New herbicide models from benzoxazinones: Aromatic ring functionalization effects

Article abstract of DOI:10.1021/jf062709g

The utility of benzoxazinones and some of their synthetic derivatives in the search for new leads for herbicide model development has been widely discussed. As the benzoxazinone skeleton contains three different potential areas for functionalization (C-2,

Full text of DOI:10.1021/jf062709g

J. Agric. Food Chem. 2006, 54, 9843−9851 9843
New Herbicide Models from Benzoxazinones: Aromatic Ring
Functionalization Effects
FRANCISCO A. MACIÄAS,* JOA˜ O M. DE SIQUEIRA,† NURIA CHINCHILLA,
DAVID MARIÄN, ROSA M. VARELA, AND JOSEÄ M. G. MOLINILLO
Grupo de Alelopat´ıa, Departamento de Qu´ımica Orga´nica, Universidad de Ca´diz, Avenida Repu´blica
Saharaui s/n, 11510 Puerto Real, Ca´diz, Spain
The utility of benzoxazinones and some of their synthetic derivatives in the search for new leads for
herbicide model development has been widely discussed. As the benzoxazinone skeleton contains
three different potential areas for functionalization (C-2, N-4, and aromatic protons H-5, H-6, H-7,
and H-8), and the first two have already been optimized, the main objective of this work was the
substitution of aromatic protons for different substituent types and the study of the effects of the
prepared chemicals on selected standard target species (STS) and weeds. Thus, different combina-
tions of aromatic substituents, including methoxy, methoxycarbonyl, fluorine, chlorine, and trifluo-
romethyl, were introduced at different positions. Phytotoxicity results were successfully correlated
with steric and electronic molecular parameters, the resulting molecular volume (V) and dipole moment
(µ) being the most influential ones. Halogenations at C-6 and fluorination at C-7 were the most
successful modifications. Compounds 6-fluoro-(2H)-1,4-benzoxazin-3(4H)-one (6F-D-DIBOA), 7-fluoro-
(2H)-1,4-benzoxazin-3(4H)-one (7F-D-DIBOA), and 6-chloro-(2H)-1,4-benzoxazin-3(4H)-one (6Cl-
D-DIBOA) had the highest phytotoxic activities. The dose-response profiles on wheat and two of its
most common weeds (Lolium rigidum Gaud. and Avena fatua L.) were compared by means of a
proposed selectivity index, which displayed 7F-D-DIBOA as the most selective of the tested
benzoxazinones.
KEYWORDS: Benzoxazinones; halobenzoxazinones; dipole moment; structure-activity relationship
(SAR); herbicide models; phytotoxicity
INTRODUCTION
benzoxazinone analogues with various substituents at C-2, N-4,
and some aromatic protons. A preliminary structure-activity
relationship study with these derivatives along with natural
benzoxazinones and their degradation metabolites was a first
successful approach with respect to enhancing the phytotoxicity
of benzoxazinones and discovering the potential ecological role
of the degradation products (2-4). These works provided the
optimal heterocyclic structure for phytotoxicity enhancement.
Taking advantage of the hydroxyl group at N-4 present in the
most active models, and by systematic esterification, some
interesting derivatives with optimal lipophilicity were obtained
(5). The different modification possibilities for benzoxazinone
model 2,4-dihydroxy-(2H)-1,4-benzoxazin-3(4H)-one (D-DI-
BOA) are summarized in Figure 1. The most relevant structural
features of D-DIBOA were the lack of substitution at C-2
(especially with regard to the role of hydroxyl group of DIBOA
and DIMBOA in their degradability) and the presence of the
hydroxyl group at N-4 (preserving the integrity of the natural
hydroxamic acid moiety). In general terms the presence of
aromatic substituents lowered the phytotoxic effects of the tested
chemicals.
The preparation of derivatives from natural products as leads
for new bioactive chemicals has been a commonly followed
strategy in recent times (1). In most of the cases, the design of
these new chemicals has been carried out by following two
different strategies: employment of the molecular fragments
responsible for the observed bioactivity of the natural product
or functionalization of the basic molecular backbone. These
simplifications are necessary in most of the cases, as the
structures of bioactive natural products are usually complicated,
mainly due to the presence of one or more stereogenic centers.
Their synthetic preparation at the multigram scale, needed for
greenhouse and field bioassays, is therefore difficult and
expensive.
In the context of new herbicide model development from
benzoxazinones, their unusual chemical simplicity, in compari-
son to most secondary metabolites, eliminates the need to use
molecular fragments. Functionalization led to several synthetic
* Author to whom correspondence should be addressed (telephone +34-
956.016.370; fax +34-956.016.295; e-mail famacias@uca.es).
† Permanent address: DFB, CCBS, Universidade Federal de Mato Grosso
do Sul, CP 549, 79070-900, Campo Grande, MS, Brazil.
Research on the chemistry and bioactivity of benzoxazinones
yielded a wide variety of natural and synthetic derivatives, with
10.1021/jf062709g CCC: $33.50
© 2006 American Chemical Society
Published on Web 11/29/2006

9844 J. Agric. Food Chem., Vol. 54, No. 26, 2006
Mac´ıas et al.
Figure 1. Structure and potential modification places for 4-hydroxy-(2H)-2,4-benzoxazin-3(4H)-one (D-DIBOA).
Table 1. Natural Benzoxazinones Substituted at C-7 and/or C-8
Figure 2. Molecular structure for 2-O-
â
-D-glucopyranosyl-5-chloro-7-
methoxy-(2H)-1,4-benzoxazin-3(4H)-one (5-Cl-HMBOA-Glc).
R1
R2
R3
name
DIBOA
DIBOA-Glc
DIMBOA
DIMBOA-Glc
DIM2BOA
DIM2BOA-Glc
TRIBOA
H
H
H
H
H
H
H
Glc
H
Glc
H
Glc
H
No bioactivities were described for this benzoxazinone
chloride. Its isomer, 8-chloro-4-hydroxy-2H-1,4-benzoxazin-
3(4H)-one, was synthesized by employing the latter procedure
and using 6-chloro-2-nitrophenol as starting material, being
further evaluated from the point of view of its antifungal activity
(14). Soon after this patent, O¨ zden and collaborators (15)
synthesized this derivative in an analogous manner, employing
the Zn/NH4Cl reductive cyclization procedure, and evaluated
its antifungal activity. It was described as slightly active. With
regard to polyhalogenated benzoxazinones, the derivative 6,8-
dichloro-4-hydroxy-2H-1,4-benzoxazin-3(4H)-one (6,8-diCl-D-
DIBOA, Table 2) was synthesized in previous studies about
plant growth regulators chemistry (16). This work is older than
the discovery of natural benzoxazinones DIBOA and DIMBOA,
and the synthesis is made in the context of the preparation of
several derivatives of aryloxyalkanecarboxylic acids such as 2,4-
dichlorophenoxyacetic acid (the herbicide 2,4-D). The catalytic
hydrogenation of the 6-nitro derivative of 2,4-D with Raney
nickel afforded the desired substituted benzoxazinone. This
chemical has also been patented as a fungicide (17).
CH3O
CH3O
CH3O
CH3O
OH
CH3O
CH3O
H
different substituent combinations in the aromatic ring. In
addition to the natural 7-methoxybenzoxazinone DIMBOA and
its glucoside, there are a few natural benzoxazinones containing
aromatic methoxy groups at C-8 (6), such as DIM2BOA and
its glucoside (Table 1), which have been isolated from maize.
Although it is not widely distributed on benzoxazinone producer
plans, Le-Van and Wratten (7) isolated a 5-chloro-substituted
benzoxazinone (Cl-HMBOA-Glc), which was demonstrated to
be naturally occurring (Figure 2).
There are several studies on the chemistry and potential
utilities of some other halo-substituted benzoxazinones, obtained
by means of synthetic procedures as they are not naturally
occurring chemicals. For example, 7-fluoro-4-hydroxy-(2H)-
1,4-benzoxazin-3(4H)-one (7F-D-DIBOA) (Table 2) was syn-
thesized by Geng et al. (8) by adapting the method previously
described by Quiroz and Niemeyer, (9) which afforded several
chloro-substituted derivatives. These methods were based on
direct chlorinations of synthetic benzoxazinones D-DIBOA and
4-hydroxy-(2H)-1,4-benzoxazin-3(4H)-one (D-DIMBOA). Chlo-
rination at C-5 was achieved for D-DIMBOA, whereas D-
DIBOA was functionalized at C-5 and C-7. Logically, these
procedures were needed for the previous synthesis of the starting
benzoxazinones (10).
The antialgal, antifungal, and antimicrobial activities of 7F-
D-DIBOA were evaluated on several organisms (Chlorella
xanthella, Candida albicans, Escherichia coli, and Staphylo-
coccus aureus), with medium or low activities, depending on
the organism tested (11, 12).
Other relevant contributions to chlorobenzoxazinone’s chem-
istry were made by Hashimoto et al. (13). In this case, a suitable
side chain was added to readily available chloronitrophenols,
yielding a chlorobenzoxazinone [6-chloro-4-hydroxy-(2H)-1,4-
benzoxazin-3(4H)-one, 6Cl-D-DIBOA, Table 2] after reductive
cyclization.
There are some precedents about the usage of benzoxazinone
derivatives as active ingredients of commercial herbicides. The
benzoxazinyltetrahydrobenzotriazol oxide derivatives (18) and
the benzylpyridones (19) were proposed as candidates for it.
The radicals at position N-4 were defined as alkyl or alkenyl
with one or more halogen atoms and successfully tested on
several weed species. There is also a commercial herbicide
belonging to the phenylphthalymides group, featuring a ben-
zoxazinone heterocycle as a substituent (Flumioxazin) (20).
Taking into consideration the functionalization possibilities
of benzoxazinones (Figure 1), further research on the effects
of modifications to the aromatic substitution should help to
complete our knowledge about the optimal structural require-
ments for maximum phytotoxicity, in addition to yielding new
derivatives with potential utility in the development of natural
herbicide models, with a minimum modification of the natural
benzoxazinone moiety. Thus, 12 benzoxazinones with different
combinations of functional groups linked to the aromatic ring

Herbicide Models from Benzoxazinones
J. Agric. Food Chem., Vol. 54, No. 26, 2006 9845
Table 2. Benzoxazinone Derivatives Tested
R1
R2
R3
compound
H
H
CH3O
H
CH3O
H
CH3O
H
H
H
CH3O
H
H
H
H
4-hydroxy-(2H)-1,4-benzoxazin-3(4H)-one (D-DIBOA)
4-hydroxy-7-methoxy-(2H)-1,4-benzoxazin-3(4H)-one (D-DIMBOA)
4-hydroxy-6-methoxy-(2H)-1,4-benzoxazin-3(4H)-one (6-MeO-D-DIMBOA)
4-hydroxy-7,8-dimethoxy-(2H)-1,4-benzoxazin-3(4H)-one (7,8-diMeO-D-DIMBOA)
6-methoxycarbonyl-(2H)-1,4-benzoxazin-3(4H)-one (6-MeOCO-D-DIMBOA)
7-methoxycarbonyl-(2H)-1,4-benzoxazin-3(4H)-one (7-MeOCO-D-DIMBOA)
6-fluoro-(2H)-1,4-benzoxazin-3(4H)-one (6F-D-DIMBOA)
H
CH3OCO
H
F
H
H
CH3OCO
H
F
7-fluoro-(2H)-1,4-benzoxazin-3(4H)-one (7F-D-DIMBOA)
H
H
CF3
Cl
H
H
F
H
H
H
H
F
F
H
H
Cl
Cl
8-fluoro-(2H)-1,4-benzoxazin-3(4H)-one (8F-D-DIMBOA)
7,8-difluoro-(2H)-1,4-benzoxazin-3(4H)-one (7,8-diF-D-DIMBOA)
6-trifluoromethyl-(2H)-1,4-benzoxazin-3(4H)-one (6CF3-D-DIMBOA)
6-chloro-(2H)-1,4-benzoxazin-3(4H)-one (6Cl-D-DIMBOA)
8-chloro-(2H)-1,4-benzoxazin-3(4H)-one (8Cl-D-DIMBOA)
6,8-dichloro-(2H)-1,4-benzoxazin-3(4H)-one (6,8-diCl-D-DIMBOA)
Cl
Synthesis of Aromatic Ring-Modified Benzoxazinones. The
chemicals were obtained according to a previously reported methodol-
ogy (Figure 3) (10) by a sequence of nucleophillic substitutions (side-
chain linkage) and reductive cyclization (benzoxazinone ring formation).
The starting materials employed were 2-nitrophenols substituted in the
adequate positions: 5-methoxy-2-nitrophenol and 5,6-dimethoxyphenol
were synthesized from 3-methoxyphenol and 2,3-dimethoxyphenol
(Sigma-Aldrich Co.), respectively, by following the previously reported
procedure for aromatic nitration (22). These methods afforded D-
DIMBOA and 6-MeO-D-DIBOA, respectively. The remaining chemi-
cals were obtained directly by transforming the suitable nitrophenols,
as they were commercially available.
(methoxy, methoxycarbonyl, fluorine, trifluoromethyl, and
chlorine) were synthesized according to the previously described
methodology (10). They were evaluated on standard target
species (STS) Lepidium satiVum L. (cress), Allium cepa L.
(onion), Lactuca satiVa L. (lettuce), Lycopersicon esculentum
Mill. (tomato), and Triticum aestiVum L. (wheat), in addition
to the common wheat weeds Lolium rigidum Gaud. (rigid
ryegrass) and AVena fatua L. (wild oat). Their effects, structure-
activity relationships, and their correlation with molecular
parameters of the tested chemicals are discussed.
MATERIALS AND METHODS
General Procedure for Nucleophilic Substitution. The starting
2-nitrophenol was dissolved in a solution of 0.1 M potassium hydroxide
in absolute ethanol (1 mol equiv of KOH). After 1 h, the solvent was
removed at reduced pressure. The resulting alkoxide was redissolved
in N,N-dimethylformamide (50 mL/g starting material), and 1.2 mol
equiv of ethyl 2-bromoacetate was added. The reaction mixture was
stirred under argon for 24 h. After this time, 50 mL/g starting material
of ethyl acetate was added, and the resulting organic solutions were
washed with five portions of distilled water. The organic layers were
combined and dried over anhydrous sodium sulfate, and the solvent
was distilled at reduced pressure. Reaction crude was chromatographed
(column chromatography, ethyl acetate/hexane, 1:4) to obtain the ethyl
2-[2′-nitrophenoxy]acetates in quantitative yield.
General Methods. The purity of the isolated and synthetic com-
pounds was determined by 1H NMR and HPLC analyses and was found
to be >98%. 1H and 13C NMR spectra were recorded using MeOH-d4
or CDCl3 as solvent in a Varian INOVA spectrometer at 399.99 or
100.577 MHz, respectively. The resonance of residual methanol for
1H was set to δ 3.30 and that for residual chloroform to δ 7.25. The
solvent peak for 13C was set to δ 49.00 (methanol) or δ 77.00
(chloroform) and used as internal reference. Mass spectra were recorded
by means of a Varian 1200L quadrupole MS/MS detector. The physical
data for all tested chemicals can be found in the Supporting Information.
Chemicals. The chemicals employed for this study are summarized
in Table 2. They can be classified into four groups: methoxy derivatives
(with methoxy groups at C-6 and C-7 or two groups at C-7 and C-8),
methoxycarbonyl derivatives (C-6 and C-7), fluorine derivatives
(monosubstituted at C-6, C-7, and C-8 or disubstituted at C-7 and C-8
and a trifluoromethyl group at C-6), and chlorine derivatives (mono-
substituted at C-6 and C-8 and disubstituted at C-6 and C-8). The
commercial herbicide Logran was used as internal standard for the
phytotoxicity study according to a previously reported bioassay
methodology (21). The starting compound for herbicide models based
on benzoxazinones, D-DIBOA, was also included in the study. It does
not have any aromatic substituents.
General Procedure for ReductiVe Cyclization. Palladium on carbon
(10% Pd, 10% w/w proportion from starting nitrophenoxyacetate) was
suspended in an aqueous solution of 1,4-dioxane (1:1) (100 mL/g
starting material). After this, sodium borohydride (2 mol equiv) was
added, and the solution was vigorously stirred. To this stirred suspension
as added dropwise a solution of the nitrophenoxyacetate in 1,4-dioxane
(0.5 g/mL). The reaction evolution was controlled by thin-layer
chromatography. Once the reaction was completed, the suspension was
vacuum filtered through Celite to remove the catalyst, and the filtrate
was treated with 10% hydrochloric acid until pH 2 was reached. This
Figure 3. General procedure for tested chemicals’ synthesis.

9846 J. Agric. Food Chem., Vol. 54, No. 26, 2006
Mac´ıas et al.
solution was further extracted with ethyl acetate (three times). The
organic layers were combined and dried over anhydrous sodium sulfate,
and the solvent was distilled at reduced pressure. The obtained residue
was chromatographed (column chromatography, ethyl acetate/hexane,
increasing polarity) to obtain the corresponding benzoxazinones.
Bioactivity Evaluation and Statistical Analysis. Target Plants.
Selection of target plants is based on an optimization process made by
us in the search for a standard phytotoxicity evaluation bioassay (21).
After this process, several STS were proposed, including monocots
wheat and onion and dicots tomato, cress, and lettuce, which were
assayed for this study. The weeds rigid ryegrass and wild oat were
also tested by employing the same methodology.
Molecular Modeling and QSAR Calculations. Three-dimensional
models for the tested chemicals were obtained from AM1 calculations
performed by Hyperchem 7.01 software (26). Dipole moments, partial
charges, polarizabilities, and molecular volumes were obtained by
employing the algorithms implemented in this software. Molecular
parameters-activity correlations were performed by means of the
suitable Microsoft Office Excel 2003 spreadsheets (27). Statistical
significance, given as probability value for correlation (P), was
calculated for volume and dipole moment adjustments to phytotoxicity,
by employing the Pearson product moment correlation method, as
implemented in SigmaStat 3.1 software (28).
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
10-2 M 2-[N-morpholino]ethanesulfonic acid (MES) and 1 M NaOH
(pH 6.0). Compounds to be assayed were dissolved in dimethyl
sulfoxide (DMSO) at different concentrations (0.2, 0.1, 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 for each compound
(10-3, 5 × 10-4, 10-4, 5 × 10-5, and 10-5 M) were reached. Additional
treatments for 5 × 10-6, 10-6, 5 × 10-7, 10-7, and 5 × 10-8 M were
prepared the same way for the most active chemicals to calculate
accurate IC50 values (concentrations at which the observed parameter
is stimulated or inhibited at 50% from control). This procedure
facilitated the solubility of the assayed compounds. The number of seeds
in each Petri dish depended on the seed size. The numbers of seeds
that were used for tomato, lettuce, cress, and onion were 25, 15 for
rigid ryegrass, and 10 were used for wheat and wild oat. Treatments,
negative controls (buffered aqueous solutions with DMSO and without
any tested compound), or internal reference solutions [commercial
herbicide Logran, selected after a comparison study previously reported
(21)] were added (5 mL) 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. The same number of replicates was used
for treatments, negative controls, and internal references.
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
facilitated the handling of the plants and allowed a more accurate
measurement of root and shoot lengths.
Bioassay Data Acquisition. Evaluated parameters (germination rate,
root and shoot length) were recorded by using a Fitomed system (23),
which allowed automatic data acquisition and statistical analysis by
its associated software.
RESULTS AND DISCUSSION
General Bioactivity Profiles. All compounds, when active,
showed inhibitory profiles for all evaluated parameters. Growth
parameters (root and shoot length) were more affected than
germination rates in most of the cases, although 6-Cl-D-DIBOA
provoked very powerful inhibitions also in this parameter (IC50
) 20.4 µM, R2 ) 0.8103, L. esculentum; and 10.3 µM, R2 )
0.9754, L. satiVa). Among growth parameters, root length was
more affected than shoot length. All tested species were
significantly affected by the tested chemicals. Among STS, cress
and onion were the most affected, whereas lettuce and wheat
were the plants most tolerant to the tested treatments. The effects
provoked by certain chemicals such as 6-Cl-D-DIBOA and 6-F-
D-DIBOA forced us to test them at lower concentrations (see
above). Results for both compounds on all tested species (root
lengths) are shown in Figure 4.
In general terms, these compounds had a more pronounced
phytotoxicity compared to the previously tested benzoxazinones
(2-4). To relate phytotoxic effects with structural features and
to compare the aromatic ring substituted benzoxazinones with
the previously tested ones, a cluster analysis was performed
(Figure 5).
The analysis, based on more than 1500 activity data, yielded
four clusters (G1, G2, G3, and G4). The latter three are grouped
in a cluster of higher order, indicating the chemicals from the
first one (6-Cl- and 6-F-D-DIBOA) to be very different from
the other ones in terms of bioactivity profiles. Groups G2 and
G3 belong to the same cluster, as similarities among the
chemicals grouped in them, in terms of bioactivity, may exist.
Chemicals are ordered by their average phytotoxic effects from
top to bottom. Thus, G1 groups the most phytotoxic compounds,
G2 contains highly phytotoxic chemicals, G3 is formed by
medium or low phytotoxicity compounds, and the last group,
G4, is formed by slightly active or nonactive chemicals.
Statistical Analysis. Data were statistically analyzed using Welch’s
test, with significance fixed at 0.01 and 0.05 and presented as percentage
differences from control. Zero represents control, positive values
represent stimulation of the studied parameter, and negative values
represent inhibition. The cluster analysis (joining/tree clustering) was
obtained by using Statistica v. 5.0 software (24), and it is presented as
a horizontal hierarchical tree plot. Germination rate, shoot length, and
root length average activity values for each concentration, for all tested
species and all chemicals, were included in the analysis to acquire an
overall view of the phytotoxicity and its relation with chemical structure.
IC50 values were obtained after the phytotoxicity data and concentra-
tions (logarithmic scale) had been adjusted to a constant slope sigmoidal
dose-response curve, defined by the equation
6-Cl- and 6-F-D-DIBOA provoked highly significant inhibi-
tions at concentrations lower than the other chemicals. This
behavior can be quantified by calculating the corresponding IC50
values (micromolar), shown in Table 3. Blank spaces indicate
a bad fit to the dose-response model (R2 < 0.85).
Both chemicals provoked on root length with IC50 values
close to or below 1 µM, as recorded for A. cepa and L. satiVum.
The effects on the weed A. fatua are especially relevant. None
of their IC50 values are above 60 µM. Their structures have a
halogen atom at C-6 in common.
With regard to the chemicals placed in group G2, all of them
have in common the presence of fluorine in their structures,
this atom being placed at C-7 and C-8 or as a trifluoromethyl
group at C-6. IC50 values for these chemicals varied over a wide
range of concentrations. Wheat was the less sensitive plant
(248.3-735.1 µM), whereas the most pronounced inhibitions
were obtained for onion and cress with IC50 values below 50
µM. It is interesting to compare the effects on wheat and its
Ymax - Ymin
Y ) Ymin
+
50-X
1 + 10logIC
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 determination coefficient (R2). The adjustment and the
R2 were obtained by using GraphPad Prism software v. 4.00 (25).

Herbicide Models from Benzoxazinones
J. Agric. Food Chem., Vol. 54, No. 26, 2006 9847
Figure 4. Phytotoxic effects of 6-chloro-4-hydroxy-(2H)-1,4-benzoxazin-3(4H)-one (6-Cl-D-DIBOA) and 6-fluoro-4-hydroxy-(2H)-1,4-benzoxazin-3(4H)-
one (6-F-D-DIBOA) on standard target species and common wheat weeds (root length, percent of control). 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 5. Cluster analysis (phytotoxicity data, germination and growth parameters, all tested species) and molecular volumes of the tested chemicals.
common weeds, as the effects recorded for both species (A. fatua
and L. rigidum) are higher. A quantitative description of this
selective behavior is provided below.
between D-DIMBOA and D-DIBOA, being closer in its effects
to the latter one as they belong to the G3 cluster.
All chemicals belonging to group G4 provoked very low
effects on the tested plants. All IC50 values for root lengths are
above 100 µM, especially for methoxycarbonyl, dimethoxy, and
dichloride derivatives.
The phytotoxicity of some chemicals belonging to group G3
has been already described (3, 4). Their relative phytotoxic
character is maintained also when they are compared with much
more phytotoxic halobenzoxazinones. It is interesting to point
out the effects of 6-MeO-D-DIBOA, as the situation of the
methoxy group greatly affects the bioactivity: 7-methoxyben-
zoxazinones were already found to be less active than the
nonmethoxylated ones (3, 4). The 6-methoxy derivative is placed
A. fatua Wheat Selectivity. After the bioactivity data had
been acquired and the corresponding IC50 values obtained, A.
fatua and wheat IC50 were compared to find the most selective
chemicals of the tested series. Herein we propose the employ-

9848 J. Agric. Food Chem., Vol. 54, No. 26, 2006
Mac´ıas et al.
T. aestivum
Table 3. Selected IC50 Values (Micromolar, Root Length), with Indication of Clustering Groups
A. fatua
L. sativum
A. cepa
L. sativa
L. rigidum
L. esculentum
G1
G2
6Cl-D-DIBOA
6F-D-DIBOA
10.9
8.7
0.1
1.1
0.5
0.9
12.2
55.7
29.0
28.1
4.9
8.6
25.0
43.6
7F-D-DIBOA
8F-D-DIBOA
6CF3-D-DIBOA
7,8-diF-D-DIBOA
69.7
148.4
250.9
108.8
8.9
40.4
48.8
41.7
18.8
28.7
167.6
190.2
64.4
220.1
195.0
194.9
159.3
304.2
389.0
162.9
89.5
735.1
318.4
248.3
453.4
191.6
G3
G4
8Cl-D-DIBOA
D-DIBOA
6MeO-D-DIBOA
Logran
154.7
194.3
428.4
242.2
149.2
128.9
147.9
129.5
1481.0
151.8
374.0
100.0
218.8
10.5
2681.0
862.1
835.4
390.3
1092.0
176.0
465.9
55.2
605.5
892.3
914.1
951.3
942.5
58.8
288.4
D-DIMBOA
212.1
211.9
6-MeOCO-D-DIBOA
7-MeOCO-D-DIBOA
7,8-diMeO-D-DIBOA
6,8-diCl-D-DIBOA
709.7
1140.0
642.4
578.9
1796.0
627.1
362.2
977.8
776.2
947.1
730.1
691.8
636.9
319.8
935.5
1009.0
114.1
266.1
Steric Effects. Molecular volume is one of the most im-
mediate parameters for the description of a bioactive chemical,
because it determines the possibility to interact with the
corresponding molecular target site of action. The presence of
substituents at the benzoxazinone aromatic ring provokes a
volume change, which was correlated with phytotoxicity data.
A qualitative description of the volume effects is shown in
Figure 5.
As the molecular volume increases, the phytotoxicity decays.
An approach to a quantitative analysis is shown in Figure 7.
This correlation displays a close relationship of activity with
molecular volume, as two zones can be found: one for groups
G1 and G2, where low volume changes cause high activity
decreases, and other one for groups G3 and G4, in which
changes in volume do not affect the phytotoxic effects. Thus,
molecular volume seems to be a determinant parameter for
groups G1 and G2.
Figure 6. Selectivity index (SI50) for selected chemicals on Avena fatua
and Lolium rigidum.
ment of the quotient of both IC50 values as a “selectivity index”
(SI):
According to the cluster analysis shown above, all chemicals
substituted at C-6 are more active than their respective isomers
with substituents placed at C-7 and/or C-8. Then, not only the
overall molecular volume but the location of the different
substituents seem to affect bioactivity. This general tendency
is less clear with fluorine derivatives, as the steric hindrances
provided by fluorine and hydrogen atoms are very similar. These
differences are much clearer for chlorine, methoxy, and meth-
oxycarbonyl derivatives. Then, no good correlations could be
observed by treating the chemicals all together, but discarding
for correlation the ones substituted at C-7 or C-8 (Figure 8).
wheat
IC50
SI50
)
wheat weed
IC50
Thus, the higher the value, the more selective toward the weed
the compound is. This parameter was calculated for all of the
chemicals with values fitting the dose-response model correctly
for both species (Figure 6).
The commercial herbicide Logran was found to be the most
selective treatment on L. rigidum. Among the tested benzox-
azinones, the most remarkable behavior was the one recorded
for 7F-D-DIBOA and D-DIMBOA. SI50 values on A. fatua are
higher, and the most selective benzoxazinones were 7F-D-
DIBOA, D-DIMBOA, and 6F-D-DIBOA. Thus, 7F-D-DIBOA
These data fitted correctly a second-order polynomial func-
tion, in which the decay of the phytotoxic effect with volume
is clearly displayed. In conclusion, a general correlation of the
activity with molecular volume could be observed, the location
of the substituents being also crucial. Compounds substituted
at C-7 and C-8 were less active than the ones substituted just
at C-6, the size of the radical at this position being also
determinant in the activity observed.
has special interest as its effects on wheat are very low (IC50
)
735.1 µM) (29). Its effects are the most remarkable from the
point of view of both weed species tested. Further research will
be needed to clarify the selectivity of the tested chemicals in
an agricultural environment, but we propose the SI50 quotient
as a useful and easy-to-calculate parameter to select or discard
chemicals for further modifications and/or higher level bioassays
(greenhouse, field).
Electronic Effects. A relationship between benzoxazinone’s
bioactivity and the electron density at N-4 has been suggested
(13). Thus, to evaluate the role of molecular electronic distribu-
tion, some electronic densities for heterocycle atoms were
calculated (Table 4). No significant variation (see standard
deviation, σ) was found for any of these chemicals, which
included active and nonactive. Thus, the aromatic substitution
pattern seems not to affect the electron density of the heterocycle
and the differences in activity cannot be attributed to these
parameters. Then, a representative of the overall electronic
Quantitative Structure-Activity Relationships (QSAR).
To find the structural requirements for maximum phytotoxicity,
the activity data were correlated with several molecular property
descriptors. For that purpose, all tested chemicals were modeled
by means of AM1 calculations, and their three-dimensional
structures were obtained.

Herbicide Models from Benzoxazinones
J. Agric. Food Chem., Vol. 54, No. 26, 2006 9849
Figure 7. Phytotoxicity (Avena fatua, root length) versus average molecular volume for each cluster.
Figure 8. Phytotoxicity−
molecular volume correlation. R 2, determination coefficient; P, statistical significance for correlation (variables correlated if P <
0.05).
Table 4. Electron Densities for Selected Atoms of 6-Substituted
Benzoxazinones
O (hydrox-
amic acid)
H (hydrox-
amic acid)
O
N-4
C-3
(carbonyl)
D-DIBOA
6-MeO-D-DIBOA
6-F-D-DIBOA
6-Cl-D-DIBOA
6-CF3-D-DIBOA
6-MeOCO-D-DIBOA
0.552
0.554
0.553
0.551
0.549
0.553
−0.259
−0.258
−0.260
−0.259
−0.260
−0.260
0.215
0.215
0.216
0.215
0.215
0.216
0.346
0.344
0.346
0.348
0.349
0.346
−0.773
−0.774
−0.772
−0.771
−0.770
−0.772
σ
0.002
0.001
0.001
0.002
0.001
Figure 9. Representative direction and module values of dipole moment
) for the tested chemicals.
population of the molecule was needed to correlate effects with
electronic parameters.
(
µ
Dipole moment (µ) is a frequently used parameter for QSAR
studies (30). It describes the overall polarity of the molecule,
which is closely related to transport phenomena and interactions
with target sites of action. Thus, the dipole moment modules
for all chemicals tested were calculated by using their three-
dimensional structures (Figure 9), with the purpose of correlat-
ing its magnitude with the phytotoxic effects observed.
Dipole moments were obtained from the three-dimensional
structures found for the chemicals by means of AM1 calcula-
tions. Compounds belonging to group G4 were discarded as
the effects on the activity provoked by steric factors could distort
the dipole moment-activity correlation. Moreover, 7F-D-
DIBOA and 6-CF3-D-DIBOA data were not used for the
correlation as highly significant changes were observed for the
dipole moment orientation. From the point of view of its dipole
moment module, this chemical could belong to the group formed
by 6-F- and 6-Cl-D-DIBOA, but its orientation is more similar
to the lower activity and higher dipole moment module
chemicals (Figure 10).
Taking into consideration the approaches made above, IC50
data were correlated with dipole moment modules (Figure 11).
According to this, slight variation in dipole moment in the region

9850 J. Agric. Food Chem., Vol. 54, No. 26, 2006
Mac´ıas et al.
The correlations observed for topological and electronic
QSAR descriptors show a strong dependence of activity with
volume and dipole moment and are closely related to polariz-
ability. Taking into consideration the results shown above, the
optimal value ranges for these parameters are
parameter
optimal value
molecular volume (V)
470
2.0
16.3
−
520 Å3
−2.5 Db
dipole moment (
µ
)
polarizability (p)
−
17.5 Å3
Figure 10. Directions of dipole moments for selected chemicals.
from µ ) 2 to µ ) 2.5 (G1 and first cluster of G2) has a great
influence on activity. In addition to this, if polarizability (p)
calculations are performed for each of the chemicals, the most
active are the less polarizable ones, the ones in which
transformations in dipole moments caused by the electronic
environment are lower. The average value of this parameter for
each of the clusters is shown in Figure 12.
The overall charge distribution in the benzoxazinones tested
is a fundamental parameter to the understanding of their
phytotoxic effects. When discarding the effects of the higher
volume molecules, a quadratic dependence of activity with
dipole moment modules is observed, the most active chemicals
sharing a similar dipole moment orientation pattern. The changes
provoked by the electronic environment to the dipole moments,
quantified by means of polarity calculations, are relevant for
the most active chemicals, as the three most active groups show
similar values for this parameter.
Minimum values for the three parameters favored phytotoxic
activity. Including selectivity comparisons, the best leads for
natural herbicide model development are the chemicals placed
in groups G1 and G2. 6-Chloro-4-hydroxy-(2H)-1,4-benzoxazin-
3(4H)-one (6-Cl-D-DIBOA) and 6-fluoro-4-hydroxy-(2H)-1,4-
benzoxazin-3(4H)-one (6-F-D-DIBOA) (G1) were the most
phytotoxic chemicals and fitted the optimal structure require-
ments in terms of volume and dipole moment. 7-Fluoro-4-
hydroxy-(2H)-1,4-benzoxazin-3(4H)-one (7F-D-DIBOA) was
the most selective chemical toward A. fatua and fitted the
optimal values for the three parameters, although the differences
of overall electronic distribution, caused by changes in dipole
moment orientation, probably affect its phytotoxicity.
These results, in combination with those obtained in the
previous studies on structure-activity relationships for ben-
zoxazinones and related compounds (2-5), provide an overview
Figure 11. Phytotoxicity−
dipole moment correlation. R 2, determination coefficient; P, statistical significance for correlation (variables correlated if P <
0.05).
Figure 12. Average polarizabilities (p, Å3) for each of the clusters.

Herbicide Models from Benzoxazinones
J. Agric. Food Chem., Vol. 54, No. 26, 2006 9851
(14) Tanimura, K.; Kosegi, K.; Shimazu, H.; Yaginuma, H. 2H-1,4-
Benzoxazin-3(4H)-one derivatives as antifungal drugs. Patent
JP63185970, 1988.
(15) O¨ zden, S.; O¨ zturk, A. M.; Go¨ker, H.; Altanlar, N. Synthesis and
antimicrobial activity of some new 4-hydroxy-2H-1,4-benzox-
azin-3(4H)-ones. Il Farmaco 2000, 55, 715-718.
(16) Cavill, G. W. K.; Ford, D. L. The chemistry of plant-growth
regulators. I. 2,4-Dichloro-6-hydroxyphenoxyacetic acid and
related compounds. J. Chem. Soc. 1954, 565-568.
(17) Hodel, E. Fungicides. Patent DE1161080, 1964.
(18) Enomoto, M.; Nagano, E.; Haga, T.; Morita, K.; Sato, R.
Benzoxazinyltetrahydrobenzotriazole oxides, their production and
use as herbicides (Sumitomo Chemical Co., Ltd., Japan). Eur.
Pat. Appl., 1989; 21 pp.
(19) Uekawa, T.; Takemura, S.; Enomoto, M.; Sakaki, M.; Sato, R.;
Nagano, E. Benzoxazinyl-substituted pyridone derivatives, and
their production and use as herbicides (Sumitomo Chemical Co.,
Ltd., Japan). Eur. Pat. Appl., 1992; 29 pp.
(20) U.S. Environmental Protection Agency, Office of Prevention,
Pesticides and Toxic Substances. Pesticide Fact Sheet, http://
www.epa.gov/opprd001/factsheets/flumioxazin.pdf.
(21) Macias, 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-
2521.
(22) Hartenstein, H.; Sicker, D. An efficient synthesis of 5-methoxy-
2-nitrophenol. J. Prakt. Chem./Chem. Zeit. 1993, 335, 103-104.
(23) Castellano, D.; Mac´ıas, F. A.; Castellano, M.; Cambronero, R.
M. FITOMED (automated system for measurement of variable
lengths). Sp. Patent P9901565.
(24) Statistica for Windows 5.0; Statsoft Inc.: Tulsa, OK, 1984-
1995.
(25) PRISMA 4.0; GraphPad Software, Inc.: San Diego, CA, 2003.
(26) HYPERCHEM 7.01; Hypercube Inc.: Gainesville, FL, 2000.
(27) Microsoft Office Excel 2003; Microsoft Corp.: Redmond, WA,
1985-2003.
of the thermodynamic, steric, electronic, and functionalization
requirements for benzoxazinones’ phytotoxic effects.
Supporting Information Available: Physical data for all
chemicals tested, phytotoxic activity values for all chemicals
and all plant species, and complete tables for all calculated
QSAR parameters. This material is available free of charge via
the Internet at http://pubs.acs.org.
LITERATURE CITED
(1) Rimando, A. M.; Duke, S. O. Natural products for pest
management. ACS Symp. Ser. 2006, No. 927, 2-21.
(2) Mac´ıas, F. A.; Mar´ın, 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 standard target species (STS).
J. Agric. Food Chem. 2005, 53, 538-548.
(3) Mac´ıas, F. A.; Chinchilla, N.; Varela, R. M.; Oliveros-Bastidas,
A.; Mar´ın, 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.
2005, 53, 4373-4380.
(4) Mac´ıas, F. A.; Mar´ın, D.; Oliveros-Bastidas, A.; Castellano, D.;
Simonet, A. M.; Molinillo, J. M. G. Structure-activity (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,
1040-1048.
(5) Mac´ıas, F. A.; Mar´ın, D.; Oliveros-Bastidas, A.; Molinillo, J.
M. G. Optimization of benzoxazinones as natural herbicide
models by lipophilicity enhancement. J. Agric. Food Chem. 2006,
54, 9357-9365.
(6) Niemeyer, H. M.; Pesel, E.; Franke, S.; Francke, W. Ingestion
of the benzoxazinone DIMBOA from wheat plants by aphids.
Phytochemistry 1989, 28, 2307-2310.
(7) Le-Van, N.; Wratten, S. J. Compound 30.4, an unusual chlori-
nated 1,4-benzoxazin-3-one derivative from corn (Zea mays).
Tetrahedron Lett. 1984, 25, 145-148.
(8) Geng, X.; Li, K.; Xu, F. Investigation of hydroxamic acids as
laccase-mediators for pulp bleaching. Appl. Microbiol. Biotech-
nol. 2004, 64, 493-496.
(9) Quiroz, A.; Niemeyer, H. M. Reaction of 7-substituted 4-hy-
droxyl-1,4-benzoxazin-3-ones in strongly acidic media. Hetero-
cycles 1991, 32, 1681-1685.
(10) Mac´ıas, F. A.; Mar´ın, D.; Oliveros-Bastidas, A.; Chinchilla, D.;
Simonet, A. M.; Molinillo, J. M. G. Isolation and synthesis of
benzoxazinones and related compounds. J. Agric. Food Chem.
2006, 54, 991-1000.
(28) SigmaStat 3.1; Systat Software Inc.: San Jose, CA, 2004.
(29) Mac´ıas, F. A.; Molinillo, J. M. G.; Mar´ın, D.; Varela, R. M.;
Chinchilla, N.; De Siqueira, J. M. Halogenated derivatives of
benzoxazinones with phytotoxic activity. Sp. Patent P200602033.
(30) Karelson, M.; Lobanov, V. S.; Katritzky, A. R. Quantum-
chemical descriptors in QSAR/QSPR studies. Chem. ReV. 1996,
96, 1027-1043.
Received for review September 21, 2006. Revised manuscript received
October 25, 2006. Accepted October 25, 2006. This research was
financially supported by the Spanish Ministry of Science and Technol-
ogy (Project AGL2006-10570/AGR). Fellowships from CNPq (Brazil,
J.M.S.) and the Spanish Ministry of Science and Technology (N.C.)
and a contract from the Research, Development and Innovation Council
(Junta de Andaluc´ıa, Spain, D.M.) are also gratefully acknowledged.
(11) Bravo, H. R.; Lazo, W. J. Antialgal and antifungal activity of
natural hydroxamic acids and related compounds. J. Agric. Food
Chem. 1996, 44, 1569-1571.
(12) Bravo, H. R.; Lazo, W. Antimicrobial activity of cereal hydrox-
amic acids and related compounds. Phytochemistry 1993, 33,
569-571.
(13) Hashimoto, Y.; Ishizaki, T.; Shudo, K.; Okamoto, T. Rearrange-
ment of 4-acetoxy-2H-1,4-benzoxazin-3(4H)-one. Chem. Pharm.
Bull. 1983, 31, 3891-3896.
JF062709G