New 2-aryl-7,8-dimethoxy-3,4-dihydroisoquinolin-2-ium salts as potential antifungal agents: synthesis, bioactivity and structure-activity relationships

The title compounds can be considered as simple analogues of quaternary benzo[c]phenanthridine alkaloids (QBAs). In order to develop potent QBA-like antifungal agents, as our continuing study, a series of new title compounds were synthesized and evaluated for bioactivity against five plant pathogenic fungi by the mycelium growth rate method in this study. The SAR were also derived. The majority of the compounds showed good to excellent inhibition activity with average EC50 values of 7.87–20.0 μM for the fungi, superior to sanguinarine and cherythrine (two QBAs) and the commercial fungicide azoxystrobin. Part of the compounds were more active than commercial fungicides thiabendazole or carbendazim against F. solani, F. graminearum and C. gloeosporioides. Six compounds with average EC50 of 3.5–5.1 μg/mL possessed very great potential for development of new antifungal agents. SAR found that substitution patterns of the two aryl-rings significantly affect the activity. There exists a complex interaction effect between substituents of the two aryl-rings on the activity. Generally, the presence of electron-withdrawing groups on the C-ring can significantly increase the activity. These findings will be of great importance for the design of more potent antifungal isoquinoline agents.

Most of plant diseases are caused by plant pathogenic fungi 1 . Plant mycoses not only often result in the loss of crop yield and quality, but also are a food safety problem because some of plant pathogenic fungi can produce mycotoxins harmful to animal and human health 2 . Therefore, various fungicides have been extensively used to manage and treat plant mycosis in current agriculture. However, the persistent and incorrect use of some commercial fungicides had led to some increasingly serious problems, such as hereditable resistance, cross-resistance or environmental pollution 3 . Therefore, it is necessary and urgent to develop novel antifungal agents, especially environmentally friendly plant fungicides. In the past decades, natural product-based antimicrobial agents have attracted attention from researchers due to their lower environmental and mammalian toxicity 4 .
In addition, our study also found that the C=N + moiety in ADHIQs is a determinant for their bioactivities including antifungal properties 5,9 , and the antifungal activity of ADHIQs is closely related to the electron density distribution in its conjugated system, especially in the C=N + bond [9][10][11][12][13] . The substitution patterns of the two aryl-rings can significantly impact the activity. Meanwhile, there exists a complex effect between substituents of the two aryl-rings on the antifungal activity. These findings encouraged us to further extend the modification of 2-aryl-3,4-dihydroisoquinolin-2-ium compounds with the aim of finding more potent antifungal agents.  gloeosporioides). Sanguinarine (SA) and chelerythrine (CH), two natural model compounds, were used as reference controls. Thiabendazole (TBZ), carbendazim (CBD) and azoxystrobin (ASB), three commercial fungicide standards, were used as positive controls. The results are listed in Table 1.
Gratifyingly, almost all the compounds presented a certain inhibition activity against each fungus. For average inhibition rate of the same compound on all the five test fungi, among all the 38 tested compounds, 33 compounds gave average inhibition rates of >55%, superior to ASB (52%); 22 compounds showed average inhibition rates of >80%, higher than both TBZ (78.9%) and CH (79.6%); 12 compounds displayed average inhibition rates of 90.0% to 94.9%, more active than CBD (84.5%) and SA (89.4%). It was worth noting that except for A26-A28, all compounds were much more active than three commercial fungicides against C. lunata. On the other hand, for each fungus, the vast majority of the compounds showed the higher activity than ASB. Based on the average In order to more clearly know the difference among the activity of the more active compounds in Table 1, we further determined their activity at lower concentrations (75.0 μM) (26-38 μg/mL) against the same fungi. The results are listed in Table 2. Excitingly, 27 out of the 33 tested compounds showed the higher activity with average inhibition rates of 62.5% to 84.1% for the five test fungi.
Subsequently, the compounds with inhibition rates of >50% in Table 2 were further assayed for median effective concentrations (EC 50 ) against each fungus to explore their antifungal potential in more detail and SAR. Compound A1 without substituents on the C-ring, SA and CH were used as reference controls. TBZ and CBD were used as positive controls. The results are listed in Table 3.

Compounds
Average inibition rate ± SD (%) a  Table 2. Antifungal activity of compounds A and B at 75 μM. a The difference between the data with the different lowercase letters within a column is significant (P < 0.05). b Mean: the average values of inhibition rates of the same compound on the five test fungi.
Structure-activity relationship. By comparison of the EC 50 values of the compounds in Table 3 and their initial activity in Tables 1 and 2, we analyzed the SAR of compounds A and B. From Tables 4 and 5, it is concluded that the substitution patterns on both the C-ring and the A-ring can significantly influence the antifungal activity of the target compounds. The general trends are as follows.
A. The type of substituents on the C-ring can significantly affect the activity. Compared with A1 without substituents on the C-ring, the presence of all electron-withdrawing groups like halogen atoms (A2-A15), nitro (A16), trifluoromethyl (A17, A18) or cyano (A19) on the C-ring can significantly increase the activity against each fungus. However, the opposite was observed for electron-donating groups like hydroxyl (A26-A28), 2′-Me (A20), 2′-OMe (A23) or 4′-OMe (A25) ( Table 4). On the other hand, comparison of B2-B10 and B1 (R′ = H) showed that except for 2′-F (B2), all the substituents on the C-ring caused significant decrease on the activity of B class of compounds in most cases, which is very different from or even opposite to that of A class of compounds (Table 5). Obviously, the effect of substiutents on the C-ring also depends on the substituent on the A-ring.    Tables 1 and 2. B. For the same substituent, the site of substituents on the C-ring can also significantly influence the activity (Table 3), but this impact varies with the type of substituents and species of fungi. For example, 2′-brominated compound (A8) showed the highest activities against F. solani, V. mali and C. gloeosporioides among three isomers (A8-A10), whereas 4′-brominated compound (A10) gave the highest activity against F. graminearum and C. lunata. Unlike brominated compounds, 3′-F compound (A3) showed the highest activity against C. lunata and C. gloeosporioides among three isomers (A2-A4). For methylate-or methoxylated compounds, only 3′-Me, 4′-Me or 3′-OMe isomer can give improvement of the activity in most cases (A21, A22, A24, Table 2).
D. Comparison of the activities of various compounds B (R = Br) and the corresponding compounds A (R = H) with the same substituents on the C-ring showed that the effect of 5-Br on the A-ring on the activity varies with the substitution patterns of the C-ring group. When 2′-, 3′-or 4′-F, 2′-OMe or no substituent is present on the C-ring, the introduction of 5-Br can significantly increase the activity against all or most of the fungi (Bn vs An, n = 1, 2, 3, 4; B10 vs A23). However, the opposite is found when 4′-Cl, 3′-Br, 4′-Br or 4′-I is present (B7 vs A9; B8 vs A10; B9 vs A13). When R′ is 3′-Cl, the introduction of 5-Br only makes a small influence on the activity against all the fungi (B6 vs A7) ( Table 5).
The results above show that there is an interaction effect between substituents on the A-ring and the C-ring, and the activity of compounds depends on the combined effect of substituents on both the A-ring and the C-ring. A similar case was found for the antifungal activity of 6-chloro-ADHIQs 9 and 8-methoxy-ADHIQs 11 reported previously by us. The facts above strongly suggest that the bioactivity of the compounds should be related with the electron density distribution of the conjugated system. Therefore, it is necessary to extend the modification of the A-ring to discover more potent ADHIQs agents.   Table 5. Effects of the substituents (R′) on C-ring and 5-Br on the A-ring on the activity. Significantly increasing the activity relative to compound B1 or the corresponding compounds A with the same substituents on the C-ring. Significantly decreasing the activity; No significant effect on the activity.

Discussions
The SAR of the title compounds is similar but not completely equal to that of 8-OMe-ADHIQs recently reported by us 11 , and obviously different from that of 6-chloro-ADHIQs 9 . The main difference between the present compounds (7,8-dimethoxy-ADHIQs) and 8-OMe-ADHIQs is the position effect of substituents on C-ring. For the 8-OMe-ADHIQs, the introduction of 3′-Cl to the C-ring gave the highest activity against F. solani and F. graminearum. For the 6-Cl-ADHIQs, the introduction of 2′-Me, an electron-donating group, led to the highest activity against C. lunata and V. mali, whereas the presence of halogen atoms on the C-ring did not give significant improvement of the activity in most cases. Compared with the 8-OMe-or 6-Cl-ADHIQs, the majority of the present compounds, especially compounds A showed the higher activity against each fungus. The results may be due to the present compounds possessing the higher structural similarity to CH than the 8-OMe or 6-Cl compounds ( Fig. 1). Based on the results and analysis above, it is necessary to further explore the antifungal activity of 7,8-methylenedioxy-ADHIQs which are structurally similar to sanguinarine. At present, this work is under way in our lab.
Up to now, no report was found on antifungal mechanism of ADHIQs or QBAs. However, as an analogue of chelerythrine or sanguinarine, berberine had been proved to inhibit the fungus Aspergillus fumigatus by targeting ergosterol biosynthesis pathway 20 . Based on the structural similarity, we conjecture that chelerythrine, sanguinarine and ADHIQs may have the same or similar action mechamism to berberine. Ergosterol biosynthesis process includes a Δ 14 reduction reaction involving a tertiary carbon cation intermediate (Fig. 3) 21 . Quaternary ADHIQ ions or QBAs have very high structural similarity to the intermediate. Therefore, we think that ADHIQs or QBAs may be inhibitors of Δ 14 reductase in ergosterol biosynthesis. This also is what we are going to do next.
As natural compounds, sanguinarine and chelerythrine possess very high safety to mammal such as pigs or mice 22,23 . QBAs have been used as antimicrobial agents to treat oral disease 24 , dermatomycosis 25 or expectorant 26 . Although we don't know whether ADHIQs have high safety to mammal like QBAs at present, our previous study showed that ADHIQs have lower toxicity to normal cells than cancer cells 5,16 . Additionally, our study also proved that ADHIQs have no effect on seed germination and seeding growth of plants (Panicum miliaceum L. and Brassica campestris L.) 17,18 . Interestingly, some ADHIQs were found to have growth-promoting action for plants. Therefore, ADHIQs can be considered as promising and potent antifungal agents.
In conclusion, a series of new 2-aryl-7,8-dimethoxy-3,4-dihydroisoquinolin-2-ium bromides were designed, synthesized and evaluated for antifungal activity in vitro against the five plant pathogenic fungi in the present study. The majority of the compounds showed good to excellent inhibition activity with average EC 50 values of 7.87-20.0 μM for the five fungi, superior to their natural model compounds SA and CH. Part of the compounds showed the higher activity against F. solani, F. graminearum and C. gloeosporioides than commercial fungicides TBZ or CBD. Compared with other series 2-aryl-3,4-dihydroisoquinolin-2-iums with or without substituents on the A-ring, the present compounds showed the highest activity in most cases. Compounds A8, A10, A11, A13, A15 and B2 possessed very great potential to be developed as new antifungal agents. In addition, SAR was derived also. It was found that substitution patterns of both the C-ring and A-ring significantly affect the activity. There exists an interaction effect between substituents on the A-ring and the C-ring, and the activity of compounds depends on the combined effect of substituents on the two aryl-rings. The presence of electron-withdrawing groups on the C-ring can significantly increase the activity. These findings will be of great importance for the design of more potent antifungal 2-aryl-3,4-dihydroisoquinolin-2-ium agents. It is necessary to conduct more extensive structural modification, especially for the A-ring, determination of an antifungal spectrum and the in vivo activity of these compounds.  Instruments. Melting points were determined on an mp420 automatic melting point meter (Hanon Instrument, Beijing, China) and uncorrected. NMR spectra were recorded on a BrukerAvance III 500 MHz instrument. Chemical shift values (δ) were given in parts per million (ppm). Coupling constant values (J) were given in Hz. High resolution mass spectra (HR-MS) and low resolution mass spectra (LR-MS) were carried out with Micromass Auto Spec-3000 instrument and Thermo FisherLCQ Fleet instrument, respectively.
Synthesis. Synthesis of 2-(2-bromo-4,5-dimethoxyphenyl)acetic acid (1). Bromine (17.60 g, 0.11 mol) in 55 mL anhydrous CH 2 Cl 2 was dropwise added into a solution of 2-(3,4-dimethoxyphenyl)acetic acid (19.62 g, 0.10 mol) in 60 mL dry CH 2 Cl 2 under ice bath. The resulting mixture were stirred at room temperature for 24 h, and extracted with saturated Na 2 SO 3 aqueous solution (3 × 80 mL). To the aqueous phase was added 2 M HCl aqueous solution until the product was completely precipitated. The precipitates were filtered off and dried at 55 °C in vacuum to yield 1 as white solids (25. (2). To a solution of 1 (13.76 g, 0.05 mol) in 150 mL dry THF was portion-wise added NaBH 4 (5.67 g, 0.15 mol) under ice bath. After no gas was liberated, iodine (12.65 g, 0.05 mol) in 80 mL dry THF was dropwise added into the mixture, and then stirred at 40 °C until the reaction was completed. The reaction was quenched with methanol until no gas was generated. After evaporated under reduced pressure, to the residue was added 100 mL water, and then extracted with CH 2 Cl 2 (3 × 80 mL). The organic phase was sequentially washed with saturated sodium sulfite solution (3 × 80 mL) and saturated brine (80 mL), and dried over anhydrous MgSO 4 . After filtration and concentration, crude 2 was obtained and directly used for the following reaction.

Synthesis of 5-bromo-7,8-dimethoxyisochromane (3)
. Paraformaldehyde (1.65 g, 55 mmol) was fully dissolved in 50 mL trifluoroacetic acid with assistance of ultrasonic wave. The resulting solution was poured into 2 and stirred at room temperature until the reaction was completed. To the solution was added 100 mL water, extracted with CH 2 Cl 2 (3 × 100 mL), sequentially washed with saturated sodium sulfite solution (150 mL), saturated sodium bicarbonate solution (2 × 100 mL) and saturated brine (100 mL). The organic phase was dried over anhydrous MgSO 4 . After filtration and concentration, the residue was purified by column chromatography on silica gel using petroleum ether-ethyl acetate (30:1) to yield 3 as white solids (10.29 g) in 75% yield for two steps. 1 (5). To a solution of 4 (5.79 g, 29.8 mmol) in 60 mL dry CH 2 Cl 2 was added DDQ (7.94 g, 35 mmol) and anhydrous methanol (1.5 mL, 35.8 mmol), and then stirred under argon at room temperature for 24 h. After the reaction was completed, saturated sodium bicarbonate (80 mL) was added into the solution. The organic phase was collected, and the aqueous phase was extracted with CH 2 Cl 2 (3 × 60 mL). The combined organic phase was washed with water (2 × 150 mL) and saturated brines (1 × 150 mL), and dried over anhydrous MgSO 4 . After filtration, the solution was evaporated under reduced pressure. The residue was chromatographed over silica gel using petroleum ether-ethyl acetate (20:1) to yield 5 as pale yellow oils (4.14 g) in 62% yield. The structure of 5 was unconfirmed and directly used for the following reaction.

2-(3-Bromophenyl)-7,8-dimethoxy-3,4-dihydroisoquinolin-2-ium bromide (A9
were screened for antifungal activity in vitro against five plant pathogenic fungi. Briefly, a solution of the tested compound (36 or 18 mmol) in 10 mL sterile water containing 0.6 mL DMSO was fully mixed with 230 mL of 50 °C melted PDA agar to provide a culture medium containing 150 μM or 75 μM tested compound and 0.25% (v/v) DMSO, and then poured into a sterilized Petri dish (ca. 16 mL each plate). No observable effect on the growth of fungi was proved for 0.25% (v/v) DMSO in the culture medium. TBZ, CBD and ASB were used as positive controls and 0.25% DMSO was used as a blank control. SA iodine and CH iodine were used as reference controls. A fungus disc (d = 5 mm) cut from subcultured petri dishes was placed at the center of the just solidified medium in petri dish. The dishes were kept in an incubator at 25 °C for 72 h. Each experiment was carried out in triplicate. The diameters (in mm) of a fungal colony were measured in three different directions, and the growth inhibition rates were calculated according to the method reported previously 9 . Duncan multiple comparison test was performed on the data to evaluate significant difference between the activities of various compounds at the same concentration.
The compounds with higher initial activities were further assayed for EC 50 values according to the method described above. Based on the screening results, a series of test concentrations of the compound was set and tested for inhibition rate against the fungi. Antifungal toxicity regression equations were established according to the method previously reported 12 . EC 50 values were calculated from the equations by using PRISM software ver. 5.0 (GraphPad Software Inc., San Diego, CA, USA).