2Cl2 (1:1 mixture).
Keywords: Eudesmanolide, Molecular Dynamics, Acyl
Derivative, Natural Product, Caspase-3
[1]
[2]
D. Camp, A. Garavelas, M. Campitelli, J. Nat. Prod. 2015,
78, 1370–1382.
Synthesis of 4-fluorobenzoic gummiferolic anhydride (21).
Gummiferolic acid (20) (100 mg, 2.50·10–4 moles) was dissolved
in dry pyridine (2 mL) under an inert atmosphere. Five equivalents
of 4-fluorobenzoyl chloride (1.25·10–3 moles, 147.5 L) were
added and the reaction mixture was stirred for 72 h at room
temperature. The mixture was extracted three times with
saturated aqueous CuSO4 and the organic phase was washed
three times with 0.1 M NaOH solution. Chromatographic
separation was carried out with a gradient from 0% to 40%
(Hexane:AcOEt) to give the target compound in 22% yield. NMR
data see Table S13, Figure S28 and S29. Calculated m/z for
[C32H38F2O5]- 521.2682, obtained 521.2703. IR (cm–1): 2932.1,
2871.2, 1797.0, 1730.9, 1713.7, 1602.6, 1507.8, 1455.6, 1381.8,
1235.3, 1156.8, 997.2, 847.7. UV (CH3CN), max: 203 nm.
E. W. Coronado-Aceves, C. Velázquez, R. E. Robles-
Zepeda, M. Jiménez-Estrada, J. Hernández-Martínez, J. C.
Gálvez-Ruiz, A. Garibay-Escobar, Pharm. Biol. 2016, 54,
2623–2628.
[3]
[4]
L. Liu, D. Liu, C. Xiang, W. Dai, B. Li, M. Zhang, Nat. Prod.
Res. 2020, 34, 1563–1566.
L. F. Ding, J. Su, Z. H. Pan, Z. J. Zhang, X. N. Li, L. D.
Song, X. De Wu, Q. S. Zhao, Phytochemistry 2018, 155,
182–190.
[5]
S. Lim, S. J. Lee, K. W. Nam, K. H. Kim, W. Mar, Arch.
10
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000783
ChemMedChem
FULL PAPER
Pharm. Res. 2013, 36, 485–494.
Adv. Drug Deliv. Rev. 2001, 46, 3–26.
[6]
[7]
[8]
A. Ham, D. Kim, K. Ho, S. Lee, K. Oh, J. Shin, W. Mar,
[23]
K. Donthiboina, P. Anchi, P. V. Sri Ramya, S. Karri, G.
Srinivasulu, C. Godugu, N. Shankaraiah, A. Kamal, Bioorg.
Chem. 2019, 86, 210–223.
Brain Res. 2013, 1524, 54–61.
S. Choodej, K. Pudhom, T. Mitsunaga, Planta Med. 2018,
84, 329–335.
[24]
[25]
[26]
P. Prayong, S. Barusrux, N. Weerapreeyakul, Fitoterapia
2008, 79, 598–601.
R. Amici, C. Bigogno, R. Boggio, A. Colombo, S. M.
Courtney, R. Dal Zuffo, G. Dondio, F. Fusar, S. Gagliardi,
S. Minucci, et al., ChemMedChem 2014, 9, 1574–1585.
B. J. Schultz, S. Y. Kim, W. Lau, E. S. Sattely, J. Am.
Chem. Soc. 2019, 141, 19231–19235.
[9]
K. R. A. Abdellatif, H. A. H. Elshemy, S. A. Salama, H. A.
H. Wang, J. Zhu, L. Jiang, B. Shan, P. Xiao, J. Ai, N. Li, F.
Omar, J. Enzyme Inhib. Med. Chem. 2015, 30, 484–491.
Qi, S. Niu, BMC Complement. Med. Ther. 2020, 20, 180.
[10]
K. S. Gavale, S. R. Chavan, N. Kumbhar, S. Kawade, P.
Doshi, A. Khan, D. D. Dhavale, Bioorganic Med. Chem.
2017, 25, 5148–5159.
[27]
[28]
S. Elmore, Toxicol. Pathol. 2007, 35, 495–516.
C. L. Yaws, in Transp. Prop. Chem. Hydrocarb., William
Andrew Inc., 2009, pp. 502–593.
[11]
[12]
[13]
J. Y. Cho, A. R. Kim, J. H. Jung, T. Chun, M. H. Rhee, E. S.
Yoo, Eur. J. Pharmacol. 2004, 492, 85–94.
[29]
A. Yiallouris, I. Patrikios, E. O. Johnson, E. Sereti, K.
Dimas, C. De Ford, N. U. Fedosova, W. F. Graier, K.
Sokratous, K. Kyriakou, et al., Cell Death Dis. 2018, 9, 764.
W. H. Wu, T. Y. Chen, R. W. Lu, S. T. Chen, C. C. Chang,
Phytochemistry 2012, 83, 110–115.
[30]
[31]
[32]
J. G. Zorrilla, C. Rial, R. M. Varela, J. M. G. Molinillo, F. A.
S. K. Steffensen, H. A. Pedersen, K. B. Adhikari, B. B.
Laursen, C. Jensen, S. Høyer, M. Borre, H. H. Pedersen,
M. Borre, D. Edwards, et al., J. Agric. Food Chem. 2016,
64, 8235–8245.
Macías, Phytochem. Lett. 2019, 31, 229–236.
C. Rial, P. Novaes, R. M. Varela, J. M. . Molinillo, F. A.
Macias, J. Agric. Food Chem. 2014, 62, 6699–6706.
[14]
E. Flores-Giubi, N. Geribaldi-Doldán, M. Murillo-Carretero,
C. Castro, R. Durán-Patrón, A. J. MacÍas-Sánchez, R.
Hernández-Galán, J. Nat. Prod. 2019, 82, 2517–2528.
S. C. C. Oliveira, C. K. Z. Andrade, R. M. Varela, J. M. G.
Molinillo, F. A. Macías, ACS Omega 2019, 4, 2362–2368.
[33]
[34]
R. Zangi, ACS Omega 2018, 3, 18089–18099.
[15]
[16]
[17]
[18]
[19]
[20]
S. Lim, S.-J. Lee, K.-W. Nam, K. H. Kim, W. Mar, Arch.
Pharm. Res. 2013, 36, 485–494.
V. Zoete, M. A. Cuendet, A. Grosdidier, O. Michielin, J.
Comput. Chem. 2011, 32, 2359–2368.
N. Gören, H. Woerdenbag, C. Bozok-Johansson, Planta
Med. 1996, 62, 419–422.
S. K. Poole, C. F. Poole, J. Chromatogr. B Anal. Technol.
Biomed. Life Sci. 2003, 797, 3–19.
L. G. Danielsson, Y. H. Zhang, TrAC - Trends Anal. Chem.
1996, 15, 188–196.
C.-M. Sun, W.-J. Syu, M.-J. Don, J.-J. Lu, G.-H. Lee, J. Nat.
Prod. 2003, 66, 1175–1180.
A. C. Beekman, H. J. Woerdenbag, W. Van Uden, N. Pras,
A. W. T. Konings, H. V. Wikström, T. J. Schmidt, J. Nat.
Prod. 1997, 60, 252–257.
[21]
[22]
D. Siriwan, T. Naruse, H. Tamura, Fitoterapia 2011, 82,
1093–1101.
C. A. Lipinski, F. Lombardo, B. W. Dominy, P. J. Feeney,
11
This article is protected by copyright. All rights reserved.

10.1002/cmdc.202000783
ChemMedChem
FULL PAPER
Entry for the Table of Contents
Acyl Eudesmanolides
Physicochemical Modulation
•
•
•
Cytotoxic selectivity
Natural products derivatives
Cross-membrane process
improvement
Acyl derivatives of different natural product families, have been synthesized and tested against tumoral and non-tumoral cell lines to identify
selective derivatives with reduced negative impact upon application. The mode of action of these compounds was analyzed by anti-caspase-3
assays and molecular dynamics simulations provide evidence that lipophilicity plays a key role and the 4-fluorobenzoyl derivative can pass
easily through the cell membrane.
12
This article is protected by copyright. All rights reserved.