Antiviral mechanism study of gossypol and its Schiff base derivatives based on reactive oxygen species (ROS)

Abstract

A series of amino acid gossypol Schiff bases were designed, synthesized and evaluated for their anti-TMV activities, and a new method of preparing hydrophilic amino acid gossypol Schiff bases, which are difficult to prepare using existing methods, was developed. Most of these compounds displayed good activities. Structure–activity relationship studies revealed that both the carboxy group and R group in amino acids are important to their anti-TMV activities. Further mechanism studies based on ROS-producing ability indicated that the O₂˙⁻ production rates of gossypol Schiff bases showed a positive correlation with their anti-TMV activity, suggesting that O₂˙⁻ is more important than H₂O₂ at the primary stage of TMV inoculation. However, the gossypol Schiff bases derived from D and L-amino acids, respectively, showed different anti-TMV activities (4 > 16, 6 > 17, 7 > 18, 11 > 19) although they had the same ROS-producing abilities. Both the D and L-amino acid gossypol Schiff bases stimulate O₂˙⁻ and H₂O₂ production in tobacco leaves. The current study provides fundamental support for the development of gossypol alkaloids as potential inhibitors of plant viruses.

---

Introduction

The use of natural products as lead compounds is one of the most effective pesticide design methods because natural products not only decompose rapidly, therefore reducing their adverse environmental impact,^1–3 but also possess novel structures and mechanisms. Gossypol (1), a polyphenolic natural product (Fig. 1), displays various drug properties, including antifertility, anti-parasitic,⁴ anti-malarial,⁵ antibacterial⁶ and anticancer.⁷ Gossypol is also reported to have antiviral activity toward enveloped viruses such as HIV,⁸ H5N1,⁹ HSV-2,¹⁰ and the influenza virus,¹¹ but not toward nonenveloped viruses. However, recently, a patent demonstrated that a (±)-gossypol fatty acid mixture shows antiviral activity against tobacco mosaic virus (TMV), which is a nonenveloped virus.¹² In our previous studies, we developed a series of (±)-gossypol Schiff bases with aromatic amines^13a and aliphatic amines^13b which not only displayed high inhibitory activity against TMV, but also had less toxicity to rats than (±)-gossypol. Therefore, gossypol Schiff bases can serve as good lead compounds for further development of anti-TMV reagents. However, the anti-TMV mechanism of these compounds has not been studied. We also proposed that it may be the improved ROS-producing abilities of these compounds that accounts for their higher anti-TMV activities compared to gossypol. However, this idea has not yet been confirmed.

Fig. 1 Structures of (±)-gossypol 1, amino acid gossypol Schiff bases and (±)-gossypol Schiff bases.

Polyphenol compounds have been demonstrated to have the ability to generate reactive oxygen species (ROS), e.g. O₂˙⁻ and H₂O₂.¹⁴ For example, Marklund and Xu found that pyrogallol could autoxidize rapidly under alkaline conditions, generating O₂˙⁻.^15a,b Akagawa et al. reported that both catechins and gallic acid can generate H₂O₂ by autoxidation under neutral-alkaline conditions.¹⁶ Doke et al. demonstrated that an O₂˙⁻ generating system was involved in the induction of necrotic lesions on tobacco leaves infected by TMV.^17a However, Hafez et al. proposed that an early external application of ROS (O₂˙⁻ and H₂O₂) in tobacco may contribute to the development of resistance to TMV infection, while the same ROS treatments did not inhibit TMV titers if applied much later.^17b,c As polyphenolic compounds, gossypol and its Schiff bases may also have ROS producing abilities,¹⁸ which may be an important factor in their anti-TMV activities. In this paper, we aim to design a series of gossypol Schiff bases with different ROS-producing abilities and determine the relationship between their structures, their ROS-producing abilities, and their anti-TMV activities.

As we can see in Fig. 1, all gossypol Schiff bases have the same skeleton (in green background); the only difference between those Schiff bases is their R groups. Therefore, we deduce that if their R groups have different antioxidative and prooxidative abilities, the Schiff bases may also have different ROS-producing abilities.

Amino acids not only have better water solubility, but also show different antioxidative and prooxidative abilities.^19–21 Therefore, we suppose that introducing amino acids into gossypol Schiff bases may not only improve their water solubility but may also improve their ROS-producing abilities.

Thus, we designed and synthesized a series of amino acid gossypol Schiff bases with different amino acids and some other related amino compounds; we then systematically investigated the possible influencing factors for anti-TMV activity, such as carboxylic groups (relates to water solubility), R groups, and the D or L configuration of amino acids (Fig. 1). Furthermore, in order to verify whether the ROS-producing abilities of gossypol and its derivatives can result in anti-TMV activity at the primary stage of TMV inoculation, as well as which ROS, O₂˙⁻ or H₂O₂, is more important to these anti-TMV activities, we performed corresponding research and summarized these relationships.

Materials and methods

General reagents were purchased from commercial sources and were used as received. Reaction progress was monitored by thin-layer chromatography (TLC) on silica gel GF-254 with detection by UV. ¹H NMR and ¹³C NMR spectra were recorded on a Bruker Ascend 300/400 MHz spectrometer. Chemical shift values (δ) are given in parts per million and are downfield from internal tetramethylsilane. High-resolution mass spectra (HRMS) were recorded on FT-ICR MS (Ionspec, 7.0 T). Absorbances were recorded on a BioTek Synergy 4 enzyme-labeled instrument.

General procedure for the synthesis of amino acid (±)-gossypol Schiff bases

Procedure A. Procedure A (suitable for 3, 4, 8, 9, 10) a mixture of (±)-gossypol acetic acid (0.50 g, 0.86 mmol), sodium hydroxide (0.069 to 0.103 g, 1.73 to 2.59 mmol), and amino acid (1.73 mmol) in ethanol (20 mL) was stirred and heated at reflux for 1 to 3 h, with the progress of the reaction being monitored by TLC. When the reaction reached completion, the mixture was cooled to room temperature. The resulting solid was filtered, washed with ethanol, and recrystallized to give the desired gossypol derivatives.

Procedure B. Procedure B (suitable for 3–12, 16–19). A mixture of (±)-gossypol acetic acid (0.50 g, 0.86 mmol), sodium hydroxide (0.069 to 0.103 g, 1.73 to 2.59 mmol), and amino acid (1.73 mmol) in methanol (20 mL) was stirred and heated at reflux for 1 to 3 h, with the progress of the reaction being monitored by TLC. When the reaction was complete, the solution was cooled to room temperature. Then, dichloromethane or acetone was added to the solution to precipitate a solid. The resulting solid was filtered and washed with ethanol to give the desired gossypol derivatives.

Procedure C. Procedure C (suitable for 13, 14, 15). A mixture of (±)-gossypol acetic acid (0.50 g, 0.86 mmol), sodium hydroxide (0.034 g, 0.86 mmol), and amino-containing compound in ethanol (20 mL) was stirred and heated at reflux for 1 to 3 h, with the progress of the reaction being monitored by TLC. When the reaction reached completion, the mixture was cooled to room temperature. The resulting solid was filtered, washed with ethanol, and recrystallized to give the desired gossypol derivatives.

The four isomers of the gossypol Schiff bases of tryptophan methyl ester (22–25) were prepared by the method reported by Jiang's group.²²

General procedure for H₂O₂ and O₂˙⁻ determination

H₂O₂ determination. The H₂O₂ concentration of each amino acid gossypol Schiff base aqueous solution was determined by the colorimetric method. 200 μL aqueous solution consisting of ammonium ferrous sulfate (200 μmol L⁻¹), sulfuric acid (20 mmol L⁻¹), xylenol orange (160 μmol L⁻¹), and sorbitol (160 mmol L⁻¹) was added to 200 μL gossypol amino acid Schiff base solution (500 μg mL⁻¹). After incubating at 25 °C for an hour, the absorbance at 560 nm was read using an enzyme-labeled instrument. The values obtained from this assay were plotted against a standard curve with 1 to 50 μmol L⁻¹ concentrations of H₂O₂. H₂O₂ standards were made fresh from a 30% H₂O₂ stock solution. The H₂O₂ concentrations were calculated by the equation below (see ESI†).

In the above equation, A₅₆₀ stands for the absorbance data of the mixture at 560 nm; constant A is the slope of the standard curve of H₂O₂; constant B is the absorbance of blank sample (c(H₂O₂) = 0), which is a background of this method; constant C represents the A₅₆₀ of the gossypol Schiff base. The absolute values of H₂O₂ are expressed in micromoles per liter (μmol L⁻¹).²³

O₂˙⁻ determination. 90 μL phosphate buffer (pH 7.8, 65 mmol L⁻¹) and 10 μL hydroxyl ammonium chloride (10 mmol L⁻¹) were added to 100 μL of gossypol amino acid Schiff base solution (500 μg mL⁻¹), and the solution was maintained at 25 °C for 20 min. 200 μL sulfanilic acid aqueous solution (17 mmol L⁻¹) and 200 μL α-naphthylamine solution (7 mmol L⁻¹) were added separately to the above solution, and the solution was mixed well by shaking. After standing at 25 °C for 20 min, the optical density of the mixture was determined at 530 nm on an enzyme-labeled instrument. The values obtained from this assay were plotted against an adjusted standard curve with 1 to 50 μmol L⁻¹ concentrations of NaNO₂, and the O₂˙⁻ concentration divided by time is the O₂˙⁻ production rate.

The equation of the O₂˙⁻ production rate could be described as below (see ESI†).

In this equation, A₅₃₀ stands for the absorbance data of the mixture at 530 nm; constant D is the slope of the standard curve of NaNO₂; constant E is the absorbance of blank sample (c(NaNO₂) = 0), which is a background of this method; constant F represents the A₅₃₀ of the gossypol Schiff base. The absolute values of the O₂˙⁻ production rates are expressed in nanomoles per liter per hour (nmol h⁻¹ L⁻¹).²⁴

ROS determination in tobacco leaves. A piece of tobacco leaf was cut into small pieces (about 1 × 1 cm²), and each piece was separately immersed in a solution of D or L amino acid gossypol Schiff base (500 μg mL⁻¹) for 15 min. A piece of tobacco leaf immersed into double-distilled water was used as a mock (negative control) sample. Each piece was washed with saline, and the tissue homogenate was prepared at a certain proportion (9 mL saline to 1 g fresh tissue) after homogenizing and centrifuging (10 000 rpm for 10 min at 4 °C). The supernatant was used for protein and H₂O₂ concentration determination. For determination of the O₂˙⁻ production rate, we added the same volume of ethyl ether to the supernatant and separated them by centrifugation (3500 rpm for 10 min). The water phase was used for O₂˙⁻ determination. Protein content was measured using a Protein Assay Kit (Suzhou Comin Biotechnology Co. Ltd., China).

The absolute concentration values of H₂O₂ are expressed in micromoles per gram protein (μmol g⁻¹), and the O₂˙⁻ production rate values are expressed in nanomoles per gram protein per min (nmol min⁻¹ g⁻¹).²⁵

Results and discussion

Chemistry

Development of the synthetic procedure of amino acid gossypol Schiff bases. Yan and Yang have reported the preparation of some amino acid gossypol Schiff bases.^7d,26 However, the starting material of those methods was limited to (±)-gossypol. Compared with (±)-gossypol (1), (±)-gossypol acetic acid (2) is not only much more stable, but also less expensive. Therefore, we developed a method (procedure A) using (±)-gossypol acetic acid as the starting material. Furthermore, the method is easier to perform: (±)-gossypol acetic acid was treated with 2 equivalents of corresponding amino acid and 2 to 3 equivalents of sodium hydroxide in refluxing ethanol for 1 to 3 h (Scheme 1). The amino acid Schiff base disodium salts had poor solubility in ethanol; therefore, they were simply collected by filtration. Unfortunately, when preparing the gossypol Schiff bases of tryptophan (5, 20), threonine (6, 17), asparagine (11, 19) and glutamine (12), procedure A, Yan's method, and Yang's method do not afford pure products. To solve this problem, variable factors such as temperature, time, solvent and amount of base were explored; however, good results were not obtained. Quite accidentally, we found that when methanol was used as a solvent to prepare compound 5, no desired precipitate was formed initially; however, when dichloromethane was added to the solution, a light yellow precipitate formed whose ¹H NMR spectrum was much cleaner than that of the precipitate produced by procedure A. By this method, pure amino acid gossypol Schiff bases were obtained in disodium salt form.

Scheme 1 General procedure for the synthesis of amino acid (±)-gossypol Schiff bases.

Physi-chemical properties of amino acid gossypol Schiff bases. Unsurprisingly, the amino acid (±)-gossypol Schiff bases prepared by this procedure were in their disodium salt forms and highly hydroscopic; therefore, it is more favorable to store the compounds in a cold (below −15 °C) and dry environment. These Schiff bases also showed obviously greater water solubility than (±)-gossypol, (±)-gossypol acetic acid, and derivatives 13–15 and 22–25; the sodium salts can be totally dissolved in water to a solution of 500 μg mL⁻¹ at 25 °C, whereas the others require the addition of organic solvent such as acetone or N,N-dimethylformamide (Fig. 2).

Fig. 2 Target compounds 3–26.

Identification of the dr and existing form of amino acid gossypol Schiff bases. The amino acid gossypol Schiff bases (1–21 and 26, not including 14) are pairs of diastereomers due to the racemic starting material of (±)-gossypol acetic acid and the chiral amino acid. In order to confirm the dr (diastereomeric ratio) of the Schiff bases, it is first necessary to identify the characteristic peaks of both diastereomers. Taking the isomers of the D-tryptophan methyl ester gossypol Schiff base as an example, as we can see in Fig. 3, for compound 22, the peaks which appear at 9.76 ppm (d, J = 12.5 Hz, assigned to H²) are coupling with the peaks at 13.48 ppm (dd, J = 4.8, 12.5 Hz, assigned to H¹). Analogously, for compound 23, the peaks of H² appeared at 9.64 ppm (d, J = 12.5 Hz) and also coupled with the peaks at 13.48 ppm (dd, J = 4.8, 12.5 Hz). This is also reliable evidence that the gossypol Schiff bases exist in the enamine–enamine form.^27,28 From the NMR spectrum of (±)-gossypol Schiff base (13), the signals of H² appear both at 9.76 ppm (d, 1H, J = 12.5 Hz) and 9.65 ppm (d, 1H, J = 12.5 Hz); therefore, the signal of H² can be a characteristic peak when confirming the dr of the Schiff bases. The amino acid gossypol Schiff bases also show the same characteristic peaks between 9 and 10 ppm. Taking D-phenylalanine gossypol Schiff base (4) as an example, characteristic peaks emerge at 9.69 pm and 9.50 pm with a ratio of about 1 : 1. It is worth mentioning that the dr of the other gossypol Schiff bases is also about 1 : 1 as determined by ¹H NMR, and the Schiff bases all exist in the enamine–enamine form in solution.

Fig. 3 (a) ¹H NMR spectrum (300 MHz, DMSO-d₆) of D-tryptophan methyl ester (+)-gossypol Schiff base (22). (b) ¹H NMR spectrum (300 MHz, DMSO-d₆) of D-tryptophan methyl ester (−)-gossypol Schiff base (23). (c) ¹H NMR spectrum (400 MHz, DMSO-d₆) of D-tryptophan methyl ester (±)-gossypol Schiff base (13). (d) ¹H NMR spectrum (400 MHz, DMSO-d₆) of D-phenylalanine gossypol Schiff base 4.

Structure–activity relationship against TMV

Antiviral bioassays of the compounds against TMV were carried out, and the results are listed in Table 1. To evaluate the antiviral potency of the synthesized compounds, the commercially available plant virucide ribavirin was used as the control. As we can see from Table 1, the anti-TMV activities of these compounds in vitro were in accordance with their activities in vivo. (±)-Gossypol (1) and (±)-gossypol acetic acid (2) displayed similar in vitro anti-TMV activities (32% and 28%, respectively); however, they were less effective than ribavirin (36%) at 500 μg mL⁻¹. In contrast, all the amino acid (±)-gossypol bases showed antiviral activities equivalent to or even higher than ribavirin both in vitro and in vivo; especially, compounds 4, 6, 7, 11 and 19 displayed obviously higher anti-TMV activities than ribavirin at a concentration of 500 μg mL⁻¹, and their anti-TMV activities at 100 μg mL⁻¹ even reached the level of ribavirin at 500 μg mL⁻¹.

Table 1 In vitro and in vivo anti-TMV activities of compounds 1–26 against TMV^a

Compd	Concn (μg mL^−1)	In vitro curative effect^b (%)	In vivo activity
			Inactivation effect^b (%)	Curative effect^b (%)	Protection effect^b (%)
a Compounds 4, 6, 7, 11 and 19, in red, show obviously higher anti-TMV activities than ribavirin at a concentration of 500 μg mL^−1.b Average of three replicates. All results are expressed as mean ± SD.
1	500	32 ± 1	35 ± 1	29 ± 2	29 ± 2
	100	0	0	0	0
2	500	28 ± 1	32 ± 1	30 ± 1	33 ± 2
	100	0	0	0	0
3	500	33 ± 1	41 ± 1	43 ± 2	36 ± 1
	100	6 ± 2	15 ± 1	8 ± 1	10 ± 1
4	500	55 ± 1	63 ± 1	57 ± 1	60 ± 1
	100	21 ± 1	31 ± 1	17 ± 1	25 ± 1
5	500	38 ± 1	42 ± 2	45 ± 2	47 ± 1
	100	7 ± 2	9 ± 2	14 ± 2	11 ± 2
6	500	58 ± 2	68 ± 2	63 ± 1	64 ± 2
	100	27 ± 1	30 ± 1	20 ± 2	34 ± 1
7	500	53 ± 2	62 ± 1	64 ± 1	58 ± 2
	100	15 ± 1	19 ± 2	22 ± 1	32 ± 1
8	500	35 ± 1	45 ± 1	40 ± 1	39 ± 1
	100	0	20 ± 2	12 ± 3	16 ± 1
9	500	50 ± 2	50 ± 1	46 ± 2	41 ± 1
	100	22 ± 1	16 ± 1	10 ± 1	7 ± 1
10	500	43 ± 2	50 ± 1	54 ± 2	56 ± 2
	100	27 ± 2	11 ± 1	20 ± 1	17 ± 1
11	500	67 ± 2	69 ± 2	58 ± 1	72 ± 1
	100	38 ± 2	24 ± 2	36 ± 1	33 ± 1
12	500	41 ± 2	48 ± 1	46 ± 1	51 ± 2
	100	16 ± 2	22 ± 1	13 ± 1	18 ± 1
13	500	46 ± 2	37 ± 2	29 ± 1	23 ± 2
	100	0	0	0	0
14	500	22 ± 2	27 ± 1	35 ± 2	37 ± 2
	100	0	6 ± 4	0	0
15	500	20 ± 1	31 ± 2	27 ± 2	21 ± 1
	100	0	0	0	0
16	500	49 ± 2	49 ± 1	51 ± 1	41 ± 1
	100	18 ± 2	25 ± 1	19 ± 1	13 ± 1
17	500	60 ± 1	57 ± 1	50 ± 2	55 ± 1
	100	14 ± 2	24 ± 1	21 ± 2	29 ± 2
18	500	39 ± 1	47 ± 1	43 ± 1	54 ± 1
	100	9 ± 1	17 ± 1	11 ± 1	15 ± 1
19	500	56 ± 1	65 ± 1	68 ± 1	59 ± 1
	100	28 ± 2	33 ± 1	27 ± 1	36 ± 1
20	500	57 ± 1	52 ± 2	50 ± 1	59 ± 1
	100	24 ± 2	17 ± 2	27 ± 1	29 ± 1
21	500	39 ± 1	36 ± 1	34 ± 1	46 ± 1
	100	8 ± 1	0	11 ± 1	6 ± 2
22	500	22 ± 1	32 ± 1	27 ± 1	24 ± 1
	100	0	0	0	0
23	500	44 ± 1	48 ± 1	51 ± 1	57 ± 1
	100	9 ± 2	27 ± 1	12 ± 1	18 ± 1
24	500	32 ± 1	47 ± 1	39 ± 1	36 ± 1
	100	9 ± 2	23 ± 1	16 ± 1	6 ± 1
25	500	31 ± 2	37 ± 1	29 ± 1	43 ± 1
	100	6 ± 2	11 ± 1	15 ± 1	18 ± 1
26	500	51 ± 1	53 ± 1	58 ± 2	42 ± 2
	100	16 ± 2	24 ± 1	20 ± 1	16 ± 2
Ribavirin	500	36 ± 1	39 ± 2	33 ± 1	40 ± 1
	100	0	19 ± 1	8 ± 1	13 ± 1

---

The significance of the carboxyl group. The bioassays show that compound 5 exhibited much higher anti-TMV activities than compounds 13 and 14. Similarly, compound 4 showed higher anti-TMV activities than compound 15. This suggests that the carboxyl group in amino acid should be important for the anti-TMV activities of the compounds. We also found that compounds 4 and 5 showed obviously greater solubility than 13, 14 and 15; compounds 4 and 5 can be totally dissolved in water to a solution of 500 μg mL⁻¹ at 25 °C, while conversely, 13, 14 and 15 cannot be dissolved in water at this concentration.

Therefore, we proposed that the anti-TMV activities of these compounds were somehow consistent with their water solubilities (4 > 15, 5 > 13, 14) and that the carboxyl groups (in sodium carboxylate form) of the amino acid gossypol Schiff bases were beneficial to both their solubilities and anti-TMV activities, whereas methyl ester and hydroxymethyl groups were unfavorable (H, CH₂OH, CO₂Me < CO₂Na).

R-Group properties of amino acids. The R group of the amino acid moiety determines the physical–chemical characteristics of the Schiff bases, such as polarity (hydrophobic or hydrophilic) and charge (acidic or positive versus basic or negative).

As shown in Table 1, D-valine gossypol Schiff base 9 exerted moderate anti-TMV activity at 500 μg mL⁻¹. Although D-leucine is just one carbon atom longer than D-valine, the D-leucine gossypol Schiff base 10 showed slightly higher activity than D-valine gossypol Schiff base 9. However, when comparing 11 with 12, 11 showed higher activity. This suggests that the carbon chain length of the amino acid also affects the activity, and the lengths of leucine 10 and asparagine 11 are favorable.

Furthermore, D-threonine gossypol Schiff base 6 showed better activity than 9, which indicates that the anti-TMV activity can be improved by substituting the methyl group of D-valine with a hydroxyl group.

Serine can be regarded as replacing the methyl group in threonine by an H atom. The serine gossypol Schiff base 7 also displayed a high inhibitory effect. However, cysteine gossypol Schiff base 8, which could be regarded as a derivative of 7, exhibited a lower inhibitory effect than 7; the possible reason for its low activity may be the relatively poorer water solubility of 8 compared to 7.

Comparing the R groups in the amino acid Schiff bases, substituents which show hydrophilic properties are beneficial to the anti-TMV activities of the corresponding compounds, whereas hydrophobic substituents are unfavorable (10 < 11; 9 < 6; 8 < 7).

Relationship between anti-TMV activity and ROS-producing activity

Hafez has reported that externally applied ROS-producing compounds that infiltrate into tobacco leaves at the primary stage of inoculation can suppress the replication of TMV.^17b,c On that basis, we supposed that the anti-TMV activities of gossypol and its Schiff bases may be related to their ROS-producing activities. To date, there is little scientific evidence available regarding the H₂O₂ or O₂˙⁻ producing ability of gossypol; there is also little evidence regarding which ROS, H₂O₂ or O₂˙⁻, is more important to its anti-TMV activities. Therefore, a series of D-amino acid gossypol Schiff bases with different oxidative and antoxidative abilities were synthesized, and their ROS-producing abilities were tested.

As shown in Fig. 4a, most gossypol Schiff bases showed obviously higher H₂O₂ concentrations than gossypol, except for compounds 13, 14 and 15. However, there is no obvious correlation between their anti-TMV activities and their H₂O₂ concentrations. We also noticed that the concentration of H₂O₂ is very low in fresh solutions of the gossypol Schiff bases. As the time increased from 5 h to 10 h, the concentration of H₂O₂ increased slowly. Hafez et al. demonstrated that applying ROS at the early stage of TMV infection is necessary to their anti-TMV activities; therefore, we suppose the H₂O₂ concentration may not be high enough in the primary stage of infection.^17b,c To confirm this view, the anti-TMV activities of H₂O₂ at different concentrations were measured. It can be seen from Fig. 5 that H₂O₂ at 10 mmol L⁻¹ exerts moderate anti-TMV activity; however, no activity was observed at a concentration of 100 μmol L⁻¹, which explains why the H₂O₂ producing abilities of those gossypol Schiff bases do not positively correlate with their anti-TMV activities.

Fig. 4 (a) H₂O₂ concentration vs. anti-TMV activity at 500 μg mL⁻¹: H₂O₂ concentration is not positively correlated with anti-TMV activity, and compounds 5 and 8 show relatively higher H₂O₂ concentrations. (b) O₂˙⁻ production rate vs. anti-TMV activity at 500 μg mL⁻¹: O₂˙⁻ production rate is positively correlated with anti-TMV activity. Compounds 4 and 5, which contain carboxyl groups, not only displayed higher O₂˙⁻ production but also had higher anti-TMV activities than compounds 15, 13 and 14, separately. The compounds marked with * were measured in mixed solution comprising 10% DMF and 90% double distilled water, and the rest were measured in a solution of 100% water. The means of three independent experiments ± SD are shown.

Fig. 5 Anti-TMV activities (in vivo) of compound 11 (500 μg mL⁻¹), SOD (120 units per mL), CAT (200–500 units per mL), compound 11 (500 μg mL⁻¹) with SOD (120 units per mL), and compound 11 (500 μg mL⁻¹) with CAT (200–500 units per mL); anti-TMV activities of H₂O₂ at concentrations of 10 mmol L⁻¹, 1 mmol L⁻¹, 100 μmol L⁻¹ and 50 μmol L⁻¹. Blank refers to double distilled water. The means of three independent experiments ± SD are shown.

By contrast, the O₂˙⁻ production rate can reach a fixed level within an hour. All the amino acid gossypol Schiff bases showed higher O₂˙⁻ production rates than gossypol (Fig. 4b). Additionally, there is a significantly positive correlation between their O₂˙⁻ production rates and anti-TMV activities (in vivo).

To further verify the importance of H₂O₂ and O₂˙⁻, the most potent compound, 11, was mixed with superoxide dismutase (SOD) or catalase (CAT) when spraying it on the tobacco leaves (Fig. 5). The results of this study demonstrated that application of SOD significantly reduced the anti-TMV activities of 11, which suggests that O₂˙⁻ is involved in the TMV-inhibiting activity. By contrast, the addition of CAT did not reduce the anti-TMV activity as much as SOD, which implies that H₂O₂ may also be involved in the anti-TMV mechanism but may not be as important as O₂˙⁻. However, SOD and CAT did not totally suppress the anti-TMV activity of 11. The reason for this is that SOD and CAT themselves display anti-TMV activities (Fig. 5).^17a,c

Water solubility vs. ROS-producing activity. Compound 5 shows an obviously higher O₂˙⁻ production rate and higher anti-TMV activity than 13 and 14 (Fig. 4b). The same result is observed when comparing compound 4 with compound 15 (Fig. 4b). This result indicates that carboxyl groups in the amino acid Schiff bases not only can improve their water solubility, but also can improve their anti-TMV activities and O₂˙⁻ producing abilities.

As mentioned above, water solubility and O₂˙⁻ production rate both appear to be related to the anti-TMV activities of these compounds. We wished to determine which is more important: water solubility, ROS producing ability, or both.

For this purpose, the relationship between the water solubility and anti-TMV activities of different gossypol amino acid Schiff bases were explored; the results revealed that the water solubilities of the compounds are not positively correlated with their anti-TMV activities (Fig. 3, ESI†).

The O₂˙⁻ producing abilities of polyphenol compounds are higher under alkaline conditions than in neutral conditions. On the basis of this, we propose that the carboxyl group in compounds 4 and 5, which are in form of sodium salts, offered relatively more alkaline conditions than gossypol Schiff bases 13, 14, and 15, and that these alkaline conditions facilitate the production of O₂˙⁻ (Fig. 4b). In order to verify this, acidic conditions were used to convert some percentage of the carboxyl groups from the sodium salts to carboxylic acid groups. The study revealed that when compounds 4, 5, 13, 14 and 15 were dissolved in solvent comprising 10% DMF and 90% NaAc–HAc buffer solution (pH = 4), the anti-TMV activities and O₂˙⁻ production rates of these compounds all decreased (Fig. 6). The differences between compounds 4 and 15 and between compound 5 and compounds 13 and 14 in this solvent are smaller than those in solvent containing 10% DMF and 90% water (Fig. 4b).

Fig. 6 Anti-TMV activities of compounds 4, 5, 13, 14 and 15 in mixed solvent comprising 10% DMF and 90% buffer (pH = 4). Blank refers to mixed solvent comprising 10% DMF and 90% buffer (pH = 4). The means of three independent experiments ± SD are shown.

Relationship between anti-TMV activity and chirality

Each pair of (±)-gossypol Schiff bases with chiral amino acids, such as 4 and 16, 6 and 17, 7 and 18, and 11 and 19, differs only in the optical configurations of the amino acid. As we can see in Fig. 7, compounds (+)-4 and (−)-16 and compounds (−)-4 and (+)-16 are enantiomers, separately; therefore, they theoretically have the same ROS-producing abilities. Accordingly, compounds 4 and 16, which are composed of (+)-4 and (−)-4 and (+)-16 and (−)-16, respectively, have the same ROS-producing abilities. Therefore, if their anti-TMV activities are only related to their ROS-producing abilities, their anti-TMV activities should also be the same. However, as is shown in Table 1, the (±)-gossypol Schiff bases of D-phenylalanine, D-threonine, D-serine and D-asparagine all displayed higher anti-TMV activities than the L-amino acid Schiff bases (4, 6, 7, 11 > 16, 17, 18, 19). This reminds us that, in addition to the ROS-producing abilities of these compounds, chirality may be another factor influencing their anti-TMV activities. One reason for this effect may be the different responses of tobacco to D and L amino acid derivatives, which may influence the multiplication of TMV. Another reason may be the slower rate of degradation of the D amino acids compared to their corresponding L-isomers.²⁹ We found that the four isomers of the methyl tryptophan gossypol Schiff bases (22, 23, 24, and 25) displayed different anti-TMV activities; compound 23, which is prepared from methyl D-tryptophan and (−)-gossypol, exerted the highest anti-TMV activity, even higher than that of the D-tryptophan gossypol Schiff base (5). Therefore, the (+) and (−)-configurations of gossypol should also be considered as one of the variables that could affect the anti-TMV activities of these compounds.

Fig. 7 Compound 4, composed of (+)-4 and (−)-4, and compound 16, composed of (+)-16 and (−)-16.

Use of ROS as an index to evaluate the influence of chirality. The differences between the anti-TMV activities of each pair of diastereomers should be considered as a reflection of different interactions between the compounds and tobacco leaves, or between the compounds and TMV. In this article, we used ROS as an index to evaluate the interactions between the compounds and tobacco leaves.

ROS_-plant = ROS_-exo + ROS_-endo (1)

ROS_-endo = ROS_-before + ROS_-after (2)

ROS_-plant = ROS_-exo + ROS_-before + ROS_-after (3)

As we can see in eqn (1), ROS_-plant is the only ROS data we can obtain experimentally. This value comprises ROS_-exo (the ROS produced by the gossypol Schiff base itself) and ROS_-endo (the ROS related to tobacco). ROS_-endo is composed of ROS_-before (the ROS in tobacco before spraying the compound on the leaf) and ROS_-after (the ROS produced by the interaction between the compound and tobacco).

To make ROS_-plant a useful index, we must simplify eqn (3) in a reasonable way. As is known, the ROS_-before values are different in different leaves. To simplify eqn (3), we used the same leaf to perform all the experiments; thus, the ROS_-before values of the different gossypol Schiff bases are the same, and eqn (3) can be transferred into eqn (4). We also notice that, in this case, the value of ROS_-before is equal to that of ROS_-plant of the mock (ROS_-mock) sample.

ROS_-plant = G + ROS_-exo + ROS_-after (G = ROS_-before = ROS_-mock) (4)

As we can see in eqn (4), the differences between the ROS_-plant values of the gossypol Schiff bases and the ROS_-mock value of the sample are equal to the sum of ROS_-exo and ROS_-after, and they reflect the changes of the ROS in tobacco after the compounds have been sprayed on the tobacco leaf.

When the amino acid gossypol Schiff bases have the same ROS-producing abilities, ROS_-exo can be regarded as another constant H; therefore, eqn (4) can be converted into eqn (5).

ROS_-plant = G + H + ROS_-after (G = ROS_-before, H = ROS_-exo) (5)

According to eqn (5), for example, the difference between the ROS_-plant values of compounds 11 and 18 is equal to that of ROS_-after; therefore, this difference can reflect the different responses of tobacco to the D and L amino acid gossypol Schiff bases.

As we can see in Fig. 8, all the H₂O₂ concentrations and O₂˙⁻ production rates are higher than that of the mock sample, except for the H₂O_2-plant of 18. This implies that the application of amino acid gossypol Schiff bases could stimulate the production of ROS in tobacco leaves. The asparagine gossypol Schiff bases (compounds 11 and 19) exerted the highest ROS levels among the amino acid Schiff bases. It also can be seen that the H₂O_2-plant and the O₂˙⁻_-plant data have opposite trends: the gossypol Schiff bases derived from D-amino acids exerted higher H₂O₂ concentrations than those derived from L-amino acids. In contrast, the O₂˙⁻ production rates of the D-amino acid gossypol Schiff bases were lower than those of the L-amino acid gossypol Schiff bases.

Fig. 8 (±)-Gossypol Schiff bases of D-tryptophan, D-threonine, D-serine and D-asparagine (5, 6, 7, 11, respectively) have higher H₂O_2-plant values than the corresponding Schiff bases of L-amino acids (20, 17, 18, 19). In contrast, the rule for their O₂˙⁻_-plant production rates appears to follow the opposite trend (20, 17, 18, 19 > 5, 6, 7, 11). The means of three independent experiments ± SD are shown.

From these experiments, we propose that the gossypol Schiff base first increases the O₂˙⁻ production rate in tobacco; then, the SOD inside the tobacco leaves converts O₂˙⁻ into H₂O₂.

Mⁿ⁺–SOD + O₂˙⁻ + 2H⁺ → M⁽ⁿ⁺¹⁾–SOD + H₂O₂

The SOD in the tobacco leaf treated with D-amino acid gossypol Schiff base seems more active than that in the leaf treated with L-amino acid gossypol Schiff base.

Conclusions

Carboxyl groups, which are related to water solubility, not only improve the anti-TMV activities of amino acid gossypol Schiff bases but also enhance their ROS-producing activities. Furthermore, the properties of R groups (length and hydrophilic) in amino acids also seem to be important factors for the anti-TMV activities of these compounds. Comparing their ROS-producing abilities with their anti-TMV activities, we found that the O₂˙⁻ producing abilities of these compounds are more correlative with their anti-TMV activities than their H₂O₂-producing abilities. These findings suggest that O₂˙⁻ is effective at the beginning of the TMV replication process, which results in its positive correlation with anti-TMV activity. Chirality should also be a factor affecting the anti-TMV activities of these compounds. We further used ROS as an index to explore the interactions between gossypol Schiff bases and tobacco leaf. The results also demonstrate that optically pure gossypol Schiff bases can be considered for further development as a new class of plant protection agents.

Acknowledgements

We are grateful to the National Natural Science Foundation of China (21132003, 21421062, 21372131, 21672117), the Specialized Research Fund for the Doctoral Program of Higher Education (20130031110017) and the Tianjin Natural Science Foundation (16JCZDJC32400) for generous financial support for our programs.

Notes and references

1. X. H. Qian, P. W. Lee and S. Cao, J. Agric. Food Chem., 2010, 58, 2613–2623.
2. J. N. Seiber, J. Agric. Food Chem., 2011, 59, 1–21.
3. Y. M. Li, L. H. Wang, S. L. Li, X. Y. Chen, Y. M. Shen, Z. K. Zhang, H. P. He, W. B. Xu, Y. L. Shu, G. D. Liang, R. X. Fang and X. J. Hao, Proc. Natl. Acad. Sci. U. S. A., 2007, 104, 8083–8088.
4. D. J. Newman and G. M. Cragg, J. Nat. Prod., 2012, 75, 311–335.
5. (a) D. L. Vander Jagt, L. M. Deck and R. E. Royer, Curr. Med. Chem., 2000, 7, 479–498; (b) R. E. Royer, L. M. Deck, N. M. Campos, L. A. Hunsaker and D. L. Vander Jagt, J. Med. Chem., 1986, 29, 1799–1801.
6. T. S. Lin, R. Schinazi, B. P. Griffith, E. M. August, B. F. Eriksson, D. K. Zheng, L. A. Huang and W. H. Prusoff, Antimicrob. Agents Chemother., 1989, 33, 2149–2151.
7. (a) P. Yildirim-Aksoy, C. Lim, M. K. Dowd, P. J. Wan, P. H. Klesius and C. Shoemaker, J. Appl. Microbiol., 2004, 97, 87–92; (b) S. D. Jolad, R. M. Wiedhopf and J. R. Cole, J. Pharm. Sci., 1975, 64, 1889–1890; (c) M. D. Shelley, L. Hartley, P. W. Groundwater and R. G. Fish, Anticancer Drugs, 2000, 11, 209–216; (d) F. Yan, X. X. Cao, H. X. Jiang, X. L. Zhao, J. Y. Wang, Y. H. Lin, Q. L. Liu, C. Zhang, B. Jiang and F. A. Guo, J. Med. Chem., 2010, 53, 5502–5510.
8. (a) L. M. Deck, R. E. Royer, B. B. Chamblee, V. M. Hernandez, R. R. Malone, J. E. Torres, L. A. Hunsaker, R. C. Piper, M. T. Makler and D. L. Vander Jagt, J. Med. Chem., 1998, 41, 3879–3887; (b) B. Polsky, S. J. Segal, P. A. Baron, J. W. M. Gold, H. Ueno and D. Armstrong, Contraception, 1989, 39, 579–587.
9. (a) J. Yang, G. Chen, L. L. Li, W. Pan, F. Zhang, J. Yang, S. Wu and P. Tien, Bioorg. Med. Chem. Lett., 2013, 23, 2619–2623; (b) J. Yang, F. Zhang, J. R. Li, G. Chen, S. W. Wu, W. J. Ouyang, W. Pan, R. Yu, J. X. Yang and P. Tien, Bioorg. Med. Chem. Lett., 2012, 22, 1415–1420.
10. (a) K. Wichmann, A. Vaheri and T. Luukkainen, Am. J. Obstet. Gynecol., 1982, 142, 593–594; (b) R. J. Radloff, L. M. Deck, R. E. Royer and D. L. Vander Jagt, Pharmacol. Res. Commun., 1986, 18, 1063–1067.
11. S. A. Vichkanova, A. I. Oifa and L. V. Goryunova, Antibiotiki, 1970, 15, 1071–1073.
12. Y. F. Wu, CN Pat., 1489904A, 2004.
13. (a) L. Li, Z. Li and Q. M. Wang, J. Agric. Food Chem., 2014, 62, 11080–11088; (b) L. Li, Z. Li, K. L. Wang, Y. X. Liu and Q. M. Wang, Bioorg. Med. Chem., 2016, 24, 474–483.
14. (a) Y. H. Miura, I. Tomita, T. Watanabe, T. Hirayama and S. Fukui, Biol. Pharm. Bull., 1998, 21, 93–96; (b) T. Nakayama, Y. Enoki and K. Hashimoto, Food Sci. Technol. Int., 1995, 1, 65–69; (c) L. H. Long, M. V. Clement and B. Halliwell, Biochem. Biophys. Res. Commun., 2000, 273, 50–53; (d) M. Mochizuki, S. Yamazaki, K. Kano and T. Ikeda, Biochim. Biophys. Acta, 2002, 1569, 35–44.
15. (a) S. Marklund and G. Marklund, Eur. J. Biochem., 1974, 47, 474; (b) C. Xu, S. Liu and Z. Q. Liu, Anal. Chim. Acta., 2013, 793, 53–60.
16. M. Akagawa, T. Shigemitsu and K. Suyama, Biosci., Biotechnol., Biochem., 2003, 67, 2632–2640.
17. (a) N. Doke and Y. Ohashi, Physiol. Mol. Plant Pathol., 1988, 32, 63–175; (b) R. Bacsó, Y. Hafez, Z. Király and L. Király, Acta Phytopathol. Entomol. Hung., 2011, 46, 1–10; (c) Y. M. Hafez, R. Bacsó, Z. Király, A. Künstler and L. Király, Phytopathology, 2012, 102, 848–856.
18. (a) P. Kovacic, Curr. Med. Chem., 2003, 10, 2711–2718; (b) R. Zaidi and S. M. Hadi, Biochem. Int., 1992, 28, 1135; (c) M. J. Laughton, B. Halliwell, P. J. Evans and J. R. S. Hoult, Biochem. Pharmacol., 1989, 38, 2859; (d) R. Zaidi and S. M. Hadi, J. Biochem. Toxicol., 1992, 7, 213; (e) D. Wu and Y. Yu, Proc. Chin. Acad. Med. Sci. Peking Union Med. Coll., 1986, 1, 150.
19. N. Suzuki, M. Kōchi, N. Wada, S. Mashiko, T. Nomoto and B. Yoda, Biosci., Biotechnol., Biochem., 1992, 56, 409–411.
20. S. Matsushita and F. Buki, Agr. Bioi. Chern., 1965, 29, 792–795.
21. E. Meucci and M. C. Mele, Amino Acids, 1997, 12, 373–377.
22. H. X. Jiang, X. X. Cao, H. Huang and B. Jiang, Tetrahedron: Asymmetry, 2007, 18, 2437–2441.
23. Y. Jiang, A. C. S. Woollard and S. P. Wolff, FEBS Lett., 1990, 268, 69–71.
24. A. G. Wang and G. H. Luo, Plant Physiol. Commun., 1990, 84, 2895–2898.
25. K. L. Wang, B. Su, Z. W. Wang, M. Wu, Z. Li, Y. N. Hu, Z. J. Fan, N. Mi and Q. M. Wang, J. Agric. Food Chem., 2010, 58, 2703–2709.
26. (a) B. Tian, S. W. Wu and J. Yang, CN Pat., 101844994A, 2010; (b) J. Yang, S. W. Wu, L. L. Li, W. Pan, F. Zhang and J. R. Li, CN Pat., 102775320, 2012.
27. P. Przybylski, A. Huczynski, K. Pyta, B. Brzezinski and F. Bartl, Curr. Org. Chem., 2009, 13, 124–148.
28. (a) S. A. Matlin, S. Roshdy, Q. B. Cass, L. C. G. Freitas, R. L. Longo and I. Malvestiti, J. Braz. Chem. Soc., 1990, 1, 128–133; (b) P. Przybylski, K. Pyta, M. Ratajczak-Sitarz, A. Katrusiak and B. Brzezinski, Polym. Chem., 2009, 83, 747–759; (c) P. Przyby lski, W. Schilf, B. Kamienski, B. Brzezinski and F. Bartl, Magn. Reson. Chem., 2008, 46, 534–544; (d) N. S. Ilkevych, G. Schroeder, V. I. Rybachenko, K. Y. Chotiy and R. A. Makarova, Spectrochim. Acta, Part A, 2012, 86, 328–335.
29. S. M. Rodriguez, A. M. Gomez, F. R. Vico and J. C. Jimene, Chem. Biodiversity, 2010, 7, 1531–1548.

---

Footnote

† Electronic supplementary information (ESI) available: ¹H NMR, ¹³C NMR, and HRMS spectra of compounds 3–26. See DOI: 10.1039/c6ra14015g

---