Synthesis, Herbicidal Evaluation, and Structure-Activity Relationship of Benzophenone Oxime Ether Derivatives

A novel series of benzophenone oxime ether derivatives with tertiary amine groups were synthesized and their herbicidal activities of 24 compounds againstOryza sativa, Sorghum sudanense, Brassica chinensis, andAmaranthusmangostanus L. were also evaluated. Most of these compounds exhibited significant inhibitory effect on root growth at 20 ppm. Based on the herbicidal activity data, computational Three-Dimensional Quantitative Structure-Activity Relationship (3D-QSAR) analysis and molecular docking were undertaken. CoMFA contour maps were generated for the design of benzophenone oxime ether analogues with enhancing activity.


Introduction
Pesticides are extensively used in agriculture and undoubtedly play a pivotal role in retaining high production and quality of crop [1].Nevertheless, with the escalating demand of environment protection and food safety, the economic, health, and environmental costs of pesticides have to be considered [2].This requires that pesticides must be selective, effective, of less residue, and safe.On the other hand, most pesticides could lose efficacy due to resistance after long-term usage.Thus, continuous effort has been made to develop new pesticides with broad spectrum, low dosage and cost, long efficacy duration, less environmental pollution, and safety to humans [3,4].
However, very few results on the herbicidal activities of oxime ether compound were reported [19].In this paper, a series of aryl oxime ether compounds with tertiary amino groups which are conducive to enhancing environmental compatibility [17,20] were designed and synthesized.The herbicidal activities of these compounds on Oryza sativa, Sorghum sudanense, Brassica chinensis, and Amaranthus mangostanus L. were investigated.Computational Three-Dimensional Quantitative Structure-Activity Relationship (3D-QSAR) analysis and molecular docking were also undertaken.

General Procedure for the Synthesis of Substituted Benzophenones.
Benzoyl chloride (0.12 mol) was added dropwise to a mixture of substituted benzene (0.1 mol) and AlCl 3 (0.12 mol) at room temperature.After complete addition, the mixture was refluxed for another 6 h.Then, the reaction was quenched with ice-water and extracted with chloroform (3 × 30 mL).The combined organic phase was washed with water, Na 2 CO 3 solution, and brine successively, dried over MgSO 4 , filtered, and evaporated.The crude product was used directly for next step without any further purification.

General Procedure for the Synthesis of Substituted
Benzophenone Oximes.Aqueous NaOH solution (10 mL, 20 M) was added to a mixture of substituted benzophenone (0.05 mol) and hydroxylamine hydrochloride (0.1 mol) in ethanol (50 mL) at room temperature.Then, the reaction was heated to 75 ∘ C and stirred until the starting material was completely consumed as indicated by TLC.The mixture was cooled to room temperature and filtered.The filtrate was evaporated and dissolved in chloroform, washed with water.
The crude compound was recrystallized by ethanol to give a white solid [21].

General Procedure for the Synthesis of Substituted Benzophenone-O-(2-bromoethyl)
Oxime.Aqueous NaOH solution (10 mL, 25 M) was added to a mixture of substituted benzophenone oxime (0.025 mol), 1,2-dibromoethane (0.03 mol), and tert-butylammonium bromide (0.2 g) in toluene (20 mL).The mixture was stirred at room temperature and monitored by TLC.After the reaction was completed, the resulting solution was separated.The organic phase was washed with water until neutral, dried over MgSO 4 , and filtered.The filtrate was evaporated and purified by column chromatography on silica gel to give substituted benzophenone-O-(2-bromoethyl) oxime [22].

General Procedure for the Synthesis of Target Compounds.
A mixture of substituted benzophenone-O-(2-bromoethyl) oxime (0.01 mol), amine (0.012 mol), and Na 2 CO 3 (0.01 mol) in acetone (20 mL) was stirred at room temperature.The reaction was monitored by TLC.After the reaction was completed, the resulting suspension was filtered.The filtrate was evaporated to remove acetone and the residue was dissolved in EtOAc (20 mL), washed with water, dried over MgSO 4 , filtered, and evaporated.The crude product was purified by column chromatography on silica gel [23].
2.3.Biological Testing.The in vivo herbicidal activities of compounds 5a-5x against monocotyledon (Oryza sativa, Sorghum sudanense) and dicotyledon (Brassica chinensis, Amaranthus mangostanus L.) were examined according to the plating method in agricultural industry standard of China pesticide indoor bioassay test criteria (herbicide) (NY/T 1155.1,2006).The weeds were soaked in 25 ∘ C water for 12 h and then transferred to moist gauze, which was put in manual climatic box at 28 ∘ C to germinate.0.12 g of the target compounds was diluted with 3 mL DMF.0.5 mL of the diluted solution was diluted to 100 mL with 0.1% aqueous Tween-80 solution to prepare a mother solution (200 g/L).The test solutions were then prepared by diluting suitable mother solution to 100 mL with 0.1% aqueous Tween-80 solution.10 germinating seeds were selected and put in the Petri dish matted with a filtered paper, to which 9 mL of testing solution was added.0.1% aqueous Tween-80 solution was used as blank control.The Petri dishes were put in manual climatic box setting temperature as 25 ∘ C and humidity as 98% in dark condition.The root length was measured after 5 days and the inhibitory rate was calculated with the following equation.All the samples were repeated for 3 times.The abnormal data was got rid of by SPSS19.0.Consider is growth inhibitory rate (%);  0 is the root length of control;  1 is the root length of testing sample.
A coordinate plot was built with concentration as  axial and inhibitory rate as  axial.Linear fitting was processed for each compound to give the univariate linear regression equation (the correlation data was attached in the supporting information).IC 90 was calculated by the linear equation.pIC 90 was the value of log IC 90 .

Molecular Modeling.
The analysis is based on SYBYL 7.3 (Tripos Inc., USA).A training set of 19 compounds was used to construct the 3D-QSAR models.Considering the distribution of the structural diversity, 5 compounds were randomly selected as prediction test set to evaluate the obtained 3D-QSAR model.
Since the acceptor is unknown, the low energy conformation was chosen as the active conformation [24,25].Molecules are minimized by Tripos force field and the Powell conjugate gradient algorithm with a convergence criterion of 0.05 kcal/(mol Å) to give the low energy conformation, and then all compounds were aligned by Database Alignment method.Compound 5k was selected as template molecule [26].Other molecules are overlapped to reduce the rootmean-square derivation.The common skeleton was shown in Figure 1.
The composite data and pIC 90 data were imported to calculate the CoMFA field parameters by Tripos Standard force field (the steric and electrostatic field).Default parameters such as dielectric constant were used to obtain the molecular force parameters [27,28].Partial Least-Square (PLS) methodology was employed to construct the relationships between the target compound and biological activity.The leave-oneout (LOO) cross-validation was performed to obtain the cross-validation correlation coefficient ( 2 ) and the optimum number of components ().A non-cross-validated (NV) analysis was carried out to obtain the non-cross-validated correlation coefficient ( 2 ) and  test values and standard error of estimate (SEE).

Results and Discussion
3.1.Herbicidal Activity Evaluation.After a preliminary study, the concentration of testing compounds was set as 1, 2.5, 5, 10, 20, and 40 mg/L in the herbicidal activity investigation.The root growth was observed and the collected data was analyzed by SPSS to calculate the inhibitory rate.The results showed that the inhibitory effect became stronger as the concentration was increased and most of these compounds' inhibitory rates were up to 80% above at 20 mg/L (see Supporting Information).It indicated that these oxime ether compounds with tertiary amines significantly inhibited the crop root growth, which is another example to support the point that nitrogen functional groups are significant for the biological activity in most agrochemicals [29].IC 90 , the concentration inhibiting 90% of activity, was calculated according to the correlation analysis between the inhibitory rate on root growth and the concentration of the testing solution (see Supporting Information).Compound 5f exhibited the best herbicidal activity against Oryza sativa with an IC 90 of 13.70 mg/L.Compound 5o effectively inhibited the root growth of Sorghum sudanense with an IC 90 of 11.27 mg/L.Compounds 5k and 5i showed the highest herbicidal activity against Brassica chinensis and Amaranthus mangostanus L. with an IC 90 of 16.83 and 11.22 mg/L, respectively.The herbicidal activities of these compounds are comparable to those of tribenuron, which is a commonly used commercial herbicide [30] (Table 2).
It was noteworthy that the benzophenone oxime ethers containing piperidine or piperazine moiety (5e-5o) showed significant inhibitory activities.Increase of the concentration of these kinds of compounds to 20 mg/L caused plants' fatality.The compounds containing oxygen atom on amino moiety (5p-5r) show less inhibitory effect than that of those without oxygen (5a-5o).In particular, incremental carboxylic acid group on piperidine ring (5s-5x) strongly reduced the herbicidal activities.The inhibitory effects of the testing compound on Oryza sativa were generally stronger than those of Sorghum sudanense in terms of monocotyledon, while the inhibitory effects on Amaranthus mangostanus L. were better than those of Brassica chinensis as for dicotyledon.

Comparative Molecular Field Analysis (CoMFA).
The CoMFA analysis results for Oryza sativa and Amaranthus mangostanus L. were listed in Table 3.The activities data of Sorghum sudanense and Brassica chinensis did not show satisfactory regularity and CoMFA model was not analyzed.The CoMFA model for Oryza sativa has a  2 value of 0.567 and  2 value of 0.997.It has an  value of 751.322 and an SEE value of 0.020.The CoMFA model for Amaranthus mangostanus L. has a  2 value of 0.621 and  2 value of 0.879.It has an  value of 53.325 and an SEE value of 0.185.According to the literature [31], the  2 ( 2 > 0.5) and  2 ( 2 > 0.6) values illustrated that the resulted models have good robustness and internal prediction ability.Both models revealed that the electrostatic field (55.4%, 51.9%) has a little bit more contribution than that of steric fields (44.6%, 48.1%).This indicated that the electronic effect of testing compounds structure has more influence on the inhibitory activities against the plants' roots.
The 3D-QSAR models established with the training set were further validated with the test set.The CoMFA model gave reasonable predictions of both training and test set compounds.The experimental activity and predicted activity of the compounds and their residuals are listed in Tables 4   and 5.In both models, most of the predicted pIC 90 values are pretty close to the corresponding experimental values.All the deviations are smaller than 1 log unit and the maximum deviation is only 0.321.

Contour Map Analysis.
The 3D contour maps were generated as scalar products of coefficients and standard deviation associated with each CoMFA.The active compound 5k was superimposed with the CoMFA contour maps for Oryza sativa and Amaranthus mangostanus L.
In Figure 2, yellow region near benzene ring shows that substituents at this position have unfavorable steric interaction.This is verified by the observation that compound 5h with o-Cl on benzene exhibited less inhibition than that of 5f with p-Cl.Blue region below the chain especially around the amino groups indicates that electronegative groups are unfavored in this region.This is rightly explained by the fact that compounds 5s-5x with extra carboxylic acid group The compound marked with * was selected for test set.
80.000 20.000 showed worse inhibitory effect than those of 5f and 5g.The inhibitory activity on Oryza sativa may increase if electronpositive groups were introduced.In Figure 3, green region near benzene ring and blue region near amino groups manifested that increasing the size of substituent on benzene ring and importing electronpositive group on tertiary amino part are beneficial for The compound marked with * was selected for test set.
80.000 20.000 enhancing the inhibitory effect on Amaranthus mangostanus L. Actually, compounds 5a-5o showed stronger inhibitory effect than that of 5p-5r with morpholine tail and 5s-5x with carboxylic acid on piperidine ring.On the whole, the structural insights obtained from molecular docking and 3D-QSAR contour maps are consistent with the experimental data, indicating that the molecular docking and the developed 3D-QSAR models are reliable to some extent.The 3D-QSAR contour maps show that the electronic effect contributes to the strong herbicidal activity.Introduction of electron-positive group to the amino group of the test compounds may facilitate improving the inhibitory effect on Oryza sativa and Amaranthus mangostanus.

Conclusions
A series of benzophenone oxime ether derivatives with tertiary amines were synthesized and characterized.These compounds exhibited good herbicidal activities to both monocotyledon and dicotyledon.Based on the experimental results, the combined 3D-QSAR modeling and molecular docking analysis was performed with the data of Oryza sativa and Amaranthus mangostanus L. Due to the insufficient compounds structure diversity and variable influence factor, the models were assumed to be predictive but not perfectly reliable.However, these results threw a light on the research and development on the oxime ethers with amino groups to be herbicides.

Table 1 :
Chemical structure of the target compounds 5a-5x.

Table 4 :
Observed and predicted activities for training and test set of Oryza sativa CoMFA model.

Table 5 :
Observed and predicted activities for training and test set of Amaranthus mangostanus L. CoMFA model.