2.1 Chemicals and instruments

All chemicals are purchased without any further purification from Aladdin (Aladdin Company, CHN). The raw materials needed for the reaction are directly used without further purification. Under the irradiation of ultraviolet lamp, the reaction was monitored on silica gel plate by thin layer chromatography. Melting point instrument (Shanghai INESA Optical Instrument Co., Ltd., China), without temperature correction. Using DMSO‑d₆ or CDCl₃ as solvent and TMS as internal standard, the ¹H and ¹³C NMR spectra of the compounds were recorded on a Bruker DPX-400 or 500 spectrometers. High-resolution mass spectrometry (HRMS) was performed by Thermo Scientific Q-Exactive (Thermo, USA). The seeds of all weeds are purchased commercially and disinfected with 75 % alcohol.

2.2 Virtual screening based on ligands

The polar head quinone is the key structure for the interaction between Q_B and protein. Therefore, quinone was used as ligand and conducted virtual screening based on ligand on BRUSELAS (https://bio-hpc.ucam.edu/Bruselas/) (Banegas-Luna et al., 2019). The calculation mode is similarity search, and the algorithm selects pharmacophore and 3D scanning. The scanned databases include DIA-DB (186 compounds), Drug Bank (1,760 compounds), CHEMBL (1,578,131 compounds) and KEGG (23,667 compounds) (Wishart et al., 2006; Gaulton et al., 2012; Kanehisa et al., 2016; Pereira et al., 2019). The scoring function takes a weighted average, and the similarity threshold is 0.6 for screening.

2.3 Virtual screening based on structures

The program PyRx was used for virtual screening of the obtained small molecules based on D1 protein structure. The crystal structure of the D1 protein (PDB: 5XNL from Pisum sativum) was acquired from the RCSB Protein Data Bank (Battaglino et al., 2021), and further treated with the addition of hydrogen atoms and removal of water molecules with PyMOL. The binding behavior of the top 20 small molecules was analyzed according to the default score. Meanwhile, the physical and chemical properties including logP, pKa, hydrogen bond acceptor (HBAs), hydrogen bond donor (HBDs), and aromatic ring (ARs) of compounds were calculated with Discovery Studio v3.5, and the electronegativity was predicted by SYBYL-X 2.0 (Zhao et al., 2023).

2.4 Binding mode-based ligand design

In order to discover more active quinone derivatives, the binding mode of lead structure and D1 was determined by LeDock compared with Q_B (Zhang and Zhao, 2016). According to the molecular interaction modes of Q_B and lead structure binding to D1 protein (Fig. 4), some active substructures were introduced to lead structure, and the binding model and interaction energy of each ligand binding with Q_B site of D1 protein was performed by LeDock and Autodock Vina, respectively. The binding models of high active compounds to D1 were performed using the same method.

Close

Fig. 4

Simulated surface representation of Atrazine (green) and Lawsone (yellow) binding to D1 protein (A). In PsPS II D1 protein, Q_B (green) and Lawsone (yellow) combined with surrounding amino acids (B).

Export to PPT

2.5 General procedures for the preparation of compounds 1a–1e

As shown in Fig. 5, the lead compound lawsone (1.72 mmol) reacted with K₂CO₃ (2.58 mmol) in N, N-dimethylformamide (DMF) solution for 1 h. Subsequently, chlorine reagent was added and react at 50℃ for 1–2 h. After the reaction was completely monitored by TLC, the distilled water was added, and the mixture was extracted with ethyl acetate (3 × 50  mL) and the organic phase was dried over anhydrous Na₂SO₄, and then purified by column chromatography eluted with ethyl acetate/petroleum ether (1:5, V/V) to give target compounds 1a–1e.

Close

Fig. 5

Synthesis route of the target compounds 1a − 1e.

Export to PPT

2.6 General procedures for the preparation of compounds 2a–2e, 3a–3e and 4a–4 k

As shown in Fig. 6, lawsone (1.44 mmol) and dry Et₃N (1.72 mmol) were added to the dry CH₂Cl₂, and then the aromatic sulfonyl chloride or acyl chloride (1.72 mmol) was added and stirred at 0 ℃ for 30 min, the mixture was allowed to reach room temperature and then stirred for 2 h. Then the solvent was removed under vacuum and the mixture was purified by column chromatography eluted with ethyl acetate/petroleum ether (1:10, V/V) to give target compounds 2a–2e, 3a–3e and 4a–4 k.

Close

Fig. 6

Synthesis route of the target compounds 2a − 2e, 3a − 3e and 4a − 4 k.

Export to PPT

2.7 Petri dish assay

The herbicidal activities of target compounds were evaluated on P. oleracea and E. crusgalli using the reported methods (Sun et al., 2020; Cao et al., 2017). The seeds used in the experiment were disinfected and sterilized, then kept in a constant temperature incubator until they germinated. Dissolve the compound with 1 mL of DMF and prepared the required solution with 0.1 % tween-80. Subsequently, the germinated seeds are placed in 12-well plates, with 4–6 seeds in each hole (two filter papers are placed in each hole to keep moisture). Immediately after that, 1 mL of solution with a concentration of 100 μg/mL or 10 μg/mL was added to each well. The blank control group was a mixed solution of DMF and 0.1 % Tween-80 and the lead structure lawsone and commercial herbicides atrazine were used as the positive control group. After added drugs, the 12-well plates were placed in a manual climate incubator, and were automatically circulated in the light for 16 h (27 ℃) and in the dark for 8 h (25 ℃), and the relative humidity was 75 %. Each experiment had three parallel groups, and each experiment was repeated three times. After 5 days of observation, the length of stems and roots of weed seeds were measured and calculated the inhibition rate with follows formula.

Inhibition rate (%) = (L_control–L_treatment)/L_control × 100 %.

2.8 Post-emergence herbicidal activity

The herbicidal activity of the target compounds against monocotyledonous grass (Digitaria sanguinalis, E. crusgalli, and Setaria viridis) and dicotyledonous grass (P. oleracea, A. retroflexus, and A. theophrasti) was performed with the previous method (Wang et al., 2005). The specific operation is to dissolve the compound in 1 mL DMF and dilute it to the required concentration with 0.1 % Tween −80. Lawsone and atrazine were used as the positive controls, and a mixture of DMF and 0.1 % Tween-80 was used as the blank control group. The sprouting weeds are planted in 8 cm plastic pots with two-thirds of which are organic soil. The compound is sprayed during the growth of monocotyledonous plants to the two-leaf stage and dicotyledonous plants to the three-leaf stage and cultivated in the greenhouse. After 15 days, the inhibition rate was visually measured compared with the blank control group. Each treatment was repeated three times.

2.9 Chlorophyll fluorescence measurements

The leaves of A. retroflexus were infiltrated with the 0.1 % Tween-80 solution of 3e and 4d with a concentration of 100 μM and kept in darkness for 30 min at room temperature. Then the initial fluorescence (F_O) and the maximum fluorescence (F_M) of the dark adaptation blade were measured with MINI-PAM-II fluorometer (Heinz Walz GmbH, Germany) under the setting of saturated pulse mode when the fluorescence value of the passing current reached a stable level. When the actinic light reached the steady-state fluorescence (F_S), a saturation pulse was given to determine the maximum fluorescence (F'_M) in the light-adapted leaves. Then the actinic light was removed, and the minimum fluorescence (F'_O) in the leaves adapted to light was measured with far-red light for 3 sec. Finally, the maximum quantum efficiency (F_V/F_M) and photochemical quenching coefficient (qP) of PS II are calculated with follows formula (F_M–F_O)/F_M and (F'_M−F_S)/(F'_M−F'_O), respectively (Chen et al., 2008; Thuillier et al., 2024).

2.10 Molecular dynamic simulation

The water molecules were removed from the PsPS II D1 protein (PDB: 5XNL), and then it was hydrogenated and its amino acid residues were corrected. The complexes of compound 4d, lawsone and atrazine with PsPS II D1 were performed with molecular dynamic simulation (MDS) by AMBER 18 (Adcock and McCammon, 2006; Yang et al., 2020). A water box was added around the ligand–protein complexes with a distance of 10 Å, and Cl^- and Na⁺ were introduced to balance the system to electric neutrality. After energy minimization, the ligand–protein system was heated to 300 K and the pressure was raised to 0.1 MPa. Eventually, the MDS of each complex was executed lasting 60000 ps.

3.1 Virtual screening

According to previous report (Battaglino et al., 2021), at the bottom of Q_B binding site, polar groups of His215, Ser264 and Phe265 form hydrogen bonds with carbonyl groups at the head of PQ, and Phe255, Ser264 and Leu271 in D1 participate in PQ binding and electron transfer in PS II. Therefore, on the basis of the structural characteristics of Q_B and the structure of quinones, we first conducted ligand-based virtual screening to screen thousands of molecules from the databases DIA-DB, DrugBank, ChEMBL, and KEGG, and selected the top 100 molecules. Secondly, the program PyRx was used to hydrogenate for the D1 (PDB: 5XNL) protein, remove water treatment, and carry out virtual screening based on protein structure to obtain the top 20 molecules in the score (Fig. 3). It can be seen that they are all of quinone derivatives and their affinity, physical and chemical properties were listed in Table 1. As is known that the physical and chemical properties of compounds affect their pharmacological efficacy and interaction with target enzymes (Zhao et al., 2023). From Table 1, we found that the compounds MW were within 136.15 to 287.11. Meanwhile, the logp, HBAs and ARs of these 20 compounds were similar. Compared with other compounds, lawsone, lapachol, α-aminoan-thraquinone, alizarin, chrysazin, and 1-hydroxy anthraquinone had HBDs, which indicated that they might form strong interactions with surrounding amino acids. In all the compounds, lawsone has the lowest logP and pKa values of 1.21 and 1.92, respectively, which are beneficial to the transmission and absorption of plants. Lawsone was identified as a promising lead structure for developing new PS Ⅱ inhibitors with the results of the pre-seedling herbicidal activity evaluation (Table 2).

Table 1 Physical and chemical properties of quinones.

compound	affinity (kcal/mol)	MWa	LogPa	pKaa	HBAsa	HBDsa	ARsa	electronegativityb
1	−6.7	158.15	1.59	2.75	2	0	1
2	−6.6	158.15	1.48	2.74	2	0	1
3	−6.6	172.18	1.98	3.35	2	0	1
4	−6.5	188.18	1.43	2.56	3	0	1
5	−6.4	174.15	1.21	1.92	3	1	1
6	−6.4	256.30	2.93	4.13	3	1	1
7	−6.3	242.27	2.42	2.31	3	0	1
8	−6.3	136.15	1.22	3.97	2	0	0
9	−6.3	164.20	1.85	3.02	2	0	0
10	−6.3	234.25	3.27	4.68	2	0	1
11	−6.2	223.23	2.30	3.50	2	1	2
12	−6.2	240.21	2.02	3.10	4	2	2
13	−6.2	240.21	2.04	3.09	4	2	2
14	−6.2	222.24	3.01	4.67	2	0	2
15	−6.2	208.21	2.64	4.15	2	0	2
16	−6.1	208.21	2.55	4.26	2	0	2
17	−6.1	224.21	2.39	3.54	3	1	2
18	−6.1	182.17	2.05	3.72	2	0	2
19	−6.1	287.11	3.08	4.90	2	0	2
20	−6.0	258.27	3.57	5.60	2	0	3

a The parameters were calculated by DS.

b The electronegativity was predicted by SYBYL-X 2.0.

3.2 Binding mode-based ligand design

PS II is a protein complex that includes many co-factors, of which, the hetero-dimer of D1 and D2 homologous protein is the core (Zabret et al., 2021). It is known that PS II herbicides interrupt the electron from Q_A to Q_B by binding the Q_B site between helix IV and V of D1 protein from Phe211 to Leu275 (Oettmeier, 1999). Based on the model of action between D1 protein and Q_B, the binding model of lawsone to D1 protein was performed by LeDock. The results indicated that lawsone can completely bind in the active pocket formed by Q_B (Fig. 4A). As well as Q_B, the two carbonyl and hydroxyl of lawsone form hydrogen-bond interaction to D1 protein with amino acids His 215 and His 252, respectively. Meanwhile, the quinone ring forms π-Alkyl action to amino acid Leu271. Surprisingly, the benzene ring of lawsone forms two π-Alkyl and π-π interactions with Leu218, Leu271, and Phe255 (Fig. 4B). It indicated that lawsone can inhibit electron transfer in PS II by occupying the binding site of Q_B. In order to discover the novel lawsone derivatives for the development of PS Ⅱ inhibitor herbicides, the aryl carboxylic acid substructures were introduced into the hydroxyl to design compounds 1a–1e, 2a–2e, 3a–3e, and 4a–4 k compounds on the basis of the model of action between D1 protein and lawsone. The results of the binding affinity of them showed that the order is 4a–4 k > 3a–3e > 2a–2e > 1a–1e, and stronger than the commercial PS Ⅱ inhibitor herbicide atrazine (Table 3). Among them, compound 4d has the strongest binding affinity with a value of 7.0 kcal/mol, which was obviously superior to lawsone (6.4 kcal/mol) and atrazine (5.4 kcal/mol). Theoretically, the better binding with D1 protein means more blocking electron transfer and the better herbicidal activity. However, the current conclusion is obtained by computer virtualization, and the actual results need to be verified by weeding activity tested.

3.3 Chemistry

The compounds 1a–1e, 2a–2e, 3a–3e and 4a–4k via binding mode-based ligand design were synthesized. As shown in Fig. 5, the target compounds 1a–1e was obtained by etherification of starting materials lawsone and chlorine reagents (directly purchased) with the assistance of K₂CO₃ in DMF, respectively. In addition, the lead structure lawsone was esterified with substituted sulfonyl chloride/acyl chloride to obtain the target compounds 2a–2e, 3a–3e and 4a–4 k, respectively (Fig. 6). The structures of the target compounds were confirmed by ¹H NMR, ¹³C NMR, and HRMS data, as listed in the Supporting Information.

3.4 Herbicidal activity

In order to further confirm the lead compound, the pre-seedling herbicidal activities of 20 molecules obtained by virtual screening were tested using Petri dish and the results are shown in Table 2. As can be seen from Table 2, when the concentration is 100 μg/mL, most of them inhibit the growth of P. oleracea roots by more than 80 %. Especially, compounds 1–6, 10, 15–17 exhibited 100 % inhibitory rate on P. oleracea, which was similar to atrazine. Meanwhile, compounds 3, 5, 16, 17 showed more than 80 % than inhibitory effect at the concentration of 10 μg/mL. It indicated that the inhibitory effect of roots were obviously better than stems. In addition, compounds 5, 8, 12, 15, 16 showed remarkable inhibitory (more than 60 %) effect on the root growth of E. crusgalli, which was as well as atrazine. Overall, lawsone (compound 5) showed excellent inhibitory activity on the root and stem growth of P. oleracea and E. crusgalli at 100 and 10 μg/mL. It proved that lawsone is a promising lead structure for the development of herbicides.

As shown in Table 3, compounds 1a–1d, 2a–2c, 2e, 3a, 3e, 4a–4d, 4f, 4i and 4 k showed 100 % inhibition on the root growth of P. oleracea at 100 μg/mL, which were similar to lawsone and atrazine. Meanwhile, compounds 1a, 3a, and 4d exhibited more than 90 % inhibitory effect on stem growth of P. oleracea, which were better than lawsone (64.9 %) and atrazine (63.6 %). Even at the concentration of 10 μg/mL, compounds 4a–4d had 100 % inhibitory activity against the growth roots of P. oleracea, better than lawsone (82.4 %) and atrazine (91.9 %). Only compound 4d showed more than 80 % inhibitory activity on the growth stem of P. oleracea. In addition, compounds 3e (82.4 %), 4d (91.3 %), and 4f (80.6 %) showed better inhibitory effects on root of E. crusgalli than lawsone (67.6 %) and atrazine (65.9 %) at 100 μg/mL (Fig. 7). It is worth noting that compound 4d (82.1 %) had better inhibitory effects on stem of E. crusgalli than lawsone (45.4 %) and atrazine (46.1 %). Furthermore, compound 4d showed 54.6 and 49.2 % inhibition on the roots and stems of E. crusgalli at 10 μg/mL, superior to lawsone (31.6 and 26.9 %) and atrazine (4.2 and 5.3 %). To sum up, compound 4d showed excellent inhibitory effects on P. oleracea and E. crusgalli. As shown in Table 4, compounds 3e and 4d showed 100 % post-emergence herbicidal activity against P. oleracea, A. retroflexus, and A. theophrasti at 450 g a.i./ha, which were similar to atrazine but significantly superior to lawsone (lower than 10 %). Compound 4d exhibited over 100 %, 100 % and 90 %, 70 % inhibitory effect on A. retroflexus and A. theophrasti at a dose of 300 and 150 g a.i./ha, respectively. It indicated that compound 4d can be used as pre- and post-emergence candidate herbicide.

Close

Fig. 7

Symptoms of Echinochloa crusgalli treated by compound 4d at 100 μg/mL, compared with the positive control (Atrazine), lead compound (Lawsone) and control check (CK).

Export to PPT

3.5 Structure-activity relationship

In terms of herbicidal activity data, compounds 3e and 4d have good herbicidal activity against P. oleracea, A. retroflexus and A. theophrasti. The structure–activity relationships were analyzed according to the inhibitory activities of compounds on the growth stem of E. crusgalli at 100 μg/mL. When X is sulfonyl or acyl, electron-donating aryl favor for inhibitory effect, such as 2b (R=4-CH₃-Phenyl, 51.3 %) > 2a (R=Phenyl, 46.1 %) > 2d (R=4-F-Phenyl, 41.7 %) > 2e (R=4-NO₂-Phenyl, 30.9 %) and 3b (R=4-OCH₃-Phenyl, 60.7 %) > 3a (R=Phenyl, 48.7 %) > 3c (R=4-F-Phenyl, 41.5 %). At the same time, the introduction of pyridyl and furyl is beneficial to the activity, and the introduction of thienyl is unfavorable to the activity, which is consistent with the orde 4c (R=pyridine-3-yl, 59.7 %) > 4a (R=furfuran-2-yl, 51.8 %) > 3a (R=Phenyl, 48.7 %) > 4b (R=thiophene-2-yl, 40.6 %). Among the compounds 4a–4d containing pyridinyl, the compound with 2-Cl-pyridine showed the better activity than others, such as 4d (R=2-Cl-pyridine-3-yl, 91.3 %) > 4 h (R=2-CF₃-pyridine-3-yl, 55.3 %) ≈ 4 g (R=2-F-pyridine-3-yl, 53.6 %). It indicated that introducing the chlorine-containing pyridine groups is an effective strategy to discover compounds with high herbicidal activity.

3.6 Chlorophyll fluorescence measurements

Environmental stress including thermal damage, photo-inhibition and others factors can lead to the structural change of PS II pigment level and then affect the F_O level (Krause and Weis, 1984). The F_O and F_M values were determined when all reaction centers of PS II are opened after 30 min of dark reaction and turned off by intense pulsed light illumination, respectively. As shown in Table 5, the ratio of F_V/F_M has no significant difference of compounds 3e, 4d, lawsone, and atrazine, which indicates that these compounds have no influence on the efficiency of light energy transfer between antenna pigment molecules or from these molecules to PS II reaction center. In addition, the qP value indicates the proportion of excitons captured by open traps and converted into chemical energy at the PS II reaction center, which depends on the presence of the oxidation state Q_A Chen et al., 2008; Krause and Weis, 1984) The qP values of compounds 3e (0.71), 4d (0.66) and atrazine (0.43) were significantly lower than CK (0.95), which indicated that they affected the efficiency of the entire photochemical process and the functional state of PS II. The present works demonstrated that the lawsone derivative 4d can be considered as a promising PS II inhibitor.

3.7 Molecular docking analysis

The molecular docking was performed and shown in Fig. 8. It can be seen from Fig. 8A, atrazine binds to Q_B binding site in D1 protein, and N of aromatic ring forms hydrogen bond with His215 (3.4 Å). Besides, aromatic rings form π-Alkyl hydrophobic interactions with Leu218 (4.5 Å) and Leu271 (4.5 Å). Lawsone can bind to the D1 protein (Fig. 8B), the two carbonyl groups form hydrogen bonds with His215 (3.0 Å) and His252 (3.0 Å) and the aromatic rings form one π-π and two π-Alkyl hydrophobic interactions with Phe255 (4.3 Å) and Leu271 (4.6 Å of both), respectively. As well as lawsone, the D1 protein was intensively bound with compound 4d (Fig. 8C), which forms two hydrogen bonds His215 (3.3 Å) and His252 (3.3 Å) with two carbonyl groups, respectively. Meanwhile, the aromatic rings of 4d form one π-π and three π-Alkyl hydrophobic interactions with Phe255 (4.5 Å), Leu 218 (5.1 Å), and Leu271 (4.6 and 4.8 Å), respectively. In addition, the pyridine ring of 4d forms a new π-π interaction on Phe211 (4.4 Å) with D1 protein.

Close

Fig. 8

Binding models of the commercial herbicide Atrazine (A), Lawsone (B) and compound 4d (C) with PsD1 (PDB: 5XNL). The purple line indicates hydrophobic interactions between compounds and surrounding residues. The red line indicates hydrophilic hydrogen-bond interactions.

Export to PPT

3.8 Molecular dynamic simulation

During the simulation of 60,000 picoseconds, the RMSD mode of protein and ligand was shown in Fig. 9. Obviously, at the beginning, D1 protein has some fluctuations of 0.8 Å displacement. The RMSDs fluctuates about 0.2 Å–1 Å around a certain thermal average during the whole simulation time, which accurately explains the equilibrium simulation. Lawsone maintains roughly the same RMSD levels as compound 4d, while atrazine had slightly higher mean levels than compound 4d. In general, the fluctuation of protein and ligand remains unchanged after 5000 picoseconds.

Close

Fig. 9

RMSD trajectories of D1-ligand complex during 60000 ps simulations.

Export to PPT