Study on the bZIP-Type Transcription Factors NapA and RsmA in the Regulation of Intracellular Reactive Species Levels and Sterigmatocystin Production of Aspergillus nidulans

Basic leucine zipper (bZIP) transcription factors play a crucial role in the environmental stress response of eukaryotes. In this work, we studied the effect of gene manipulations, including both deletions and overexpressions, of two selected bZIP transcription factors, NapA and RsmA, in the oxidative stress response and sterigmatocystin production of Aspergillus nidulans. We found that NapA was important in the oxidative stress response by negatively regulating intracellular reactive species production and positively regulating catalase activities, whereas RsmA slightly negatively regulated catalase activities. Concerning sterigmatocystin production, the highest concentration was measured in the ΔrsmA ΔnapA double deletion mutant, but elevated sterigmatocystin production was also found in the OErsmA OEnapA strain. Our results indicate that NapA influences sterigmatocystin production via regulating reactive species level whereas RsmA modulates toxin production independently of the redox regulation of the cells.

To study the role of NapA in A. nidulans both OEnapA and ∆napA mutants have been characterized ( Figure 1) [13]. In the OEnapA strain, reduced production of several secondary metabolites and imbalance in asexual/sexual development have been observed [13].
Manipulation of napA gene also affected the oxidative stress defense system of A. nidulans; namely, deletion of napA increased the tBOOH sensitivity of the fungus, with concomitantly higher reactive species (RS) production, and decreases in catalase activity in 10 h cultures and higher glutathione peroxidase activity in 24 h cultures ( Figure 1) [2,13]. RS production was also higher in the OEnapA strain compared to the wild type without disturbing the catalase and glutathione peroxidase activities [13]. In the study of Mendoza-Martínez et al. [14], it was confirmed that napA is induced at high ROS levels and nuclear localization of NapA is induced by H 2 O 2 , menadione and osmotic stress, glucose starvation, and growth on ethanol. NapA is also required for conidiation but represses fruiting body formation ( Figure 1) [14]. NapA controls several genes involved in detoxification and drug efflux, which protect the fungus during conidiation, e.g., NapA directly activates catalase B (catB), the thioredoxin system, and glutathione reductase [15]. NapA is also crucial in the induction of conidiation through the oxidative stress response in the presence of redox metabolites, e.g., phenazine produced by Pseudomonas aeruginosa [16]. The impact of NapA on the oxidative stress defense was counterbalanced by another transcription factor, RsrA, by repressing napA and some NapA activated genes, such as glrA, trxA, and catB [17].

Stress Sensitivity Phenotypes of the RsmA and NapA Mutants
As previously shown, deletion of napA yielded oxidative stress phenotypes in the presence of all tested oxidative stress generating agents, e.g., diamide, MSB, tBOOH, and H2O2, independently of the rsmA gene manipulation. Figure 2, Supplementary Table S1). The effect of napA on the oxidative stress-sensitive phenotype was dependent on the applied stressor, namely, in the presence of MSB decreased, and in the presence of diamide increased, the sensitivity of the mutants, whereas H2O2 sensitivity of the mutants was dependent on the rsmA gene. (Figure 2, Supplementary Table S1). The effect of rsmA deletion or overexpression on the stress sensitivity of the mutants was based on the type of the stress-generating agent and the napA gene ( Figure 2, Supplementary Table S1). RsmA, (restorer of secondary metabolism A) has also been described as a Yap-like protein and is involved in the control of secondary metabolite production of A. nidulans [18]. Overexpression of rsmA restores sterigmatocystin production of A. nidulans in two Velvet complex deletion mutants, ∆laeA and ∆veA, both of which greatly reduce sterigmatocystin synthesis as single deletions (Figure 1) [18]. In a previous study by Yin et al. [13], functions of RsmA in the regulation of secondary metabolism, sexual development, and stress responses were studied ( Figure 1). Overexpression of rsmA increased sterigmatocystin production 100-fold and resulted in disturbances in ascospore formation in A. nidulans. RsmA activates sterigmatocystin production by binding to aflR (coding for a transcription factor positively regulating sterigmatocystin biosynthesis) promoter regions [19]. Interestingly, the ∆rsmA gene deletion strain also produced sterigmatocystin at slightly higher concentrations than the wild type, suggesting a complex regulatory role of this protein [18]. In the human pathogenic fungus A. fumigatus, overexpression of rsmA increased the concentration of twelve gli cluster metabolites in the culture medium and, consequently, gliotoxin production in A. fumigatus-infected mice. The supernatant of OErsmA with higher gliotoxin concentration compared to the control inhibited human neutrophil chemotaxis in vivo [20]. Interestingly, the OErsmA mutant showed growth retardation at 25 • C and increased menadione tolerance in comparison to the wild-type strain [20].
Contrarily, overexpression of the rsmA ortholog AflrsmA showed increased sensitivity to menadione in A. flavus, whereas deletion of AflrsmA resulted in menadione tolerance when compared to the wild type strain [21]. Concerning aflatoxin biosynthesis, the overexpression of AflrsmA increased the production of this mycotoxin as expected. Following stress treatment with menadione and tBOOH, aflatoxin production decreased in the ∆AflrsmA mutant. These observations suggest that, in A. flavus, AflrsmA regulates aflatoxin biosynthesis via oxidative stress signaling, although the possibility that AflrsmA can bind to aflR promoter regions was not examined in this work [21].
In this study, we characterized NapA and RsmA functions by construction of deletion and overexpression mutants prepared in all combinations.

Stress Sensitivity Phenotypes of the RsmA and NapA Mutants
As previously shown, deletion of napA yielded oxidative stress phenotypes in the presence of all tested oxidative stress generating agents, e.g., diamide, MSB, tBOOH, and H 2 O 2 , independently of the rsmA gene manipulation. Figure 2, Supplementary Table S1). The effect of napA on the oxidative stress-sensitive phenotype was dependent on the applied stressor, namely, in the presence of MSB decreased, and in the presence of diamide increased, the sensitivity of the mutants, whereas H 2 O 2 sensitivity of the mutants was dependent on the rsmA gene. (Figure 2, Supplementary Table S1). The effect of rsmA deletion or overexpression on the stress sensitivity of the mutants was based on the type of the stress-generating agent and the napA gene ( Figure 2, Supplementary Table S1).

Biomass, Specific RS and Specific Catalase Enzyme Productions
Deletion or overexpression of the napA and/or rsmA genes significantly affected the growth of the strains. In general, all mutant strains grew slower than the wild-type, both in untreated control and in tBOOH-exposed cultures (Supplementary Table S2).
In unstressed conditions, significant increases in RS production were only observed in the OErsmA ∆napA strain as compared to the control ( Figure 3A, Supplementary Table S3). tBOOH treatment extensively increased the specific RS production of the ∆napA mutant ( Figure 3A, Supplementary Table S3). Both rsmA deletion and overexpression reduced RS production in the napA gene deletion background but were unable to re-establish the RS level of the control strain ( Figure 3A, Supplementary Table S3). RS production was higher but not statistically different in the OErsmA strain when compared to the wild-type strain after tBOOH treatment ( Figure 3A, Supplementary Table S3).
Catalase production significantly increased in the ∆napA mutant in comparison to the control strain without oxidative stress treatment. Neither rsmA overexpression nor rsmA deletion carried out in the napA gene deletion background significantly changed the catalase production measured in the ∆napA strain ( Figure 3B, Supplementary Table S3). Deletion of rsmA alone highly increased the catalase production of the fungus but RS production was similar to the control strain ( Figure 3B, Supplementary Table S3). Because overexpression of napA alone also increased the catalase production, a remarkably high specific catalase activity was measured in the ∆rsmA OEnapA strain ( Figure 3B, Supplementary Table S3). The catalase activity of the ∆rsmA strain was comparable to that of the double overexpression mutant ( Figure 3B, Supplementary Table S3). Interestingly, it was observed that tBOOH treatment did not influence the specific catalase activities within a given strain, and remained high in the ∆rsmA, ∆rsmAOE napA, and OErsmA OEnapA strains ( Figure 3B, Supplementary Table S3).

Figure 2.
Oxidative stress sensitivity of the control and mutant A. nidulans strains. The oxidative stress tolerances of the mutants were tested on nutrient agar stress plates. A quantity of 10 5 freshly grown conidia were spotted on minimal nitrate medium with 0.05 mg/L pyridoxine agar plates, which were supplemented with one of the stress0generating agents: diamide 2.0 mmol/L, MSB 0.12 mmol/L, tBOOH 0.8 mmol/L, H2O2 6.0 mmol/L. The stress plates were incubated at 37 °C for 5 days. cause overexpression of napA alone also increased the catalase production, a remarkably high specific catalase activity was measured in the ΔrsmA OEnapA strain ( Figure 3B, Supplementary Table S3). The catalase activity of the ΔrsmA strain was comparable to that of the double overexpression mutant ( Figure 3B, Supplementary Table S3). Interestingly, it was observed that tBOOH treatment did not influence the specific catalase activities within a given strain, and remained high in the ΔrsmA, ΔrsmAOE napA, and OErsmA OEnapA strains ( Figure 3B, Supplementary

ST Production
The manipulation of rsmA alone did not change the sterigmatocystin production of the fungus without stress treatment ( Figure 4, Table S4). The highest ST production was observed in the double deletion mutant, whereas deletion of napA alone also significantly increased ST production. Overexpression of rsmA increased the production of this toxin in the ΔnapA genetic background, and increased ST production was also measured in the double overexpression mutant when compared to the control ( Figure 4, Table S4).
Similar to the catalase activities, no statistically significant differences in ST production were detected between the unstressed and tBOOH-exposed cultures within a given strain ( Figure 4, Table S4). After tBOOH treatment, the highest ST levels were measured in the ΟΕrsmA ΔnapA mutant (Figure 4, Table S4).

ST Production
The manipulation of rsmA alone did not change the sterigmatocystin production of the fungus without stress treatment ( Figure 4, Table S4). The highest ST production was observed in the double deletion mutant, whereas deletion of napA alone also significantly increased ST production. Overexpression of rsmA increased the production of this toxin in the ∆napA genetic background, and increased ST production was also measured in the double overexpression mutant when compared to the control ( Figure 4, Table S4).
Similar to the catalase activities, no statistically significant differences in ST production were detected between the unstressed and tBOOH-exposed cultures within a given strain ( Figure 4, Table S4). After tBOOH treatment, the highest ST levels were measured in the OErsmA ∆napA mutant (Figure 4, Table S4).
increased ST production. Overexpression of rsmA increased the production of this toxin in the ΔnapA genetic background, and increased ST production was also measured in the double overexpression mutant when compared to the control (Figure 4, Table S4).
Similar to the catalase activities, no statistically significant differences in ST production were detected between the unstressed and tBOOH-exposed cultures within a given strain (Figure 4, Table S4). After tBOOH treatment, the highest ST levels were measured in the ΟΕrsmA ΔnapA mutant (Figure 4, Table S4).

Expression Patterns
We also examined the effect of rsmA on the expression of wild type napA and vice versa ( Figure 5, Table S5). The pairwise comparison of ∆rsmA, control, and OErsmA strains showed that OE of rsmA increased the transcription of napA, whereas OE of napA resulted in elevated rsmA expression ( Figure 5, Table S5). Deletion of either rsmA or napA had no significant effect on the transcription of the other gene ( Figure 5, Table S5). In the tBOOHtreated cultures of the control strain, the expression of rsmA increased in comparison to the untreated cultures (Table S5)

Expression Patterns
We also examined the effect of rsmA on the expression of wild type napA and vice versa ( Figure 5, Table S5). The pairwise comparison of ΔrsmA, control, and OErsmA strains showed that OE of rsmA increased the transcription of napA, whereas OE of napA resulted in elevated rsmA expression ( Figure 5, Table S5). Deletion of either rsmA or napA had no significant effect on the transcription of the other gene ( Figure 5, Table S5). In the tBOOH-treated cultures of the control strain, the expression of rsmA increased in comparison to the untreated cultures (Table S5).

Interaction between NapA or RsmA Expressions and Specific Catalase Activities, DCF Formation, or ST Production
The increase in transcriptional activity of the napA gene tended to be associated with an increase in specific catalase activities and a decrease in RS and ST production ( Figures  3, 4 and 6, Tables S3 and S4). In contrast, rsmA transcription hardly affected RS production (Figures 3 and 6). Surprisingly, both increased and reduced rsmA transcriptions were associated with elevated catalase activities or ST production in some cultures (Figures 3, 4  and 6, Tables S3 and S4).

Interaction between NapA or RsmA Expressions and Specific Catalase Activities, DCF Formation, or ST Production
The increase in transcriptional activity of the napA gene tended to be associated with an increase in specific catalase activities and a decrease in RS and ST production (Figures 3, 4 and 6, Tables S3 and S4). In contrast, rsmA transcription hardly affected RS production (Figures 3 and 6). Surprisingly, both increased and reduced rsmA transcriptions were associated with elevated catalase activities or ST production in some cultures (Figures 3, 4 and 6, Tables S3 and S4).

Discussion
There is strong evidence that secondary metabolite production is associated with oxidative stress, which is co-regulated by various transcription factors [7,[22][23][24][25]. In this study we constructed a series of gene deletion and overexpression strains of napA and rsmA, either alone or in combination, to understand how these bZIP-type transcription factors contribute to the regulation of the stress tolerance and secondary metabolite production in A. nidulans.
In line with previous studies, NapA is a key player in the regulation of the oxidative stress response of A. nidulans [2,13,14]. Deletion of napA largely increased the oxidative

Discussion
There is strong evidence that secondary metabolite production is associated with oxidative stress, which is co-regulated by various transcription factors [7,[22][23][24][25]. In this study we constructed a series of gene deletion and overexpression strains of napA and rsmA, either alone or in combination, to understand how these bZIP-type transcription factors contribute to the regulation of the stress tolerance and secondary metabolite production in A. nidulans.
In line with previous studies, NapA is a key player in the regulation of the oxidative stress response of A. nidulans [2,13,14]. Deletion of napA largely increased the oxidative stress sensitive phenotype of the fungus, even in the presence of pyridoxine (Figure 2), which has some antioxidant features [3]. According to gene expression studies, both rsmA and napA have an impact on each other's expressions ( Figure 5) but neither the deletion nor the overexpression of rsmA was unable to mitigate the absence of NapA. The increased stress sensitivity of ∆napA also occurred with increased RS production but unaltered catalase activities after tBOOH treatment (Figure 3, Table S3).
bZIP transcription factors may also modulate secondary metabolite production in filamentous fungi via regulating ROS levels [6,7,24] or directly by binding to promoters of biosynthetic genes responsible for secondary metabolite production [8,19]. In this study we also examined the possible relationship between oxidative stress and secondary metabolite production.
The oxidative stress-sensitive ∆napA mutant was characterized by increased sterigmatocystin production but, unexpectedly, the highest sterigmatocystin level was observed in the ∆napA ∆rsmA double deletion mutant ( Figure 4, Table S4). As reported before by Shaaban et al. [18], the deletion of rsmA alone may also slightly increase ST production. Previously, Yin et al. [19] reported that overexpression of rsmA resulted in an elevated ST production. Our data suggest that both the reduced and the elevated rsmA can lead to increased ST levels ( Figure 4, Table S4).
Based on these observations, we can assume that NapA regulates ST biosynthesis via modulating oxidative stress in A. nidulans (Figure 4, Table S4) similarly to ApyapA (A. parasiticus), Aoyap1 (A. ochraceus), and AfyapA (A. fumigatus) [5,7,[22][23][24]. Furthermore, RsmA seems to support ST production at low intracellular RS levels [18,19] and, therefore, both oxidative stress-dependent and -independent regulatory elements are likely to modulate ST production in A. nidulans. Additionally, RsmA has a direct impact on ST synthesis by actively binding to and promoting AflR activity, the ST pathway specific regulator [19]. Considering the complex regulatory patterns of NapA and RsmA on the oxidative stress response and secondary metabolite production of A. nidulans, in addition to their impact on each other's expressions, we can hypothesize that NapA and RsmA are likely to interact with each other either genetically or even physically to coordinate ST production ( Figure 5, Table S5).
It is well known that bZIP type transcription factors may form heterodimers and coordinate a wide array of cellular processes, including the oxidative stress response, in addition to secondary metabolite biosynthesis [26]. For example, AtfA, AtfB, AtfC, and AtfD physically interact with each other in A. fumigatus to coordinate stress response and virulence of this opportunistic human pathogenic fungus [27]. Further research is needed to shed light on the nature of the hypothesized interaction between NapA and RsmA in A. nidulans.
Interestingly, rsmA-regulated secondary metabolite production was also stress related in A. nidulans, A. fumigatus, and A. flavus [20,21]. However, the A. fumigatus OErsmA mutant was less sensitive to MSB [16] and the A. flavus OEAflrsmA strain was more sensitive to MSB [21], suggesting that the co-regulation of secondary metabolite production and oxidative stress response has species-specific components, even in fungal species belonging to the same genus. Again, this versatile regulatory pattern of bZIPs observable in the Kingdom of Fungi may be the consequence of their multilevel, easily variable, and highly flexible interactions.

Reactive Species Production and Catalase Activity
To determine the physiological parameters in submerged cultures, strains were pregrown in Erlenmeyer flasks (500 mL) containing 100 mL minimal-nitrate medium (pH 6.5) and also supplemented with 0.05 mg/L pyridoxine. Culture media were inoculated with 10 6 conidia/mL and incubated at 37 • C and at 3.7 Hz shaking frequency. Oxidative stress was induced by the addition of tBOOH (at 0.1, 0.2, and 0.4 mmol/L concentrations) to exponential growth phase (18 h) cultures. For dry cell mass (DCM) determinations, samples were taken just before the stress treatment (0 h) and at every 12 h after the stress exposure for up to 48 h. Dry cell mass (DCM) of the samples was determined as described previously [13].
At all the incubation times tested, 2 ,7 -dichlorofluorescin diacetate was added to 20 mL aliquots of the cultures at a final concentration of 30 µmol/L, and after incubating them further for 1 h in 100 mL culture flasks, mycelia were harvested by centrifugation. DCF productions were determined spectrofluorimetrically [32,33].
Changes in the specific catalase activities were also recorded in separate experiments. Submerged cultures were treated with tBOOH (0.2 mmol/L) at 18 h culture time. Samples were taken at 5 h after tBOOH treatment, and mycelia harvested by filtration were washed with distilled water and resuspended in ice-cold 0.1 M potassium phosphate buffer (pH 7.5). In these cases, cell-free extracts were prepared by disrupting mycelia with 0.5 mm glass beads (5000 rpm, 30 s) and centrifugation [32]. Catalase activities were determined spectrophotometrically, measuring H 2 O 2 decomposition and NADPH diminution rates [34]. Protein contents of the cell-free extracts were measured by a modification of the Lowry method [35].

rRT-PCR Assays
Total RNA was isolated from lyophilized mycelia according to Chomczynski [37] and RT-qPCR experiments were carried out as described earlier [38]. The applied primer pairs are summarized in Table S6. Relative transcription levels were quantified with the ∆∆CP value (mean ± S.D. calculated from 4 biological replicates), which was defined as ∆CP treated − ∆CP control , where ∆CP treated = CP reference gene − CP tested gene measured in stresstreated cultures, ∆CP control = CP reference gene − CP tested gene measured in untreated cultures, and CP values represent the rRT-PCR cycle numbers of crossing points. As a reference gene, actA (AN6542) was used [38].

Statistical Analysis
The effects of napA and rsmA gene manipulations on the colony diameters in stresstreated and untreated surface cultures were analyzed with Dunnett's test. The interaction between gene manipulation(s) and stress treatment was studied by two-way ANOVA.
In the case of the tBOOH-treated or the untreated submerged cultures, the effects of gene manipulations on catalase activity, DCF formation, and ST production, and on the relative transcription of the rsmA and napA genes, were analyzed by one way ANOVA followed by Tukey's post-hoc test. When tBOOH-treated and untreated cultures were compared, the Student's t-test with Holm's p-value correction was applied. In all cases, the difference between the mean values were regarded as significant if the (adjusted) p-value was less than 0.05.