A p53-like transcription factor, BbTFO1, contributes to virulence and oxidative and thermal stress tolerances in the insect pathogenic fungus, Beauveria bassiana

The p53-like transcription factor (TF) NDT80 plays a vital role in the regulation of pathogenic mechanisms and meiosis in certain fungi. However, the effects of NDT80 on entomopathogenic fungi are still unknown. In this paper, the NDT80 orthologue BbTFO1 was examined in Beauveria bassiana, a filamentous entomopathogenic fungus, to explore the role of an NDT80-like protein for fungal pest control potential. Disruption of BbTFO1 resulted in impaired resistance to oxidative stress (OS) in a growth assay under OS and a 50% minimum inhibitory concentration experiment. Intriguingly, the oxidation resistance changes were accompanied by transcriptional repression of the two key antioxidant enzyme genes cat2 and cat5. ΔBbTFO1 also displayed defective conidial germination, virulence and heat resistance. The specific supplementation of BbTFO1 reversed these phenotypic changes. As revealed by this work, BbTFO1 can affect the transcription of catalase genes and play vital roles in the maintenance of phenotypes associated with the biological control ability of B. bassiana.


Introduction
The p53-like transcription factor (TF) superfamily includes the NDT80/PhoG-like DNA-binding family (http://pfam.xfam.org/family/PF05224), which has only been discovered in unikont lineages [1]. NDT80-like proteins have different functions in different fungi, such as regulation of sexual development, meiosis, filamentation, virulence, drug resistance, programmed cell death and the response to nutrient stress [1,2]. The NDT80 protein of Saccharomyces cerevisiae is considered to be the founding member of the NDT80 family within the p53-like superfamily [1] and a key regulator of sporulation and meiosis [3]. XprG, also known as an NDT80-like protein, positively regulates mycotoxin generation, carbon depletion-induced apoptosis, and extracellular protease expression in Aspergillus nidulans [4]. This protein is similar to Neurospora crassa VIB-1, which plays a role in gene expression related to the positive regulation of extracellular protease generation as well as heterokaryon incompatibility-related apoptosis [2,3,5]. In Candida albicans, the roles of three NDT80-like proteins have been extensively studied. As regulators, CaNDT80 and CaRep1 control transporters (CDR1 and MDR1), which can eliminate drugs [6,7]. Mutations in CaNDT80 and CaRON1 affect the induction of hyphal growth [7]. The CaNDT80 mutant also shows weakened virulence in a mouse model of infection [8]. NDT80 homologue mutations have certain impacts on secondary metabolism, including asexual or sexual spore pigmentation. N. crassa Δvib-1 mutants show pinkish conidia but not orange conidia (asexual spores); by contrast, mutants of XprG in A. fumigatus and A. nidulans possess pale conidia [3,9]. Moreover, NDT80-like proteins are involved in fungal nutrition sensing. For example, suppressor of fusion (Suf), the Ndt80 homologue in Fusarium oxysporum, plays a vital role in regulating anastomosis depending on the nutrient [10]. As far as Trichoderma reesei is concerned, this protein also plays an important role in activating the N-acetylglucosamine (GlcNAc) gene cluster [2,11]. C. albicans Rep1 and A. nidulans XprG are also required for GlcNAc catabolism [7]. The genus Beauveria, including the species B. bassiana, represents a strong insect pathogen with a wide range of hosts and has been prepared as a candidate mycoinsecticide with environmental friendliness [12]. As an important filamentous insect fungal pathogen, B. bassiana is the typical model for examining the pathogenic mechanism and growth of fungi together with their interactions with the host [13,14]. Recently, some TFs in B. bassiana have been characterized. Typically, the BbMbp1 gene facilitates the transition of B. bassiana morphology with regard to saprophytic and pathogenic growth through a variety of genetic pathways [15]. In addition, in B. bassiana, the MADS-box TF Bbmcm1 plays a vital role in the regulation of the cell cycle, virulence and integrity [16]. Moreover, BbOhmm affects the homeostasis of reactive oxygen species (ROS) combined with the available oxygen to control BbSre1 transcriptional activity, thus regulating the homeostasis of mitochondrial iron, respiration-related genes and haem production to adapt to hypoxic conditions [14].
In our study, it is shown that the NDT80-like protein in B. bassiana plays an important role in oxidative stress (OS) resistance by impacting the expression of key catalase genes. Therefore, this protein was designated TFO1 (transcription factor regulating oxidation) in B. bassiana. This study was conducted to examine how BbTFO1 affects B. bassiana phenotypic changes related to its biological control ability through multi-phenotypic analyses of specific gene deletion or complementation strains combined with the parental WT. The results indicate that BbTFO1 plays an important role in regulating antioxidant effects, thermotolerance and pathogenesis in insect fungal pathogens.

Phylogenetic analysis of BbTFO1
In NCBI databases, using the S. cerevisiae NDT80 sequence as a query, BLAST analysis based on NCBI was carried out to identify the genome from ARSEF 2860, the wild-type B. bassiana strain (WT), and some typical filamentous fungi. Next, each NDT80 homologous protein sequence found in the fungi was aligned to compare the structures and analyse the phylogenetics using MEGA7 (http://www.megasoftware.net).

Construction of BbTFO1 deletion and complementation mutants
BbTFO1 deletion together with a complementation mutant was created via Agrobacteriummediated transformation [17]. The synthetic plasmids for homologous transformation were transformed by Agrobacterium tumefaciens and mobilized into B. bassiana according to a previous procedure [18]. Using the paired primers (Table 1), PCR was carried out to clone BbTFO1 5 0 or 3 0 coding/flanking fragment sequences based on the WT genome, with TaKaRa La Taq 1 (TaKaRa, Japan) used as the catalyst, followed by specific enzyme digestion and insertion to the target plasmid to form target gene deletion plasmids. BbTFO1 was knocked out in the WT by homogenously recombining the 5 0 and 3 0 fragments with a bar marker. The gene deletion mutant was initially screened by herbicide (200 μg ml −1 phosphinothricin) resistance [19].
For the reverse complement of the BbTFO1 deletion mutants, using Gateway BP Clonase™ II Enzyme Mix (Shanghai, China), primers (Table 1) were used to clone the entire BbTFO1 coding sequence based on the WT genome, followed by insertion into the p0380-sur-gateway for exchange for gateway fragments, and finally, p0380-sur-BbTFO1 was formed [20]. The plasmid was transformed in deletion via Agrobacterium-mediated transformation to form complementation strains of BbTFO1, and a putative complementary strain was selected by chlorimuron ethyl (10 μg ml −1 ) tolerance. Quantitative real-time PCR (qPCR) and PCR were used to detect the expected deletion and complementation strains.

Multi-phenotypic assays
Phenotypic assays were performed to detect the potential heterogeneities in the phenotype of ΔBbTFO1 compared with controls (WT and ΔBbTFO1/BbTFO1). All assays were repeated three times.
To better assess how OS affected deletion mutant growth, an appropriate amount of conidium suspension (1 μl) was added to CPZ media containing gradient menadione doses (5-50 μM) and H 2 O 2 (1-4 mM). The diameters of all colonies under OS and the control were measured to calculate the colony area after 7 days of incubation at 25˚C. A minimum inhibitory concentration (MIC 50 ) value of H 2 O 2 and menadione that suppressed 50% colony growth was adopted as the OS resistance index in every strain.
Several conidial properties related to biocontrol potential were detected. The conidiation on SDAY was measured according to a related description [19]. Briefly, an appropriate amount of conidial suspension (100 μl) was uniformly added onto SDAY media, followed by incubation at 25˚C. Then, three rounded cultures were collected from every plate every day by a cork borer (6 mm diameter) from day 5 onwards. The conidia on rounded cultures were completely added to 0.02% Tween-80 (1 ml). Later, conidial suspension content was assessed, followed by conversion to conidia count in the plate culture per unit area. In germination broth (GB: 0.5% peptone and 2% sucrose in 0.02% Tween-80), the median germination time (GT 50 , h) necessary to achieve 50% conidial germination under shaking at 110 rpm and 25/ 30˚C was evaluated to be the conidial viability index at different temperatures [21].
The yield of blastospores was quantified in nitrogen-limited broth (NLB: 0.4% NH 4 NO 3, 4% glucose, 0.3% MgSO 4 , 0.3% KH 2 PO 4 ). Briefly, 40 μl conidial suspensions were added to 20 ml of NLB and cultured at 25˚C with shaking (110 rpm). The number of blastospores was counted after 3 days, and then the blastospore production was transformed into cell count per ml culture medium.
Conidial thermotolerance was evaluated using a previously described method [22]. Conidia tolerance to high temperature was defined as the median lethal time (LT 50 , min) when conidia were treated with 0-120 min heat stress exposure at 45˚C. In this process, 100 μl samples were taken from each tube at 15 min intervals, and then 1 ml of GB was added. Afterwards, the sample was subjected to 24 h of shaking at 110 rpm and 25˚C, and a cytometer was used to evaluate the conidial germination rate microscopically.

Fungal virulence
Galleria mellonella larvae were used to assay fungal virulence according to immersion and injection approaches. The larvae were soaked in conidia suspension for 10 seconds and dry paper towels were used to remove excess liquid from the larvae. For intrahaemocoel injection, 1 μl conidial suspension (5 × 10 5 conidia ml −1 ) was added to the larval haemocoel. The two control groups were exposed to sterile 0.02% Tween-80. Each experiment was carried out three times, with 30 larvae being adopted in every experiment. Moreover, the number of dead larvae was measured at intervals of half a day. After analysing the time-mortality trend, the LT 50 (d) for every strain was determined against every larval group. After curing under the optimal conditions for 6 days, the fungal growth on the surface of the mummified cadavers was photographed. During fungal virulence assessment, haemolymph samples were extracted from larvae that survived 72 h after injection, and hyphal bodies were formed in vivo, as seen by laser scanning confocal microscopy (LEICA DMi8, Germany).

Transcript profile analysis of antioxidant-related genes
The CAT family, SOD family and Bbohmm involved in OS responses were analysed by qPCR. In brief, 100 μl of conidial suspensions were inoculated into CPZ medium containing H 2 O 2 (2 mM). After growing at 25˚C for 5 days, an RNAiso Plus Kit (TaKaRa, Dalian, China) was utilized to extract total RNA from ΔBbTFO1 and the WT strain. Then, a PrimeScript RT reagent kit (TaKaRa) was utilized for cDNA synthesis by reverse transcription. The resultant cDNA was then diluted at 10 ng ml -1 and adopted as a template in qPCR by the use of paired primers (Table 1) [23]. For B. bassiana, 18S rRNA was applied as the internal standard of each transcript. The gene transcriptional level was determined and repeated three times. The relative transcription level of the antioxidant-related genes was calculated as the fold change of the mutant strain transcript relative to the WT strain transcript in stress culture.

Statistical methods
One-way ANOVA, as well as Tukey's HSD test, were used to analyse phenotype heterogeneities in the WT compared with control strains.

BbTFO1 characteristics in and generation of BbTFO1 mutants
Using the NDT80 sequence of S. cerevisiae as a query, the annotated B. bassiana ARSEF 2860 strain genome was searched against the NCBI database, generating the most highly related sequence of 502 amino acids (identity: 31%; E-value: 3e-05). Phylogenetic analysis (S1A Fig) showed that the BbTFO1 sequence had the closest relationship with NDT80 from N. crassa (XP 011394327) and T. reesei (XP 006961996) because they were located on the same branch. Similar to the transcription factor NDT80 of S. cerevisiae, BbTFO1 had an NDT80/PhoG-like DNA-binding domain (residues 131-318). Successful BbTFO1 deletion and complementation were verified through PCR and qPCR using paired primers to examine correct recombination events (S1B-S1D Fig). Additionally, BbTFO1 transcription was not detected in the deletion strains, but similar transcription levels were detected in the WT and complementary strains (S1D Fig).

Contribution of BbTFO1 to chemical stress tolerance
The growth of deletion, complementation, and WT strains in different nutrient and chemical stress media is shown in Fig 1. The colony size of deletion strains was similar to that of control strains, and there was no significant difference with those in the rich medium SDAY, nutrientdeficient medium 1/4 SDAY, or minimal medium CPZ (Tukey's HSD, P > 0.05) (Fig 1). Similar results were also found in CPZ medium containing NaCl and Congo red (Tukey's HSD, P > 0.05). Nevertheless, ΔBbTFO1 showed markedly higher sensitivity to OS than the control strains (Tukey's HSD, P < 0.05). The above findings indicate that BbTFO1 plays an essential role in maintaining OS tolerance but not in responding to other nutritional and chemical stresses.

Transcriptional changes in antioxidant-related genes in ΔBbTFO1
The colony size of deletion strains decreased by 41% and 63% compared with the control in CPZ medium containing menadione and H 2 O 2 , respectively (Fig 2A). However, the colony size of the deletion strain was not significantly different from that of the control in CPZ medium (Fig 2A). ΔBbTFO1 showed markedly high sensitivity to OS compared with controls. According to the MIC 50 (Fig 2B), the deletion mutant also displayed high sensitivity to the oxidants H 2 O 2 (19%) and menadione (23%). To explain the reduced tolerance of ΔBbTFO1 to oxidative stress, the transcriptional levels of some antioxidant-related genes were analysed by qPCR. Among them, four of six CAT family genes and Bbohmm showed transcriptional downregulation compared with WT. Significantly, two key catalases, cat2 and cat5, were transcriptionally repressed by 68% and 54%, respectively (Fig 2C), in response to OS in ΔBbTFO1 [23]. However, three of five SOD family genes showed transcriptional upregulation compared with WT. These results suggest that ΔBbTFO1 sensitivity to OS is accompanied by the transcriptional downregulation of these two key catalases.

Role of BbTFO1 in conidiation and conidial quality
For conidiation, the ΔBbTFO1 strain was decreased by 44%, 22% and 21% on days 5, 6, and 7 ( Fig 3A), respectively, which was significantly different from the controls (Tukey's HSD, P < 0.05). However, final ΔBbTFO1 conidiation was similar to that of the control strain (Tukey's HSD, P > 0.05). These results showed that the deletion of BbTFO1 affected conidiation growth but not conidiation ability. Moreover, the production of blastospores at 25˚C showed no significant difference in ΔBbTFO1 compared with the corresponding controls (Tukey's HSD, P > 0.05, Fig 3B).
Despite having no impact on conidiation capacity and the production of blastospores, deletion of BbTFO1 exerts a marked influence on conidial germination and thermotolerance. The ΔBbTFO1 mutant showed GT 50 values of 2.1 h and 2.5 h longer than those of the WT at 25˚C and 35˚C, respectively (Fig 3C). The colony sizes of ΔBbTFO1 decreased by 26% and 65% compared with WT at 30˚C and 35˚C, respectively (Fig 3D). Moreover, heat stress resistance of ΔBbTFO1 at 45˚C decreased significantly compared with that of the WT strain, and its LT 50 decreased by 67% (Fig 3E). These data revealed the role of BbTFO1 in conidial germination and its contribution to thermotolerance but not in the production of blastospores.

Contribution of BbTFO1 to insect virulence
For normal infection via the cuticular penetration method, time-mortality trends and means (± SD) LT 50 of G. mellonella larvae following conidial suspension immersion are shown in Fig  4A and 4B. On the ninth day after immersion, the mortality of the WT and complementation strain was close to 100%, while that of ΔBbTFO1 was approximately 80% (Fig 4A). The mean (± SD) LT 50 values of the WT and complementation strains against larvae were 6.0 ± 0.2 and 5.9 ± 0.2 days via normal cuticle penetration (Fig 4B), respectively. The mean (± SD) LT 50 values of the ΔBbTFO1 mutant against larvae were extended until 7.3 ± 0.3 days (Fig 4B). For intrahemocoel injection, the mortality of the WT and complementation strains were close to 100% on the fifth day after injection (Fig 4C); the mean (± SD) LT 50 values of the WT and complementation were 3.6 ± 0.2 and 3.4 ± 0.1 days, respectively; and the mean LT 50 of ΔBbTFO1 strains against the larvae increased to 4.0 ± 0.1 days (Fig 4D). After 6 days of preservation under the optimal conditions, the fungal growth on larvae infected with the control strain was greater than that on the larvae killed by ΔBbTFO1 (Fig 4E).
To probe the potential reason for the reduced virulence by cuticle-bypassing infection with ΔBbTFO1, haemolymph samples obtained from larvae that survived 72 h after injection were analysed. As a result, there were rich hyphal bodies in the control strains but rare hyphal bodies within the haemolymph samples from the larvae following injection with ΔBbTFO1 conidia (Fig 4F).

Discussion
Compared with the controls, the ΔBbTFO1 strains showed significant defects in tolerance to OS, together with certain typical defects in phenotype related to the biological control potential of fungi, such as virulence and conidial quality. Typically, two catalase genes were markedly First, B. bassiana was subjected to biotic and abiotic OS in the field [24], which negatively affects the insect host infection capacity of fungi [25,26]. Recently, more than 24 antioxidant enzymes showing dispersion within diverse families have been analysed for B. bassiana [27]. Some SODs in entomopathogenic fungi have been demonstrated to be crucial factors of fungal biocontrol potential [28]. Five members of the SOD family characterized in B. bassiana showed that Sod2 and Sod3 played a dominant role in fungal SOD activity [28,29]. Each type of peroxide can be decomposed by members of the peroxidase (POD) family, including larger hydrogen peroxide, H 2 O 2 and xenobiotics. KatG1, a bifunctional catalase-peroxidase, has been suggested to display POD and CAT activities, which play a vital role in chemical OS tolerance in Metarhizium acridum [24]. Both the GRX-GLR and TRX-TRR families have been examined to analyse their impacts on the antioxidation of B. bassiana [30,31]. Some strong antioxidant genes have been adopted for entomopathogenic fungal genetic engineering for the sake of enhancing biological control abilities [32]. Overexpression of cytoplasmic manganese core superoxide dismutase (BbSod2) in B bassiana significantly increased antioxidant capacity, UV-B virulence and tolerance [33].
In addition, catalase family members decompose H 2 O 2 . Knockout mutants of catalases have shown discrete phenotypes in susceptibility to oxidative, heat, and UV-B stress in B. bassiana [34]. Among them, cat5, cat1 and cat4 are influential virulence factors, while cat1 and cat4 show equal importance to the regulation of UV-B resistance of conidia, with cat2 ranking second. In addition, cat1 is the most important regulator of conidial heat tolerance. It is believed that the CAT family affects fungal virulence by changing other virulence-related phenotypes of B. bassiana [23,27]. Prior studies have shown that total catalase activity decreased by 89% and 56% in Δcat2 and Δcat5, respectively, while this decrease was only 9-12% in additional B. bassiana catalase knockout mutants [23]. cat2 and cat5 exert viral parts in B. bassiana antioxidation of catalase family [23]. Moreover, BbTFO1 deletion led to four of six CAT family genes showing transcriptional downregulation compared with the WT. In particular, the two key catalases cat2 and cat5 were transcriptionally repressed by 68% and 54%, respectively. In ΔBbTFO1, the defect of colony growth under OS may be due to the repression of catalase gene expression. Therefore, it was speculated that BbTFO1 can contribute to fungal antioxidation by affecting the expression of catalase genes. Beyond expectation, the absence of BbTFO1 did not increase the sensitivity to cell wall stressors in B. bassiana, while for additional fungi, such as C. albicans, NDT80 deletion resulted in enhanced sensitivity to Congo red and SDS [2]. The lack of NDT80 resulted in a denser biofilm relative to WT and might initiate the TEC1-dependent compensatory response, mainly through the TEC1-ROB1 pathway [35].
Second, according to a previous report in C. albicans [2], NDT80 deletion mutants also showed sensitivity to heat stress in B. bassiana. Heat stress could promote cellular ROS generation, thereby damaging certain biomolecules, such as DNA, lipids or proteins [33,36]. Antioxidant enzymes can scavenge diverse intracellular ROS [27,37]. Typically, SOD can decompose superoxide anions and produce H 2 O 2 and oxygen [38], while H 2 O 2 can be additionally decomposed to oxygen and water via catalase [39]. In the present work, 4 catalase gene expression levels, especially cat2 and cat5, in ΔBbTFO1 were repressed, resulting in a decrease in fungal ability to decompose H 2 O 2 , which naturally caused the reduced tolerance of ΔBbTFO1 conidia to high temperature. In addition, BbTFO1 knockout mutants were demonstrated to suffer defects in the development of conidiation and conidial germination. Similarly, the XprG mutant was slightly slower to initiate conidiation development in A. nidulans [4]. Conidial quality is also considered an essential virulence factor of B. bassiana [40]. In this study, defects in conidial quality may be one of the reasons for the decrease in virulence.
Finally, BbTFO1 mutants showed reduced virulence to hosts in B. bassiana by immersion and injection, as observed in NDT80 mutants of C. albicans. The virulence loss of NDT80 mutants in C. albicans may be associated with the critical part of NDT80 in the control of hyphal growth and stress resistance [2,8]. An NDT80 homologue in F. oxysporum (Suf, suppressor of fusion) is associated with horizontal transport of virulence gene-carrying small chromosomes under conditions with limited nutrients [10]. For B. bassiana, the host immune response will produce cytotoxic molecules such as ROS when it infects insects [41]. Host phagocytes also generate reactive oxygen intermediates (ROIs), which are toxic to a variety of microorganisms [42]. Based on this experiment, it was believed that the disruption of BbTFO1 leads to a reduction in the transcription levels of four catalases and complex phenotypic changes, which are responsible for the reduction in fungal virulence. Catalase is considered a potential virulence factor in pathogenic fungi [23]. Overexpression of one catalase led to increased virulence of B. bassiana [43]. Catalase deletion within A. fumigatus mycelium and conidia leads to a significant decrease in virulence to mice [44]. In Magnaporthe grisea, the absence of cat2 enhances fungal sensitivity to OS and significantly decreases virulence to barley [45]. An NDT80-like protein was further examined in A. fumigatus, a pathogenic fungus in animals, showing that XprG makes no difference in the virulence of mice [9]. The deletion of BbTFO1 leads to defects in antioxidant capacity and germination defects, indicating that BbTFO1 is a critical virulence factor of B. bassiana.

Conclusion
Our results confirmed the previously reported functional diversity of NDT80-like proteins in different fungi [7]. In addition, BbTFO1 plays a crucial role in antioxidant activity, thermotolerance, conidiation, conidial quality and virulence in B. bassiana. According to transcript analysis, BbTFO1 can positively affect the expression of two key catalases to influence the antioxidant capacity of fungi, following the impact of the B. bassiana biological control potential.