Design and Synthesis of N-phenyl Phthalimides as Potent Protoporphyrinogen Oxidase Inhibitors

Protoporphyrinogen oxidase (PPO) has been identified as one of the most promising targets for herbicide discovery. A series of novel phthalimide derivatives were designed by molecular docking studies targeting the crystal structure of mitochondrial PPO from tobacco (mtPPO, PDB: 1SEZ) by using Flumioxazin as a lead, after which the derivatives were synthesized and characterized, and their herbicidal activities were subsequently evaluated. The herbicidal bioassay results showed that compounds such as 3a (2-(4-bromo-2,6-difluorophenyl) isoindoline-1,3-dione), 3d (methyl 2-(4-chloro-1,3-dioxoisoindolin-2-yl)-5-fluorobenzoate), 3g (4-chloro-2-(5-methylisoxazol-3-yl) isoindoline-1,3-dione), 3j (4-chloro-2-(thiophen-2-ylmethyl) isoindoline-1,3-dione) and 3r (2-(4-bromo-2,6-difluorophenyl)-4-fluoroisoindoline-1,3-dione) had good herbicidal activities; among them, 3a showed excellent herbicidal efficacy against A. retroflexus and B. campestris via the small cup method and via pre-emergence and post-emergence spray treatments. The efficacy was comparable to that of the commercial herbicides Flumioxazin, Atrazine, and Chlortoluron. Further, the enzyme activity assay results suggest that the mode of action of compound 3a involves the inhibition of the PPO enzyme, and 3a showed better inhibitory activity against PPO than did Flumioxazin. These results indicate that our molecular design strategy contributes to the development of novel promising PPO inhibitors.

Among the protox herbicides, N-phenyl phthalimides, which exhibit broad structural diversity [20], have attracted considerable attention; their representative commercial products, Cinidon-ethyl, Flumiclorac-pentyl and Flumioxazin, were identified as a result of a stepwise optimization procedure from Chlorphthalim. Their common structural feature consists of two parts: An N-substituted phenyl group and tetrahydrophthalimide, which can interact with key active centre residues of the or π-π stacking interactions [21]. These findings have contributed to the development of PPO inhibitors.
To obtain novel PPO inhibitors, a series of N-phenyl phthalimides were designed by molecular docking using the mtPPO as a target, and by using Flumioxazin as a lead (Figure 1) [22,23], after which they were synthesized and characterized by NMR and High resolution mass spectrometry (HR-MS). Their herbicidal activities were also evaluated against Brassica campestris (B. campestris), Amaranthus retroflexus (A. retroflexus) and Digitaria sanguinalis (D. sanguinalis) to verify our molecular design strategy.

Docking Analysis
The affinity values of molecular docking between the ligand-mtPPO complexes were determined, which demonstrated that the affinity between 3a, 3c and 3s and mtPPO exhibited the highest values with −10.0, −10.1, 10.2 Kcal/mol (Table 1), respectively, suggesting that these compounds may have good herbicidal activity. Their docking models were displayed by Pymol software (Figure 2), which indicated that no hydrogen bonding was found between these compounds and mtPPO, which differs from the results of Flumioxazin (three hydrogen bonds). However, the phthalimide rings of compounds 3a, 3c and 3s shape π-π stacking interactions with Phe392, which is conserved in plant PPO enzymes. The phenyl rings of compounds 3a, 3c and 3s were sandwiched by the residues Leu356 and Leu372. These results suggest that these compounds, with a phenyl ring substituting phthalimide rings, could be PPO inhibitors that exhibit good herbicidal activity.

Docking Analysis
The affinity values of molecular docking between the ligand-mtPPO complexes were determined, which demonstrated that the affinity between 3a, 3c and 3s and mtPPO exhibited the highest values with −10.0, −10.1, 10.2 Kcal/mol (Table 1), respectively, suggesting that these compounds may have good herbicidal activity. Their docking models were displayed by Pymol software (Figure 2), which indicated that no hydrogen bonding was found between these compounds and mtPPO, which differs from the results of Flumioxazin (three hydrogen bonds). However, the phthalimide rings of compounds 3a, 3c and 3s shape π-π stacking interactions with Phe392, which is conserved in plant PPO enzymes. The phenyl rings of compounds 3a, 3c and 3s were sandwiched by the residues Leu356 and Leu372. These results suggest that these compounds, with a phenyl ring substituting phthalimide rings, could be PPO inhibitors that exhibit good herbicidal activity.

Chemistry
The starting material 1 and 2 could be commercially available. Compounds 3 were prepared by nucleophilic substitution reaction between phthalic anhydrides 1 and amines 2 in glacial acetic acid, with yields ranging from 28% to 87% ( Table 2). The reaction solvent and temperature were the key conditions for the reaction yield. The yield of the reaction stirred at 110 °C was much better than that of 80 °C. Besides, the yield of the reaction in glacial acetic acid was higher than that of other solvents, such as ethanol.

Chemistry
The starting material 1 and 2 could be commercially available. Compounds 3 were prepared by nucleophilic substitution reaction between phthalic anhydrides 1 and amines 2 in glacial acetic acid, with yields ranging from 28% to 87% ( Table 2). The reaction solvent and temperature were the key conditions for the reaction yield. The yield of the reaction stirred at 110 • C was much better than that of 80 • C. Besides, the yield of the reaction in glacial acetic acid was higher than that of other solvents, such as ethanol.

Herbicidal Activity
The herbicidal activity against B. campestris, A. retroflexus and D. sanguinalis at 200 mg/L of compound 3 were evaluated by the small cup method, and the results are shown in Table 3. The data showed that 3a at 200 mg/L displayed 92% growth inhibition against B. campestris roots, which was better than that of the positive controls Chlortoluron (85%), Atrazine (80%), and Flumioxazin (85%); 3a also showed 61% growth inhibition against stems of B. campestris, which was better than that of Atrazine (51%) and similar to that of Chlortoluron (58%). During the test, the B. campestris plants treated with 3a became chlorotic, and blade yellowing was also observed. In addition, with 87% growth inhibition, 3a exhibited better efficacy against A. retroflexus stems than did Atrazine, which was similar to that of Flumioxazin (88%); the leaves of A. retroflexus were wrinkled. Furthermore, 3a displayed 68% and 83% growth inhibition against the roots and stems of D. sanguinalis, respectively, which were better than the inhibition due to Atrazine. Compounds 3d and 3o exhibited high activity against the roots of A. retroflexus, with 81% and 83% growth inhibition, respectively, while Atrazine showed only 32% inhibition under the same conditions; 3o also showed 89% growth inhibition activity against B. campestris roots, which was slightly better than that of the three positive controls. 3g showed excellent efficacy against D. sanguinalis roots and stems, with 91% and 83% growth inhibition, respectively, these percentages were better than those due to Atrazine. The results of the evaluation of the post-emergence herbicidal activities are shown in Table 4. The data showed that compounds 3a and 3d exhibited 82% and 73% fresh weight growth inhibition at 90 g ai/ha against A. retroflexus, respectively. Therefore, 3a and 3d were also chosen for further pre-emergence herbicidal activity tests. As shown in Table 5, 3a at 90 g ai/ha exhibited a 98% inhibitory effect against A. retroflexus, and the effects were not significantly different from those of Flumioxazin.  In summary, 3a exhibited excellent herbicidal activity and should be further developed. Moreover, its herbicidal activity conformed to the predictions of the molecular docking studies.

Crystal Structure Determination of Compound 3a
The X-ray diffraction structure of compound 3a, cultured from the mixture of ethanol and chloroform, was shown in Figure 3, and the data were available at the Cambridge Crystallographic Data Centre (CCDC 1923372).

Crystal Structure Determination of Compound 3a
The X-ray diffraction structure of compound 3a, cultured from the mixture of ethanol and chloroform, was shown in Figure 3, and the data were available at the Cambridge Crystallographic Data Centre (CCDC 1923372).

PPO Enzyme Assays
The PPO enzyme activity of D. sanguinalis, treated with 3a and Flumioxazin at 90 g ai/ha, were measured, and the control was sprayed with blank solution without any compounds. The results, shown in Figure 4, indicated that the PPO activity values of the 3a treatment and the control Flumioxazin were affected; in addition, 3a showed a stronger effect on the PPO enzyme, with 33.01%, compared to that of the positive control Flumioxazin, with 21.80%.

PPO Enzyme Assays
The PPO enzyme activity of D. sanguinalis, treated with 3a and Flumioxazin at 90 g ai/ha, were measured, and the control was sprayed with blank solution without any compounds. The results, shown in Figure 4, indicated that the PPO activity values of the 3a treatment and the control Flumioxazin were affected; in addition, 3a showed a stronger effect on the PPO enzyme, with 33.01%, compared to that of the positive control Flumioxazin, with 21.80%.

Molecular Docking
The structure of mtPPO was available at the National Center for Biotechnology Information (NCBI) database. The modelled complexes of mtPPO and ligands (Flumioxazin, Chlorotoluron, Atrazine and designed target compounds) were prepared by using AutoDockTools 1.5.6 (Molecular Graphics Laboratory, La Jolla, CA 92037-1000, USA) and analysed via the AutoDock Vina program (Molecular Graphics Laboratory, La Jolla, CA 92037-1000, USA) [24,25]. The structures of the complexes were generated by the Pymol tool 2.2.0 [26,27].

Equipment and Materials
The melting points of the new compounds were measured in a microfusion melting point apparatus (X-4) (Taike, Beijing, China) and uncorrected. 1 H-NMR and 13 C-NMR spectra were recorded on Varian 400 spectrometer at 400 MHz and 101 MHz using tetramethylsilane as internal standard (solvent CDCl3 or DMSO-d6). HR-MS date were detected on an FTICR-MS Varian 7.0T FTICR-MS equipment (Agilent, Lexington, MA, USA). Crystal structure was recorded on a Bruker SMART 1000CCD diffraction meter.

Molecular Docking
The structure of mtPPO was available at the National Center for Biotechnology Information (NCBI) database. The modelled complexes of mtPPO and ligands (Flumioxazin, Chlorotoluron, Atrazine and designed target compounds) were prepared by using AutoDockTools 1.5.6 (Molecular Graphics Laboratory, La Jolla, CA 92037-1000, USA) and analysed via the AutoDock Vina program (Molecular Graphics Laboratory, La Jolla, CA 92037-1000, USA) [24,25]. The structures of the complexes were generated by the Pymol tool 2.2.0 [26,27].

Equipment and Materials
The melting points of the new compounds were measured in a microfusion melting point apparatus (X-4) (Taike, Beijing, China) and uncorrected. 1 H-NMR and 13 C-NMR spectra were recorded on Varian 400 spectrometer at 400 MHz and 101 MHz using tetramethylsilane as internal standard (solvent CDCl 3 or DMSO-d 6 ). HR-MS date were detected on an FTICR-MS Varian 7.0T FTICR-MS equipment (Agilent, Lexington, MA, USA). Crystal structure was recorded on a Bruker SMART 1000CCD diffraction meter.

General Synthetic Procedure for Compounds 3
The starting materials 1 and 2 were commercially available (Energy Chemical, Shanghai, China). Compound 2 (3.72 mmol) was added to a stirred solution of compound 1 (3.38 mmol) in glacial acetic acid (10 mL). The reaction mixture was then stirred at 110 • C for 4 h. After completion of the reaction, the solvent was evaporated, and the residue was purified on a silica gel column chromatography and eluted with ethyl acetate/petroleum ether (bp 60-90 • C) (1:3, v/v) to give compounds 3. Among them, compounds 3k, 3l, 3n, 3x and 3y were published [28][29][30][31][32], while their herbicidal activities were not studied. Their yields, physical properties, 1 H-NMR, 13  the post-emergence method. After 3 days, D. sanguinalis leaves (0.2 g) were collected and dissolved in extraction medium (1.5 mL) in an ice bath, and the mixture was then centrifuged at 4 • C × 12,000 rpm for 15 min. A polyphenol oxidase kit (G0113W, Suzhou Grace Bio-technology Co., Ltd., Suzhou, China) was obtained from commercial sources to determine the PPO activity. The change in absorbance (every 5 min) was measured on a POLARstar Optima/Galaxy instrument (BMG) (Shanghai Microplate Co., Ltd., Shanghai, China) at 420 nm [37][38][39]. One unit of PPO activity was defined as a change in absorbance of 0.01 per minute.

Conclusions
In summary, a series of phthalimide derivatives were designed by molecular docking and by using Flumioxazin as a lead, then synthesized and characterized by NMR, HR-MS, and the typical crystal structure was determined by X-ray diffraction. The herbicidal activities of these compounds were assessed against B. campestris, A. retroflexus and D. sanguinalis, by the small cup, pre-emergence, and post-emergence methods, respectively. Most of the synthesized compounds exhibited good to excellent herbicidal activities, and especially 3a displayed the same efficacy against A. retroflexus and B. campestris to commercial standards of Flumioxazin. Further PPO activity assays confirmed that the mode of action of 3a is similar to PPO inhibitors. These results suggest that our molecular design strategy is effective.