Nitroreductase Increases Menadione-Mediated Oxidative Stress in Aspergillus nidulans

ABSTRACT

Nitroreductases (NTRs) catalyze the reduction of a wide range of nitro-compounds and quinones using NAD(P)H. Although the physiological functions of these enzymes remain obscure, a tentative function of resistance to reactive oxygen species (ROS) via the detoxification of menadione has been proposed. This suggestion is based primarily on the transcriptional or translational induction of an NTR response to menadione rather than on convincing experimental evidence. We investigated the performance of a fungal NTR from Aspergillus nidulans (AnNTR) exposed to menadione to address the question of whether NTR is really an ROS defense enzyme. We confirmed that AnNTR was transcriptionally induced by external menadione. We observed that menadione treatment generated cytotoxic levels of O₂^•−, which requires well-known antioxidant enzymes such as superoxide dismutase, catalase, and peroxiredoxin to protect A. nidulans against menadione-derived ROS stress. However, AnNTR was counterproductive for ROS defense, since knocking out AnNTR decreased the intracellular O₂^•− levels, resulting in fungal viability higher than that of the wild type. This observation implies that AnNTR may accelerate the generation of O₂^•− from menadione. Our in vitro experiments indicated that AnNTR uses NADPH to reduce menadione in a single-electron reaction, and the subsequent semiquinone-quinone redox cycling resulted in O₂^•− generation. We demonstrated that A. nidulans nitroreductase should be an ROS generator, but not an ROS scavenger, in the presence of menadione. Our results clarified the relationship between nitroreductase and menadione-derived ROS stress, which has long been ambiguous.

IMPORTANCE Menadione is commonly used as an O₂^•− generator in studies of oxidative stress responses. However, the precise mechanism through which menadione mediates cellular O₂^•− generation, as well as the way in which cells respond, remains unclear. Elucidating these events will have important implications for the use of menadione in biological and medical studies. Our results show that the production of Aspergillus nidulans nitroreductase (AnNTR) was induced by menadione. However, the accumulated AnNTR did not protect cells but instead increased the cytotoxic effect of menadione through a single-electron reduction reaction. Our finding that nitroreductase is involved in the menadione-mediated O₂^•− generation pathway has clarified the relationship between nitroreductase and menadione-derived ROS stress, which has long been ambiguous.

INTRODUCTION

Reactive oxygen species (ROS) are by-products of aerobic metabolism generated endogenously from immune cells (1, 2), or exogenously by drug exposure (3–6). ROS include the superoxide anion (O₂^•−), hydrogen peroxide (H₂O₂), and hydroxyl radicals (OH^•), all of which confer reactivity to different biological targets such as lipids, proteins, and DNA. In both prokaryotic and eukaryotic organisms, high levels of oxidative stress generated by intracellular ROS are involved in many pathological processes, damaging living cells (2, 7). Therefore, investigations into cell sensitivity, adaptivity, and resistance to cytotoxic ROS are important.

Most studies of oxidative stress at the molecular level have used free-radical-generating compounds. Menadione (2-methyl-1,4-naphthoquinone) is a useful O₂^•− generator due to its water solubility and ease of diffusion (8, 9). Menadione appears to generate ROS by reducing one-electron quinone to semiquinone (9, 10). The semiquinone is then autoxidized back to quinone under aerobic conditions. The by-product of this reaction is O₂^•−. In biological systems, quinones are generally reduced by coenzyme NAD(P)H in the presence of cellular reductases, including NADH ubiquinone oxidase, NADH cytochrome b₅ reductase, and NADPH cytochrome P-450 reductase (10). These three reductases therefore appear to be responsible for the generation of ROS by menadione.

Nitroreductases are a family of evolutionarily related proteins involved in reducing nitroaromatic and nitroheterocyclic compounds using flavin adenine dinucleotide (FAD) or flavin mononucleotide (FMN) as the cofactor and NAD(P)H as a reducing equivalent (11). Bacterial nitroreductases can be classified as oxygen insensitive (type I) or oxygen sensitive (type II) according to their biochemical characteristics (11). The oxygen-insensitive nitroreductases catalyze the two-electron reduction of nitro-compounds to produce nitroso and hydroxylamine and finally primary amines. The oxygen-sensitive nitroreductases catalyze the one-electron reduction of nitro groups to form a nitro anion radical, which subsequently reacts with oxygen, producing O₂^•− and regenerating the initial nitro-compound. As well as reducing nitro-compounds, nitroreductases have been shown to be efficient quinone reductases. Oxygen-insensitive bacterial nitroreductases have been reported to reduce lawsone to its hydroquinone, which can function as a redox mediator in the subsequent reduction of azo compounds outside the cells (12). Therefore, bacterial nitroreductases have received considerable attention because they can be used for the biodegradation of environmental pollutants (11). However, these compounds may not be the physiological substrates of nitroreductases because they are the products of human activities. The real physiological functions of nitroreductases remain obscure.

Some studies have shown that nitroreductases may be involved in the response to anti-oxidative stress. Bacterial nitroreductases, which include nfsA in Escherichia coli (13, 14), snrA in Salmonella enterica serovar Typhimurium, and nprA in Rhodobacter capsulatus are induced at high levels by the oxidant-producing herbicide paraquat (13–15). In Saccharomyces cerevisiae, disruption of two nitroreductases, Frm2p and Hbn1p, increased the sensitivity of yeast to nitro-compound-derived oxidative stress (16). However, the definitive conclusion that nitroreductases (NTRs) are antioxidants could not be drawn only from these transcriptional and phenotypic findings. Another indication that nitroreductases are involved in defense against oxidative stress is the observation that the deazaflavin-dependent nitroreductase from Mycobacterium tuberculosis protects mycobacteria from oxidative stress by catalyzing the two-electron reduction of menadione to less-toxic hydroquinone, competing with the one-electron reduction pathway to avoid O₂^•− formation (17). However, deazaflavin-dependent nitroreductase is a special case among nitroreductases and only spreads in a few species of archaea and bacteria (17, 18). In fungi, one of the Aspergillus nidulans nitroreductase-like proteins, AnNTR (encoded by AN2343), was identified and shown to be significantly induced under menadione-mediated oxidative stress conditions (19–21). However, it is not clear whether fungal nitroreductase really plays a role in antioxidative stress because the effect on menadione resistance of deleting this gene was not investigated.

Filamentous fungi are important in industry, having important roles in pharmacology and medicine. Fungi frequently encounter high levels of oxidative stress, generated by host-pathogen defense and fermentation (22, 23). A. nidulans is an excellent model fungus with which to study such important processes in genetics and cell biology. Therefore, the elucidation of a novel antioxidant system in A. nidulans is valuable, both because it provides a deeper understanding of the biology of this important group of organisms and because of the potential industrial and medical applications that may arise from this work.

According to the current literature, A. nidulans nitroreductase appears to play a direct role in resisting menadione-derived oxidative stress. However, when we characterized the phenotype of the AN2343 deletion mutant, we found that a deficiency of AN2343 instead enhanced the strain’s resistance to menadione. That observation triggered our current effort to investigate in depth the mechanisms underlying the promotion of menadione-mediated cytotoxicity by AnNTR.

RESULTS

AnNTR had a deleterious effect on protection against menadione.

The nitroreductase-like protein AnNTR, which is encoded by AN2343, has been identified as a protein induced in menadione-exposed A. nidulans cells (20, 21). However, there is no evidence for the involvement of AN2343 in protection against oxidative stress. To investigate the relationship between AnNTR and oxidative stress, we investigated the transcription profile of AN2343 in response to exposure to external menadione. AN2343 expression was significantly induced by 0.8 mM menadione (Fig. 1A), suggesting the possibility that AN2343 is involved in menadione metabolism. We deleted AN2343 to examine the disrupted strain’s sensitivity to oxidative stress induced by menadione and H₂O₂. There was no detectable difference in growth between ΔAN2343 and the control strain (wild type [WT]) (Fig. 1B) under normal conditions. Menadione was toxic to both strains in a dose-dependent manner. However, the deletion of AN2343 appeared to relieve the cytotoxicity caused by menadione (Fig. 1B). The application of 75 μM menadione inhibited the cell growth of WT more strongly than that of ΔAN2343, and the difference was amplified when the concentration of menadione was increased to 100 μM. Subsequent reintroduction of AN2343 fused with green fluorescent protein (GFP; ΔAN2343-com), flanked by its native promoter and terminator, restored the sensitivity to menadione compared to the WT (Fig. 1C), confirming the adverse function of AnNTR in the detoxification of menadione. The function of AnNTR-GFP, as well as the whole-cell fluorescence imaging described in Fig. 1C, indicated that AnNTR is localized to the cytosol. In a control experiment, AN2343 did not increase H₂O₂ damage because the deletion of AN2343 did not affect the sensitivity of the strain to H₂O₂ (see Fig. S2A). These observations do not support the proposition that AnNTR may act as an antioxidant enzyme that protects the cell against menadione toxicity but suggested an opposing hypothesis: that AnNTR may be involved in the conversion of menadione to toxic metabolites in A. nidulans.

FIG 1.

Response of AN2343 to external menadione. (A) Q-RT-PCR analysis of AN2343 induced by menadione. After preculture for 16 h, the WT strain was exposed to 0.8 mM menadione (MEN) for 3 h. The relative mRNA levels were normalized to that of actA. Error bars represent the standard deviations (SD) of three independent experiments. (B) Deleterious effect of AnNTR on protection against menadione. After incubation of A. nidulans conidiospores on agar plates containing various concentrations of menadione at 37°C for 48 h, the colony morphology was examined to determine the sensitivities to menadione. (C) Recovery of menadione sensitivity by the expression of AnNTR-GFP in ΔAN2343 and the subcellular localization of AnNTR-GFP. The AnNTR::GFP fusion protein was expressed using the native promoter and terminator of AN2343. Conidiospores (1 × 10⁵) from the WT, ΔAN2343, and ΔAN2343-com (ΔAN2343 complemented with AnNTR-GFP) strains were cultivated on minimal medium agar plates supplemented with 125 μM MEN at 37°C for 48 h. The fluorescence from strains cultivated in liquid MM without menadione for 16 h was observed by using confocal laser scanning microscopy.

Disruption of AN2343 decreased intracellular O₂^•− derived from menadione.

As an O₂^•−-producing agent, menadione is believed to trigger cellular oxidative stress. However, its physiological effects may be more extensive. For example, it may destroy cellular 4Fe-4S proteins, resulting in the production of deleterious OH^• radicals (24), and directly affecting the GSH pool of cells (25). It may also chemically modify cell components (26), generating nonoxidative stress in cells. The question of whether O₂^•− originating from menadione results in cytotoxicity has not been addressed experimentally. We quantified the menadione-derived intracellular O₂^•− using dihydroethidium (DHE), a membrane-permeable probe that reacts with O₂^•− to form the highly fluorescent ethidium cation. In the absence of menadione, there was only weak fluorescence when the cells were loaded with DHE (Fig. 2A). The fluorescence was completely quenched when the ROS scavenger N-acetyl-l-cysteine (NAC; 10 mM) was applied, indicating the existence of a small amount of intracellular O₂^•− production (Fig. 2A), a by-product of cellular respiration under normal physiological conditions (27). The application of 300 μM menadione induced a significant rise in fluorescence, and the menadione-induced elevation of ROS was completely prevented by the presence of NAC (Fig. 2A). These results provide evidence for the generation of O₂^•− in response to menadione. We estimated the oxidative stress level caused by menadione by observing the phenotypes of sodA, prxA, and catB deletion mutants on menadione-containing plates because the fungus generally eliminates ROS using these well-known antioxidants. The gene sodA encodes a copper-zinc superoxide dismutase, the key O₂^•− dismutase responsible for superoxide dismutation during oxidative stress (28). The genes prxA and catB encode a key peroxiredoxin and a major catalase, respectively, and are indispensable for defense against H₂O₂ (29–31). Compared to the WT phenotype, the growth of cells treated with concentrations of menadione as low 50 μM was significantly inhibited (Fig. 2B). The growth of ΔsodA cells under menadione concentrations of 100 μM was completely blocked. Menadione also inhibited ΔprxA and ΔcatB strains in a dose-dependent manner (Fig. 2B). This observation indicated that ROS induced by menadione, including O₂^•− and its decomposition product H₂O₂, produced considerable oxidative stress in cells.

FIG 2.

AnNTR is an efficient O₂^•−-producing enzyme in A. nidulans in the presence of menadione. (A) Images of intracellular O₂^•− levels using an O₂^•− specific fluorescent probe. After 12 h of cultivation, the strains were treated with or without menadione (MEN; 300 μM), followed by incubation with 10 μM dihydroethidium (DHE) for another 30 min, and then observed using fluorescence microscopy. The O₂^•− scavenger NAC (10 mM) was added to block O₂^•− generation as a control experiment. (B) ROS-resistant enzymes were involved in the menadione stress defense. Conidia from WT, ΔprxA, ΔsodA, and ΔcatB strains were spotted onto MM plates with or without the indicated concentrations of menadione, followed by incubation at 37°C for 48 h. (C) Effects of AN2343 deletion on intracellular O₂^•− generation. After 16 h of cultivation, the mycelia of the WT and ΔAN2343 strains were exposed to 0.8 mM menadione for another 6 h, followed by a further 1 h of incubation with DHE (10 μM). Mycelia were disrupted by grinding in liquid nitrogen, and the fluorescence in the supernatant was measured. Values (means ± the SD of three independent experiments) represent relative fluorescence units (RFU) per mg of total cell protein. (D) Menadione-induced cellular oxidative damage is reflected by the inhibition of the activity of intracellular aconitase. After menadione treatment, the mycelia of the WT and ΔAN2343 strains were disrupted, and the activities in the cell extracts were measured. The data are the means ± the SD of three independent experiments. One-way ANOVA was used to test for significant differences among the means (*, P < 0.05; **, P < 0.01).

There are two plausible mechanisms for the way in which AnNTR could increase menadione-mediated oxidative stress in A. nidulans: (i) suppression of the expression of ROS resistant genes and (ii) direct involvement in menadione-derived ROS generation. We compared the transcriptional profiles of the ROS resistance genes, including sodA, catB, and prxA, in response to menadione in WT and ΔAN2343 strains (see Fig. S3). Treatment with 0.8 mM external menadione induced the expression of all of these genes to different extents, without obvious differences between the WT and ΔAN2343 strains. This finding appears to exclude the possibility that AnNTR participates in the transcriptional regulation of ROS resistance genes. To investigate whether AnNTR is directly involved in menadione-dependent O₂^•− production in A. nidulans cells, we compared the changes in intracellular O₂^•− levels before and after exposure of WT and ΔAN2343 strains to menadione by measuring the fluorescence intensity of DHE. We found that the absence of AnNTR did not change O₂^•− accumulation under nonstressed conditions but decreased the level of O₂^•− by one-third compared to that of the WT under menadione stress conditions (Fig. 2C), suggesting that AnNTR is an efficient menadione-dependent O₂^•− generator in A. nidulans.

We estimated the extent of oxidative damage to cells caused by O₂^•− derived from menadione converted by AnNTR. Aconitase is a major target of ROS because of its particularly sensitive 4Fe-4S cluster (32). No difference in cellular aconitase activity was observed between WT and ΔAN2343 cells under normal conditions (Fig. 2D). Treatment with 0.8 mM menadione inhibited the activity of cellular aconitase in the WT to a greater extent than in the ΔAN2343 strain (Fig. 2D), indicating that menadione-derived O₂^•− converted by AnNTR led to substantial cellular oxidative damage in A. nidulans.

Recombinant AnNTR exhibited menadione reductase activity.

To investigate the mechanisms by which AnNTR catalyzes the generation of O₂^•− from menadione, we expressed and purified recombinant AnNTR using an E. coli expression system. AnNTR can only be produced successfully in the E. coli thioredoxin N-terminal tagged form (AnNTR-Trx, 35 kDa) (see Fig. S4A). After purification, the Trx tag was removed from the recombinant protein to facilitate analysis of its activity. To indicate the occurrence of the reduction reaction, we used a functional assay based on the reduction of MTT to formazan by reduced substrates (33). Formazan has a characteristic absorption peak at 590 nm. Using NADPH as an electron donor, we found that, in the absence of menadione, the addition of FMN, but not FAD, resulted in a significant increase in absorbance at 590 nm (Fig. 3A). Under the same reaction conditions, NADH did not produce any change in absorbance (Fig. 3A). These findings indicate that AnNTR is an efficient NADPH-dependent FMN reductase. Adding menadione to the existing reaction solution further promoted the generation of formazan (Fig. 3A). Replacing AnNTR with its protein tag TrxA did not facilitate dye generation, excluding the possibility that trace amounts of TrxA were involved in the reaction as a purification contaminant (Fig. 3A). These results indicated that menadione is a good substrate for AnNTR when NADPH is used as an electron donor and FMN as a cofactor.

FIG 3.

AnNTR drives the one-electron metabolism of menadione, leading to ROS generation via redox cycling. (A) Reduction activity of recombinant AnNTR toward menadione. MTT was used as the ultimate electron acceptor of menadione reduction, and the MTT reduction product formazan was measured at 590 nm to measure the reduction due to menadione. The reaction mixture was 0.5 ml of sodium phosphate buffer (50 mM [pH 7.4]), NADPH (100 μM), DTPA (100 μM), FMN (10 μM), MTT (0.5 mM), and AnNTR (1.5 μg). The arrow indicates the time point of menadione (MEN; 50 μM) addition. As three controls, TrxA (2.6 μg) replacing AnNTR, NADH (100 μM) replacing NADPH, and FAD (10 μM) replacing FMN were added to the reaction solution in the presence of menadione. (B) No changes in menadione concentration were observed before or after menadione reduction catalyzed by AnNTR. After incubation for 70 min at 25°C, the reaction mixture was analyzed by using HPLC. The mixture without AnNTR was the control. (C) Confirmation of O₂^•− generation during menadione reduction procedure by EPR spectroscopy. DMPO was used as an O₂^•− trapper, and the four successive peaks are the characteristic spectrum of a DMPO–O₂^•− adduct. EPR spectra of the spin adduct of the reaction mixture obtained in the absence or presence of AnNTR or AnNTR plus SOD are shown. (D) H₂O₂ generation during the menadione reduction procedure. H₂O₂ was measured using hydrogen peroxide assay kits, and the absorbance was measured at 540 nm. Catalase (CAT) was employed to eliminate H₂O₂ in the reaction solution. The data are the means ± the SD of three independent experiments.

We analyzed the reaction mixture using high-pressure liquid chromatography (HPLC) to determine the fate of the reduced menadione catalyzed by AnNTR (Fig. 3B). Menadione had a retention time of 14.8 min, and chromatography profiles showed no time-dependent decrease in the substrate peaks. These results suggested that the metabolism of menadione by AnNTR should be a one-electron reductive pathway, in which an unstable semiquinone radical is first generated, and subsequently reoxidized to menadione via redox cycling under aerobic conditions. Back-oxidation of menadione from semiquinone usually generates O₂^•− without menadione consumption (34), a process that could explain the nonquantitative changes in menadione observed in the reaction mixture. Another additive agent, FMN, which has a retention time of 13.9 min, was also detected (Fig. 3B). The amounts of FMN after the reaction were not substantially reduced, which should be a property of an electron transfer mediator in redox reactions.

To confirm the generation of O₂^•−, the reaction products of the menadione reduction were analyzed using EPR spectroscopy after combination with DMPO [5,5-dimethyl-1-pyrroline-N-oxide], an O₂^•− trapper (Fig. 3C). This is one of the most widely used methods for the determination of free radicals (35). Without AnNTR, no spectra were detectable in the reaction solution. However, the addition of AnNTR to the reaction mixture produced a strong EPR signal corresponding to the DMPO–O₂^•− adduct. This signal was completely quenched by the exogenous superoxide radical scavenging enzyme SOD (Fig. 3C), indicating that menadione-derived O₂^•− generation was catalyzed by AnNTR. O₂^•− is a highly reactive molecule and can undergo spontaneous dismutation to H₂O₂, providing the basis for the sensitivity of ΔprxA and ΔcatB mutants to menadione (Fig. 2B). To estimate the extent of the oxidative stress attributable to O₂^•−-derived H₂O₂, we measured H₂O₂ levels in the reaction solution. As shown in Fig. 3D, a large amount of H₂O₂ appeared in the AnNTR-catalyzed menadione reduction reaction mixture and was completely decomposed by catalase. Our data demonstrated that AnNTR drives the one-electron metabolism of menadione leading to ROS generation via redox cycling. We proposed that the catalytic process proceeds as follows: AnNTR catalyzes the reduction of menadione to produce semiquinone by accepting one electron from NADPH. The resulting unstable semiquinone is released from AnNTR and quickly reoxidized aerobically to menadione, with concomitant generation of O₂^•−. Another electron from NADPH participates in the next round of reduction of menadione in the same way. Therefore, the whole reaction appears to be a futile cycle, except for the incessant NADPH consumption and O₂^•− generation.

E. coli NTR is responsible for cell growth defects caused by menadione.

Recombinant E. coli NTR (NfsB) can catalyze menadione to produce O₂^•− in vitro, a reaction which has been utilized in the development of an O₂^•− generation system for biochemical and biomedical applications (9). We compared the efficiency of O₂^•− generation catalyzed by bacterial and fungal NTRs and found that the initial velocity of reaction of NfsB was higher than that of AnNTR under the same assay conditions, although the final levels of the product were similar (Fig. 4A). Given the high activity of menadione-dependent O₂^•− production catalyzed by purified NfsB, we speculated that NfsB might be an efficient generator of cellular O₂^•− in E. coli.

FIG 4.

Deleterious effect of NfsB on menadione detoxification in E. coli. (A) Comparison of the efficiency of O₂^•− generation catalyzed by A. nidulans and E. coli NTRs. The reaction mixture was the same as that described in Fig. 3A. The reaction was initiated by the addition of 3.6 μM AnNTR or NfsB, respectively. (B) NTRs from either E. coli or A. nidulans promotes E. coli sensitivity to menadione. For the phenotype analysis, E. coli-WT (wild-type E. coli) and ΔnfsB (E. coli nfsB disruption) strains were transformed with empty pET32a. The ΔnfsB-AN2343^C strain was constructed by introducing the pET32a-AnNTR plasmid into ΔnfsB (cross-species complemented with AN2343 in ΔnfsB). After induction for 4 h, both strains were diluted to the same concentrations (OD = 0.45) with Luria-Bertani medium (LB) containing 0.8 mM IPTG. Growth curves of both strains in LB with or without 0.3 mM menadione (MEN) were recorded at 1 h intervals at OD₆₀₀. (C) Expression of NfsB in A. nidulans ΔAN2343 restored its sensitivity to menadione. Conidia from the WT, ΔAN2343, and ΔAN2343-nfsB^C (cross-species complementation of E. coli nfsB gene in ΔAN2343) strains with serial dilutions were spotted on minimal media plates with or without the indicated concentrations of menadione, followed by incubation at 37°C for 48 h. The data are the means ± the SD of three independent experiments.

To test this hypothesis, the nfsB knockout E. coli BL21(DE3) (ΔnfsB) was constructed using CRISPR/dCas9-mediated base editing technology (36), and the phenotype was investigated (Fig. 4B). No growth rate difference was observed between the ΔnfsB mutant and the E. coli BL21(DE3) control strain (WT) under normal conditions. Exposure to external menadione at a concentration of 0.3 mM produced more survival stress to the WT than to the ΔnfsB strain (Fig. 4B), confirming that E. coli NfsB had an adverse effect on a menadione detoxification pathway. To investigate possible functional complementarity between the fungal and bacterial NTRs, we heterologously expressed the two NTRs in ΔnfsB E. coli and ΔAN2343 A. nidulans, respectively, and characterized their phenotypes. We found that heterologously expressing the two NTRs significantly promoted the menadione sensitivity of both NTR knockout strains (Fig. 4B and C), suggesting that the function of cellular NTRs in inducing cytotoxicity by menadione is evolutionarily conserved.

AnNTR has broad substrate specificity but only transcriptionally responds to aromatic compounds.

Nitroreductase has been reported to metabolize a wide range of substrates, including nitroaromatic and nitroheterocyclic compounds and quinone derivatives (5, 11, 13, 37). To characterize the substrate specificity of AnNTR, we tested its capability to catalyze the reduction of various nitro-compounds containing different nitro groups in the benzene or heterocyclic rings, as well as the other type of quinone. Activity assays were performed with NADPH as the electron donor and MTT as the chromogenic electron acceptor, and the colored products were quantified. As shown in Table 1, different activity levels were observed for AnNTR with various nitro-compounds, indicating that AnNTR possesses nitroreductase activity in vitro. Plumbagin has been reported to be another cytotoxic quinone compound, and is often used as an electrophile of cellular GSH, to produce oxidative stress (38). AnNTR had a clear preference for plumbagin over menadione in reduction activity (Table 1). These results indicated that AnNTR has broad substrate specificity, as is the case for most nitroreductases (11).

TABLE 1.

Activity of recombinant AnNTR with different substrates

Substrate	Mean activity (ΔA590 min−1 mg protein−1) ± SDa
Nitroaromatics
2-Nitrotoluene	0.13 ± 0.09
4-Nitrotoluene	0.09 ± 0.08
1,3-Dinitrobenzene	0.15 ± 0.09
2-Chloronitrobenzene	0.09 ± 0.08
4-Chloronitrobenzene	0.01 ± 0.00
2-Nitrobenzaldehyde	1.21 ± 0.09
3-Nitrobenzaldehyde	0.28 ± 0.08
4-Nitrobenzaldehyde	1.18 ± 0.09
2-Nitrobenzoate	0.21 ± 0.09
3-Nitrobenzoate	0.37 ± 0.08
4-Nitrobenzoate	0.24 ± 0.08
Ethyl p-nitrobenzoate	0.09 ± 0.05
4-Nitrophenol	0.27 ± 0.08
2,4,6-Trinitrotoluene	0.51 ± 0.03
Chloramphenicol	0.44 ± 0.22
Nitroheterocyclic derivatives
Nitrofurantoin	2.20 ± 0.23
Metronidazole	1.73 ± 0.22
4-Nitroquinoline N-oxide	3.09 ± 0.22
Quinones
Menadione	0.64 ± 0.22
Plumbagin	7.68 ± 0.35

^aThe activities of recombinant AnNTR with various nitro- and non-nitro compounds (50 μM) as the substrates are shown. Functional assays were performed with MTT as a chromogenic electron acceptor and NADPH as an electron donor. The activity was measured as the increase of MTT formazan (590 nm) in the first 0- to 60-s reaction at 25°C. Mean values are given for three replicates after subtraction of substrate blanks.

We investigated the transcriptional responsiveness of AnNTR gene to the several nitro-compounds listed in Table 1, to further confirm the status of AnNTR as a functional nitroreductase. We found that nitroaromatics such as 2,4,6-trinitrotoluene (TNT) and ethyl p-nitrobenzoate (E-PNB) increased the expression of AN2343 by 6- to 50-fold (Fig. 5). AN2343 was less responsive to the nitroheterocyclic derivatives 4-nitroquinoline N-oxide (4-NQO) and metronidazole (MTZ) (Fig. 5). This observation indicated that nitroaromatics are more potent inducers of the AnNTR gene, and the substrate preference is due to the backbone structure rather than the nitro group of the compounds. We next investigated the association between cytotoxicity and AnNTR, using the nitroaromatic compound TNT and the nitroheterocyclic compound 4-NQO. Phenotypic analysis showed that both TNT and 4-NQO produced cytotoxicity in A. nidulans, causing growth defects in strains exposed to 40 μg/ml TNT or 0.3 μg/ml 4-NQO (see Fig. S2B and C). However, deletion of AnNTR did not alter their susceptibilities to these nitro-compounds, indicating that AnNTR is not the key enzyme in the metabolism of these nitro-compounds, in contrast to findings obtained using bacterial and yeast NTRs (16, 39–42).

FIG 5.

Transcription levels of AN2343 in response to nitro-compounds and amino acids. After preculture for 16 h in minimal medium, the WT strain was incubated in the presence of a series of nitro-compounds (0.3 mM) or amino acids (5 mM) for 3 h. The relative mRNA expression levels are shown compared to the untreated strain and normalized to actA. E-PNB, ethyl p-nitrobenzoate; TNT, 2,4,6-trinitrotoluene; 4-NQO, 4-nitroquinoline N-oxide; MTZ, metronidazole; PLU, plumbagin; L-Phe, l-phenylalanine; L-Tyr, l-tyrosine; L-Glu, l-glutamate; L-Ala, l-alanine. The data show the means ± the SD of three independent experiments.

The significant induction of AnNTR expression triggered by menadione and nitroaromatics, rather than by nitroheterocyclic derivatives, led us to hypothesize that the aromatic nucleus is responsible for the transcriptional upregulation of AnNTR. To test this hypothesis, we investigated the transcriptional responsiveness of AN2343 to exposure to aromatic amino acids (Fig. 5). AN2343 expression was upregulated 8-fold after 3 h of incubation with external phenylalanine and 13-fold after incubation with tyrosine. Other amino acids, such as glutamate or alanine, did not evoke this response, confirming our hypothesis that the aromatic group triggers the induction of AnNTR gene expression.

DISCUSSION

It has been suggested that NTRs participate in defense against oxidative stress in cells, acting as ROS-resistant enzymes in many living organisms, including Aspergillus. However, our results indicated that although the transcription of NTR from A. nidulans is increased in response to menadione-derived ROS, the actual cellular behavior of AnNTR is that of accelerating, rather than inhibiting, ROS generation in the presence of menadione. Subsequent in vitro characterization of the menadione reduction catalyzed by AnNTR identified a mechanistic link between ROS generation and the function of AnNTR, in which AnNTR drives the one-electron metabolism of menadione, resulting in O₂^•− generation via redox cycling. Our results did not support the contention that NTR is an antioxidant enzyme, protective against quinone toxicity, but indicated that NTR is a key generator of ROS in response to menadione.

Three types of reductases, including NADH ubiquinone oxidase, NADH cytochrome b₅ reductase, and NADPH cytochrome P-450 reductase, have been reported to be responsible for menadione-dependent ROS generation in biological systems (34). We propose that NTR is the fourth menadione reductase capable of generating ROS, based on the performance of AnNTR inside and outside of fungal cells. Moreover, our present data showed that menadione-derived ROS generation activity is not limited to fungal NTR but also occurs in E. coli NfsB, suggesting a conserved function among some NTRs. NfsB has been classified as a type I NTR (O₂-insensitive type) for its two-electron reduction of nitro-compounds. However, acts as a type II (O₂-sensitive type) NTR when reducing menadione (Fig. 4). The noticeably different reaction mechanisms between the reduction of nitro-compounds and quinones catalyzed by NfsB (11), leading us to suggest that the classification of NTRs based on the biochemical properties of nitro-compound reduction may be not completely adequate to define quinone reductases.

Although there are multiple quinone detoxification enzymes in mammalian, yeast, and bacterial cells, their presence in filamentous fungus is still unconfirmed. Mammalian NAD(P)H:quinone oxidoreductase 1 (NQO1) is known to metabolize quinones to less toxic hydroquinones by two-electron reduction reactions and thus is an endogenous cellular detoxifying enzyme in mammalian cells (34). However, no genes homologous to NQO1 have been found using BLAST against the genomic DNA of A. nidulans. The Fqr proteins of Mycobacterium tuberculosis catalyze an F₄₂₀-specific obligate two-electron reduction of endogenous quinones. They thus compete with the one-electron reduction pathway and prevent the formation of harmful cytotoxic semiquinones, protecting mycobacteria against quinone-produced oxidative stress (17). This menadione detoxification pathway probably does not exist in A. nidulans due to the absence of the basic components of the reaction systems. Yeast Pst2p has been identified from S. cerevisiae cultivated under menadione stress. Pst2p is a nitroreductase-like protein, induced by oxidative stress, whose function is unknown (43, 44). A BLAST search against the A. nidulans genomic database identified a protein encoded by AN0297 that shows high homology to yeast pst2p. AN0297 is induced by menadione at both transcriptional and translational levels (20, 21), but no further information about this gene has been published. The physiological function of AN0297 and its corresponding protein will be investigated in the future to determine whether it is a candidate gene for menadione detoxification

Given that AnNTR accelerates the release of O₂^•− from menadione, it is unclear why the fungal NTR gene is induced in the presence of menadione. Transcription of AnNTR occurs not only in response to menadione and nitroaromatic compounds but also in response to aromatic amino acids (Fig. 5). We therefore speculated that it is the aromatic nucleus that is responsible for the upregulation of AnNTR. A similar phenomenon has also been observed in the regulation of the bacterial NTR in Rhodobacter capsulatus, whose transcription was reported to be upregulated by both nitroaromatic compounds and aromatic amino acids (45). Although R. capsulatus NTR (NprA) shows nitroreductase activity with a wide range of nitroaromatic compounds, such as the universal NTRs, NprA may act in vivo as a dihydropteridine reductase. In R. capsulatus, dihydropteridine reductase catalyzes the reduction of dihydropterine to tetrahydrobiopterin, which is a cofactor for aromatic amino acid hydroxylase, and thus may be involved in the assimilation of aromatic amino acids. Upregulation of NprA increased cell sensitization to some nitroaromatic substrates (45). In view of these data, we speculated that the response of AnNTR gene expression to menadione may be an obligatory misoperation due to the misrecognition of the aromatic amino acids by NTR transcriptional regulation systems. However, the potential involvement of AnNTR in the metabolism of aromatic amino acids requires further experimental confirmation.

MATERIALS AND METHODS

Strains and growth conditions.

Fungal strains (Table 2) were grown at 37°C in minimal medium (MM; 10 mM NaNO₃, 10 mM KH₂PO₄, 2 mM MgSO₄, 7 mM KCl, 1% glucose, and 2 ml/liter Hunter’s trace metals) supplemented appropriately (0.4 mg/liter biotin, 0.5 g/liter uracil, 0.6 g/liter uridine, and 0.4 mg/liter pyridoxine). E. coli DH5α was used for gene cloning. E. coli BL21(DE3) was used for protein expression and phenotype analysis.

TABLE 2.

Aspergillus nidulans strains used in this study

Strain	Genotype	Source or reference
ABPU1	biA1 pyrG89 wA3 argB2 pyroA4	32
A6	nicA2 veA+	FGSC
ΔAN2343	biA1 pyrG89 wA3 argB2 pyroA4 ΔAN2343::argB	This study
ΔAN2343-com	biA1 pyrG89 wA3 argB2 pyroA4 ΔAN2343::argB AN2343::GFP::pyrG	This study
ΔsodA	biA1 pyrG89 wA3 argB2 pyroA4; ΔsodA::argB	This study
ΔprxA	biA1 pyrG89 wA3 argB2 pyroA4 ΔprxA::argB	29
ΔcatB	biA1 pyrG89 wA3; argB2 pyroA4 ΔcatB::argB	Unpublished
ΔAN2343-nfsBC	biA1 pyrG89 wA3 argB2 pyroA4 ΔAN2343::argB PAN2343-nfsB-Trpc-pyrG::pyrG	This study

Disruption of AnNTR and SodA encoding genes of A. nidulans.

The disruption primers are listed in Table 3. The AN2343 disruption cassette was constructed to delete the whole AN2343 encoding region using the A. nidulans argB gene as the selection marker, flanked by 1.1 kb of the 5′and 3′ UTRs of the AN2343 gene. PCR was used to amplify the 5′- and 3′-flanking sequences of the AN2343 gene with the specific primer pairs ΔNTR-1/ΔNTR-2 or ΔNTR-3/ΔNTR-4. The marker gene argB was amplified using A. nidulans A6 genomic DNA with the primers argB-F and argB-R. These DNA fragments were mixed and fused by overlap extension PCR using the nested primers ΔNTR-nest-F and ΔNTR-nest-R. The sodA disruption cassette was constructed using the same methods and corresponding primers. The two disruption cassettes were transformed into the ABPU1 strain to obtain ΔAN2343 and ΔsodA mutants, respectively. The resulting disruptants were confirmed using colony PCR with the appropriate primers (Table 3), as shown in Fig. S1.

TABLE 3.

Primers used in this study

Primer	Gene	Nucleotide sequence (5′–3′)a
Gene disruption
ΔNTR-1	AN2343	GAGGAAAGTTCTTGGGATAAGATG
ΔNTR-2		aaaaccgcgaaataaagcttAAATAGAAATCATGAAGAGGGGTAG
ΔNTR-3		cgcaatggctgtaggtcgacTTCTGAATAGCATCATAGACGCCG
ΔNTR-4		ACCTTCCACCCCGGAGTCAAC
ΔNTR-nest-F		ATTGGCTATTTTGCTCGTTGTG
ΔNTR-nest-R		CGCCATAGCAACTCTTGTCCA
ΔsodA-1	sodA	CAAGTTCGCCGCCGACGCTG
ΔsodA-2		aaaaccgcgaaataaagcttACCTGAAAGTGATGTGAGGATGG
ΔsodA-3		cgcaatggctgtaggtcgacGAGAGCTAGGTAGCGCAAAACTGTC
ΔsodA-4		TCCTGAACGACGATCTCCGGTAAC
ΔsodA-nest-F		AAGGTCCAGGCGTAGATGAATGAAC
ΔsodA-nest-R		ACAAGAAGTTCGCGGAGGAATTCG
sodA-check-F		ATGGTCAAGGCTGGTAAGTTGTG
sodA-check-R		TACGCAGCAATACCGATGACACC
argB-F		AAGCTTTATTTCGCGGTTTTTTG
argB-R		GTCGACCTACAGCCATTGCG
argB-check-3-F		AGTCGTCCTAGCCAAGGTAG
argB-check-5-R		AAGTGTCTTCGGAGTCAACC
GFP-tagged AnNTR
pyrG-F		GCTCTAGAGAATTCGATACCTGTCGAAAG
PyrG-R		GCTCTAGATCAGTGCTTGTCTACCAG
gfp-F		GGAGCTGGTGCAGGCGCTG
gfp-R		TTATTTGTATAGTTCATCCATGCCA
NTRgfp-1		GGGTCAGGTCTCGAAACTTTCTAGG
NTRgfp-2		ccagcgcctgcaccagctccCGCGGACTTGCCAAACACCTTG
NTRgfp-3		tggatgaactatacaaataaATTCTGAATAGCATCATAGACGCCG
NTRgfp-4		GCAAGTCTTCCATTATCAAGCCCTG
NTRgfp-nest-F		AACTGCAGTTGTGAATATGCCATAATACAGTGC
NTRgfp-nest-R		AACTGCAGATTCTATCCCTACAATCCTTCCCTC
PAN2343-nfsB
PAN2343-F		TCTAGAGTCGACCTGCAGACTTGCAATTGAACCAGTTGGTT
PAN2343-R		ATCCATTGTGACTTGCGCTGCTAGAGAC
trpC-F		CCTTAACCGAAGTGTAATGATTTAATAGCTCCATGTCAACAAG
trpC-R		GCTATGACCATGATTACGCCAAAGAAGGATTACCTCTAAACAAGTG
nfsB-F		CAGCGCAAGTCACAATGGATATCATTTCTGTCGCCTTA
nfsB-R		TCATTACACTTCGGTTAAGGTGATGTTTT
pUC-pyrG-F		GGCGTAATCATGGTCATAGC
pUC-pyrG-R		CTGCAGGTCGACTCTAGAG
q-RT-PCR
q-RT-NTR-F	AN2343	CGCTCTCAAGGACATCAAGG
q-RT-NTR-R		AAGTACTGAGACATGGCATTGG
q-RT-catB-F	catB	GGAAGCTCAGCAAATTTCTGG
q-RT-catB-R		CACGTTAAGCTCCCACTCAG
q-RT-sodA-F	sodA	GATAAGCTGATCAAGCTCATTGG
q-RT-sodA-R		GCCAAGGTCATCAGTACCAG
q-RT prxA-F	prxA	CCCCGCTGACGTTGTCTTC
q-RT-prxA-R		GAGGGCGAAGAGGATGACC
q-RT-actA-F	actA	GAAGTCCTACGAACTGCCTGATG
q-RT-actA-R		GACCAAGAACGCTGGGCTGG
Gene cloning
AnNTR-F	AN2343	CCGGAATTCATGGTCGAGTTCAAGAACCCCGC
AnNTR-R		ATAAGAATGCGGCCGCTTACGCGGACTTGCCAAACACC
nfsB-F′	nfsB	CGCGGATCCATGGATATCATTTCTGTCGCCTTAAAG
nfsB-R′		CCCAAGCTTTTACACTTCGGTTAAGGTGATGTTTTGC
trxA-F	trxA	CATCATCATCATCATTAATCTGGTCTGGTGCCA
trxA-R		TGGCACCAGACCAGATTAATGATGATGATGATG

^aAll restriction enzymes sites in the primer sequences are underlined. Sections indicated by lowercase letters were designed to overlap the 5′- and 3′-terminal sequences of the marker and GFP genes, respectively.

Disruption of the nfsB gene of E. coli BL21(DE3).

The CRISPER/dCas9-mediated cytidine base editing (CBE) system can be used to inactivate eukaryotic genes through the induction of STOP codons in gene open reading frames (36). Recently, this CBE system was developed in E. coli by our lab (unpublished data) and was used in this study to produce the ΔnfsB strain. The mutant stain was confirmed by sequencing (see Fig. S5).

Construction of A. nidulans strains expressing GFP-tagged AnNTR or E. coli NfsB.

The corresponding primers are listed in Table 3. The AN2343::gfp cassette was constructed as follows. The marker gene pyrG was amplified using PCR with A. nidulans A6 genomic DNA and the primers pyrG-F and pyrG-R. The PCR product was digested with XbaI and then inserted into the XbaI site of pUC19 to construct pUC19-pyrG. The gfp gene was amplified from the pUC19-egfp plasmid produced in our lab using the primers gfp-F and gfp-R. The DNA fragment, including the native promoter of AN2343 (1-kb 5′ UTR of AN2343) and the AN2343 encoding region, was amplified using PCR with A. nidulans A6 genomic DNA and the primers NTRgfp-1 and NTRgfp-2. The native terminator region of AnNTR (1-kb 3′ UTR of AN2343) was amplified using the primers NTRgfp-3 and NTRgfp-4. The resulting DNA fragments were fused by an overlap PCR using the primers NTRgfp-nest-F and NTRgfp-nest-R, both of which contain the PstI restriction site, and were cloned into pUC19-pyrG to produce the GFP-tagged AnNTR expression plasmid pNTRgfp. The pNTRgfp plasmid was transformed into the ΔAN2343 strain to obtain a strain harboring GFP-tagged AnNTR.

A strain expressing NfsB in ΔAN2343 was constructed as follows. The DNA fragments containing the AN2343 native promoter (2-kb 5′ UTR of AN2343) and the Trpc terminator were amplified using PCR with A6 genomic DNA and two primer pairs, P_AN2343-F/P_AN2343-R and trpC-F/trpC-R. The nfsB gene was amplified using the primers nfsB-F and nfsB-R. pUC19-pyrG was linearized using PCR with the primers pUC-pyrG-F and pUC-pyrG-R. These DNA fragments were connected using In-Fusion HD cloning kits (TaKaRa Bio, Shiga, Japan). The resultant plasmid, pP_AN2343-nfsB-Trpc-pyrG, was transformed into the ΔAN2343 strain to obtain a strain expressing E. coli NfsB.

Fluorescence microscopy.

A. nidulans conidiospores (5 × 10⁴) were inoculated onto a glass-bottom cell culture dish (NEST Biotechnology, Wuxi, China) dipped in 200 μl of MM medium and grown at 37°C for 12 h. The hyphae were then washed twice with phosphate-buffered saline (PBS; pH 7.4) after removal of the MM medium and observed using confocal laser scanning microscopy (TCS SP8; Leica, Wetzlar, Germany). To examine the O₂^•− produced by menadione, 300 μM menadione was added to 200 μl of MM for 1 h after 12 h of cultivation, followed by treatment with 10 μM dihydroethidium (DHE) and incubation at 37°C for 30 min. Subsequently, dishes with attached mycelia were washed twice with PBS (pH 7.4), and the intracellular O₂^•− levels were monitored by the formation of ethidium from DHE. The fluorescence of the GFP and the O₂^•− specific products for DHE were imaged using excitation with a 488-nm laser and a 405-nm laser, respectively.

Quantitative PCR.

Strains were cultured in MM for 16 h and then treated with the indicated concentrations of various compounds for 3 h. Mycelia were harvested by filtration, and total RNA was extracted using EZ-10 DNAaway RNA Mini-Prep kits (Sangon, China). cDNAs were reverse transcribed using ReverTra Ace qPCR RT Master Mix with gDNA remover (Toyobo Co., Ltd., Osaka, Japan). The primer pairs q-RT-NTR-F/q-RT-NTR-R, q-RT-catB-F/q-RT-catB-R, q-RT-sodA-F/q-RT-sodA-R, q-RT-prxA-F/q-RT-prxA-R, and q-RT-actA-F/q-RT-actA-R (Table 3) were used to amplify the AN2343, catB, sodA, prxA, and actA genes, respectively. The relative mRNA levels were normalized to that of actA, which was used as a reference gene.

Measurement of O₂^•− and H₂O₂.

Electron paramagnetic resonance (EPR) was used to measure the O₂^•− generated by the AnNTR-catalyzed menadione reduction. All EPR spectra were recorded at room temperature on a Bruker BioSpin EMX-8/2.7 (Bruker Corporation, Billerica, MA) at 9.8 GHz with a 100-kHz modulation frequency and a 1.0-G modulation amplitude. The reaction mixture for EPR analysis was 0.5 ml sodium phosphate buffer (50 mM [pH 7.4]) containing AnNTR (1.5 μg), menadione (50 μM), NADPH (100 μM), diethylenetriaminepentaacetic acid (DTPA; 100 μM), FMN (10 μM), and DMPO (40 mM). The reaction was initiated by the addition of the enzyme and quickly transferred into a quartz capillary. H₂O₂, as a disproportionation reaction product of O₂^•− from the above-mentioned reaction, was analyzed using hydrogen peroxide assay kits (Beyotime Biotech, Haimen, China), and the the absorbance was measured at 540 nm using a U-5100 spectrophotometer (Hitachi, Tokyo, Japan). The rates of O₂^•− generation catalyzed by AnNTR and E. coli NfsB were assayed in a reaction mixture containing a final volume of 0.5 ml of 50 mM sodium phosphate buffer (pH 7.4), 400 μM NADPH, 100 μM DTPA, 10 μM FMN, 200 μM menadione, 50 μM DHE, and 3.6 μM AnNTR or NfsB. The fluorescence of DHE oxidized by O₂^•− was monitored by using a fluorescence spectrophotometer (F-4600; Hitachi).

The intracellular O₂^•− levels were measured as follows. Approximately 1 × 10⁸ conidiospores of the control and ΔAN2343 strains were inoculated into 250-ml Erlenmeyer flasks containing 100 ml of medium and cultured at 37°C for 16 h at 200 rpm. All mycelium samples were filtered using a textile filter (Miracloth; Calbiochem, San Diego, CA) and transferred immediately into 100 ml of medium with or without 0.8 mM menadione. After menadione treatment for 6 h, 10 μM DHE was added for 1 h, and the fluorescence was tested. Mycelia were harvested by filtration, immediately ground into a powder with liquid nitrogen, and then suspended in 50 mM sodium phosphate buffer (pH 7.4). The supernatant of the disrupted mycelia was used for fluorescence detection.

Cloning, expression, and purification of AnNTR and E. coli NfsB.

The cDNA of AnNTR was subcloned using the primers AnNTR-F and AnNTR-R (Table 3). The PCR-amplified product was digested using the appropriate restriction enzymes and inserted into pET32a. The nfsB gene was amplified with the primers nfsB-F′ and nfsB-R′ (Table 3) and subcloned into pET32a to create the overexpression vector pET32a-nfsB. To express the TrxA tag (13.2 kDa), a stop codon (TAA) was introduced at the multiple-cloning site just before the thrombin site in the empty pET32a by using site-directed mutagenesis with the primers trxA-F and trxA-R (Table 3). E. coli BL21(DE3) cells containing each plasmid were cultured at 37°C in the presence of 50 μg/ml ampicillin to an optical density (OD) of 0.5. Protein expression was induced by the addition of 0.5 mM IPTG (isopropyl-1-thio-β-d-galactopyranoside) for 12 h at 30°C. Recombinant AnNTR, NfsB, and TrxA were purified from these fractions using a HisTrap FF column (GE Healthcare, Chicago, IL) and characterized using SDS-PAGE analysis.

Enzyme assays.

The nitro-compound and quinone reductase activities of recombinant AnNTR were measured spectrophotometrically by monitoring MTT reduction at 590 nm, as previously described (33). The reaction mixture contained 0.5 ml of sodium phosphate buffer (50 mM [pH 7.4]), NADPH (100 μM), DTPA (100 μM), FMN (10 μM), MTT (0.5 mM), various substrates (50 μM), and AnNTR (1.5 μg). After subtraction of the substrate blanks, the activities were measured as the change in absorbance at 590 nm (ΔA₅₉₀).

Aconitase activity was detected using Cis-Aconitase (ACO) activity detection kits (Solarbio Life Sciences, Beijing, China). The production of 1 nmol of cis-aconitate per min per mg of protein was defined as 1 U of aconitase activity.

Menadione reduction products.

The products of menadione reduction were analyzed using HPLC (Agilent 1260; Agilent Technologies, Santa Clara, CA) with a reversed-phase column (TSKgel ODS-100V, 4.6 × 15 cm; Tosoh Co., Ltd., Tokyo, Japan). The HPLC conditions were as follows. The column was preequilibrated with 95% B for 20 min; 0 to 15 min, using a linear gradient from 95% A to 5% A; 15 to 17 min, a linear gradient from 5% A to 95% A; 17 to 20 min, constant with 95% A; the flow rate was maintained at 0.3 ml/ml, and the eluent was monitored using a UV detector at 300 nm. Solvent A was 0.1% formic acid in H₂O, and solvent B was 0.1% formic acid in acetonitrile.

Data availability.

All necessary data required to assess our findings are available in this manuscript or its supplementary materials.