HapX, an Indispensable bZIP Transcription Factor for Iron Acquisition, Regulates Infection Initiation by Orchestrating Conidial Oleic Acid Homeostasis and Cytomembrane Functionality in Mycopathogen Beauveria bassiana

Conidial maturation and germination are highly coupled physiological processes in filamentous fungi that are critical for the pathogenicity of mycopathogens. Compared to the mechanisms involved in conidial germination, those of conidial reserves during maturation are less understood. The insect-pathogenic fungus Beauveria bassiana, as a representative species of filamentous fungi, is important for applied and fundamental research. In addition to its conserved roles in fungal adaptation to iron status, the bZIP transcription factor HapX acts as a master regulator involved in conidial virulence and regulates fatty acid/lipid metabolism. Further investigation revealed that the Δ9-fatty acid desaturase gene (Ole1) is a direct downstream target of HapX. This study reveals the HapX-Ole1 pathway involved in the fatty acid/lipid accumulation associated with conidial maturation and provides new insights into the startup mechanism of infection caused by spores from pathogenic fungi.

U nlike unicellular yeasts, filamentous fungi develop through an asexual sporulation process (1,2), which produces numerous conidia and promotes fungal dispersal, survival, and evolution in ecosystems (3)(4)(5). Conidial germination is critical for the propagation of filamentous fungi and is indispensable for host infection by pathogenic fungi (6). Conidia are dormant cells whose cytoplasm is enclosed by the plasma membrane and cell wall (7). Two dormant forms (i.e., endogenous and exogenous dormancy) exist in fungal spores; the exit from exogenous dormancy to germination is triggered by water and involves the mobilization of endogenous nutrients (8). For instance, in Aspergillus niger, conidial germination occurs in the presence of water and requires additional sugars (e.g., glucose and mannose) for the formation of germ tubes (9). Conidia form and mature on the conidiophore, which is a process accompanied by the storage of nutrients (e.g., fatty acids [FAs] and lipids) (10). Fatty acids (FAs) act as the synthetic precursors of storage lipids and membranes (11). In addition to bound forms, various types of free FAs also accumulate in conidia (12,13). The major pathways for FA synthesis are conserved among fungal species, including yeasts and molds (11), but the transcriptional regulation mechanisms underlying biosynthesis are divergent. For example, the OLE pathway, which involves the Δ9-fatty acid desaturase gene (Ole1), is critical for the synthesis of unsaturated fatty acids (UFAs). In Saccharomyces cerevisiae, the Ole1 gene is controlled by the transcription factor Mga2, which contains an IPT/TIG domain (14). In Aspergillus flavus, the Zn 2 -Cys 6 transcription factor FarA is required for activation of the Ole1 ortholog (15). Filamentous fungi occupy a wide range of ecological niches (1). Thus, there is high diversity expected in transcription regulatory mechanisms of FA storage, particularly in conidia.
Filamentous entomopathogenic fungi (e.g., Beauveria bassiana and Metarhizium robertsii) are important biotic factors that regulate the arthropod populations of ecosystems and have enormous potential for the biological control of insect pests (16). Aerial conidia are the major dispersive and infectious propagules, both in nature and as the active ingredient in practical applications (17). Infection is initiated by conidial attachment to the host cuticle, followed by conidial germination (18). The infective hyphae breach the host cuticle and enter the hemocoel, where they develop into in vivo hyphal bodies through dimorphic transition (19). When the host is killed, the growing hyphae cover the cadaver and produce thick conidia responsible for subsequent infection cycles (20,21). The conidia of entomopathogenic fungi germinate on oligotrophic insect cuticles, which represent special niches of host-fungus interaction. B. bassiana is a well-investigated filamentous entomopathogenic fungus, and its conidia perform exogenous dormancy and germinate after contact with water by mobilizing its endogenous nutrients (22). B. bassiana accumulates many FAs in its conidia (23). Therefore, B. bassiana could be an ideal representative species of filamentous fungi for exploring the mechanisms involved in FA accumulation in the conidia of mycopathogens.
In cells, gene transcription is regulated by various transcription factors (TFs). The basic leucine zipper (bZIP) family of TFs is present in all eukaryotic organisms and involved in extensive cell physiologies. The bZIP domain consists of the basic and leucine zipper regions (24). Among diverse bZIP-type TFs, HapX is characterized by a CCAAT binding complex (CBC)-interacting domain at the N terminus (25). The CBC is a highly conserved transcriptional initiation regulator in eukaryotes. In the filamentous fungus Aspergillus nidulans, the core structure of the CBC consists of three subunits: HapB, HapC, and HapE (or in the case of S. cerevisiae, Hap2/3/5, respectively) (26). The A. nidulans CBC recruits HapX, which shares a domain with the CBC-recruited Hap4 in S. cerevisiae (25). In S. cerevisiae, Hap4 is essential for transcriptional activation of respiration (27). Based on the investigations of Aspergillus species, the CBC contributes to diverse processes (e.g., metabolism and development) in filamentous fungi (28). The roles of HapX and action mode have been best characterized in Aspergillus fumigatus, a human mycopathogen. HapX mediates iron homeostasis, which is essential for fungal physiology. When iron abundance is in excess, HapX activates vacuolar iron storage to enhance fungal resistance to the high-iron conditions (29). During iron limitation or starvation, HapX represses the iron-consuming pathways and enhances the iron-uptake pathways by activating siderophore synthesis. HapX loss attenuates the virulence of A. fumigatus, which indicates that HapX is required for fungal adaption to iron starvation in host niches (30). Additionally, transcription factor HapX has been investigated in other fungal species: A. nidulans (25), Aspergillus terreus (31), Arthroderma benhamiae (32), Fusarium oxysporum (33), Verticillium dahliae (34), Candida albicans (35), and Cryptococcus neoformans (36). In all these species, HapX is required for adaptation to iron starvation and in most also for iron resistance (except for C. albicans). In A. fumigatus and C. albicans, direct repression of genes involved in iron-consuming pathways during iron starvation has been revealed via chromatin immunoprecipitation (29,35). However, robust data regarding the roles of HapX in entomopathogenic fungi are still not available.
In this study, a B. bassiana bZIP-type TF gene (BBA_06576) was recognized as gene HapX (BbHapX) and functionally analyzed. The HapX role in fungal adaptation to iron availability is conserved for B. bassiana. Importantly, our data revealed that BbHapX regulates conidial UFA storage and membrane functionality required for fungi to successfully infect insect hosts.

RESULTS
BbHapX is crucial for fungal virulence. A BLASTP search using the amino acid sequence of A. nidulans HapX (AnHapX) (25) revealed a single significant match (1eϪ071), BBA_06576, in B. bassiana (37). B. bassiana HapX (BbHapX) is encoded by an open reading frame (ORF) of 1,950 bp with a deduced protein of 649 amino acids. BbHapX contains two typical traits in this class of transcription factors, including a CBC-interacting domain and a bZIP domain (see Fig. S1A and B in the supplemental material). Also, BbHapX contains a cysteine-rich-region containing an iron-sensing motif (CGFCN5CNC) which is prevalent in the HapX proteins from filamentous fungi (29) (Fig. S1C). Sequence similarity between two CBC-binding proteins was rather low, and BbHapX displays only 7.40 to 46.8% overall identity to the orthologs from other fungi (Fig. S1D).
To elucidate the roles of BbHapX in B. bassiana, a target deletion mutant of BbHapX was generated by homologous recombination, and the effect of gene loss was rescued by ectopic integration of the entire BbHapX gene into the ⌬BbHapX strain (Fig. S2).
Ablation of BbHapX severely attenuated fungal pathogenicity. In the cuticle-invasion route (Fig. 1A), the insects infected with the ΔBbHapX strain exhibited very low mortality (Ͻ5%) at 12 days postinfection (DPI), whereas both wild-type (WT) and complementation strains killed all insects. The survival curve of the ΔBbHapX strain was extremely significant compared to those of the WT and complementation strains (P Ͻ 0.0001). The median lethal time (LT 50 ) values for the WT and complementation strains were 6 and 7 days, respectively. When bypassing the host cuticle, ΔBbHapX mutant displayed 31% mortality at 12 DPI, which was highly significantly different from those of the WT and complementation strains (100%) (P Ͻ 0.0001) (Fig. 1B). The LT 50 values for the latter two strains were 4 and 5 days, respectively. Additionally, conidia of the WT strain germinated well and developed into germ tubes on the host cuticles; however, most conidia of the ΔBbHapX mutant did not germinate (Fig. 1C). In host hemocoel, conidia of the WT strain developed into hyphal bodies 4 DPI, whereas the ΔBbHapX mutant formed hyphal bodies at 8 DPI (Fig. 1D). These results demonstrated that BbHapX is critical for fungal growth in the host hemocoel and, in particular, for the fungal ability to penetrate the host exoskeleton.
At the main level of functional categories ( Fig. 2A), the upregulated DEGs were significantly overrepresented in the terms of metabolism, protein synthesis, and cell rescue (Data Set S2), whereas the downregulated DEGs were enriched in the categories of metabolism and cellular transport (Data Set S3). In the overlapped subcategory of metabolism ( Fig. 2B), the upregulated DEGs were enriched only in the term of secondary metabolism, while the downregulated DEGs had additional terms of lipid/fatty acid/isoprenoid metabolism and carbohydrate metabolism. In budding yeast, Ole1 encodes a Δ9-fatty acid desaturase which is indispensable for the biosynthesis of UFA (14). In B. bassiana, there are two orthologs of Ole1 (i.e., BBA_07664 and BBA_09419) whose expression was significantly repressed in the ⌬BbHapX mutant (Data Set S1).
BbHapX is required for the metabolic homeostasis of FAs. Ablation of BbHapX resulted in a significant reduction in the content of four FAs (Fig. 2C). The reductions for stearic and palmitic acids were 51% and 57%, respectively, whereas the reductions for oleic and linoleic acids were 87% and 84%, respectively. Exogenous UFAs could partially restore conidial germination on water agar (WA) plates (Fig. 2D) and vegetative growth on nutrient plates (Fig. 2E); however, the saturated fatty acids had no significant effects. In addition, glucose had no restoration effects on the germination percentage of the BbHapX mutant strain grown on the WA plates. This suggests that the BbHapXmediated FA metabolism is not involved in energy supply during conidial germination.
Exogenous OA recovers membrane integrity and virulence. When topically inoculating conidia, oleic acid (OA) significantly enhanced the virulence of the ΔBbHapX mutant, with approximately 20% mortality at 12 DPI (Fig. 3A). Its survival curve was extremely distinct from those of the WT and complementation strains (P Ͻ 0.0001). The LT 50 value for the latter two strains was 6 days. Strikingly, when adding oleic acid in the intrahemocoel injection bioassay, the ΔBbHapX mutant killed all insect hosts at 6 DPI, whereas the WT strain displayed 100% mortality at 4 DPI (Fig. 3B). The LT 50 values for the WT and mutant strains were 3.5 and 4.5, respectively. These results indicated that exogenous oleic acid decreased the difference in mortality and LT 50 value between the WT and ΔBbHapX mutant strains.
Loss of BbHapX led to poor membrane integrity in germlings and in vivo blastospores (Fig. 3C). Without oleic acids, nearly all germlings of ⌬BbHapX mutant were stained by SYTOX, while only 1% of the WT cells were permeable to this dye. However, exogenous oleic acids significantly reduced the percentage of the stained cells (PSC) (ϳ50%) (Fig. 3D). Stearic acid could not significantly reduce the PSC value for the ΔBbHapX mutant. These data show that disruption of BbHapX results in impaired membrane integrity, which could be restored by supplying oleic acids.
BbHapX contributes to fungal adaptation to iron starvation. On the minimal nutrient agar (MMA) plates ( Fig. 4A and B), The ΔBbHapX mutant displayed a colony which was significantly smaller than that of the WT strain. This difference was not changed by supplementing sufficient iron (MMAϩFe). Notably, in the presence of the iron chelator bathophenanthroline disulfonate (BPS), the WT strain was not significantly influenced, but the ΔBbHapX mutant did not form a visible colony. All these defects could not be recovered by exogenous stearic acid (Fig. 4A). Under submerged conditions (Fig. 4C), the relative levels of biomass (RLB) between the WT and mutant strains were similar between MMB and MMBϩFe broth. The RLB value was dramatically reduced in the presence of BPS. Additionally, on the WA plates containing 30 M ferrous iron, the germination percentages for the WT and ΔBbHapX mutant strains were 51.7% Ϯ 3.5% (mean Ϯ standard deviation) and 30.4% Ϯ 3.2%, respectively. The difference between two strains was similar to that on WA plates. On the WA plates including stearic acid and BPS, the germination percentages for the WT and ΔBbHapX mutant strains were 58.7% Ϯ 4.6% and 38.7%, respectively. The difference between two strains was not distinct from that on the WA plates with stearic acid. These results indicated that ambient iron availability had no significant effects on conidial germination and radial growth of the ΔBbHapX mutant.
BbOle1 is a direct target of transcription factor BbHapX. Only the gene of BBA_07664 (named BbOle1) was successfully disrupted. Biochemical analysis indicated that BbOle1 was required for the homeostasis of FA (Fig. 5A). Disruption of BbOle1 resulted in a slight decrease in linoleic acid content (24%) and a modest reduction in the content of palmitic acids (58%) and oleic acids (63%). However, BbOle1 loss had no significant effect on conidial storage of stearic acids. Disruption of BbOle1 resulted in 90% of the cells being stained by SYTOX. After oleic acid feeding, there was no significant difference in the PSC values among the WT, gene disruption mutant, and complementation strains (Fig. 5B). This result indicated that BbOle1 contributes to membrane integrity, which could be restored by additional oleic acids. A significantly lower germination of the ΔBbOle1 mutant conidia (35.3% Ϯ 3.2%) was observed on the WA plates compared to the WT parent (50.0% Ϯ 3.6%) (Fig. 5C). On Sabouraud dextrose agar (SDAY) plates, the disruption mutant displayed a slight growth defect. Nevertheless, oleic acid feeding completely eliminated the growth difference between the WT and gene disruption mutant strains (Fig. 5D). BbOle1 ablation significantly weakened conidial virulence in two bioassays. In the topical bioassay (Fig. 5E), survival curves for the WT and ΔBbOle1 mutant strains were extremely significant (P Ͻ 0.0001), and the LT 50 values for these two strains were 6 and 8 days, respectively. Though oleic acids did not decrease the LT 50 for the ΔBbOle1 mutant, the mutant strain killed all insects at 12 DPI (Fig. 5G). The LT 50 values for the WT and ΔBbOle1 mutant strains were 6 days. In the intrahemocoel injection bioassays (Fig. 5F), the LT 50 values for the WT and ΔBbOle1 mutant strains were 4 and 5 days, respectively. Adding oleic acids resulted in the Conidial suspension was inoculated on the plates and cultured at 25°C. (B) Seven days later, the colony diameter was examined. (C) Submerged growth in the defined broth. The agarose in MMA was omitted, generating MMB broth. The chemicals were included as mentioned above. After inoculation, the broth was cultured at 25°C with shaking. Six days later, mycelia were harvested and dried. The number on the column is the biomass ratio of WT to ΔBbHapX mutant. Scale bars in panel A represent 1 cm. reduced difference in LT 50 between the WT (3.5 days) and ΔBbOle1 mutant (4 days) (Fig. 5H). Apparently, lack of BbOle1 results in significantly lower attenuation compared to lack of BbHapX, which indicates that the role of BbHapX in virulence of B. bassiana is not limited to regulation of BbOle1.
The fluorescence intensity in WT was significantly higher than that in the ΔBbHapX mutant (Fig. 6A). Furthermore, BbHapX was prepared in a heterologous expression system with a fused tag of thioredoxin (Fig. S3). During electromobility shift assay (EMSA), the fused protein bound to the promoter region of BbOle1, while thioredoxin had no binding activity to the promoter (Fig. 6B). This result indicated that the BbOle1 gene functions as a direct target of the BbHapX factor. BbHapX and BbOle1 are required for phospholipid homeostasis. Lipid molecules can exist in their cationic or anionic forms, and certain lipids exist in both forms (e.g., PC16:0/18:1) (Data Set S4). In the cationic profiles (Fig. 7A), the contents of five phospholipids with oleic acid (OA) as an FA chain (e.g., phosphatidylethanolamine [PE] and phosphatidylcholine [PC]) were notably reduced in the ΔBbHapX mutant by approximately 75 to 95% compared with WT. Meanwhile, the BbOle1 loss resulted in a slight to modest reduction (approximately 20 to 30%) in all detected lipids. As for anionic forms (Fig. 7B), BbHapX mutation caused significant reductions in the contents of 11 phospholipids (e.g., PE and PC). In the ΔBbOle1 mutant strain, seven phospholipids (e.g., PE18:1/18:2) had slight decreases and the other four lipids (e.g., PC18:1/18:2) had no significant changes. These results showed that BbHapX and BbOle1 had overlapping roles in phospholipid generation, but BbHapX had a greater effect than BbOle1.
As shown in Fig. 7C, nearly all germlings of the WT strain were not stained by SYTOX, and exogenous phospholipids had no effect on the PSC values. Feeding tests indicated that exogenous phospholipids could reduce the PSC values in two gene disruption mutants. In the ΔBbHapX mutant, lysophosphatidylcholine (LPC) and PC resulted in an approximately 48% and 43% reduction in PSC values, respectively. As for the ΔBbOle1 strain, these two chemicals caused an approximately 36% and 26% reduction in the PSC, respectively. PE had a greater effect in reducing the PSC values, with an approximately 56% and 63% reduction in the ΔBbHapX and ΔBbOle1 mutant strains, respectively. These data indicated that exogenous phospholipids could reduce cytomembrane permeability to SYTOX, and the decreasing degrees of the PSC values were very similar between two gene disruption mutants fed with the indicated phospholipid.

DISCUSSION
In filamentous fungi, transcription factor HapX plays a vital role in the fungus' ability to handle iron starvation in the environment (38) by orchestrating complex metabolic pathways (e.g., siderophore synthesis) (39,40). In B. bassiana, BbHapX is also indispensable for fungal survival under iron starvation. These facts reinforce that HapX is a conserved regulator for fungal iron acquisition, and the question as to whether the HapX-mediated pathways are present in B. bassiana is still open. Conidial germination is an indispensable process in the initial stage of fungal pathogen infection and is highly related to the conidial maturation process (6). Strikingly, besides its roles in fungal adaptive capacity to ambient iron status, B. bassiana HapX acts as an essential transcriptional regulator of infection initiation by regulating FA/phospholipid synthesis during conidial maturation.
In filamentous pathogenic fungi (e.g., Magnaporthe grisea), the reserve lipids are hydrolyzed into FAs and glycerol for conidial germination, which is critical for fungal virulence (41). However, the regulatory mechanisms for lipid/FA storage are still unknown for conidia. In this study, HapX, a bZIP transcription factor, is established as an essential regulator for FA homeostasis in conidia and fungal virulence. A previous study indicated that B. bassiana conidia contain stearic, palmitic, oleic, linoleic, and linolenic acids (23). Our B. bassiana strain accumulates only the first four FAs, which further suggests that accumulation of lipids/FAs varies among different cell types and fungal species (13). B. bassiana conidia can germinate upon contact with water in the absence of external nutrients (22), but external FAs can significantly enhance conidial germination. Conidia require nutrients for germination, and in some phytopathogenic fungi, germination initiates after the conidia adhere to the hydrophobic surface (42) and requires the chemicals exuded by the host plant (43). In B. bassiana, HapX is required for conidial germination on oligotrophic host cuticle, and its loss could be recovered by exogenous UFAs. These results suggest that germination initiation is rather complicated and varies among fungi. During interaction of the fungus and the host, BbHapX contributes to its almost complete virulence during topical infection and partial virulence in intrahemocoel infection. The surface of the insect cuticle is deficient in nutrients (22), whereas the host hemolymph is rich in nutrients, including FAs (e.g., oleic acid) (44). These facts readily explain why exogenous oleic acid restores partial fungal virulence during cuticle penetration and complete virulence via direct infection. Oleic acid could recover the full conidial germination rate but only partial virulence in the topical application bioassay. This might be due to the complexity of host cuticles (18) and the multiple roles of BbHapX during fungal pathogenic growth. BbHapX contributes to fungal survival under the iron-limitation condition. The weakened virulence of the ΔBbHapX mutant might partially result from the impaired growth at low iron availability during fungal interaction with the host. Even so, these findings indicate that BbHapX functions as a crucial regulator for infection initiation by partially orchestrating FA reserves in conidia. Many transcription factors are involved in conidial germination (e.g., Forkhead-box protein in Magnaporthe oryzae and C 2 H 2 zinc-finger protein in Colletotrichum gloeosporioides) (45,46). Nevertheless, few transcription factors link conidial reserves with their germination capacity. In Aspergillus oryzae, atfA (an ATF/CREB-type transcription factor) controls conidial germination via regulating carbohydrate and amino acid reserves (47). Our study reveals a transcription factor engaged in conidial FA accumulation. Both atfA and HapX belong to the superfamily of bZIP transcription factors. This superfamily mediates comprehensive physiological processes in filamentous fungi. For example, in Podospora anserina, IDI-4 is required for heterokaryon incompatibility (48). In Sordaria macrospora, JLB1 regulates autophagy and is required for vegetative growth and development (49). The yeast AP-1-like transcription factor contributes to the oxidation tolerance of many fungi (e.g., A. nidulans and B. bassiana) (50,51). Our results update the understanding of the transcriptional control of reserve accumulation in conidia and also extend the roles of bZIP-type transcription factors.
In B. bassiana, BbHapX has a marked transcriptional influence on metabolism during fungal development and acts as a regulator of oleic acid biosynthesis ( Fig. 2A). BbHapX regulates two FA desaturases of the OLE pathway which is responsible for oleic acid synthesis. In yeast, activation of the OLE pathway involves transcription factors Mga2 and Spt23 (14). The transcription factor Mga2 is analogous to the sterol regulatory element-binding protein transcription factors in mammalian cells (52). In mammalian cells, FA synthesis is under the control of the basic-helix-loop-helix-leucine zipper family of transcription factors (53). In A. flavus, FarA, a Zn 2 -Cys 6 transcription factor, positively controls the stearic acid desaturase genes required for the oleic acid biosynthesis (15). In our study, BbHapX acts as a direct transcriptional regulator of the BbOle1 gene. Desaturase involved in FA metabolism (e.g., Ole1) requires iron as a cofactor for catalytic activity (54). In B. bassiana, exogenous iron has no recovery effects on the impaired growth of the HapX-deficient mutant grown on stearic acid. This result reveals that BbHapX mediates the desaturation of stearic acid, which is independent of the HapX-mediated iron uptake. The present study revealed a new transcription factor controlling oleic acid biosynthesis and suggested that the UFA synthesis pathway is conserved while its transcriptional regulation process varies among organisms.
As revealed, the HapX-Ole1 pathway (HOP) contributes to the maintenance of the plasma membrane integrity of the germinating conidia ( Fig. 3D and Fig. 5B). Additionally, the plasma membrane is a critical barrier for conidial survival in nature, and any impairment to the membrane results in the loss of conidial viability (55,56). For instance, chitosan inhibits conidial germination via disruption of conidial membrane integrity (57). Membrane fluidity is determined by the lipid type, acyl chain composition, and sterol content, which ultimately determine membrane integrity and cell viability (58). In this study, the impaired membrane integrity was restored by exogenous UFAs, particularly by oleic acid. Thus, the HOP controls membrane integrity mainly by maintaining the homeostasis of oleic acid. Similar results have been observed in other microorganisms. For example, Helicobacter pylori, a Gram-negative pathogenic bacterium, relies on UFA biosynthesis for maintaining membrane architecture and function (59). The ratio of saturated to unsaturated acyl chains is an important factor determining membrane fluidity (14). To date, it is well known that UFAs contribute to plasma membrane integrity and determine cell responses to environmental stresses (e.g., hypersaline stress) (60). In budding yeast, Ole1 expression is significantly upregulated by cadmium stress, and its overexpression confers enhanced cadmium resistance to cells resulting from increased membrane fluidity and integrity (61). In Candida parapsilosis, a yeast pathogen of humans, Ole1 significantly contributes to fungal virulence through regulating the UFA production for cell membrane biosynthesis (62). Our data provide evidence that UFA homeostasis is required for membrane integrity during conidial germination and fungal infection processes. This finding helps explain why the UFAs supplied by the biosynthesis pathway are critical for the pathobiology of filamentous mycopathogens.
Phospholipids mainly play structural roles in membranes, including phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylinositol, phosphatidylserine, phosphatidylglycerol, and phosphatidic acid (63). The phospholipid constituents shape the physicochemical traits of the membrane, including fluidity, thickness, rigidity, etc. (64). For example, PC is an important membrane component of vegetative cells and critical for maintaining membrane integrity. In M. robertsii, the PC biosynthesis pathway and its products are required for membrane integrity and contribute to fungal development and virulence (65). Triacylglycerol biosynthesis supplies phosphatidic acid and diacylglycerol for phospholipid synthesis and stabilizes the phospholipid compositions of plasma membranes during fungal growth, differentiation, and virulence of M. robertsii (66). In B. bassiana, the HOP has a substantial effect on the steady state of membrane phospholipids (e.g., PC and PE) (Fig. 7). Defects in the membrane integrity of the HOP-null mutants are due to the loss of phospholipid homeostasis. It is well known that FAs enter into the triacylglycerol biosynthesis pathway in the form of acyl coenzyme A (acyl-CoA) and then appear as hydrophobic chains in phospholipids (67). Together, these results develop an overview of maintenance of phospholipid homeostasis in filamentous fungi, involving three coupled processes of FA, diacylglycerol, and phospholipid synthesis. BbHapX adopts the HOP as a principal route to balance conidial levels of FAs and phospholipids. However, BbHapX has a greater influence on conidial germination, lipid homeostasis, and virulence than BbOle. This implies that as a transcription factor, BbHapX might regulate other potential pathways beyond the HOP to maintain membrane homeostasis and conidial physiology.
FA/lipid metabolism has been thought to form a coevolutionary network between insect host and entomopathogenic fungi (68). In addition to its conserved roles in fungal adaptation to ambient iron availability, transcription factor HapX is a master determinant for virulence that acts to initiate fungal infection via regulating the UFA biosynthesis pathway to maintain conidial lipid homeostasis and membrane functionality (Fig. 8). The virulence contribution of HapX is due to its participation in conidial germination and growth of invasive hyphae. The surface of the insect cuticle is a nutrient-deficient environment. Involvement of endogenous nutrients in infection initiation is an evolutionary adaptation to such selective pressure. This study not only establishes the transcriptional regulation of FA metabolism but also expands our understanding of the diverse mechanisms involved in the initial stage of fungal invasion of hosts.

MATERIALS AND METHODS
Strains and culturing conditions. The WT strain of B. bassiana ARSEF2860 was routinely maintained on Sabouraud dextrose agar at 25°C (22). For plasmid propagation, Escherichia coli DH5␣ (Invitrogen, Carlsbad, CA, USA) was cultured in Luria-Bertani medium supplemented with an appropriate antibiotic as the selection reagent. During the fungal transformation procedure, Agrobacterium tumefaciens AGL-1 was used as a donor strain and cultured in yeast extract broth medium. Czapek-Dox agar was used as a defined medium in this study. When omitting iron, the medium was named as minimal nutrient agar (MMA).
Bioinformatics identification of BbHapX and generation of its mutants for functional analysis. A. nidulans HapX (GenBank no. AB052971) was used to query potential homologs in B. bassiana (37) with a local BLAST program. The homologs in other fungi were downloaded from the NCBI database, and their domains were recognized through the online portal SMART (http://smart.embl-heidelberg.de) (69).
Disruption and complementation of BbHapX were accomplished as previously described (22). The gene disruption vector was constructed as follows: the upstream (1.46-kb) and downstream (1.54-kb) sequences of the BbHapX open reading frame (ORF) were amplified by PCR using the primer pairs P H 1/P H 2 and P H 3/P H 4 (see Table S1 in the supplemental material), respectively. The PCR products were purified and individually ligated into the EcoRI/BamHI and BglII sites of p0380-bar (conferring resistance to ammonium glufosinate), using the ClonExpress II one-step cloning kit (Vazyme Biotech, Nanjing, China). The resultant vector was designated p0380-BbHapX-KO. For gene complementation, the full ORF sequences of BbHapX, together with their corresponding promoter regions (amplified with the primer pair P H 7/P H 8), were cloned into the vector p0380-sur-gateway (conferring resistance to chlorsulfuron), which was then transformed into the ΔBbHapX mutant strain to generate the complemented strain.
All fungal transformation experiments were performed by Agrobacterium-mediated methods. The putative mutants were screened by PCR with the primer pair P H 5/P H 6 and further confirmed by Southern blot analysis with a DIG DNA labeling and detection kit (Roche, Germany). A fragment (319 bp) amplified with the primer pair P H 9/P H 10 (Table S1) was used as a template to prepare the probes.
Transcriptomic analysis of the BbHapX-mediated gene expression. A global transcriptomic analysis was performed on the WT and ΔBbHapX mutant strains as previously described (20).
Total RNA was extracted from the mycelia cultured on SDAY plates for 3 days at 25°C and constructed into two libraries which were analyzed on the Illumina HiSeq X Ten platform at Vazyme Biotech Co., Ltd. (Nanjing, Jiangsu, China). Each library was repeated twice in independent experiments.
All clean reads were mapped onto the genome database of Bb2860 (37) using the HISAT program (70). All mapped genes were normalized in terms of the expected number of fragments per kilobase of transcript sequence per million base pairs sequenced, using Cufflinks software (71). The Cuffdiff method was used to search the differentially expressed genes (DEGs) between two libraries, using the threshold of the q-value of Ͻ0.05 (5% false discovery rate) and an absolute value of log 2 ratio (fold change) of Ͼ1 (72). Enrichment analysis was performed on the DEGs using the online FungiFun2 portal (https://elbe .hki-jena.de/fungifun/) (73), using the threshold of the corrected P value Ͻ0.05.
FA extraction and analysis. Conidial free FAs were extracted and analyzed as previously described (10). Briefly, 100 mg conidia (7 days old on SDA plates) was dispersed in 2.5 ml water, and then 5.0 ml chloroform and 2.5 ml methanol were added to suspension. The resulting suspension was kept at Ϫ20°C for 2 h. Then, 1 ml chloroform and 2 ml methanol were added to the suspension, followed by mixing. After stratification for 1 h, the bottom layer of the liquid was transferred into a tube and dried on a Termovap sample concentrator. The sample was dissolved in 1 ml n-hexane and 0.25 ml sodium methoxide (0.5 M), and the methylation reaction was performed at 55°C for 30 min. The n-hexane layer was collected and dried. Finally, the methylated FAs were dissolved in 120 l n-hexane.
Total fatty acid methyl esters (FAME) were analyzed on a Focus series gas chromatograph (Thermo Scientific) coupled to a DSQ2 mass selective detector (Thermo Scientific). Peak areas were recorded using HP ChemStation software (version D.01.02. 16,2004). FAME were identified and quantified using Supelco 37 component FAME mixture as a standard (catalog no. 47885U) (Sigma).
Lipidomic analysis. Total FAs and lipids were extracted following the above methods. Samples were separated by ultrahigh-performance liquid chromatography and analyzed on a Q Exactive Plus mass spectrometer (Thermo Scientific). Lipid secondary identification and quantification were performed using the software LipidSearch v.4.1.30 (Thermo Scientific). The mass content of each identified molecule was quantified using a primary ion mass spectrum. The differentially changed molecules between the WT and mutant strains were screened with Student's t test at a threshold of P Ͻ 0.05. There were three independent replicates for each strain.
Phenotypic assays and feeding tests with FAs/phospholipids. Fungal phenotypes were evaluated as described previously (22).
(i) Conidial germination. To view conidial germination on the host cuticle, a conidial suspension (10 7 cells/ml) was sprayed on the locust hindwings. Water agar (WA) plates containing 1.5% agarose were used to mimic oligotrophic conditions. Conidial suspensions (100 l) were smeared on the WA plates. After 1 day of incubation at 25°C, the germination percentage was examined.
(ii) Fungal growth on plate and in culture broth. For aerial growth, aliquots of 1 l conidial suspension (10 6 spores/ml) were inoculated on the nutrient plates and incubated at 25°C. Seven days later, the colony diameter was examined. To control the iron level, MMA was used as the iron-depleted conditions. The iron chelator bathophenanthroline disulfonate (BPS) (0.2 mM) and ferrous iron ( were included in MMA to generate iron starvation and iron-sufficient conditions, respectively. To determine the effects of stearic acid on fungal growth, the final concentration used was 0.3%. For preparing culture broth, the agar of MMA was omitted, generating MMB broth. The ion levels were controlled as described above. Conidia were inoculated into broth at final concentration of 10 6 cells per ml and cultured at 25°C with constant shaking (150 rpm). Six days later, mycelia were separated from broth by filtering and dried by the oven-drying method.
(iii) Conidial pathogenicity. Two bioassay methods were used on Galleria mellonella larvae. In the cuticle infection assay, the insects were inoculated by immersing the larvae in the conidial suspension (10 7 conidia/ml) for 15 s. In the intrahemocoel injection assay, 5-l aliquots of conidial suspension (10 5 conidia/ml) were injected into the host hemocoel. All bioassay experiments were repeated three times, and 30 to 35 larvae were used in each replicate. The survival data were recorded and plotted as Kaplan-Meier curves. A log-rank test was used to determine statistical difference between the paired curves.
(iv) Membrane integrity of fungal cells. SYTOX Green nucleic acid staining was used as previously described (74). The in vitro germlings were cultured by spreading 100 l conidial suspension (10 7 conidia/ml) on SDA plates for 12 h at 25°C. The in vivo blastospores were prepared by injecting 5 l of conidial suspension (10 7 conidia/ml) into G. mellonella for 2 days at 25°C. The blastospores were isolated from the hemolymph by centrifuging at 4°C. Cells were stained with 5 mM SYTOX Green for 10 min at 25°C in darkness. The fluorescent signals were observed under a laser scanning confocal microscope, and the percentage of stained cells (PSC) was recorded.
(v) Feeding test with FAs/lipids. To determine whether the impaired phenotypes of the ⌬BbHapX and ΔBbOle1 mutants resulted from the reduced FA generation, we supplemented FAs in the above phenotypic assays. When examining conidial germination on WA plates, glucose was used as a favored energy source control. Additionally, ferrous ion (30 M) or BPS plus stearic acid was included in WA plates and used to evaluate conidial germination. Stearic and palmitic acids were each adjusted to a final concentration of 0.3% (vol/vol). The concentration of oleic and linoleic acids was 0.2% and 0.5m (vol/vol), respectively. Three phospholipids (lysophosphatidylcholine [LPC], phosphatidylethanolamine [PE], and phosphatidylcholine [PC]) were set as 0.005 mg/ml. SDA plates without additional phospholipids were used as the blank control, and SDA plates with stearic acid were used as an FA control. In insect bioassay, OA was mixed with conidial suspension by vigorous vortexing.
Functional analyses of ⌬9-fatty acid desaturase BbOle1. To probe the potential targets involved in FA desaturation, a significantly repressed Δ9-fatty acid desaturase (locus tag BBA_07664) was deemed a candidate gene and named BbOle1 due to its high similarity with Ole1 in budding yeast (GenBank accession no. P21147).
The biological roles of BbOle1 in B. bassiana were analyzed through the same strategy used on the BbHapX gene. All required primers are also listed in Table S1. All phenotypes of interest were examined using the same methods described under "Phenotypic assays and feeding tests with FA/phospholipids." Detecting the requirement of BbHapX for activation of BbOle1. The promoter of BbOle1 was amplified with the primer pair P o 11/P o 12 (Table S1) and fused to the 5= terminus of the green fluorescent protein gene (Gfp). The hybrid fragment was cloned into the XmaI/BamHI site in plasmid pBbTEF-MCS-sur (22). The resulting vector was individually transformed into the WT and ΔBbHapX mutant strains. The green fluorescence intensity in mycelia was determined by Image J software (ImageJ 1.52v; National Institutes of Health, USA) and used to calculate the relative expression level of Gfp between two strains.
Additionally, electrophoretic gel mobility shift assay (EMSA) was used to detect the direct binding activity of BbHapX to the promoter of BbOle1. To increase its solubility, BbHapX was fused to the thioredoxin gene. The BbHapX coding sequence was amplified from cDNA by PCR using the primer pair P H F/P H R (Table S1) and cloned into the BamHI and EcoRI sites in pET32a (Novagen) expression vector. The resultant plasmid was transformed into the E. coli Rosetta DE3 strain. The protein was purified through Ni-chelating affinity chromatography and stored at Ϫ80°C until use. EMSAs were performed according to previously established methods (75,76). Briefly, the BbOle1 promoter for the EMSA was prepared by PCR with the primers P o 13/P o 14 (Table S1). The reaction system (20 l) was established in a 1:10 dilution of binding buffer [poly(dI-dC), 50% glycerol, 1% OP-40,1 M KCl, 100 mM MgCl, 200 mM EDTA (pH 8.0)], promoter (400 ng), and purified BbHapX protein (0 to 2.7 g). The mixture was incubated for 30 min at 37°C and analyzed on 6% polyacrylamide gels. The nucleic acids were stained with ethidium bromide, and the shifted bands were visualized under a UV lamp.
Data availability. The sequence data have been deposited in the NCBI Gene Expression Omnibus (accession no. GSE141409).

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.