Optimization of benzoxazinones as natural herbicide models by lipophilicity enhancement

Article abstract of DOI:10.1021/jf062168v

Benzoxazinones are plant allelochemicals well-known for their phytotoxic activity and for taking part in the defense strategies of Gramineae, Ranunculaceae, and Scrophulariceae plants. These properties, in addition to the recently optimized methodologies for their large-scale isolation and synthesis, have made some derivatives of natural products, 2,4-dihydroxy-(2H)-1,4- benzoxazin-3-(4H)-one (DIBOA) and its 7-methoxy analogue (DIMBOA), successful templates in the search for natural herbicide models. These new chemicals should be part of integrated methodologies for weed control. In ongoing research about the structure-activity relationships of benzoxazinones and the structural requirements for their phytotoxicity enhancement and after characterization of the optimal structural features, a new generation of chemicals with enhanced lipophilicity was developed. They were tested on selected standard target species and weeds in the search for the optimal aqueous solubility-lipophilicity rate for phytotoxicity. This physical parameter is known to be crucial in modern drug and agrochemical design strategies. The new compounds obtained in this way had interesting phytotoxicity profiles, empowering the phytotoxic effect of the starting benzoxazinone template in some cases. Quantitative structure-activity relationships were obtained by bioactivity-molecular parameters correlations. Because optimal lipophilicity values for phytotoxicity vary with the tested plant, these new derivatives constitute a more selective way to take advantage of benzoxazinone phytotoxic capabilities.

Full text of DOI:10.1021/jf062168v

J. Agric. Food Chem. 2006, 54, 9357−9365 9357
Optimization of Benzoxazinones as Natural Herbicide Models
by Lipophilicity Enhancement
FRANCISCO A. MAC ´ı AS,* DAVID MAR ´ı N, ALBERTO OLIVEROS-BASTIDAS,† AND
JOS EÄ M. G. MOLINILLO
Grupo de Alelopat ´ı a, Departamento de Qu ´ı mica Org a´ nica, Universidad de C a´ diz, C/Rep u´ blica
Saharaui s/n, 11510 Puerto Real, C a´ diz, Spain
Benzoxazinones are plant allelochemicals well-known for their phytotoxic activity and for taking part
in the defense strategies of Gramineae, Ranunculaceae, and Scrophulariceae plants. These properties,
in addition to the recently optimized methodologies for their large-scale isolation and synthesis, have
made some derivatives of natural products, 2,4-dihydroxy-(2H)-1,4-benzoxazin-3-(4H)-one (DIBOA)
and its 7-methoxy analogue (DIMBOA), successful templates in the search for natural herbicide
models. These new chemicals should be part of integrated methodologies for weed control. In ongoing
research about the structure-activity relationships of benzoxazinones and the structural requirements
for their phytotoxicity enhancement and after characterization of the optimal structural features, a
new generation of chemicals with enhanced lipophilicity was developed. They were tested on selected
standard target species and weeds in the search for the optimal aqueous solubility-lipophilicity rate
for phytotoxicity. This physical parameter is known to be crucial in modern drug and agrochemical
design strategies. The new compounds obtained in this way had interesting phytotoxicity profiles,
empowering the phytotoxic effect of the starting benzoxazinone template in some cases. Quantitative
structure-activity relationships were obtained by bioactivity-molecular parameters correlations.
Because optimal lipophilicity values for phytotoxicity vary with the tested plant, these new derivatives
constitute a more selective way to take advantage of benzoxazinone phytotoxic capabilities.
KEYWORDS: Allelochemicals; DIBOA; DIMBOA; D-DIBOA; lipophilicity; herbicides; SAR
INTRODUCTION
cinmethylin (a derivative from natural monoterpene 1,4-cineole,
which is widely distributed in plants) (8) and the herbicide
family of triketones, which derive from leptospermone, a natural
product from the plant genus Callistemon (9), exemplify the
potential utility of plant allelochemicals for weed management.
The intensive use of synthetic herbicides during the past 50
years has provoked different problems mainly connected to
environmental impact (1, 2) and weed resistance to herbicides
(3, 4). In the context of an increasing demand for high-quality
Hydroxamic acids with a (2H)-1,4-benzoxazin-3(4H)-one
skeleton have attracted the attention of phytochemistry and
allelopathic researchers since the discovery of 2,4-dihydroxy-
feeding products and environmental concerns, it is urgent to
change the philosophy applied to the research and development
of strategies useful in crop protection. Chemical control of crop
pests and diseases is one of the areas in which new methodolo-
gies mainly focused on the understanding and application of
ecological phenomena should contribute to solve the problems
created in the past (5).
The employment of allelochemicals (natural plant toxins) as
templates for the discovery of new phytotoxic agents useful in
agriculture has developed into a convenient approach to
chemical low-dose treatments with less environmental impact,
higher selectivity, and alternative modes of action, which could
decisively contribute to solve the problems associated with
chemical weed control (6, 7). Commercial herbicides such as
(2H)-1,4-benzoxazin-3(4H)-one (DIBOA) (10) and 2,4-dihy-
droxy-7-methoxy-(2H)-1,4-benzoxazin-3(4H)-one (DIMBOA)
(11) (Figure 1). They were first isolated from rye and maize,
respectively, being preserved as â-D-glucosides prior to their
release (12-14). Soon after their discovery, they were found
in a wide variety of plants mainly belonging to Gramineae,
Ranunculaceae, and Scrophulariceae families (15). Their anti-
fungal (16), antimicrobial (17), and phytotoxic effects (18-
22), in addition to their interesting ecological role (23), raised
interest in their isolation and synthesis in the search for an
accurate explanation of the defense mechanisms of benzoxazi-
none producer plants. Moreover, the discovery of microbial
degradation phenomena (24, 25), which led to the detection of
new chemicals with different structures, effects, ecological roles,
and potential utilities, increased the interest in the development
of analytical methods (26) and the design of experiments in
*
Author to whom correspondence should be addressed (telephone +34-
9
56.016.370; fax +34-956.016.193; e-mail famacias@uca.es).
†
Permanent address: Grupo de Qu ´ı mica Ecol o´ gica, Departamento de
Qu ´ı mica, Universidad de Los Andes-U.L.A. N u´ cleo Universitario Pedro
Rinc o´ n Guti e´ rrez, La Hechicera, M e´ rida 5101-A, Venezuela.
1
0.1021/jf062168v CCC: $33.50
© 2006 American Chemical Society
Published on Web 11/14/2006

9358 J. Agric. Food Chem., Vol. 54, No. 25, 2006
Mac ´ı as et al.
Figure 1. Molecular structures of benzoxazinones DIBOA, DIMBOA,
D-DIBOA, and 6-MeO-D-DIBOA.
Table 1. Molecular Parameters Found To Be Relevant in
Pharmaceutical and Herbicidal Activities, with their Optimal Values
optimal value
Lipinski
Tice
parameter
(pharmaceuticals)
(postemergence herbicides)
H-bond donors
H-bond acceptors
molecular weight
cLog P
e
e
g
e
5
e
g
g
e
e
3
10
500
5
2 and
150 and
4
12
e
12
e
500
rotatable bonds
which all of the chemicals present in the plant-plant interaction
involving benzoxazinones could be present (23). Finally, the
availability of all those instruments makes benzoxazinones some
of the most attractive groups of natural products to be employed
in the design of new chemicals with potential use in weed
management (27).
In the context of the new methodologies for bioactive
compounds’ design, quantitative structure-activity relationships
studies (QSAR) constitute the key for a systematic analysis of
structure and bioactivity properties. An adequate correlation of
them to search for the most significant structural requirements
is needed to achieve the desired physical, chemical, and
biological properties (26) with the aid of combinatorial chemistry
techniques (29).
QSAR methodology has been extensively employed for drug
discovery, and its applications on new agrochemicals’ design
are starting to rise, as the potential applicability of the physical-
chemical parameters employed in QSAR design of pharmaceu-
ticals has been discussed (30). According to those studies, four
parameters should be taken into consideration to generate leads
for herbicide development: molecular mass, number of hydrogen-
bond donors, number of hydrogen-bond acceptors, and log P
Figure 2. Structures and acronyms of the tested benzoxazinone esters.
important for pre-emergence herbicides (33). Systemic trans-
portation would be favored by a higher aqueous solubility, but
some lipophilicity will be necessary to cross cell membranes
and to reach the target site of action.
In this work, the algorithm included in OSIRIS Property
Explorer (34) for cLog P calculation (35) was found to be the
most adequate one, as it fits the previously reported values for
the empirical rules mentioned above.
Herein we report the correlation of phytotoxic effects of two
series of benzoxazinone ester derivatives with increasing
lipophilicity in the search for the optimal value for log P. The
lipophilicity was increased by preparing esters with increasing
side chain length. This was found to be the most chemically
convenient way, taking advantage of the hydroxyl group at N-4
present in the hydroxamic acids with a (2H)-1,4-benzoxazin-
3(4H)-one skeleton selected for this study (Figure 1). In our
previous papers on structure-activity relationship studies on
benzoxazinones and related compounds (20-22), several req-
uisites for phytotoxicity enhancement were displayed: the lack
of functional groups at the aromatic ring, the lack of a hydroxyl
group at C-2, and the presence of a hydroxamic acid moiety.
4-Hydroxy-(2H)-1,4-benzoxazin-3(4H)-one (D-DIBOA, Figure
1) was proposed as the optimal template for herbicide model
development. Its derivative, 4-hydroxy-6-methoxy-(2H)-1,4-
benzoxazin-3(4H)-one (6-MeO-D-DIBOA, Figure 1), was
included in the study to compare the effects on transport
modulation in two structurally related chemicals, in order to
assume or discard intrinsic phytotoxic effects from the ester
side chains. Thus, 14 ester derivatives of D-DIBOA (Figure 2)
were prepared and tested on the same species as in our previous
studies. Their correlation to Tice’s modification to Lipinski’s
rule of 5 (31), in the context of Hansch’s transport theory (37),
is discussed. These benzoxazinones constitute a second genera-
tion of chemicals from previously reported SAR studies of
benzoxazinones and related compounds (20-22).
(logarithm of the octanol/water partition coefficient), which
constitutes a measurement of the aqueous solubility-lipophi-
licity ratio for a given chemical. On the basis of an empirical
study on more than 2000 pharmaceuticals, Lipinski et al.
proposed the optimal values for those parameters. Log P,
previously employed to analyze the assimilation capacity of oral
usage drugs (30), could also help to determine the bioavailability
of a given organic chemical for a plant, as was stated by Tice
(31), after the fitting of those parameters to the herbicidal and
insecticidal activity requirements, and the addition of a new
parameter (number of rotatable bonds). The optimal values of
those parameters are shown in Table 1. These studies have been
greatly influential as they provide useful clues for drug and
agrochemical development on the basis of physicochemical
parameters, which are easy to calculate.
The barriers that a crop protection agent must cross to reach
its molecular target site may vary: Postemergence herbicides
must cross the leaf cuticle to reach the whole plant, having more
potent effects when being phloem mobile (32). On the other
hand, root and shoot uptake and xylem mobility are more
MATERIALS AND METHODS
General Experimental Procedures. The purity of the compounds
1
to test was determined by H NMR and HPLC analyses and found to

Lipophilic Benzoxazinones
J. Agric. Food Chem., Vol. 54, No. 25, 2006 9359
be >98%. H and 13C NMR spectra were recorded using CDCl
1
as
After the addition of seeds and aqueous solutions, Petri dishes were
sealed with Parafilm to ensure closed-system models. Seeds were further
incubated at 25 °C in a Memmert ICE 700 controlled-environment
growth chamber, in the absence of light. Bioassays took 4 days for
cress, 5 days for lettuce, tomato, rigid ryegrass, wild oat, and wheat,
and 7 days for onion. After growth, plants were frozen at -10 °C for
24 h to avoid subsequent growth during the measurement process. This
helped in the handling of the plants and allowed a more accurate
measurement of root and shoot lengths.
3
solvent in a Varian INOVA spectrometer at 399.99 and 100.577 MHz,
respectively. The resonance of residual chloroform was set to δ 7.25.
1
3
The solvent peak for C was set to δ 77.00 (chloroform) and used as
internal reference. UV-vis spectra were acquired by means of a Varian
Cary 50 BIO spectrophotometer, chloroform being used as solvent.
Mass spectra (EIMS) were recorded by means of a Voyager Termoquest
equipment. FTIR spectra were obtained by means of a Spectrum BX
Perkin-Elmer FTIR system. Frequency values are given in cm-1
.
The commercial herbicide Logran, a combination of N-(1,1-
dimethylethyl)-N′-ethyl-6-(methylthio)-1,3,5-triazine-2,4-diamine (Ter-
butryn, 59.4%) and 2-(2-chloroethoxy)-N-[[(4-methoxy-6-methyl-1,3,5-
triazin-2-yl)amino]carbonyl]benzenesulfonamide (Triasulfuron, 0.6%),
was used as internal reference, according to a comparison study
Starting Material. 4-Hydroxy-(2H)-1,4-benzoxazin-3(4H)-one (D-
DIBOA, Figure 1) was synthetically obtained as previously reported
(25). Acyl chlorides were purchased from Sigma-Aldrich Co. and used
as received, except undecanoyl and tridecanoyl chlorides; these were
synthesized as described below. 6-Methoxy-D-DIBOA was synthesized
by using 4-methoxy-2-nitrophenol (purchased from Acros Organics,
used as received) as starting material and following the same procedures
as for D-DIBOA synthesis.
-
3
previously reported (38). It was used at the same concentrations (10
,
-
4
-4
-5
-5
5
× 10 , 10 , 5 × 10 , and 10 M) and in the same conditions as
the compounds in study. Negative control samples (buffered aqueous
solutions with DMSO and without any tested compound) were used
for all of the plant species assayed.
Undecanoyl and Tridecanoyl Chloride Obtention. These acyl
chlorides were obtained by means of the typical thionyl chloride
procedure. Five hundred milligrams of undecanoic or tridecanoic acid
Bioassay Data Acquisition. Evaluated parameters (germination rate,
root length, and shoot length) were recorded by using a Fitomed system
(purchased from Acros Organics Inc. and Sigma-Aldrich Co., respec-
(39), which allowed automatic data acquisition and statistical analysis
tively, used as received) were dissolved in thionyl chloride and refluxed
for 4 h. After this, the excess thionyl chloride was distilled at reduced
pressure, and the oily residue, containing the corresponding pure acyl
chloride, was dissolved in dry THF (0.1 N assuming complete
conversion) prior to their use.
by its associated software.
Statistical Analysis. Data were statistically analyzed using Welch’s
test, with significance fixed at 0.01 and 0.05. Results are expressed in
bar charts in which the null value represents control, negative values
represent inhibition, and positive values represent stimulation of the
studied parameter (39). Statistical significance is expressed by means
of letters, “a” meaning significantly different from control with 0.01
confidence and “b” meaning significantly different from control with
a confidence from 0.01 to 0.05. The absence of a letter indicates no
significant difference from control values. Phytotoxic activities ex-
pressed in this way can be found in the Supporting Information for all
chemicals and species.
Once the germination and growth data had been acquired, cluster
analysis was used to group compounds with similar phytotoxicity
behaviors and associate them with their molecular structure. Complete
linkage was used as an amalgamation rule, and the distance measure-
ment was based on squared Euclidean distances (41), given by the
equation
General Esterification Method. One hundred milligrams of D-
DIBOA or 6-methoxy-D-DIBOA was dissolved in 25 mL of dry
pyridine, and 1.2 mol equiv of the suitable acyl chloride (commercial
or readily synthesized and dissolved in THF for undecanoyl and
tridecanoyl) was added dropwise at 0 °C (ice bath) and under a dry
argon atmosphere. After the addition, the flask was allowed to warm
to room temperature; 12 h later, 25 mL of EtOAc was added, and the
mixture was transferred to an extraction funnel and washed with 50
4
mL of 0.1 N HCl (3 times) and 50 mL of 0.1 N NaHSO (three times).
The organic phases were further combined and dried over anhydrous
magnesium sulfate, and the solvent was removed (rotatory evaporator).
2
The dry residue was chromatographed (CC, SiO , EtOAc/hexane 1:10)
to obtain the benzoxazinone ester derivatives.
Calculation of IC50 and Log P. The phytotoxicity data were fitted
to a sigmoidal dose-response model (constant slope) by employing
the GraphPad Prism v. 4.00 software package (34). cLog P data were
acquired by using the OSIRIS property explorer (35). This software
uses the Chou and Jurs algorithm, based on computed atom contribu-
tions (36).
2
d(x,y) )
(x - y )
i i
∑i
where d(x,y) is the squared Euclidean distance (i-dimensional), i
represents the number of variables, and x and y are the observed values.
The cluster was obtained by using Statistica v. 5.0 software (41).
Germination rate, shoot length, and root length effects, for all tested
species, were included in the analysis to acquire an overall view of the
phytotoxicity and its relationship with chemical structure.
EC50 values were obtained after the phytotoxicity data had been
adjusted to concentration (logarithmic scale), to a constant slope
sigmoidal dose-response curve, defined by the equation
Bioassays. Target Plants. Selection of target plants is based on an
optimization process made by us in the search for a standard
phytotoxicity evaluation bioassay (38). After this process several
standard target species (STS) were proposed, including monocots
Triticum aestiVum L. (wheat) and Allium cepa L. (onion) and dicots
Lycopersicon esculentum Will. (tomato), Lepidium satiVum L. (cress),
and Lactuca satiVa L. (lettuce), which were assayed for this study. The
weeds Lolium rigidum Gaud. (rigid ryegrass) and AVena fatua L. (wild
oat) were also tested by employing the same methodology.
Ymax - Ymin
Y ) Ymin
+
logEC50-X
Methodology. Bioassays used Petri dishes (90 mm diameter) with
one sheet of Whatman no. 1 filter paper as substrate. Germination and
growth were conducted in aqueous solutions at controlled pH by using
1 + 10
where X indicates the logarithm of concentration, Y indicates the
response (phytotoxicity), and Ymax and Ymin are the maximum and
minimum values of the response, respectively. Goodness of fit is
described by the determination coefficient (r ). The adjustment and the
r were obtained by using GraphPad Prism software v. 4.00 (39).
-
2
1
0
M 2-[N-morpholino]ethanesulfonic acid (MES) and 1 M NaOH
pH 6.0). Compounds to be assayed were dissolved in DMSO (0.2,
.1, 0.02, 0.01, and 0.002 M), and these solutions were diluted with
buffer (5 µL of DMSO solution/mL of buffer) so that test concentrations
(
0
2
2
for each compound (10-3, 5 × 10 , 10 , 5 × 10 , and 10 M)
were reached. This procedure facilitated the solubility of the assayed
compounds. The number of seeds in each Petri dish depended on the
seed size; 25 seeds were used for tomato, lettuce, cress, and onion, 15
seeds were used for rigid ryegrass, and 10 seeds were used for wheat
and wild oat. Five milliliters of treatment, control, or internal reference
solution was added to each Petri dish. Four replicates were used for
tomato, cress, onion, lettuce, and rigid ryegrass; 10 replicates were used
for wheat and wild oat.
-4
-4
-5
-5
RESULTS AND DISCUSSION
Synthesis of D-DIBOA, 6-Methoxy-D-DIBOA, and Their
Esters. D-DIBOA was obtained according to the previously
reported methodology. 6-Methoxy-D-DIBOA was obtained
following the same methodology with 4-methoxy-2-nitrophenol
as starting material, with a 65% yield after two reaction steps.
Quantitative yield was achieved in all esterification reactions.

9360 J. Agric. Food Chem., Vol. 54, No. 25, 2006
Mac ´ı as et al.
strongly inhibited root and shoot length. Low values of IC50
were reached, especially on shoot length inhibition [IC50 ) 82.4
and 87.0 µM for valeryl and heptanoyl derivatives, respectively,
Table 2. Molecular Parameter Values Found for D-DIBOA and Its
Estersa
molecular
weight
rotatable
bonds
H-bond
donors
H-bond
acceptors
2
with excellent fitting to the dose-response model (R > 0.99
cLog P
in both cases)]. Valeryl D-DIBOA also provoked a significant
D-DIBOA
ABOA
165.15
207.19
221.21
235.23
249.25
263.27
277.29
291.31
305.33
319.35
333.37
347.39
361.41
375.43
0.42
1.18
1.65
2.11
2.58
3.04
3.5
3.97
4.43
4.89
5.36
5.82
6.29
6.75
0
2
3
4
5
6
7
8
9
10
11
12
13
14
1b
0
0
0
0
0
0
0
0
0
0
0
0
0
4
5
5
5
5
5
5
5
5
5
5
5
5
5
2
effect on germination rate (IC50 ) 580 µM, R ) 0.9374),
maintaining significant activity values at all concentrations when
D-DIBOA was found to be active only at the most concentrated
treatment.
Pr-D-DIBOA
Bu-D-DIBOA
Val-D-DIBOA
Hex-D-DIBOA
Hept-D-DIBOA
Oct-D-DIBOA
Non-D-DIBOA
Dec-D-DIBOA
Undec-D-DIBOA
Lau-D-DIBOA
Tridec-D-DIBOA
Mir-D-DIBOA
With regard to the phytotoxicity on weeds, D-DIBOA and
ABOA were the most active chemicals on L. rigidum if joint
effects on roots and shoots were considered. Nevertheless,
heptanoyl and octanoyl derivatives were also active on roots,
with IC50 values around 100 µM (123.4 and 107.4 µM,
2
respectively, with R > 0.95 in both cases). No significant effects
were observed for germination rate.
Among the different weed species tested to date, AVena fatua
(wild oat) has been the one most affected by benzoxazinone
treatments, according also to our previous research (21, 23).
Wild oat is the only plant of all that were studied that showed
a significant effect on its germination rate, this effect being
closely related to lipophilicity (Figure 3). The effects drastically
increase from Pr-D-DIBOA (cLog P ) 1.65) to Hept-D-DIBOA
a
Values outside the optimal ranges for activity are shown in boldface type.
Number of donors was found to be pH-dependent, value at pH 6.0 being displayed.
b
Table 3. Comparison of Phytotoxic Effects of D-DIBOA and Its Estersa
germination rate
root length
shoot length
L. sativum
A. cepa
L. esculentum
L. sativa
T. aestivum
L. rigidum
A. fatua
+
0
−
+
−
0
+
−
0
(
cLog P ) 3.50). A loss of the effect is observed from the
octanoyl derivative to the end of the series, this behavior being
in good accordance with the empiric rules of Tice (29). Optimal
values of cLog P are in the propanoyl-valeryl interval (1.65-
0
−
−
0
0
0
0
++
+
+
0
0
2
.58). Compounds outside the cLog P values that fit the rule
(Table 2) did not provoke any significant effect on germination
rate. Pr-D-DIBOA affected root length like no other chemical
of the series, preserving an almost complete inhibition effect
a
0
, D-DIBOA phytotoxicity not significantly affected by esterification;
−
, effects
decreased by esterification;
increased by esterification.
+, effects increased by esterification; ++, effects highly
in the first three concentrations, which provides a very low IC50
2
(
27.8 µM, R ) 0.9772). Without presentation of those so potent
values, Val-D-DIBOA is the only chemical to provoke consistent
effects on the three parameters for this plant, the effects on shoot
being of medium potency (IC50 ) 509.6 µM, R ) 0.9800).
The consistency of the effects of the tested compounds on wild
oat makes this plant the optimal one to attempt a mathematical
correlation of lipophilicity, expressed as cLog P, with phyto-
toxicity values (see next section).
The structures of all synthesized chemicals were confirmed by
spectroscopy methods (see Supporting Information).
Fitting to Lipinski’s Rule of 5 (Tice Approach for
Agrochemical Design). The parameters included in Lipinski’s
rule of 5, as they were reinvestigated by Tice as a useful
approach for agrochemicals design, are summarized in Table
2
2
for all of the obtained chemicals. All tested compounds after
Dec-D-DIBOA did not fit the rule, as they have cLog P values
5, which agrees with Lipinski’s requirements. In addition to
It is interesting to point out the effects of these chemicals on
wheat. The IC50 (root length) for the commercial herbicide
>
2
Logran was 7268 µM (R ) 0.9410) and 7026 µM for
this, Tridec- and Mir-D-DIBOA exceeded the number of
rotatable bonds. Compounds from D-DIBOA to Oct-D-DIBOA
have molecular properties adequate to reach good phytotoxicity
values, according to Tice’s optimal cLog P range.
Phytotoxic Activity. All tested compounds, when active,
showed inhibitory profiles. The esters were more or less active
than D-DIBOA depending on the tested plant and the evaluated
parameter (germination rate, root length, and shoot length). The
persistence, the increase or decrease of the phytotoxic effect
with cLog P, was also plant and parameter dependent. In general
terms, root length was the most affected parameter in all species.
A qualitative general comparison of the esters with D-DIBOA,
with regard to plant species and parameters, is given in Table
2
D-DIBOA (R ) 0.9789). They are much higher than the highest
tested dose, so they were obtained by extrapolation. All IC50
values for D-DIBOA esters are >1 mM, the maximum tested
dose. A detailed analysis of the dose-response curves for Pr-
D-DIBOA on A. fatua and wheat (root length) revealed a highly
relevant result from the point of view of the selectivity (see
below).
Lipophilicity-Activity Correlations. Taking into consid-
eration effects on all parameters and species, a cluster analysis
was made to classify the tested chemicals according to their
phytotoxicity and properties. The results are shown in Figure
4. Two main groups of compounds can be discerned according
to their phytotoxicity profiles: G1, in which the commercial
herbicide Logran, D-DIBOA, and their esters from acetyl to
heptanoyl are included, and G2, containing D-DIBOA esters
from octanoyl to miristoyl. G1 is formed by the most phytotoxic
chemicals on most of the tested species. This result clearly
agrees with the proposed dependence of phytotoxicity with cLog
P, as the chemicals fitting the proposed optimal cLog P interval
are grouped as the most active ones, except the octanoyl
derivative, which is still active in some cases (IC50 ) 316.8
3
. The addition of an ester function to D-DIBOA provoked a
significant phytotoxicity enhancement on cress (with increasing
germination rate, root length, and shoot length inhibition), as
the most affected STS, and the weeds rigid ryegrass (in which
root length inhibition increases) and especially wild oat (with a
drastic difference on germination inhibition rate and highly
significant effects on root length).
Cress was more affected by chemicals of medium polarity
of the series: Valeryl and heptanoyl D-DIBOA derivatives

Lipophilic Benzoxazinones
J. Agric. Food Chem., Vol. 54, No. 25, 2006 9361
Figure 3. Phytotoxic effects of D-DIBOA esters on A. fatua L. (germination rate). If it is not indicated, P > 0.05 for Welch’s test: (a) values significantly
different at P < 0.01; (b) values significantly different at 0.01 < P < 0.05.
Figure 4. Cluster analysis for D-DIBOA esters, made on the basis of germination and growth effects for all tested plants.
2
µM, R ) 0.9570, A. fatua, root length), but with a phytotoxicity
profile much closer to the lipophilic side of the series. Those
differences are very clearly displayed in several cases: germina-
tion rate on A. fatua and germination and root and shoot lengths
on L. satiVa and L. esculentum (all parameters, but on root length
especially). This loss of phytotoxicity with lipophilicity could
agree with the Lipinski and Tice models, as compounds with
lipophilicities higher than that of octanoyl-D-DIBOA (cLog P
)
3.97) do not fit the optimal requirements for phytotoxicity
Table 2). This fact makes difficult a mathematical correlation
(
of cLog P values with phytotoxicity results, at least if the most
lipophilic chemicals are included. These kinds of correlations
have been made recently though, yielding a lipophilicity-
activity relationship and an effect on activity related to the even
or odd character of side chains as the most interesting results
Figure 5. QSAR correlations of phytotoxic effects with lipophilicity. n
indicates number of carbons for the esters. Vertical lines indicate maximum
lipophilicity according to Tice and Lipinski models.
(
42). In that paper, the authors employ the wheat coleoptile
bioassay as a phytotoxicity prediction, finding good phytotox-
icity-Log P correlations. They also detect some influence on
the activity, depending on the odd or even nature of the ester.
Herein we present a similar correlation made by means of the
results found for a whole plant. A. fatua, which was the tested
weed in which higher phytotoxicity (21, 22) for benzoxazinones
was found, seemed to be adequate for this study. The phyto-
toxicity [expressed as log(1/IC50)] was correlated with cLog P
as shown in Figure 5. Two zones can be clearly separated: the
lower cLog P zone (fitting the Lipinski cLog P requirements
for optimal bioactivity), in which some correlations can be made,
and the zone of higher cLog P values, in which the phytotoxic

9362 J. Agric. Food Chem., Vol. 54, No. 25, 2006
Mac ´ı as et al.
Figure 6. Phytotoxic effects of 6-MeO-D-DIBOA esters on A. fatua L. (root and shoot lengths). If it is not indicated, P > 0.05 for Welch’s test: (a) values
significantly different at P < 0.01; (b) values significantly different at 0.01 < P < 0.05.
effect of the chemicals could not be adjusted to a parabolic
equation, as Hansch’s transport model proposes (37).
The phytotoxicity data for the lower cLog P zone did not fit
the quadratic model correctly when treated all together, but the
separation of even and odd side chains provided a couple of
excellent fittings. This effect was previously found and discussed
Bu-D-DIBOA are very close to the ones for hexanoyl, heptanoyl,
and also octanoyl D-DIBOA, located in the cLog P region in
which the distance from one curve to another is narrower. This
puts Bu-D-DIBOA in the middle of both clusters. In addition
to this, the even-odd effect is also observed in the cluster
analysis, if compounds other than the butyric ester are grouped
following this behavior in the G1 group. The most lipophilic
side of the series, in which the transport model has no
application (G2), does not follow this pattern. The QSAR
correlations of even and odd esters separately resulted in the
following equations:
(42), so that a relationship with the capacity of the esters’ side
chains to interact with the cell membrane lipidic bilayer was
proposed to be the cause for this phenomenon. Even and odd
side chains reasonably could behave differently, on the basis
of the side-chain lipidic bilayer van der Waals interactions that
occur. In this case, as in the previous paper, odd esters are more
active than the even ones. This could be a result of a stronger
van der Waals interaction with the lipidic bilayer of the even
ones, resulting in a low access to the corresponding target site-
odd esters
2
log(1/IC ) ) 0.151(cLog P) - 1.1685(cLog P) + 3.0612,
50
2
(s) of action. Thus, a good permeability through the lipidic
R ) 0.9929
bilayer (measured in terms of Log P), but weak interactions
with it, favor the phytotoxic effect. Logically, the closer the
side chains are to the cell membrane fatty acid length (and
lipophilicity), the better the interaction, resulting in a retention
of the phytotoxins and the subsequent lower effects. The
tendency of both even and odd series helps to explain the
disposition of butyric ester in the cluster analysis (Figure 4),
which seems not to be ordered. As can be seen, IC50 values of
even esters
2
log(1/IC ) ) 0.138(cLog P) - 1.0626(cLog P) + 2.473,
5
0
2
R ) 0.9640
Lipophilicity is not the only molecular parameter modified
by the systematic esterification of D-DIBOA. The increasing

Lipophilic Benzoxazinones
J. Agric. Food Chem., Vol. 54, No. 25, 2006 9363
Figure 7. Dose−response curves for D-DIBOA and its propanoyl ester (Pr-D-DIBOA) on wheat and wild oat (root length). Inhibitions provoked on wild
oat growth at wheat IC10 are indicated.
ester length provides a larger number of rotatable bonds, whereas
a new H-bond acceptor (carboxyl group) is added and a H-bond
donor (hydroxylic proton of the hydroxamic acid moiety) is lost.
The numbers of H-bond donors and acceptors, which are
relevant in terms of H-bond interaction of the phytotoxin with
the target site of action, remains within the model limits
throughout the series. The increasing number of rotatable bonds
lowers the conformational stability of the molecule, making it
difficult to fit the target site of action. Nevertheless, some
conformational flexibility is needed for this purpose, as com-
pounds with 12 or fewer rotatable bonds fit the rule. Thus, the
ester derivatives of D-DIBOA provide conformational flexibility
to the molecule (which lacks rotatable bonds) in addition to
lipophilicity. All of the tested chemicals fit the model in terms
of molecular weight.
obtained is slightly displaced to the lipophilic side in comparison
to the D-DIBOA series (which could be caused by the higher
aqueous solubility of the 6-methoxy derivative, with cLog P )
0.32). Neither acetyl nor propanoyl derivatives affected root
length significantly. Moreover, shoot length is the most affected
parameter instead of root length. The phytotoxic effect of the
more lipophilic compounds could be due to their accumulation
in the seed surface during the experiment, avoiding the seed
water uptake (as could happen with Mir-D-DIBOA), rather than
a “true” phytotoxic effect in which the active chemical reaches
the corresponding molecular target site. Thus, 6-methoxy-D-
DIBOA derivatives modulate the transport phenomena for this
compound, enhancing the phytotoxic effect in the case of A.
fatua shoot length, but in the context of the intrinsic activity of
this chemical, in which the methylation of the aromatic ring
makes it less active, as in the case of the previously reported
7-methoxy derivative (20, 22).
6
-Methoxy-D-DIBOA Esters. Once the lipophilicity-phy-
totoxicity relationships have been determined, and in the context
of Hansch’s transport theory (37), in which bioactivity is
described as a function of transport phenomena and “intrinsic”
bioactivity, it would be interesting to discover if the phytotoxic
effects observed for the D-DIBOA esters are provoked for a
higher bioavailability of the active part of the molecule, or the
Selective Action of D-DIBOA Esters: Pr-D-DIBOA. Pr-
D-DIBOA was the most active tested chemical on A. fatua root
2
length (IC50 ) 27.8 µM, R ) 0.9772). The introduction of an
ester function in D-DIBOA provokes a modulation of the
transport phenomena, resulting in an increase or decrease of
phytotoxicity. This effect could be different, depending on the
tested plant, as their morphological structures, then the barriers
the chemical needs to cross to reach the target site of action,
could be different.
‘intrinsic’ phytotoxicity of the fatty acid residues attached to
D-DIBOA could have. If side chains were responsible for those
phytotoxicity values, a change in the “active” part of the
molecule would not significantly affect the phytotoxicity.
6
-Methoxy-D-DIBOA (Figure 1), a benzoxazinone derivative
One of the determinant parameters in the optimization of
herbicides is the selectivity: A herbicidal treatment, when made
to a crop, involves necessarily a loss in the crop production
due to the phytotoxicity of the treatment. Thus, it is necessary
to compare the effects that a herbicide candidate can produce
on crops and their common weeds. As A. fatua is a common
with phytotoxicity values very similar to those found for
D-DIBOA (see Supporting Information), is examined through
the same methodology as D-DIBOA on A. fatua. Phytotoxicity
results are shown on Figure 6 (growth parameters). The zone
of the series in which the maximum root length effects are

9364 J. Agric. Food Chem., Vol. 54, No. 25, 2006
Mac ´ı as et al.
wheat weed, the influence of D-DIBOA and its propanoyl
derivative was studied for both plants, by means of the
comparison of their dose-response curves. The main intention
of this correlation was to discover if Pr-D-DIBOA, in addition
to a phytotoxicity increase, affects selectivity of action on wheat
and its common weed, A. fatua. Figure 7 include two sets of
curves, one for each chemical, in which the comparative effects
on wheat and wild oat can be examined. D-DIBOA was more
active on wild oat than on wheat. The D-DIBOA dose provoking
a 10% inhibition of wheat root length causes a 50% inhibition
on wild oat. D-DIBOA propanoyl ester, at the same low-activity
dose on wheat, causes an almost complete inhibition of root
length on wild oat (-85%). In other words, propanoyl ester
effects on wheat and wild oat (two Gramineae species) are much
more from one another than the ones provoked by D-DIBOA.
(6) Dayan, F.; Romagni, J.; Tellez, M.; Rimando, A.; Duke, S.
Managing weeds with natural products. Pestic. Outlook 1999,
1
0, 185-188.
(
(
(
7) Mac ´ı as, F. A.; Molinillo, J. M. G.; Galindo, J. C. G.; Varela, R.
M.; Simonet, A. M.; Castellano, D. The use of allelopathic
studies in the search for natural herbicides. J. Crop Prod. 2001,
4
, 237-255.
8) Grayson, B. T.; Williams, K. S.; Freehauf, P. A.; Pease, R. R.;
Ziesel, W. T.; Sereno, R. L.; Reinsfelder, R. E. The physical
and chemical properties of herbicide cynmethylin. Pestic. Sci.
1
987, 21, 143-153.
9) Lee, D. L.; Prisbylla, M. P.; Cromarthie, T. H. The discovery
and structural requirements of inhibitors of p-hydroxyphe-
nylpyruvate dioxygenase. Weed Sci. 1997, 45, 601-609.
(10) Virtanen, A. I.; Hietala, P. K. The structures of the precursors
of benzoxazolinone in rye plants. II. Suom. Kemistil. B 1959,
3
2, 252.
The systematic esterification of D-DIBOA was effective for
the purpose of studying the influence of transport phenomena
in the phytotoxic effect of this chemical, a useful lead for natural
herbicide model development (20-22). These derivatives
displayed a modulated phytotoxicity, which could be related to
their interactions with cell membrane lipidic bilayers, and the
subsequent capability of reaching the D-DIBOA target site of
action. Propanoyl and valeryl derivatives resulted as the most
interesting ones from the point of view of their effects,
establishing the optimal cLog P in the range from 1.65 to 2.58.
These lipophilic D-DIBOA derivatives could constitute a second
generation of leads for natural herbicide development (43), in
which lipophilicity, a determinant parameter in drug and
agrochemical design (30-31), is optimized. Pr-D-DIBOA also
possesses a much higher selectivity of action than D-DIBOA,
being then potentially able to control a common wheat weed
(
11) Hamilton, R. H.; Bandurski, R. S.; Reusch, W. H. Isolation and
characterization of a cyclic hydroxamate from Zea mays. Cereal
Chem. 1962, 39, 107-113.
12) Wahlroos, O¨ .; Virtanen, I. The precursors of 6-methoxy-
benzoxazolinone in maize and wheat plants, their isolation and
some of their properties. Acta Chem. Scand. 1959, 13, 1906-
(
1
908.
(
13) Hofman, J.; Hofmanova, O. 1,4-Benzoxazine derivatives in
plants. Eur. J. Biochem. 1969, 8, 109-112.
(
14) Hofman, J.; Hofmanova, O. 1,4-Benzoxazine derivatives in
plants: absence of 2,4-dihydroxy-7-methoxy-2H-1,4-benzoxazin-
3(4H)-one from uninjured Zea mays plants. Phytochemistry 1971,
10, 1441-1444.
(15) Sicker, D.; Schulz, M. Benzoxazinones in plants: occurrence,
synthetic access, and biological activity. Stud. Nat. Prod. Chem.
2002, 27, 185-232.
(
16) Honkanen, E.; Virtanen, A. I. Synthesis of some 1,4-benzoxazine
(A. fatua) population without significantly affecting the crop.
derivatives and their antimicrobial activity. Acta Chem. Scand.
This is the first time in which this kind of analysis is reported
with allelochemical derivatives, by using whole plants as target
sites. This work also sheds light on the usefulness of QSAR
analysis, more likely applied in drug design, to the development
of phytotoxins useful in the search for new herbicide models.
1
960, 14, 1214-1217.
(
17) Escobar, C. A.; Sicker, D.; Niemeyer, H. M. Evaluation of
DIMBOA analogs as antifeedants and antibiotics towards the
aphid Sitobion aVenae in artificial diets. J. Chem. Ecol. 1999,
2
5, 1543-1554.
(
18) Schulz, M.; Friebe, A.; Kueck, P.; Seipel, M.; Schnabl, H.
Allelopathic effects of living quack grass (Agropyron repens L.).
Identification of inhibitory allelochemicals exuded from rhizome
borne roots. Angew. Bot. 1994, 68, 195-200.
ACKNOWLEDGMENT
We acknowledge the expert assistance of Drs. Juan Carlos G.
Galindo and Ra u´ l F. Velasco.
(19) Wolf, R. B.; Spencer, G. F.; Plattner, R. D. Benzoxazolinone,
2
,4-dihydroxy-1,4-benzoxazin-3-one, and its glucoside from
Supporting Information Available: Physical data for all
Acanthus mollis seeds inhibit velvetleaf germination and growth.
J. Nat. Prod. 1985, 48, 59-63.
(20) Mac ´ı as, F. A.; Mar ´ı n, D.; Oliveros-Bastidas, A.; Castellano, D.;
Simonet, A. M.; Molinillo, J. M. G. Structure-activity relation-
ships (SAR) studies of benzoxazinoids, their degradation products
and analogues. Phytotoxicity on standard target species (STS).
J. Agric. Food Chem. 2005, 53, 538-548.
1
13
tested chemicals (FTIR, H NMR, C NMR, EIMS); phyto-
toxicity data for D-DIBOA series chemicals on all species; and
phytotoxicity data for 6-MeO-D-DIBOA series on A. fatua. This
material is available free of charge via the Internet at http://
pubs.acs.org.
(
21) Mac ´ı as, F. A.; Chinchilla, N.; Varela, R. M.; Oliveros-Bastidas,
A.; Mar ´ı n, D.; Molinillo, J. M. G. Structure-activity relationship
studies of benzoxazinones and related compounds. Phytotoxicity
on Echinochloa crus-galli (L.) P. Beauv. J. Agric. Food Chem.
LITERATURE CITED
(
1) Sullivan, T. P.; Sullivan, D. S. Vegetation management and
ecosystem disturbance: impact of glyphosate herbicide on plant
and animal diversity in terrestrial systems. EnViron. ReV. 2003,
2
005, 53, 4373-4380.
(
22) Mac ´ı as, F. A.; Mar ´ı n, D.; Oliveros-Bastidas, A.; Castellano, D.;
Simonet, A. M.; Molinillo, J. M. G. Structure-activity relation-
ship (SAR) studies of benzoxazinones, their degradation products,
and analogues. Phytotoxicity on problematic weeds AVena fatua
L. and Lolium rigidum Gaud. J. Agric. Food Chem. 2006, 54,
1
1, 37-59.
(
2) Wang, Y. S. Environmental impact of herbicide use in the
subtropics. Food Sci. Agric. Chem. 1999, 1, 165-179.
3) Preston, C. Herbicide detoxification: herbicide selectivity in
crops and herbicide resistance in weeds. ACS Symp. Ser. 2005,
No. 899, 195-204.
(
1
040-1048.
(
4) Owen, M. D. K.; Zelaya, I. A. Herbicide-resistant crops and weed
resistance to herbicides. Pest Manage. Sci. 2005, 61, 301-311.
5) Mac ´ı as, F. A.; Molinillo, J. M. G.; Oliveros-Bastidas, A.; Mar ´ı n,
D.; Chinchilla, D. Allelopathy. A natural strategy for weed
control. Commun. Agric. Appl. Biol. Sci. 2004, 69, 13-23.
(23) Mac ´ı as, F. A.; Oliveros-Bastidas, A.; Mar ´ı n, D. The allelopathic
phenomenon, a dynamic process. In Proceedings of the 4th World
Congress on Allelopathy; Harper, J. D. I., An, M., Wu, H., Kent,
J. H., Eds.; The Regional Institute Limited: Gosford, Australia,
2005; pp 77-85.
(

Lipophilic Benzoxazinones
J. Agric. Food Chem., Vol. 54, No. 25, 2006 9365
(
24) Mac ´ı as, F. A.; Oliveros-Bastidas, A.; Mar ´ı n, D.; Castellano, D.;
Simonet, A. M.; Molinillo, J. M. G. Degradation studies on
benzoxazinoids. Soil degradation dynamics of 2,4-dihydroxy-
(34) PRISMA 4.0; GraphPad Software, Inc.: San Diego, CA, 2003.
(35) Sander, T. OSIRIS Property Explorer. Actelion Property Ex-
plorer, 2001.http://www.chemexper.com/tools/propertyExplorer/
main.html.
7
-methoxy-(2H)-1,4-benzoxazin-3(4H)-one (DIMBOA) and its
degradation products, phytotoxic allelochemicals from Gramine-
ae. J. Agric. Food Chem. 2004, 52, 6402-6413.
(
(
36) Chou, J.; Jurs, P. J. Chem. Inf. Comput. Sci. 1979, 19, 172.
37) Hansh, C.; Leo, A. J. Substituent Constants for Correlation
Analysis in Chemistry; Wiley: New York, 1979.
(
25) Mac ´ı as, F. A.; Oliveros-Bastidas, A.; Mar ´ı n, D.; Castellano, D.;
Simonet, A. M.; Molinillo, J. M. G. Degradation studies on
benzoxazinoids. Soil degradation dynamics of (2R)-2-O-â-D-
glucopyranosyl-2,4-dihydroxy-(2H)-1,4-benzoxazin-3(4H)-one (DI-
BOA-Glc) and its degradation products, phytotoxic allelochem-
icals from Gramineae. J. Agric. Food Chem. 2005, 53, 554-
(
38) Mac ´ı as, F. A.; Castellano, D.; Molinillo, J. M. G. Search for a
standard phytotoxic bioassay for allelochemicals. Selection of
standard target species. J. Agric. Food Chem. 2000, 48, 2512-
2
521.
(
39) Castellano, D.; Mac ´ı as, F. A.; Castellano, M.; Cambronero, R.
M. FITOMED (automated system for measurement of variable
lengths). Sp. Patent P9901565.
5
61.
(
26) Eljarrat, E.; Guillam o´ n, M.; Seuma, J.; Mogensen, B. B.;
Fomsgaard, I. S.; Oliveros-Bastidas, A.; Mac ´ı as, F. A.; Stochmal,
A.; Oleszek, W.; Shakaliene, O.; Barcel o´ , D. First European
interlaboratory study of the analysis of benzoxazinone derivatives
in plants by liquid chromatography. J. Chromatogr. A 2004,
(
40) Mac ´ı as, F. A.; Velasco, R. F.; Vi n˜ olo, V. M. I.; Castellano, D.;
Galindo, J. C. G. Cluster analysis. A valuable tool for allelopathic
SAR studies? In Allelopathy: from Molecules to Ecosystems;
Reigosa, M. J., Pedrol, N., Eds.; Science Publishers: Enfield,
NH, 2002; pp 305-316.
1
047, 69-76.
(
27) Mac ´ı as, F. A.; Mar ´ı n, D.; Oliveros-Bastidas, A.; Chinchilla, D.;
Simonet, A. M.; Molinillo, J. M. G. Isolation and synthesis of
allelochemicals from Gramineae: benzoxazinones and related
compounds. J. Agric. Food Chem. 2006, 54, 991-1000.
28) Kubinyi, H. Strategies and recent technologies in drug discovery.
Pharmazie 1995, 50, 647-662.
(41) Statistica for Windows 5.0; Statsoft Inc.: Tulsa, OK, 1984-
1995.
(42) Mac ´ı as, F. A.; Velasco, R. F.; Castellano, D.; Galindo, J. C. G.
Application of Hansch’s model to guaianolide ester derivatives:
a quantitative structure-activity relationship study. J. Agric.
Food Chem. 2005, 53, 3530-3539.
(
(
(
29) Baum, R. M. Combinatorial approaches provide fresh leads for
medicinal chemistry. Chem. Eng. News 1994, Feb 7, 20-26.
30) Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P. J.
Experimental and computational approaches to estimate solubility
and permeability in drug discovery and development settings.
AdV. Drug DeliVery ReV. 2001, 46, 3-26.
31) Tice, C. M. Selecting the right compounds for screening: does
Lipinski’s rule of 5 for pharmaceuticals apply to agrochemicals?
Pest Manag. Sci. 2001, 57, 3-16.
32) Kleier, D. A.; Grayson, B. T.; Hsu, F. C. The phloem mobility
of pesticides. Pestic. Outlook 1998, 9, 26-30.
33) Sicbaldi, F.; Sacchi, G. A.; Trevisan, M.; Del, Re, A. A. M.
Root uptake and xylem translocation of pesticides from different
chemical classes. Pestic. Sci. 1997, 50, 111-119.
(
43) Mac ´ı as, F. A.; Molinillo, J. M. G.; Mar ´ı n, D.; Carrera, C.;
Chinchilla, N.; Oliveros-Bastidas, A. 4-Aciloyloxy-2H-1,4-
benzoxazin-3(4H)-one derivatives with phytotoxic activity. Sp.
Patent P200601039.
(
Received for review July 28, 2006. Revised manuscript received
September 27, 2006. Accepted September 28, 2006. This research was
supported by the Ministerio de Educaci o´ n y Ciencia, Spain (Project
AGL2006-10570-AGR). A contract from Junta de Andaluc ´ı a, Spain
(
(
(
D.M.), is also gratefully acknowledged.
JF062168V