Facile synthesis of anhydrojudaicin and 11,13-dehydroanhydrojudaicin, two eudesmanolide-skeleton lactones with potential allelopathic activity

Article abstract of DOI:10.1016/j.phytol.2019.04.014

Natural product anhydrojudaicin (7) is a eudesmanolide that has been synthesized from costunolide (1) in three steps with good yield, simplifying procedures available in the literature. The key step was the efficient oxidation and rearrangement using 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ). Following the same methodology, 11,13-dehydroanhydrojudaicin (8) was prepared for the first time. All compounds were characterized and tested in three specific bioassays: phytotoxicity bioassays on wheat coleoptile and on weed seeds (Amaranthus viridis, Echinochloa crus-galli and Panicum maximum); as well as bioassay of germination of parasitic weed seeds of broomrape (Orobanchaceae). The final products showed better activity profiles than the starting eudesmanolide in all bioassays. The data allow us to propose compounds 7 and 8 as leads for new natural product-based agrochemicals.

Full text of DOI:10.1016/j.phytol.2019.04.014

PhytochemistryLetters31(2019)229–236
Contents lists available at ScienceDirect
Phytochemistry Letters
journal homepage: www.elsevier.com/locate/phytol
Facile synthesis of anhydrojudaicin and 11,13-dehydroanhydrojudaicin, two
eudesmanolide-skeleton lactones with potential allelopathic activity
Jesús G. Zorrilla, Carlos Rial, Rosa M. Varela, José M.G. Molinillo, Francisco A. Macías⁎
Allelopathy Group, Department of Organic Chemistry, Institute of Biomolecules (INBIO), Campus CEIA3, School of Science, University of Cadiz, C/ Republica Saharaui, 7,
11510, Puerto Real, Cádiz, Spain
A R T I C L E I N F O
A B S T R A C T
Keywords:
Natural product anhydrojudaicin (7) is a eudesmanolide that has been synthesized from costunolide (1) in three
steps with good yield, simplifying procedures available in the literature. The key step was the efficient oxidation
and rearrangement using 2,3-dichloro-5,6-dicyano-1,4-benzoquinone (DDQ). Following the same methodology,
11,13-dehydroanhydrojudaicin (8) was prepared for the first time. All compounds were characterized and tested
in three specific bioassays: phytotoxicity bioassays on wheat coleoptile and on weed seeds (Amaranthus viridis,
Echinochloa crus-galli and Panicum maximum); as well as bioassay of germination of parasitic weed seeds of
broomrape (Orobanchaceae). The final products showed better activity profiles than the starting eudesmanolide
in all bioassays. The data allow us to propose compounds 7 and 8 as leads for new natural product-based
agrochemicals.
Allelopathy
Allelochemical
Anhydrojudaicin
Phytotoxicity
Parasitic weeds
1. Introduction
development of an efficient method to obtain 7 by organic synthesis.
This procedure will also allow to prepare for the first time 11,13-de-
Eudesmanolides constitute a large family of sesquiterpene lactones
that includes compounds with different types of bioactivities, com-
prising antibacterial (Talbi et al., 2015), phytotoxic (Da Silva et al.,
2017), cytotoxic and antiviral (Hui et al., 2018). This fact explains the
biological activities of numerous plants that produce eudesmanolides as
allelochemicals. The presence of the α-methylene-γ-lactone group in
their structure is usually correlated with some of the activities shown by
sesquiterpene lactones (Padilla-Gonzalez et al., 2016).
hydroanhydrojudaicin (8), using costunolide (1) as starting material
(Fig. 1).
Considering good activities for eudesmanolides in the literature, the
second aim of the study is the evaluation of bioactivity of compounds
prepared. Three specific bioassays are selected, focusing on the search
of new lead compounds for agrochemicals based on natural products.
Firstly, the etiolated wheat coleoptile bioassays was selected as a
quick and easy general phytotoxicity bioassay. Second bioassay will
study the phytotoxicity of synthetized compounds on three problematic
weed species: Amaranthus viridis L. (green amaranth), which can affect
crops like soybean (Kaspary et al., 2017), Echinochloa crus-galli L.
(barnyard grass), which harms rice crops (Talbert and Burgos, 2007),
and Panicum maximum Jacq. (guinea grass), which is harmful on coffee
plantations (Da Silva et al., 2017).
Many eudesmanolides are produced as secondary metabolites by
plants, and consequently their isolation usually does not provide en-
ough yields to study properly their possibilities as bioactive com-
pounds. Organic synthesis gives a solution to this limitation, when ef-
ficient synthetic routes let obtain those minor compounds isolated from
natural sources.
Anhydrojudaicin (7) is a eudesmanolide isolated from the aerial
parts of Artemisia canariensis, and it is possible to obtain only 28 mg
from 10 kg of plant for its isolation (Mansilla and Palenzuela, 1999).
The synthesis of 7 in nine steps from α-santonin has been published.
Compound 7 is an intermediate in the synthesis of several bioactive
guaianolides (Yuuya et al., 1999) but activity has not been reported for
compound 7 itself.
Bioactivities of all the compounds will be also evaluated as suitable
allelochemicals to control parasitic weed infestations. Parasitic plants
differ from common plants in relevant biological aspects. A main par-
ticularity are their seeds, which only start their germination mechanism
when they detect a specific chemical signal generated by a nearby
possible host plant. This distinctive feature leads the search of com-
pounds that stimulate the parasitic seeds germination, in order to apply
the on growing suicidal germination strategy, which is also known as
For these reasons, the first aim of the study presented here is the
⁎ Corresponding author.
E-mail address: famacias@uca.es (F.A. Macías).
https://doi.org/10.1016/j.phytol.2019.04.014
Received 7 February 2019; Received in revised form 12 April 2019; Accepted 15 April 2019
1874-3900/©2019PhytochemicalSocietyofEurope.PublishedbyElsevierLtd.Allrightsreserved.

J.G. Zorrilla, et al.
PhytochemistryLetters31(2019)229–236
Table 1
NMR spectroscopic data [1H (500 MHz) and 13C (125 MHz)] for products 7 and
8 in CDCl3 (δ in ppm, J in Hz).
Product
7
8
Position δH (J, Hz)
δC
δH (J, Hz)
δC
1
2
3
–
202.4
–
202.4
125.2
146.7
5.86 (1H, d, J = 9.9)
125.2 5.88 (1H, d, J = 10.0)
7.04 (1H, dd, J = 9.9, 146.6 7.06 (1H, dd, J = 10.0, 0.7)
0.6)
4
5
6
–
139.7
–
139.5
49.9
79.7
2.82 (1H, d, J = 10.7) 49.5
2.96 (1H, d, J = 10.8)
4.16 (1H, dd,
J = 10.7, 10.7)
1.65 (1H, m)
79.4
52.0
31.5
4.15 (1H, dd, J = 10.8, 10.8)
Fig. 1. Synthetic route to obtain anhydrojudaicin (7) and 11,13-dehy-
droanhydrojudaicin (8) from costunolide (1).
7
2.57 (1H, ddddd, J = 11.2,
11.2, 3.2, 3.2, 3.2)
2.09 (1H, m)
49.2
31.2
8a
2.06 (1H, m)
1.55 (1H, m)
8b
9b
1.66 (1H, br ddd, J = 13.8,
13.0, 3.8)
the honeypot strategy, to prevent the infestations of parasitic weeds
(Mejías et al., 2018). Three of the most harmful parasitic weeds species
of broomrape were selected for the study, as they damage different
relevant crops: Phelipanche ramosa (Le Corre et al., 2014), Orobanche
cumana (Yang et al., 2017) and Orobanche crenata (Rubiales and
Fernández Aparicio, 2012).
9a
1.97 (1H, dd,
J = 12.8, 3.3)
1.50 (1H, dd,
J = 12.8, 3.3)
–
22.4
2.18 (1H, ddd, J = 13.1, 2.8,
2.8)
20.8
1.60 (1H, m)
10
11
47.2
40.6
–
–
47.3
2.35 (1H, dq,
J = 10.0, 6.9)
–
138.1
As all compounds synthetized are structurally related, the structure-
activity relation (SAR) study will be the third objective of the research
reported herein. Bioassays carried out will be useful to identify the
features that influence the activity of the molecules for each applica-
tion.
12
178.9
12.4
–
170.3
118.0
13a
1.25 (3H, d, J = 6.9)
6.15 (1H, d, J = 3.2)
5.48 (1H, d, J = 3.2)
1.06 (3H, s)
13b
14
1.07 (3H, s)
17.8
17.8
15a
5.82 (1H, dd, J = 1.4, 122.1 5.88 (1H, d, J = 1.3)
122.4
15b 0.6)
5.54 (1H, dd, J = 1.4,
0.6)
5.56 (1H, dd, J = 1.3, 0.7)
2. Results and discussion
2.1. Synthesis of anhydrojudaicin (7) and 11,13-dehydroanhydrojudaicin
(8)
correspond with those available for 1-oxo-5αH,6,11β-eudesm-2,4(14)-
dien-6,13-olide (Ando and Takase, 1977), which is also known as an-
hydrojudaicin (Saber et al., 1964). The S configuration at C-11 was
determined by NOE 1D NMR spectroscopy, with H-6 showing a positive
NOE effect with H-11, which implies a β-orientation of H-11 (Fig. 2).
The final step of the synthetic procedure to produce 8 involved
treatment of 3 with excess DDQ to give 8 in 43% yield. A new band due
to carbonyl group was observed at 1672 cm–1 in the IR spectrum.
Comparison of the 1H NMR spectrum with that of compound 7 showed
that similar signals were displayed for the protons of the carbocyclic
moiety. These spectra only differed in the presence of two signals at δ
6.14 and 5.48 for H-13 (instead of δ 1.25) and the absence of a signal
for H-11, which indicates the presence of a methylene rather than a
methyl group in the lactonic ring. The NMR data for 8 are listed in
Table 1.
The synthetic route for compounds 7 and 8 is shown in Fig. 1. The
starting material costunolide (1) was isolated from Saussurea lappa and
the structure was confirmed by the 1H NMR and 13C NMR spectra,
which corresponded with the published data (Park et al., 2010).
The first step involved the intramolecular oxidative cyclization of
the germacranolide ring of 1 by treatment with 3-chloroperbenzoic acid
(m-CPBA) in dichloromethane as solvent at room temperature. The
reaction was complete in two hours to generate epoxide 2 (Lu and
Fischer, 1996), which was found to be unstable. The opening of the
epoxide occurred on silica gel during the chromatographic separation
to provide three eudesmanolides in different yields: santamarine (3,
59%), reynosin (4, 31%) and magnolialide (5, 7%). The spectroscopic
data for compounds 3–5 were identical to those published for santa-
marine and reynosin (Ogura et al., 1978) and for magnolialide (El-
Feraly et al., 1979), respectively.
In summary, compounds 7 has been prepared in three easy steps
with an overall yield of 19%, and this represents an improvement on
the published synthesis of 7 (Yuuya et al., 1999), which requires nine
steps from α-santonin. The synthesis of 8 is also provided by the use of
Compound 3 was used to prepare the target products. The pre-
paration of 7 required the reduction of the methylene of the lactonic
ring and this step was carried out with NaBH4 in THF at room tem-
perature to give 6 in quantitative yield in one hour. In the 1H NMR
spectrum of 6 the signal for H-13 appears as a doublet at δ 1.22, rather
than two deshielded doublets, and a new multiplet signal for H-11 was
observed at δ 2.28. The spectroscopic data correspond with those
published in the literature for 11,13-dihydrosantamarine (Clark and
Hufford, 1979).
The final step was the isomerization and oxidation of 6 by treatment
with excess DDQ in refluxing CH2Cl2 for 24 h. This reaction gave 7 in
32% yield after chromatographic separation. The IR spectrum showed
the absence of the peak assigned to the hydroxy group and the presence
of a new peak at 1672 cm–1, which corresponds to a new carbonyl
group. The 1H NMR spectrum showed modified couplings and shifts for
H-2 (from δ 2.38 and 1.94 to δ 5.86, d), H-3 (from δ 5.33 to δ 7.04, dd)
and H-15 (from δ 1.81, s, to δ 5.82 and 5.54, both dd), and the lack of
the signal for H-1 at δ 3.64. The NMR data obtained for 7 (Table 1)
Fig. 2. 3D structure using PM6 calculation of 7 and NOE 1D effects of H-6
attended to confirm β-orientation of H-11.
230

J.G. Zorrilla, et al.
PhytochemistryLetters31(2019)229–236
Fig. 3. Profiles obtained for products in the etiolated wheat coleoptile bioassay. Positive values indicate stimulation of growth vs. the control, and negative values
indicate inhibition.
DDQ for the oxidation and isomerization to obtain an exodienone
system. Compound 8 has been synthesized for the first time and this
will facilitate the study of the structure-activity relationships and the
influence on the bioactivity of the presence of a methyl or methylene
group in the lactonic ring of the eudesmanolide, along with the pre-
sence of an exodienone system in the ring A.
The feature that proved to be effective in improving the activity was
the modification of ring A. In this way, compounds 7 and 8, which
contain the exodienone system, present lower IC50 values (by around
50%) than compounds 3 and 6.
Additionally, calculated clogP values (Table 2) are generally similar,
so the structural modifications made to the molecules do not have a
significant influence on their lipophilicity. A clear relationship between
clogP and IC50 was not observed, but it is worth noting that the com-
pound with lowest clogP, 7, was the product with the lowest IC50 value.
This fact reflects a possible improvement on the activity when struc-
tural features provide a decrease on the lipophilicity of the eu-
desmanolide, being easier the transport to the active site though cell
membranes. Deeper studies should be carried out to confirm a trend in
the activity of eudesmanolides connected with the lipophilicity of the
molecule.
The availability of the starting material and the synthetic yields
obtained allowed the test of bioactivities of all eudesmanolides pre-
pared (3–7) in three different bioassays, in an effort to evaluate the
relationship between activity and structure, which was the third ob-
jective of this study.
2.2. Wheat coleoptile bioassay
This bioassay is a useful tool as a first approach to evaluate the
phytotoxicity, as this is a quick and easy assay. Results are given as
percentage coleoptile elongation in relation to that in the negative
2.3. Phytotoxicity bioassay on weed species
®
control. The commercial herbicide Logran was used as a positive
All tested compounds were active and they were therefore subjected
to the phytotoxicity bioassay on weed species. Weed species selected
were E. crus-galli, A. viridis and P. maximum. These species were chosen
since the starting material for our synthesis, i.e., santamarine (3), as
well as reynosin (4), showed high phytotoxicity on these weeds in a
previous study of our research group (Da Silva et al., 2017).
control (Macías et al., 2000). The profiles obtained are shown in Fig. 3.
All compounds showed high inhibition activities at the two highest
concentrations and it is worth highlighting 3, 7 and 8, which have
activities similar to, or higher than, the commercial herbicide with in-
hibition values close to 90% or above, even in the second dilution.
Compounds 7 and 8 showed better retained activity with dilution,
especially 7 at the lowest concentrations – with inhibition values of 42
and 30% at 30 and 10 μM, respectively. At 100 μM, compound 8 was
the most active product, with an inhibition of 68%.
The activity profiles obtained are shown in Fig. 4, where negative
differences from control indicate inhibition of growth. All compounds
showed inhibitory effects at the highest concentrations, except for 3
and 6 on shoot growth of E. crus-galli. The activities of the final products
Regarding the IC50 values of the products (Table 2), compounds 7
and 8 got the best results, 67.6 and 71.9 μM respectively, and these
values are close to that of the commercial herbicide used as a positive
control in the bioassay.
®
7 and 8 were higher than that of Logran at 1000 μM on both shoots and
roots, with inhibition values over 75% in most cases. These activities
were retained at the second concentration and, in the case of A. viridis,
at the third concentration, with values above 55% inhibition in all
cases.
Compounds 3 and 6 (139.6 and 146.4 μM, respectively) showed
very similar values and this indicates the small influence of a methylene
group in the lactonic ring. A similar effect can be observed in the cases
of 7 and 8 (67.6 and 71.9 μM respectively). This similarity indicates the
small contribution that the saturation of the bond between C-11 and C-
13 has on the activity of the molecules in this bioassay.
The IC50 values of 7 and 8 (Table 3) are better than those of the
commercial herbicide in many cases and they are the only compounds
that have significant inhibition values for E. crus-galli shoots. Further-
more, higher IC50 values than the commercial herbicide were obtained
for 8 on E. crus-galli roots (205.4 vs. 314.3 μM); 6 on A. viridis shoots
(50.0 vs. 60.9 μM); 7 on P. maximum shoots (256.6 vs. 341.7 μM); and
for all compounds on P. maximum roots. The IC50 of the commercial
herbicide on P. maximum roots would correspond with a value for a
concentration higher than those tested in the bioassay, i.e., over 1000
μM.
Table 2
Lipophilicity of products, calculated as clogP, and IC50 values calculated with
results of the wheat coleoptile bioassay.
Compound
clogP
IC50 Wheat (μM)
R2
Comparison of the IC50 values of 3 vs. 6 and 7 vs. 8 highlights the
effects of some structural changes. The presence of a methyl group in
the lactonic ring in 6 improves activity in relation to a methylene group
in 3, especially on A. viridis (from 240.2–39.0 μM in root and
287.3–50.0 μM in shoot), whereas 7 and 8 show different trends de-
pending on the weed species. Better values were obtained on E. crus-
galli (from 343.5–205.4 μM in root and 631.7–587.6 μM in shoot) for
3
4
5
6
7
8
1.18
1.18
1.38
1.05
0.95
1.02
–
139.8
284.1
191.7
146.4
67.6
0.969
0.952
0.941
0.934
0.954
0.955
0.947
71.9
®
Logran
40.1
231

J.G. Zorrilla, et al.
PhytochemistryLetters31(2019)229–236
Fig. 4. Graphs for phytotoxicity weed bioassay. Positive values indicate stimulation of growth vs. the control, and negative values indicate inhibition. Significance
levels p < 0.01 (a), or 0.01 < p < 0.05 (b).
the molecule with a methylene group (8), whereas the presence of the
methyl group in 7 improves on the IC50 value of 8 on P. maximum (from
322.5–235.1 μM in root and 415.3–256.6 μM in shoot), with A. viridis
being only slightly sensitive to this difference.
have enhanced phytotoxicity when compared to 3 on A. viridis, with
®
IC50 values close to those of Logran , from 240.2 (roots) and 287.3 μM
(shoots) to 39.0 and 50.0 μM, respectively, in the best cases (6).
Regarding P. maximum, the highest inhibitory activity was shown by 7,
which has an IC50 of 256.6 μM on shoots and the second highest IC50 on
roots (235.1 μM) – i.e., only exceeded by that of 5 (199.7 μM).
Finally, cluster analysis was carried out in order to resolve which
The final products 7 and 8 are the most phytotoxic eudesmanolides
tested on E. crus-galli growth and they show significantly improved
activities when compared to santamarine (3). Compounds 6–8 also
232

J.G. Zorrilla, et al.
PhytochemistryLetters31(2019)229–236
Table 3
IC50 values of products for root and shoot growth for the three weed species tested.
Compound IC50 A. viridis
R2
IC50 A. viridis
R2
IC50 E. crus-galli R2
IC50 E. crus-galli
R2
IC50 P. maximum
R2
IC50 P. maximum
R2
root (μM)
shoot (μM)
root (μM)
shoot (μM)
root (μM)
shoot (μM)
3
240.2
442.3
165.7
39.0
0.955 287.3
0.968 539.9
0.865 205.0
0.942 50.0
0.980 77.9
0.878 72.1
0.946 60.9
0.955 476.9
0.992 462.0
0.872 475.9
0.992 439.7
0.984 343.5
0.932 205.4
0.988 314.3
0.906
0.970
0.914
0.987
a
a
a
a
–
–
–
–
271.7
739.1
199.7
271.3
0.956
a
–
4
0.977 930.3
0.973 411.5
0.934 869.2
0.934 256.6
0.932 415.3
0.920
0.992
0.947
0.962
0.911
0.984
5
6
7
63.2
0.999 631.7
0.959 587.6
0.994 235.1
0.974 322.5
8
75.6
®
Logran
18.9
0.910
a
–
a
–
341.7
a = IC50 value over 1000 μM or not calculable.
concentrations can be explained by the phytotoxicity that these pro-
ducts may have at these relatively high doses. No activity was detected
for 8.
O. cumana seeds were not affected by any of the compounds except
for 7, which showed 75% stimulation of germination at 100 μM, and 8,
which gave 25% stimulation at 100 μM. These results again highlight
the potential of these compounds.
The tested compounds showed good profiles on P. ramosa seeds. The
presence of the exodienone system (7 and 8) led to an improvement in
the activity and 7 and 8 are more active than 3 and 6, respectively.
Compound 8 is the most active product evaluated on P. ramosa seeds.
The most active compound at lower concentrations was the natural
product reynosin (4) and this was the most active compound at the
lowest dose (0.1 μM), with an excellent stimulation of 64% observed on
seeds.
In summary, regarding synthetic considerations, the strategy used to
obtain anhydrojudaicin (7) from costunolide (1) represents a marked
improvement on the previously published method that synthesizes this
product from α-santonin. Furthermore, the synthesis of 11,13-dehy-
droanhydrojudaicin (8) is reported for the first time in only 2 steps from
costunolide.
Fig. 5. Cluster analysis for weeds bioassay.
®
compounds are as effective as commercial herbicide Logran . For this
analysis was used root and shoot length activity data of all species,
showed graphically in Fig. 4, so results of this bioassay are summarized
in the cluster. The results of the cluster analysis are shown in Fig. 5,
where products are arranged attending to activity similarities. Most
active compounds are placed on the top of the figure, being 7 the most
and 5 the least active. It can be seen that compounds 7 and 8 are ranked
in the group that exceeds the activity of the commercial herbicide, with
Considering SAR, exodienone system of targeted products (7 and 8)
improved significantly the activity in all bioassays of starting mole-
cules, that only have one double bond in ring A. Thus, IC50 of 7 and 8
were among the best obtained in all cases, being the only compounds
active on O. cumana germination.
Attending to the presence of a methyl or methylene group at the
ring C, slight differences were observed on wheat coleoptile bioassay.
However, no clear relationship could be observed on the phytotoxicity
on weeds or parasitic seed germination, except on O. cumana where the
absence of a double bond between C-11 and C-13 (compound 7) pro-
vokes a relevant enhancement of activity.
®
Logran being more similar to compound 6 in terms of activity. This
analysis supports the results discussed on previous paragraphs, de-
noting the potential of both anhydrojudaicin (7) and 11,13-dehy-
droanhydrojudaicin (8) to be proposed as lead compounds for new
herbicides based on natural products.
Results detailed in this study let conclude that the target synthetic
products, 7 and 8, could be appropriate for use as lead compounds for
new herbicides to control weeds of A. viridis, E. crus-galli and P. max-
imum and also as components in preventive herbicides to control in-
festations of the parasitic plants O. cumana and P. ramosa on crops by
the suicidal germination strategy.
2.4. Parasitic plant bioassay
This bioassay is a good tool to study the ability of products to sti-
mulate the germination of seeds of parasitic plants so that the com-
pounds can be used in preventive herbicides through the suicidal ger-
mination strategy. This technique is employed to control parasitic weed
infestations on crops, and requires products that are able to stimulate
the germination of weed seeds. These compounds are applied on the
field prior to sowing, so the parasitic seeds in question would then start
their germination but would die since there is no host plant for the
parasite.
3. Experimental
3.1. General experimental procedures
All procedures and complete elucidation of compounds were carried
out using equipment available at the University of Cadiz. Agilent
spectrometers at 400/100 and 500/125 MHz were used to record NMR
spectra. CDCl3 (ProlaboTM, VWR) was the solvent for samples and the
residual peak was used as internal reference at δ 7.26 in 1H and δ 77.0
in 13C NMR. The exact masses of compounds were measured on a UPLC-
QTOF ESI (Waters Synapt G2, Manchester, UK) high-resolution spec-
trometer (HRTOFESIMS), with mass spectra recorded by the negative-
or positive-ion method in the m/z range 100–2000, mass resolution of
The results obtained were dependent on the species studied. In all
cases (Fig. 6), compound 7 gave better profiles than 3 (eudesmane
precursor) and 8 (methylene analog).
The results (Fig. 6) indicate that O. crenata and O. cumana are very
selective species in terms of the type of compound that causes germi-
nation of their seeds. In the case of O. crenata germination, compounds
3 and 7 are remarkable as they can stimulate more than 15% germi-
nation of the seeds. The lower values obtained at the highest
233

J.G. Zorrilla, et al.
PhytochemistryLetters31(2019)229–236
Fig. 6. Graphs for parasitic weed bioassay. Significant differences with control are marked with an asterisk.
234

J.G. Zorrilla, et al.
PhytochemistryLetters31(2019)229–236
20,000 and an acceleration voltage of 0.7 kV. A Perkin-Elmer Spectrum
TWO IR spectrophotometer was used to obtain FTIR spectra. Optical
rotations were measured on a JASCO P-2000 polarimeter, using CHCl3
as solvent.
3.2.5. Synthesis of 11,13-dehydroanhydrojudaicin (8)
A solution of 3 (35 mg, 0.14 mmol) in CH2Cl2 (6 mL) was added to
DDQ (252 mg, 1.09 mmol) in a two-necked round-bottomed flask
equipped with a Liebig condenser. The reaction mixture was heated
under reflux (45 °C) for 24 h with stirring. The mixture was allowed to
cool, dried with anhydrous Na2SO4, filtered through filter paper and
concentrated under reduced pressure. The crude product was purified
by column chromatography (hexane/acetone gradient 100:0–50:50) to
give 8 as a colorless oil in 43% yield. Calculated m/z for [C15H17O3]+
Saussurea lappa root extract was purchased from Pierre Chauvet S.A.
(Seillans, France). Reagents were supplied by Merck (Darmstadt,
Germany) or Sigma-Aldrich Co. (St. Louis, Missouri). Silica gel
®
Geduran Si 60 (0.063–0.200 mm) was used for column chromato-
graphy. HPLC was carried out on 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.
25
245.1178, obtained 245.1188; [α]D = +4.6° (c 0.116, CHCl3); IR
(film) ῦmax cm–1 1772 and 1672 (carbonyl groups). For NMR data see
Table 1.
3.3. Bioassays
3.2. Synthesis of anhydrojudaicin (7) and 11,13-dehydroanhydrojudaicin
(8)
3.3.1. Wheat coleoptile bioassay
The bioassay was carried out by optimized procedures established
by our research group (Macías et al., 2004). To obtain the coleoptiles,
wheat seeds (Triticum aestivum L. cv. Catervo, provided by FITÓ S.A.,
Compounds 7 and 8 were synthesized from compound 1 in three
and two steps, respectively.
®
Barcelona, Spain) were sown in deionized water-moistened Whatman
3.2.1. Isolation of costunolide (1)
paper in Petri dishes with a diameter of 14 cm. The samples were in-
troduced into an incubation chamber, in the dark to avoid photo-
synthesis, at 25 °C for 4 days. Under a green safelight the roots and
caryopses were removed from the shoots. 4 mm lengths of the shoots
were removed using a Van der Weij guillotine to remove the apical
2 mm, thus obtaining the coleoptiles. The products tested were pre-
dissolved in DMSO (0.5%) and diluted in buffer to the final con-
centrations required for the bioassay (1000, 300, 100, 30 and 10 μM).
The buffer was a phosphate-citrate buffer containing 2% sucrose and
adjusted to pH 5.6.
Column chromatography was used to purify compound 1, eluting
first with 0.7 L of hexane and then 6 L of hexane/EtOAc 19:1, from
53.2 g of a root extract of Saussurea lappa.
3.2.2. Synthesis of eudesmanolides (3–5)
Compound 1 (180 mg, 0.77 mmol) was dissolved in CH2Cl2 (2 mL)
in a round-bottomed flask and mCPBA (199 mg, 1.15 mmol) was added
under magnetic agitation. After 2 h at room temperature, all starting
material had been consumed. One product was observed by TLC
(hexane/EtOAc 7:3) and this was believed to be epoxide 2. The mixture
was neutralized with a 0.5 M aqueous solution of NaOH. The organic
phase was extracted three times with dichloromethane, dried with an-
hydrous Na2SO4, filtered through filter paper and concentrated under
reduced pressure. The crude product was loaded onto a chromato-
graphy column by absorption onto silica gel and purified using a gra-
dient of hexane/AcOEt 100:0–80:20 to give two fractions: a mixture of
three compounds and one pure compound (product 4). The mixture was
subsequently purified by HPLC (hexane/EtOAc 1:1) to give three
compounds with different yields, in order of elution: magnolialide (5,
7%), santamarine (3, 59%) and reynosin (4, 31%).
®
The commercial herbicide Logran was used as positive control.
®
Logran is a mixture of N2-tert-butyl-N4-ethyl-6-(methylsulfanyl)-1,3,5-
triazine-2,4-diamine (terbutryn, 59.4%) and 2-(2-chloroethoxy)-N-[(4-
methoxy-6-methyl-1,3,5-triazin-2-yl)carbamoyl]benzene-1-sulfona-
mide (triasulfuron, 0.6%). Buffered aqueous solutions without any
product were used as negative controls and these contained the same
percentage of DMSO as the samples.
The bioassays were carried out in test tubes. Each concentration had
three replicates and one test tube was used per replicate. In each tube
five coleoptiles were introduced together with 2 mL of solution. The
tubes were placed horizontally in a roller tube apparatus at 0.25 rpm in
the dark for 24 h at 25 °C. Finally, coleoptiles were digitally photo-
graphed on a template to measure their elongation. Data were analyzed
by Welch’s test and results are presented as percentage elongation
against the negative control.
3.2.3. Synthesis of dihydrosantamarine (6)
NaBH4 (14 mg, 0.37 mmol) was added to a stirred solution of
compound 3 (53 mg, 0.21 mmol) in THF (5 mL) in a round-bottomed
flask. According to TLC (hexane/acetone 70:30) the reaction had fin-
ished after 1 h. The mixture was solved in ethyl acetate and poured into
water. The aqueous phase was extracted three times with ethyl acetate.
The organic layer was dried with anhydrous Na2SO4, filtered through
filter paper and concentrated under reduced pressure. The crude pro-
duct was purified by column chromatography (hexane/acetone gra-
dient 100:0–50:50) to give compound 6 with quantitative yield as a
white solid.
3.3.2. Phytotoxicity bioassay on weeds
®
Seeds were placed on moistened Whatman paper of diameter
50 mm in Petri dishes with 1.0 mL of buffer solution (2-(N-morpholine)
ethanesulfonic acid (MES) 10 mM, pH adjusted to 6.0) containing
products at concentrations of 1000, 300, 100, 30 or 10 μM. Twenty
seeds were placed in each Petri dish, with five replicates per con-
centration tested, with seeds of Amaranthus viridis L., Echinochloa crus-
galli L. and Panicum maximum Jacq. Buffer solution without any tested
®
3.2.4. Synthesis of anhydrojudaicin (7)
compound was used as negative control and the herbicide Logran as
The reaction was carried out in a two-necked round-bottom flask. A
solution of 6 (59 mg, 0.24 mmol) in CH2Cl2 (6 mL) was added to DDQ
(240 mg, 1.04 mmol). The reaction mixture was heated under reflux
(45 °C) for 24 h under magnetic agitation. The mixture was allowed to
cool to room temperature, dried with anhydrous Na2SO4, filtered
through filter paper and concentrated under reduced pressure. The
crude product was purified by column chromatography (hexane/
acetone gradient 100:0–60:40) to give 7 as a colorless oil in 32% yield.
positive control, as in the coleoptile bioassay.
Dishes were kept in darkness at 25 °C in a germination chamber: E.
crus-galli and P. maximum (11 days), A. viridis (13 days). After these
periods, the dishes were placed in a freezer for 24 h.
Measurement of the shoots and roots was performed after placing
®
seedlings on a plastic film, with the lengths measured with a Fitomed
digitizing table using a light pen. Statistical analysis was performed by
Welch’s test, with significance levels of 0.01 and 0.05 established.
Calculated m/z for [C15H19O3]+ 247.1334, obtained 247.1340; [α]D
25
= +16.3° (c 0.044, CHCl3); IR (film) ῦmax cm–1 1783 and 1672 (car-
3.3.3. Parasitic seeds bioassay
bonyl groups). For NMR data see Table 1.
The following bioassay was developed by our research group (Cala
235

J.G. Zorrilla, et al.
PhytochemistryLetters31(2019)229–236
et al., 2017; Rial et al., 2018) using seeds of parasitic species of O.
cumana (provided by Dr. Leonardo Velasco, Instituto de Agricultura
Sostenible (CSIC), University of Cordoba, Spain), O. crenata and P. ra-
mosa (both provided by Prof. Maurizio Vurro, Istituto di Scienze delle
Produzioni Alimentari (ISPA-CNR), University of Bari, Italia). Seeds
were placed on water-moistened glass fiber filter papers of diameter
10 mm in Petri dishes and samples were stored in darkness for 10 days.
The products were dissolved in acetone (1% v/v) and then diluted in
type I water to obtain solutions of 100, 10, 1 and 0.1 μM. For each
concentration, four samples were replicated, with 80 μL of solution
added per sample. A solution of water/acetone (99:1 v/v), without any
product, was employed as negative control. Synthetic strigolactone
GR24 (provided by Prof. Binne Zwanenburg, Radboud University, Nij-
megen, Netherlands) was tested in the same was as a product to provide
a positive control. The results were obtained after samples had been
stored for 7 days in darkness.
Clark, A.M., Hufford, C.D., 1979. Microbial transformations of the sesquiterpene lactone
costunolide. Sect. Title Microb. Biochem. 3022–3028. https://doi.org/10.1039/
p19790003022.
Da Silva, B.P., Nepomuceno, M.P., Varela, R.M., Torres, A., Molinillo, J.M.G., Alves,
P.L.C.A., Macias, F.A., 2017. Phytotoxicity study on Bidens sulphurea sch. Bip. as a
preliminary approach for weed control. J. Agric. Food Chem. 65, 5161–5172. https://
doi.org/10.1021/acs.jafc.7b01922.
El-Feraly, F.S., Chan, Y.M., Benigni, D.A., 1979. Magnolialide: a novel eudesmanolide
from the root bark of Magnolia grandiflora. Phytochemistry 18, 881–882.
Hui, Y., Cao, J., Lin, J., Yang, J.-N., Liu, Y.-J., Han, C.-R., Song, X.-P., Dai, C.-Y., Zhang,
X.-P., Chen, W.-H., Chen, G.-Y., 2018. Eudesmanolides and other constituents from
the flowers of Wedelia trilobata. Chem. Biodivers. 15, e1700411. https://doi.org/10.
1002/cbdv.201700411.
Kaspary, T.E., Bellé, C., Groth, M.Z., Cocco, K.L.T., Cutti, L., Casarotto, G., 2017.
Amaranthus viridis is a weed host of Meloidogyne arenariain Rio Grande do Sul State,
Brazil. Plant Dis. 101, 635. https://doi.org/10.1094/PDIS-10-16-1428-PDN.
Le Corre, V., Reibel, C., Gibot-Leclerc, S., 2014. Development of microsatellite markers in
the branched broomrape Phelipanche ramosa L. (Pomel) and evidence for host-asso-
ciated genetic divergence. Int. J. Mol. Sci. 15, 994–1002. https://doi.org/10.3390/
ijms15010994.
Lu, T., Fischer, N.H., 1996. Spectral data of chemical modification products of costuno-
lide. Spectrosc. Lett. An Int. J. Rapid Commun. 29, 437–448. https://doi.org/10.
1080/00387019608006662.
The percentages of germinated seeds were measured using a bino-
cular camera. Values were approximated to a normal frequency dis-
tribution by means of an angular transformation, and subjected to an
analysis of variance (ANOVA) with SPSS software (SPSS Inc., Chicago,
IL). Two-sided Dunnett’s tests were used to evaluate significant mean
differences between values of samples and negative-control values.
Rejection of null hypothesis was defined at level of 0.05.
Macías, F.A., Castellano, D., Molinillo, J.M.G., 2000. Search for a standard phytotoxic
bioassay for allelochemicals. Selection of standard target species. J. Agric. Food
Chem. 48, 2512–2521. https://doi.org/10.1021/jf9903051.
Macías, F.A., López, A., Varela, R.M., Torres, A., Molinillo, J.M.G., 2004. Bioactive lig-
nans from a cultivar of Helianthus annuus. J. Agric. Food Chem. 52, 6443–6447.
https://doi.org/10.1021/jf048945d.
Mansilla, H., Palenzuela, J.A., 1999. Minor eudesmanolides from Artemisia canariensis.
Phytochemistry 51, 995–997. https://doi.org/10.1016/S0031-9422(99)00013-8.
Mejías, F.J.R., López-Haro, M., Gontard, L.C., Cala, A., Fernández-Aparicio, M., Molinillo,
J.M.G., Calvino, J.J., Macías, F.A., 2018. A novel electron microscopic character-
ization of core/shell nanobiostimulator against parasitic plants. ACS Appl. Mater.
Interfaces 10, 2354–2359. https://doi.org/10.1021/acsami.7b16873.
Ogura, M., Cordell, G.A., Farnsworth, N.R., 1978. Anticancer sesquiterpene lactones of
Michelia compressa (magnoliaceae). Phytochemistry 17, 957–961. https://doi.org/10.
1016/S0031-9422(00)88656-2.
3.4. Calculation of IC50 and logP
The IC50 values for products were calculated by fitting activity data
to a sigmoidal dose-response model using GraphPad Prism v.5.00
software. ChemBioDraw Ultra 17.0 software was used for the calcula-
tion of clogP.
Padilla-Gonzalez, G.F., dos Santos, F.A., Da Costa, F.B., 2016. Sesquiterpene Lactones:
More Than Protective Plant Compounds with High Toxicity. CRC. Crit. Rev. Plant
Sci. 35, 18–37. https://doi.org/10.1080/07352689.2016.1145956.
Park, H.W., Lee, J.H., Choi, S.U., Baek, N.I., Kim, S.H., Yang, J.H., Kim, D.K., 2010.
Cytotoxic germacranolide sesquiterpenes from the bark of Magnolia kobus. Arch.
Pharm. Res. 33, 71–74. https://doi.org/10.1007/s12272-010-2227-5.
Rial, C., Gómez, E., Varela, R.M., Molinillo, J.M.G., Macías, F.A., 2018. Ecological re-
levance of the major allelochemicals in Lycopersicon esculentum roots and exudates. J.
Agric. Food Chem. 66, 4638–4644. https://doi.org/10.1021/acs.jafc.8b01501.
Rubiales, D., Fernández Aparicio, M., 2012. Innovations in parasitic weeds management
in legume crops. A review. Agron. Sustain. Dev. 32, 433–439. https://doi.org/10.
1007/s13593-011-0045-x.
Acknowledgment
This research was supported by the ‘Ministerio de Economía,
Industria y Competitividad’ (MINEICO), Spain, Project AGL2017-88-
083-R.
We would like to thank FITÓ S.A. (Barcelona, Spain) for supplying
the wheat seeds.
Saber, A.H., Abu-Shady, H., El-Antably, M.S., 1964. The chemistry of judaicin A. Bull.
Fac. Pharm. (Cairo Univ.) 3, 119–129.
Appendix A. Supplementary data
Talbert, R.E., Burgos, N.R., 2007. History and Management of Herbicide-resistant
Barnyardgrass (Echinochloa Crus-galli) in Arkansas Rice. Weed Technol. 21, 324–331.
https://doi.org/10.1614/WT-06-084.1.
Supplementary material related to this article can be found, in the
online version, at doi:https://doi.org/10.1016/j.phytol.2019.04.014.
Talbi, M., Ainane, T., Boriky, D., Bennani, L., Blaghen, M., Elkouali, M., 2015.
Antibacterial activity of eudesmanolide compounds isolated from medicinal plant
Artemisia herba-alba. J. Mater. Environ. Sci. 6, 2125–2128.
References
Yang, C., Xu, L., Zhang, N., Islam, F., Song, W., Hu, L., Liu, D., Xie, X., Zhou, W., 2017.
iTRAQ-based proteomics of sunflower cultivars differing in resistance to parasitic
weed Orobanche cumana. Proteomics 17, 1–51. https://doi.org/10.1002/pmic.
201700009.
Ando, M., Takase, K., 1977. Studies in the Syntheses of Sesquiterpene Lactones-II.
Tetrahedron 33, 2785–2789.
Yuuya, S., Hagiwara, H., Suzuki, T., Ando, M., Yamada, A., Suda, K., Kataoka, T., Nagai,
K., 1999. Guaianolides as immunomodulators. Synthesis and biological activities of
dehydrocostus lactone, mokko lactone, eremanthin, and their derivatives. J. Nat.
Prod. 62, 22–30. https://doi.org/10.1021/np980092u.
Cala, A., Molinillo, J.M.G., Fernández-Aparicio, M., Ayuso, J., Álvarez, J.A., Rubiales, D.,
Macías, F.A., 2017. Complexation of sesquiterpene lactones with cyclodextrins:
synthesis and effects on their activities on parasitic weeds. Org. Biomol. Chem. 15,
6500–6510. https://doi.org/10.1039/c7ob01394a.
236