2.1. Chemistry
Synthesis routes. The cyclic condensation reaction of β-dicarbonyl compounds with hydrazine and hydroxylamine is one efficient method for synthesizing pyrazole and isoxazole derivatives [
28]. Based on this reaction as a key step, in this work, a three-step synthesis route to the monophosphonic acid compounds
5 and
6 was designed as shown in
Scheme 1. First, diethyl vinylphosphonate (
1) as a Michael acceptor reacted with β-ketoesters or β-diketones in the presence of K
2CO
3 and benzyltriethylammonium chloride (TEBAC) to yield the intermediates
2. In the second step,
2 was used to react with hydrazine or hydroxylamine, respectively, giving the corresponding cyclization intermediates
3 and
4. Finally, the ethyl protecting groups were removed with TMSBr to yield the target products
5 and
6 in 70~82% overall yields.
Scheme 1.
Synthesis of the monophosphonic acid series 5 and 6. Reagents and conditions: (a) RCOCH2COOEt or RCOCH2COR, K2CO3, TEBAC, CH3CN, 80℃; (b) N2H4·H2O, EtOH, 75℃; (c) NH2OH·HCl, K2CO3, EtOH, 75℃; (d) i) TMSBr, CH2Cl2, 0℃ to r.t.; ii) THF/H2O, r.t.
Scheme 1.
Synthesis of the monophosphonic acid series 5 and 6. Reagents and conditions: (a) RCOCH2COOEt or RCOCH2COR, K2CO3, TEBAC, CH3CN, 80℃; (b) N2H4·H2O, EtOH, 75℃; (c) NH2OH·HCl, K2CO3, EtOH, 75℃; (d) i) TMSBr, CH2Cl2, 0℃ to r.t.; ii) THF/H2O, r.t.
A similar synthesis route was used to access the bisphosphonic acid series
12 and
13, as shown in
Scheme 2. First, a Michael acceptor tetraethyl vinylidenebisphosphonate (VBP) was synthesized through the reaction of tetraethyl methylenebisphosphonate (
7) with paraformaldehyde in methanol and followed by dehydration with the catalysis of TsOH [
29]. Then, tetraethyl VBP (
8) reacted with β-ketoesters or β-diketones to give the intermediates
9 using lithium bis(trimethylsilyl)amide (LiHMDS) as base. Cyclic condensation with hydrazine or hydroxylamine followed by deprotection gave the target compounds
12 and
13 in 75%~88% overall yields.
Scheme 2.
Synthesis of the bisphosphonic acid series 12 and 13. Reagents and conditions: (a) i) (CHO)n, Et2NH, MeOH, 65℃; (ii) TsOH, toluene, 110℃; (b) RCOCH2COOEt or RCOCH2COR, LiHMDS, THF, 0℃ to r.t.; (c) N2H4·H2O, EtOH, 75℃; (d) NH2OH·HCl, K2CO3, EtOH, 75℃; (e) i) TMSBr, CH2Cl2, 0℃ to r.t.; ii) THF/H2O, r.t.
Scheme 2.
Synthesis of the bisphosphonic acid series 12 and 13. Reagents and conditions: (a) i) (CHO)n, Et2NH, MeOH, 65℃; (ii) TsOH, toluene, 110℃; (b) RCOCH2COOEt or RCOCH2COR, LiHMDS, THF, 0℃ to r.t.; (c) N2H4·H2O, EtOH, 75℃; (d) NH2OH·HCl, K2CO3, EtOH, 75℃; (e) i) TMSBr, CH2Cl2, 0℃ to r.t.; ii) THF/H2O, r.t.
The synthesis route for isoxazolinone-containing compounds
19 and
20 is different from the previous ones and the Michaelis-Becker reaction as a key step was used, as shown in
Scheme 3. In this route, with 3-chloropivaloyl chloride (
14) as the starting material, and through the amidation with hydroxylamine followed by cyclization [
30], the intermediate 4,4-dimethylisoxazolidin-3-one (
15) was obtained. Nucleophilic substitution on
15 with 1,3-dibromopropane in the presence of NaH yielded the brominated intermediate
16. Next, the Michaelis-Becker reaction of
16 with diethyl phosphite (
1) and methylenebisphosphonate (
7) was performed to afford the intermediates
17 and
18, respectively, in the presence of Cs
2CO
3 and tetrabutylammonium iodide (TBAI). The subsequent deprotection of the ethyl protecting groups using TMSBr gave the target compounds
19 and
20 in yields of 62% and 55%, respectively.
Scheme 3.
Synthesis of the target compounds 19 and 20. (a) i) NH2OH·HCl, NaOH, H2O, 0℃ to r.t.; (ii) NaOH, Na2CO3, H2O, 45℃; (b) Br(CH2)3Br, NaH, DMF, 0℃ to r.t.; (c) HPO(OEt)2 or 7, Cs2CO3, TBAI, DMF, r.t.; (d) i) TMSBr, CH2Cl2, 0℃ to r.t.; ii) THF/H2O, r.t.
Scheme 3.
Synthesis of the target compounds 19 and 20. (a) i) NH2OH·HCl, NaOH, H2O, 0℃ to r.t.; (ii) NaOH, Na2CO3, H2O, 45℃; (b) Br(CH2)3Br, NaH, DMF, 0℃ to r.t.; (c) HPO(OEt)2 or 7, Cs2CO3, TBAI, DMF, r.t.; (d) i) TMSBr, CH2Cl2, 0℃ to r.t.; ii) THF/H2O, r.t.
Optimization of reaction conditions. In the synthesis of the target compounds, the facile connection of the phosphonate group with a heterocycle through a C-P bond is a key step, especially when a suitable linker length has to be considered in terms of the structural feature of the naturally occurring DXR-active products FOS and FR. The most common methods for simple C-P bond formation include the Michaelis-Arbuzov reaction and the Michaelis-Becker reaction [
31,
32], but both of the reactions require harsh reaction conditions and the yields are usually low. Another important strategy for synthesizing phosphonate-containing compounds involves the use of vinylphosphonate as Michael acceptor to react with a Michael donor such as ketone, β-ketoester, and Grignard reagent [
33,
34,
35]. This type of synthesis tends to be facile and mild in reaction conditions, but the subsequent utilization of the functional groups of the Michael donors for heterocycle construction is seldom reported. In addition, the reports on the reaction of vinylphosphonate and vinylidenebisphosphonate as Michael acceptors are rather few perhaps due to the relatively low reactivity in this type of Michael addition as compared to the analogous vinylcarboxylate Michael acceptors. In our case, this type of Michael addition was exploited with the reaction of vinylphosphonate and vinylidenebisphosphonate with β-dicarbonyl compounds as Michael donors, which could cyclize with hydrazine and hydroxylamine to form heterocycles, as shown in
Scheme 1 and
Scheme 2. Hence, the optimization of the reaction conditions for Michael addition between vinylphosphonates (
1) and β-dicarbonyls appeared to be crucial for the target compound synthesis.
The reaction of vinylphosphonate
1 with acetoacetate as a model reaction was first optimized and the results are shown in
Table 1. When strong bases were used, such as t-BuOK, NaH, and NaOEt, and the reaction performed in THF or EtOH at room temperature for 6 h, the yields of the Michael addition were determined to be 60, 45, and 40%, respectively, for the isolated product
2a (
entries 1~3), and heating (
entry 4) and expending the reaction time in the case of t-BuOK as base gave no increase in yield. Since these conditions gave no ideal yields, we had to sought after alternative methods. It was reported that for a typical Michael addition reaction, a moderately strong base such as Cs
2CO
3 or K
2CO
3 could be beneficial with the assistance of a phase-transfer catalyst (PTC) [
36]. Indeed, when TBAI or TEBAC used as PTC and Cs
2CO
3 or K
2CO
3 as base, the yield was significantly increased (
entries 5~8), with the best yield of 90% for the reaction performed with TEBAC as PTC and K
2CO
3 as base (
entry 8). For other β-ketoesters and β-diketones, the reaction proceeded similarly and gave the corresponding intermediate
2 in yields of 87~92%. On the other hand, vinylidenebisphosphonate as the Michael acceptor tended to react faster with β-ketoesters and β-diketones using LiHMDS as base under the reaction conditions indicated in
Scheme 2. This higher reactivity of vinylidenebisphosphonate than vinylphosphonate is due to its higher electrophilicity cause by the double electron-withdrawing effect of the two phosphonate groups on the vinyl unit. The data of yield of these target compounds are given in the Supporting Information.
Table 1.
Table 1. Optimization of the reaction conditions for Michael additiona.
Table 1.
Table 1. Optimization of the reaction conditions for Michael additiona.
Entry |
Base |
PTC |
Solvent |
Temp (℃) |
Yield (%)b
|
1 |
t-BuOK |
-- |
THF |
r.t. |
60 |
2 |
NaH |
-- |
THF |
r.t. |
45 |
3 |
NaOEt |
-- |
EtOH |
r.t. |
40 |
4 |
t-BuOK |
-- |
THF |
66 |
62 |
5 |
Cs2CO3
|
TBAI |
CH3CN |
80 |
80 |
6 |
Cs2CO3
|
TEBAC |
CH3CN |
80 |
85 |
7 |
K2CO3
|
TBAI |
CH3CN |
80 |
83 |
8 |
K2CO3
|
TEBAC |
CH3CN |
80 |
90 |
2.2. Arabidopsis Growth Inhibitory Activity
An early report described that FOS as a DXR enzyme inhibitor can disrupt the biosynthesis of essential isoprenoids for chlorophyll and carotenoids, leading to bleaching and developmental arrest in
Arabidopsis thaliana (Arabidopsis) [
37]. Thus, using FOS as the reference compound, the activity of the synthesized compounds against the growth of Arabidopsis was screened at an initial concentration of 100 mg/L, and the data are provided in
Supporting Information. For compounds with inhibition rates higher than 50% at the initial concentration of 100 mg/L, their median effective concentration (EC
50) values were determined. In the total 32 compounds, 10 compounds were found to have inhibition rates higher than 50% against Arabidopsis at the initial concentration, with their EC
50 ranging from 7.8 to 88.3 mg/L, as shown in
Table 2. Among them, the series
13 compounds, including
13a,
13d,
13e, and
13f, performed better in inhibitory activity than FOS, and compound
13e has the highest activity with EC
50 down to 7.8 mg/L, a 3.7-fold activity of FOS, which has an EC
50 value of 27.5 mg/L.
Table 2.
Median effective concentrations of target compounds against Arabidopsis growth.
Table 2.
Median effective concentrations of target compounds against Arabidopsis growth.
Comp |
EC50 (mg/L)a
|
Comp |
EC50 (mg/L) |
Comp |
EC50 (mg/L) |
5a |
>100 |
6f |
>100 |
13b |
37.7±2.8 |
5b |
>100 |
12a |
>100 |
13c |
48.4±5.3 |
5c |
>100 |
12b |
>100 |
13d |
23.1±1.9 |
5d |
>100 |
12c |
>100 |
13e |
7.8±1.2 |
5e |
>100 |
12d |
>100 |
13f |
8.7±1.3 |
5f |
>100 |
12e |
>100 |
13g |
>100 |
6a |
28.6±3.9 |
12f |
>100 |
13h |
>100 |
6b |
>100 |
12g |
>100 |
13i |
40.7±2.9 |
6c |
>100 |
12h |
>100 |
19 |
>100 |
6d |
45.7±4.8 |
12i |
88.3±4.3 |
20 |
>100 |
6e |
>100 |
13a |
21.6±3.8 |
FOS |
27.5±3.1 |
The phenotype of Arabidopsis treated with FOS and the 10 compounds at concentrations of 100, 50, 25, 10, and 1 mg/L was further tested by monitoring with digital camera and the results are shown in Figure 3. It can be seen that FOS and all the test compounds exhibit concentration-dependent growth inhibition and bleaching on Arabidopsis, suggesting the effect may arise from the inhibition on a certain pathway in the plant chloroplast, just as FOS does. At a high applied concentration of 100 mg/L, most compounds could even completely inhibit the germination of Arabidopsis. Notably, two compounds, 13e and 13f, performed superior in Arabidopsis growth inhibition and bleaching effect to the control FOS at all tested concentrations.
Figure 3.
Inhibition and bleaching effects of the 10 active compounds on Arabidopsis.
Figure 3.
Inhibition and bleaching effects of the 10 active compounds on Arabidopsis.
2.3. Pre-emergence Herbicidal Activity
Two model plants, dicot
Brassica napus L. (
BN) and monocot
Echinochloa crus-galli (
EC), were tested for the pre-emergence herbicidal activity of all the synthesized compounds using a standard Petri dish method [
38]. The EC
50 values collected on the inhibition of root and stalk for the 10 active compounds plus the 2 isoxazolinone-containing compounds
19 and
20 are shown in
Table 3 for discussion, while the data for the other compounds are given in
Supporting Information. All 12 compounds in
Table 3 displayed pre-emergence herbicidal activities with the EC
50 values comparable or superior to that of FOS, and especially, compound
13e has EC
50 of 10.7 and 2.3 mg/L and compound
13f has EC
50 values of 17.9 and 5.8 mg/L, on
BN root and stalk, respectively, while FOS only has the corresponding EC
50 of 34.7 and 32.9 mg/L. The 2 isoxazolinone-containing compounds
19 and
20, which although have no obvious effects on Arabidopsis, behaved better in inhibition of these two model plants, with the compound
20 more powerful than FOS in the inhibitory EC
50 of
BN stalk as 8.2 versus 32.9 mg/L.
Table 3.
Pre-emergence herbicidal activities of the 10 active compounds plus the compounds 19 and 20 on the root and stalk of BN and ECa.
Table 3.
Pre-emergence herbicidal activities of the 10 active compounds plus the compounds 19 and 20 on the root and stalk of BN and ECa.
Comp |
EC50 (mg/L) |
BN |
EC |
root |
stalk |
root |
stalk |
6a |
17.3 |
10.5 |
24.0 |
18.9 |
6d |
67.5 |
38.2 |
40.1 |
25.7 |
12i |
36.6 |
21.7 |
53.1 |
47.7 |
13a |
26.9 |
8.5 |
27.1 |
10.2 |
13b |
38.8 |
20.2 |
29.9 |
35.3 |
13c |
27.7 |
46.9 |
43.9 |
45.8 |
13d |
25.2 |
11.0 |
27.2 |
13.5 |
13e |
10.7 |
2.3 |
7.4 |
4.1 |
13f |
17.9 |
5.8 |
9.7 |
3.6 |
13i |
40.5 |
33.5 |
32.1 |
34.6 |
19 |
79.3 |
39.2 |
>100 |
88.0 |
20 |
34.6 |
8.2 |
40.0 |
25.2 |
FOS |
34.7 |
32.9 |
40.2 |
38.4 |
To give a better illustration of the activity difference of the 12 compounds along with FOS, four column graphs, representing separately the four sets of relative activity data on root and stalk of BN and EC with respect to that of FOS, are given in Figure 4. It can be seen from Figure 4A~B that compounds 6a, 13a, 13d~f, and 20 are more active on the root and stalk of BN than FOS, which are more evident on the inhibition of BN stalk, with 13e and 13f having 14.3- and 5.7-fold relative activities, respectively, with respect to that of FOS. For the inhibition on EC, as shown in Figure 4C~D, most of the compounds exhibited a stronger inhibitory activity than FOS. Among them, compounds 13e and 13f also have the strongest inhibitory effects on the root and stalk of EC, with the relative activities of 5.4- and 10.7-fold, respectively, with respect to that of FOS. It is also noted that both 19 and 20 have the similar activities on the inhibition of the root of BN and EC to FOS, but the compound 20 has a much stronger inhibitory effect on the stalk of BN and EC than FOS, with 4.0 and 1.5-fold relative activities, respectively, as compared to FOS.
Figure 4.
Relative activities of the 10 active compounds plus the compounds 19 and 20 on the root and stalk of the BN and EC.
Figure 4.
Relative activities of the 10 active compounds plus the compounds 19 and 20 on the root and stalk of the BN and EC.
2.6. DMAPP Rescue and Molecule Docking
DMAPP rescue. To explore whether the active compounds are acting on the DXR enzyme in the MEP pathway, a DMAPP rescue assay was conducted using the most active compound 13e and the control FOS. The principle of DMAPP rescue is that if the growth inhibition of Arabidopsis is caused by the inhibition of DXR enzyme, adding exogenous DMAPP as an intermediate downstream of DXR should restore the growth of Arabidopsis. As shown in Figure 5, the developmental arrest and bleaching of Arabidopsis by FOS treatment were rescued by the addition of DMAPP, with the green channel pixel values of the Arabidopsis images increased from 32,789 to 54,164, representing a 1.65-fold rescue with evident statistical significance. On the contrary, the addition of DMAPP to the compound 13e treatment group did not show a significant increase in pixel values, indicating that 13e may act on an herbicidal target outside of the MEP pathway.
Figure 5.
Rescue of bleaching and developmental arrest in Arabidopsis by adding exogenous DMAPP. (A) Images of 13e- and FOS-treated Arabidopsis before and after rescue with DMAPP, along with the blank control CK. The concentrations of 13e and FOS were 20 mg/L and 40 mg/L, respectively, and the DMAPP concentration was 150 mg/L. (B) Green channel pixel values of Arabidopsis images after different treatments. **p < 0.01, ns: no significance.
Figure 5.
Rescue of bleaching and developmental arrest in Arabidopsis by adding exogenous DMAPP. (A) Images of 13e- and FOS-treated Arabidopsis before and after rescue with DMAPP, along with the blank control CK. The concentrations of 13e and FOS were 20 mg/L and 40 mg/L, respectively, and the DMAPP concentration was 150 mg/L. (B) Green channel pixel values of Arabidopsis images after different treatments. **p < 0.01, ns: no significance.
Molecule docking. To further assess the possibility of the mode of action of the active compounds on the target enzyme DXR, molecular docking of compound
13e was performed using Autodock Vina software. Given the structural similarity between the reported pyridine-containing bisphosphonate (CBQ, shown in
Figure 2A) and
13e, the complex crystal structure of CBQ with
EcDXR (PDB ID: 1T1R) was chosen as a template for docking [
23]. The interactions between compound
13e and the surrounding residues of the DXR active site are depicted in
Figure 6A. The bisphosphonate group of
13e forms hydrogen bonding with Ser150, Glu151 and Lys227 within the bond length range of < 3.0 Å, while the 3-hydroxyisoxazole also interacts with Glu151 to form an additional hydrogen bond. In addition,
13e demonstrates a binding mode to Mg
2+ in the DXR active site similar to that of CBQ, through the coordination of one hydroxyl oxygen of the bisphosphonate group. In
Figure 6B, the binding conformations of
13e and CBQ could be superimposed well, with the bisphosphonate group locating in the Mg
2+ binding site, and the heterocycle part occupying the hydrophobic cavity, both of which are different from FOS that uses its hydroxamate group to bind with the metal enzyme. It is worthy of mentioning that
13e and CBQ have comparable relative activities with respect to FOS in DXR inhibition, but with much inferior IC
50 as compared to that of FOS [
23,
40]. This tendency to give low inhibition effects for
13e and CBQ may arise from the fact that their binding conformations are reversed as compared to that of FOS, which uses the hydroxamate moiety to coordinate with the metal ion and the monophosphonate to bind with the surrounding residues at the active site of DXR enzyme.
Figure 6.
The docked conformation of 13e with DXR enzyme and the interactions with the surrounding residues (A) and the conformational superimposition of 13e with CBQ and FOS in the DXR active site (B). Key residues are shown as slate sticks, the hydrogen bonds and coordinate bonds are highlighted in yellow dashed lines, and the Mg2+ ion is presented as a wheat sphere.
Figure 6.
The docked conformation of 13e with DXR enzyme and the interactions with the surrounding residues (A) and the conformational superimposition of 13e with CBQ and FOS in the DXR active site (B). Key residues are shown as slate sticks, the hydrogen bonds and coordinate bonds are highlighted in yellow dashed lines, and the Mg2+ ion is presented as a wheat sphere.
2.7. Discussion on Structure-Activity Relationships
In this work, the bioactivity evaluation of the synthesized compounds has been performed on Arabidopsis, pre-emergency model plants Brassica napus L. and Echinochloa crus-galli, post-emergency plants Amaranthus retroflexus and Echinochloa crus-galli, and the recombinant DXR enzyme. Several compounds demonstrated better herbicidal activities compared to FOS, although they might have a different target involved in the inhibiting process. The inhibition phenotype analyses revealed that the molecular structure of the synthesized compounds has a significant influence on the inhibition of different plants, and on Arabidopsis the structure-activity relationship is of representativeness.
In terms of the molecular structure and the Arabidopsis inhibitory activity, several trends could be derived. One is from the influence of the number of phosphate fragment, and bisphosphonates tended to have higher herbicidal activities on Arabidopsis than the monophosphonates, as illustrated in the EC50 ratios for the two bis-/monophosphonate pairs 13a/6a and 13d/6d, 21.6/28.6 mg/L and 23.1/45.7 mg/L, respectively. The second trend comes from the influence of heterocycle unit. For example, the 3-hydroxyisoxazole-containing monophosphonates 6a and 6d displayed moderate inhibitory activities with EC50 of 28.6 and 45.7 mg/L, while other heterocycle-containing monophosphonates gave low or even no inhibition effects. The 3-hydroxyisoxazole-containing bisphosphonates 13a~f performed even better, showing moderate to good activity with EC50 from 7.8 to 48.4 mg/L. The third trend is from the influence of fluorine-containing substituent, and -CF3 (for 13e) and -CHF2 (for 13f) on 3-hydroxyisoxazole have the highest activities, with EC50 as low as 7.8 and 8.7 mg/L, respectively, outweighing all other compounds.
Finally, compounds
19 and
20, which contain the isoxazolinone fragment like the active component of the commercial herbicide clomazone, only exhibited a weak inhibitory effect on Arabidopsis. Clomazone targets the first enzyme DXS in the MEP pathway as a type of propesticide, which is activated by metabolic oxidation in the plant to form ketoclomazone [
26]. It is speculated that compounds
19 and
20 are unable to undergo this kind of metabolism in the plant, thereby giving no activation to display an inhibitory role.
The structure-activity relationship summarized on the Arabidopsis inhibition is also applicable to that of the inhibition on other model plants. For example, 3-hydroxyisoxazole-containing 6a and 13a demonstrated good inhibitory activity not only on Arabidopsis but also on other tested model plants; introducing -CF3 and -CHF2 onto the 3-hydroxyisoxazole further improved the herbicidal activity on all plants, especially in the case of compound 13e, which has one -CF3 group on 3-hydroxyisoxazole and displayed the best activity among all the synthesized compounds.