G Protein α Subunit GpaB is Required for Asexual Development, Aflatoxin Biosynthesis and Pathogenicity by Regulating cAMP Signaling in Aspergillus flavus

The heterotrimeric G proteins are critical for signal transduction and function in numerous biological processes including vegetative growth, asexual development and fungal virulence in fungi. Here, we identified four G protein alpha subunits (GanA, GpaB, FadA and GaoC) in the notorious Aflatoxin-producing fungus Aspergillus flavus. GanA, GpaB and FadA have homologues in other fungal species, while GaoC is a novel one. Here, we showed that the loss function of gpaB displayed a defect in conidiophore formation and considerably reduced expression levels of conidia-specific genes brlA and abaA. A decreased viability of cell wall integrity stress and oxidative stress were also found in the ∆gpaB mutant. More importantly, aflatoxin (AF) biosynthesis and infection on crop seeds were severely impaired in the gpaB-deficient mutant. Further analyses demonstrated that the intracellular cAMP levels significantly reduced in the gpaB-deficient mutant compared to wildtype strains. Additionally, an alteration of PKA activities in the ∆gpaB mutant was also found. Overall, our results indicated that GpaB played diverse roles in asexual sporulation, AF biosynthesis and virulence by regulating cAMP signaling in Aspergillus flavus.


Introduction
Aflatoxins (AFs) are ranked as one of the most toxic carcinogens for human and animal known in nature. The AFs producing fungus Aspergillus flavus contaminates several important seed crops and food stuffs. Under favorable conditions, A. flavus challenged crop seeds, accumulating the toxic secondary metabolites [1], which has caused serious agriculture problems and posed a threat to human health. Therefore, it is quite important to control the contamination of this fungus on the crops both at pre-and post-harvest stages.
Accumulating studies have demonstrated that the pathogenicity of A. flavus has a close relationship with fungal growth, mycotoxins and the adaptability to environmental stresses [1,2]. In fungi, the G protein signaling, which is a key element of signal transduction pathways, engages of the Gα from the selected Aspergillus species and other pathogenic fungi were downloaded from the NCBI database. The Gα sequences were used for phylogenetic analysis, which resolved into three main groups including the group I (FadA, A. flavus), group II (GanA, A. flavus) and group III (GpaB, A. flavus) clans ( Figure 1A). Intriguingly, both A. flavus and A. oryzae possess an additional copy of the Gα subunit, GaoC, not found in other Aspergillus species (A. nidulans, A. fumigatus and A. terreus). The protein sequence analysis demonstrated that these Gα subunits only contain the unique G_alpha domain, which has GTPase activity and is responsible for binding guanyl nucleotide. FadA, GanA and GpaB in A. flavus all have the well-conserved GTP binding motif "GXGXXGKS" and GTPase domain "DXXXGQ" (Figure 1B), while the first three highly-conserved glycines (G) from the GTP binding motif in GaoC have been exchanged for aspartic acid (D), glutamic acid (E) and lysine (K), respectively, and the highly-conserved glutamine (Q) from the GTPase domain in GaoC has been exchanged for a serine residue (S) ( Figure 1B). In addition, based on the prediction from SMART, although AFLA_124830 (gaoC) encodes a G_alpha domain, this domain is probably catalytically inactive.

Generation of Target Strains
To understand the physiological function of the Gα subunit in A. flavus, the targeted gene replacement strategy was used to disrupt the Gα encoding genes from the CA14 PTs strain. Here, as GaoC was predicted dysfunctionally in A. flavus, we just generate the deficiency mutants of ganA, gpaB and fadA, but not for gaoC. The selected transformants were verified by diagnostic PCR and further confirmed by RT-PCR ( Figure 2). However, we failed to obtain deletion strains for fadA from more than 200 ectopic transformants after many attempts, indicating that fadA might be important in this fungus, making it difficult to delete. Since the ∆ganA mutant displays a wildtype phenotype (data not shown), we did not generate a complementation strain for the ∆ganA mutant. The ∆gpaB mutant, which displayed an apparent phenotype defect compared to the wildtype A. flavus, was complemented with a wildtype gene copy fused with the gfp tag at its C-terminal and was confirmed both by diagnostic PCR and RT-PCR ( Figure 2B,C).

Generation of Target Strains
To understand the physiological function of the Gα subunit in A. flavus, the targeted gene replacement strategy was used to disrupt the Gα encoding genes from the CA14 PTs strain. Here, as GaoC was predicted dysfunctionally in A. flavus, we just generate the deficiency mutants of ganA, gpaB and fadA, but not for gaoC. The selected transformants were verified by diagnostic PCR and further confirmed by RT-PCR ( Figure 2). However, we failed to obtain deletion strains for fadA from more than 200 ectopic transformants after many attempts, indicating that fadA might be important in this fungus, making it difficult to delete. Since the ∆ganA mutant displays a wildtype phenotype (data not shown), we did not generate a complementation strain for the ∆ganA mutant. The ∆gpaB mutant, which displayed an apparent phenotype defect compared to the wildtype A. flavus, was complemented with a wildtype gene copy fused with the gfp tag at its C-terminal and was confirmed both by diagnostic PCR and RT-PCR ( Figure 2B,C).

gpaB Is Involved in Pigmentation and Colonial Morphology in A. flavus
The colony size of the ∆gpaB mutant was increased compared with A. flavus wildtype (WT) and complemented strain (gpaB C ) on PDA and YGT agar medium ( Figure 3A,B). Furthermore, the ∆gpaB mutant produced white fluffy mycelium and no pigmentation compared with the somewhat woolly Kelly green colony of the WT and ∆ganA mutant ( Figure 3A). Importantly, the gpaB C complemented strain recovered all morphological defects ( Figure 3A), indicating that the phenotypic alteration in the ∆gpaB mutant was caused directly by the disruption of the gpaB gene in A. flavus. These data indicate that gpaB is important for normal fungal morphology and pigmentation in A. flavus.

gpaB Is Involved in Pigmentation and Colonial Morphology in A. flavus
The colony size of the ∆gpaB mutant was increased compared with A. flavus wildtype (WT) and complemented strain (gpaB C ) on PDA and YGT agar medium ( Figure 3A,B). Furthermore, the ∆gpaB mutant produced white fluffy mycelium and no pigmentation compared with the somewhat woolly Kelly green colony of the WT and ∆ganA mutant ( Figure 3A). Importantly, the gpaB C complemented strain recovered all morphological defects ( Figure 3A), indicating that the phenotypic alteration in the ∆gpaB mutant was caused directly by the disruption of the gpaB gene in A. flavus. These data indicate that gpaB is important for normal fungal morphology and pigmentation in A. flavus.

gpaB Is Involved in Pigmentation and Colonial Morphology in A. flavus
The colony size of the ∆gpaB mutant was increased compared with A. flavus wildtype (WT) and complemented strain (gpaB C ) on PDA and YGT agar medium ( Figure 3A,B). Furthermore, the ∆gpaB mutant produced white fluffy mycelium and no pigmentation compared with the somewhat woolly Kelly green colony of the WT and ∆ganA mutant ( Figure 3A). Importantly, the gpaB C complemented strain recovered all morphological defects ( Figure 3A), indicating that the phenotypic alteration in the ∆gpaB mutant was caused directly by the disruption of the gpaB gene in A. flavus. These data indicate that gpaB is important for normal fungal morphology and pigmentation in A. flavus.

gpaB Is Required for Conidiation and Negatively Regulates Sclerotia Formation in A. flavus
G protein signaling is important for asexual development in filamentous fungi. To know the role of gpaB functioning in A. flavus sporulation, the conidia production of WT, ∆gpaB and gpaB C strains was assayed after five days of growth on PDA agar media. As shown in Figure 4, deletion of gpaB led to a prominent decrease in conidiation in A. flavus ( Figure 4A,B). The observation of aerial conidiophores also demonstrated that ∆gpaB failed to form normal conidiophore in ∆gpaB ( Figure 4A). Further RT-qPCR analysis of conidia-specific genes' expression showed that brlA and abaA were transcriptionally downregulated in ∆gpaB mutant compared to WT strains (p < 0.01) ( Table 1). These data indicate that gpaB is essential for conidiation in A. flavus.

gpaB Is Required for Conidiation and Negatively Regulates Sclerotia Formation in A. flavus
G protein signaling is important for asexual development in filamentous fungi. To know the role of gpaB functioning in A. flavus sporulation, the conidia production of WT, ∆gpaB and gpaB C strains was assayed after five days of growth on PDA agar media. As shown in Figure 4, deletion of gpaB led to a prominent decrease in conidiation in A. flavus ( Figure 4A,B). The observation of aerial conidiophores also demonstrated that ∆gpaB failed to form normal conidiophore in ∆gpaB ( Figure 4A). Further RT-qPCR analysis of conidia-specific genes' expression showed that brlA and abaA were transcriptionally downregulated in ∆gpaB mutant compared to WT strains (p < 0.01) ( Table 1). These data indicate that gpaB is essential for conidiation in A. flavus. Colony and conidiophore morphology among the WT, ΔgpaB and gpaB C were observed after being grown on PDA agar medium for five days in the dark; (B) The number of conidia of the indicated strains was measured after being grown on PDA agar for five days; (C) Sclerotia formation of the indicated strains grown on sclerotia-inducing Wickerham (WKM) medium was detected. To visualize the sclerotia, 75% ethanol was sprayed on the WKM plates to remove the conidia; (D) The number of sclerotial was counted as in (C). Error bars represent the standard deviation from four replicates, and asterisks, "***" or "*", represent significant differences compared to the wildtype according to the t-test with p < 0.001 and p < 0.05, respectively. The experiments were conducted with four replicates for the indicated strain and were repeated three times. Genes' levels were normalized to the house keeping gene actin and calculated by the 2 −ΔΔCT method. Asterisks "**" and "***" represent significant differences compared to the wildtype for each gene according to the t-test with p < 0.01 and p < 0.001, respectively. The experiments were conducted with three replicates for the indicated strain and were repeated three times. To visualize the sclerotia, 75% ethanol was sprayed on the WKM plates to remove the conidia; (D) The number of sclerotial was counted as in (C). Error bars represent the standard deviation from four replicates, and asterisks, "***" or "*", represent significant differences compared to the wildtype according to the t-test with p < 0.001 and p < 0.05, respectively. The experiments were conducted with four replicates for the indicated strain and were repeated three times. Genes' levels were normalized to the house keeping gene actin and calculated by the 2 −∆∆CT method. Asterisks "**" and "***" represent significant differences compared to the wildtype for each gene according to the t-test with p < 0.01 and p < 0.001, respectively. The experiments were conducted with three replicates for the indicated strain and were repeated three times. In addition to conidia, the sclerotium is another important reproductive structure for A. flavus. Our former study has shown that the cAMP signaling pathway might negatively regulate sclerotia development in A. flavus. To know if the Gα subunit GpaB is involved in sclerotia formation, we assayed the sclerotia reproduction in WT, ∆gpaB and gpaB C strains. The result showed that disruption of gpaB promoted sclerotia production compared to WT and gpaB C strains (p < 0.05) ( Figure 4C,D). Additionally, the expression levels of sclerotia-related genes nsdC, nsdD and sclR were transcriptionally increased in the ∆gpaB mutant (p < 0.01) ( Table 1). These results indicate that gpaB plays a negative role in sclerotial formation in A. flavus.

gpaB Is Involved in Stress Responses in A. flavus
G protein signaling responds to multiple environmental signals including stress factors. To see the potential role of gpaB in stress response, the growth of the indicated strains was determined under different stress conditions. The results showed that the ∆gpaB mutant was much more sensitive to cell wall-damaging agent Calcofluor White (CFW) and Congo-Red (CR) than WT and gpaB C strains (p < 0.001) ( Figure 5A,B). The ∆gpaB mutant also displayed less growth viability under oxidative stress generating by 3 mM hydrogen peroxide (H 2 O 2 ) (p < 0.0001) ( Figure 5C,D). These data indicate that gpaB is involved in stress responses in A. flavus. In addition to conidia, the sclerotium is another important reproductive structure for A. flavus. Our former study has shown that the cAMP signaling pathway might negatively regulate sclerotia development in A. flavus. To know if the Gα subunit GpaB is involved in sclerotia formation, we assayed the sclerotia reproduction in WT, ∆gpaB and gpaB C strains. The result showed that disruption of gpaB promoted sclerotia production compared to WT and gpaB C strains (p < 0.05) ( Figure 4C,D). Additionally, the expression levels of sclerotia-related genes nsdC, nsdD and sclR were transcriptionally increased in the ∆gpaB mutant (p < 0.01) ( Table 1). These results indicate that gpaB plays a negative role in sclerotial formation in A. flavus.

gpaB Is Involved in Stress Responses in A. flavus
G protein signaling responds to multiple environmental signals including stress factors. To see the potential role of gpaB in stress response, the growth of the indicated strains was determined under different stress conditions. The results showed that the ∆gpaB mutant was much more sensitive to cell wall-damaging agent Calcofluor White (CFW) and Congo-Red (CR) than WT and gpaB C strains (p < 0.001) ( Figure 5A,B). The ∆gpaB mutant also displayed less growth viability under oxidative stress generating by 3 mM hydrogen peroxide (H2O2) (p < 0.0001) ( Figure 5C,D). These data indicate that gpaB is involved in stress responses in A. flavus. Error bars represent the standard deviation from four replicates, and asterisks "***" represent significant differences compared to the wildtype according to the t-test with p < 0.001. The experiments were conducted with four replicates for the indicated strain and were repeated three times.

gpaB Regulates AF Biosynthesis in A. flavus
AFs are the most important secondary metabolites in A. flavus. Previous studies have demonstrated that G protein and its downstream cAMP signaling played a negative role in AF/sterigmatocystin (ST) biosynthesis in A. nidulans. Thus, we detected the AF production in the ∆gpaB mutant, and the result showed that AF biosynthesis was severely blocked in the ∆gpaB mutant compared to WT and gpaB C strains (p < 0.01) ( Figure 6A,B). To further confirm that the loss production of AF in gpaB-deficient mutant was caused by the altered expression levels of AF biosynthesis genes, Error bars represent the standard deviation from four replicates, and asterisks "***" represent significant differences compared to the wildtype according to the t-test with p < 0.001. The experiments were conducted with four replicates for the indicated strain and were repeated three times.

gpaB Regulates AF Biosynthesis in A. flavus
AFs are the most important secondary metabolites in A. flavus. Previous studies have demonstrated that G protein and its downstream cAMP signaling played a negative role in AF/sterigmatocystin (ST) biosynthesis in A. nidulans. Thus, we detected the AF production in the ∆gpaB mutant, and the result showed that AF biosynthesis was severely blocked in the ∆gpaB mutant compared to WT and gpaB C strains (p < 0.01) ( Figure 6A,B). To further confirm that the loss production of AF in gpaB-deficient mutant was caused by the altered expression levels of AF biosynthesis genes, we performed RT-qPCR to analyze their expression levels, which demonstrated that the AF regulator coding genes aflR and aflS and its structure gene aflP were transcriptionally downregulated in the ∆gpaB mutant compared to WT strains (p < 0.001) ( we performed RT-qPCR to analyze their expression levels, which demonstrated that the AF regulator coding genes aflR and aflS and its structure gene aflP were transcriptionally downregulated in the ∆gpaB mutant compared to WT strains (p < 0.001) ( Table 1). All these results demonstrate that gpaB is important for AF biosynthesis in A. flavus. Figure 6. gpaB is required for AF biosynthesis. (A) AF production was measured by thin-layer chromatography (TLC) after being grown in YES medium in the dark at 29 °C for five days; (B) The amount of AF production in YES medium was qualified by Gene Tools analysis system software. The double asterisks "***" represent significant differences at p < 0.001. Stnd represents the AFB1 standard.
The experiments were conducted with three replicates for the indicated strain and were repeated three times.

gpaB Is Essential for A. flavus Pathogenicity on Maize Kernels
A. flavus contamination of many important seed crops has caused enormous economic losses. Here, we examined the A. flavus ∆gpaB mutant for their abilities to invade maize kernels. As shown in Figure 7, the ∆gpaB mutants grew less vigorously on maize seeds. A significant decrease in conidia production from the ∆gpaB mutant-infected maize kernels was also observed compared to the WT and gpaB C strains (p < 0.001) ( Figure 7B). Additionally, AF biosynthesis from the ∆gpaB mutant-infected maize kernels was severely blocked ( Figure 7C). All these data demonstrate that gpaB is important for A. flavus pathogenicity on maize seeds. (C) Detection of AF production of gpaB mutant on maize kernels by TLC. Error bars represent the standard deviation from three replicates, and triple asterisks "***" represent significant differences compared to the wildtype according to the t-test with p < 0.001. Stnd represents the AFB1 standard. The experiments were conducted with three replicates for the indicated strain and were repeated three times.

gpaB Regulates cAMP/PKA Signaling in A. flavus
G proteins alpha subunits are shown to operate upstream of the cAMP signaling pathway, which could regulate the activity of adenylate cyclase. To determine whether gpaB is involved in cAMP synthesis in A. flavus, the intracellular cAMP levels were assayed after 48 h of inoculation, and we found that the ∆gpaB mutant demonstrated a prominent reduction in cAMP levels compared to the WT and gpaB C strains (p < 0.001) ( Figure 8A). We further detected the effects of gpaB on  Figure 6. gpaB is required for AF biosynthesis. (A) AF production was measured by thin-layer chromatography (TLC) after being grown in YES medium in the dark at 29 • C for five days; (B) The amount of AF production in YES medium was qualified by Gene Tools analysis system software. The double asterisks "***" represent significant differences at p < 0.001. Stnd represents the AFB1 standard. The experiments were conducted with three replicates for the indicated strain and were repeated three times.

gpaB Is Essential for A. flavus Pathogenicity on Maize Kernels
A. flavus contamination of many important seed crops has caused enormous economic losses. Here, we examined the A. flavus ∆gpaB mutant for their abilities to invade maize kernels. As shown in Figure 7, the ∆gpaB mutants grew less vigorously on maize seeds. A significant decrease in conidia production from the ∆gpaB mutant-infected maize kernels was also observed compared to the WT and gpaB C strains (p < 0.001) ( Figure 7B). Additionally, AF biosynthesis from the ∆gpaB mutant-infected maize kernels was severely blocked ( Figure 7C). All these data demonstrate that gpaB is important for A. flavus pathogenicity on maize seeds. we performed RT-qPCR to analyze their expression levels, which demonstrated that the AF regulator coding genes aflR and aflS and its structure gene aflP were transcriptionally downregulated in the ∆gpaB mutant compared to WT strains (p < 0.001) ( Table 1). All these results demonstrate that gpaB is important for AF biosynthesis in A. flavus. Figure 6. gpaB is required for AF biosynthesis. (A) AF production was measured by thin-layer chromatography (TLC) after being grown in YES medium in the dark at 29 °C for five days; (B) The amount of AF production in YES medium was qualified by Gene Tools analysis system software. The double asterisks "***" represent significant differences at p < 0.001. Stnd represents the AFB1 standard.
The experiments were conducted with three replicates for the indicated strain and were repeated three times.

gpaB Is Essential for A. flavus Pathogenicity on Maize Kernels
A. flavus contamination of many important seed crops has caused enormous economic losses. Here, we examined the A. flavus ∆gpaB mutant for their abilities to invade maize kernels. As shown in Figure 7, the ∆gpaB mutants grew less vigorously on maize seeds. A significant decrease in conidia production from the ∆gpaB mutant-infected maize kernels was also observed compared to the WT and gpaB C strains (p < 0.001) ( Figure 7B). Additionally, AF biosynthesis from the ∆gpaB mutant-infected maize kernels was severely blocked ( Figure 7C). All these data demonstrate that gpaB is important for A. flavus pathogenicity on maize seeds. (C) Detection of AF production of gpaB mutant on maize kernels by TLC. Error bars represent the standard deviation from three replicates, and triple asterisks "***" represent significant differences compared to the wildtype according to the t-test with p < 0.001. Stnd represents the AFB1 standard. The experiments were conducted with three replicates for the indicated strain and were repeated three times.

gpaB Regulates cAMP/PKA Signaling in A. flavus
G proteins alpha subunits are shown to operate upstream of the cAMP signaling pathway, which could regulate the activity of adenylate cyclase. To determine whether gpaB is involved in cAMP synthesis in A. flavus, the intracellular cAMP levels were assayed after 48 h of inoculation, and we found that the ∆gpaB mutant demonstrated a prominent reduction in cAMP levels compared to the WT and gpaB C strains (p < 0.001) ( Figure 8A). We further detected the effects of gpaB on Error bars represent the standard deviation from three replicates, and triple asterisks "***" represent significant differences compared to the wildtype according to the t-test with p < 0.001. Stnd represents the AFB1 standard. The experiments were conducted with three replicates for the indicated strain and were repeated three times.

gpaB Regulates cAMP/PKA Signaling in A. flavus
G proteins alpha subunits are shown to operate upstream of the cAMP signaling pathway, which could regulate the activity of adenylate cyclase. To determine whether gpaB is involved in cAMP synthesis in A. flavus, the intracellular cAMP levels were assayed after 48 h of inoculation, and we found that the ∆gpaB mutant demonstrated a prominent reduction in cAMP levels compared to the WT and gpaB C strains (p < 0.001) ( Figure 8A). We further detected the effects of gpaB on cAMP-dependent protein kinase (PKA) activity; of interest, the ∆gpaB mutant, although showing downregulation of the intracellular cAMP levels, had a higher PKA activity compared to the WT and gpaB C strains ( Figure 8B). Taken together, these results indicated that the G protein alpha subunit GpaB functions upstream of the cAMP signaling pathway activating cAMP synthesis and regulating PKA activities in A. flavus. cAMP-dependent protein kinase (PKA) activity; of interest, the ∆gpaB mutant, although showing downregulation of the intracellular cAMP levels, had a higher PKA activity compared to the WT and gpaB C strains ( Figure 8B). Taken together, these results indicated that the G protein alpha subunit GpaB functions upstream of the cAMP signaling pathway activating cAMP synthesis and regulating PKA activities in A. flavus. A 0.8% agarose gel was used to separate protein samples. Phosphorylated peptides migrated towards the cathode (+), while nonphosphorylated peptides migrated towards the anode (−). P, phosphorylated sample control; N, non-phosphorylated sample control. The triple asterisks "***" represent significant differences compared to the wildtype according to the t-test with p < 0.001. The experiments were conducted with three replicates for the indicated strain and were repeated three times.

Discussion
The heterotrimeric G proteins are critical for signal transduction, which function in numerous biological processes, including vegetative growth, asexual and sexual development and fungal virulence from yeast to human or plant fungi [10][11][12]21,25]. G protein α subunits are the important upstream signal intermediator of cAMP signaling, for which we have recently demonstrated that it was involved in fungal development, pathogenicity and AF biosynthesis in A. flavus [23,24]. To know if Gα subunits are important for AF biosynthesis and A. flavus virulence, we here functionally characterized the Gα subunits GpaB, GanA and FadA in A. flavus. Intriguingly, the fourth Gα subunit GaoC was identified only in Aspergillus flavus/oryzae [26], while its potential role has not been studied yet. In Ustilago maydis, four distinct Gα subunits were characterized, as well, of which the novel Gpa4 has evolved dysfunctionally [27]. Here, we found that the G_alpha domain of GaoC in A. flavus was predicted to be probably inactive in A. flavus. In particular, substitutions of some conserved amino acid residues in the GTP binding motif were found in GaoC, which might affect the GTPase activity or the affinity for GDP/GTP binding.
Accumulating literature works indicated that Gα proteins play roles in regulating fungal asexual development. In A. nidulans and A. fumigatus, GanB/GpaB have been shown to be involved in conidia development and spore germination [12,19,20], and FadA was shown to regulate sporulation and mycotoxin production [9]. In M. oryzae, disruption of magC largely decreases spore production, while the magA deletion mutant does not display obvious defects in conidiation [11]. Here, we found that A 0.8% agarose gel was used to separate protein samples. Phosphorylated peptides migrated towards the cathode (+), while nonphosphorylated peptides migrated towards the anode (−). P, phosphorylated sample control; N, non-phosphorylated sample control. The triple asterisks "***" represent significant differences compared to the wildtype according to the t-test with p < 0.001. The experiments were conducted with three replicates for the indicated strain and were repeated three times.

Discussion
The heterotrimeric G proteins are critical for signal transduction, which function in numerous biological processes, including vegetative growth, asexual and sexual development and fungal virulence from yeast to human or plant fungi [10][11][12]21,25]. G protein α subunits are the important upstream signal intermediator of cAMP signaling, for which we have recently demonstrated that it was involved in fungal development, pathogenicity and AF biosynthesis in A. flavus [23,24]. To know if Gα subunits are important for AF biosynthesis and A. flavus virulence, we here functionally characterized the Gα subunits GpaB, GanA and FadA in A. flavus. Intriguingly, the fourth Gα subunit GaoC was identified only in Aspergillus flavus/oryzae [26], while its potential role has not been studied yet. In Ustilago maydis, four distinct Gα subunits were characterized, as well, of which the novel Gpa4 has evolved dysfunctionally [27]. Here, we found that the G_alpha domain of GaoC in A. flavus was predicted to be probably inactive in A. flavus. In particular, substitutions of some conserved amino acid residues in the GTP binding motif were found in GaoC, which might affect the GTPase activity or the affinity for GDP/GTP binding.
Accumulating literature works indicated that Gα proteins play roles in regulating fungal asexual development. In A. nidulans and A. fumigatus, GanB/GpaB have been shown to be involved in conidia development and spore germination [12,19,20], and FadA was shown to regulate sporulation and mycotoxin production [9]. In M. oryzae, disruption of magC largely decreases spore production, while the magA deletion mutant does not display obvious defects in conidiation [11]. Here, we found that instead of ganA, gpaB plays an important role in regulation of asexual development. The ∆gpaB mutant displayed a defect in conidiophore formation and considerably reduced expression levels of conidia-specific genes brlA and abaA. These data indicate that the role of Gα subunits, especially for the group III Gα proteins, is quite conserved in fungi. In addition to conidiation [2], here, we also found that deletion of gpaB enhanced A. flavus sclerotia formation. Conidiation and sclerotia formation seem to remain balanced as reviewed previously. In this study, we also found that deleting the gpaB gene increased the transcript levels of sclerotia formation-related genes sclR, nsdC and nsdD, which might lead to abnormal sclerotia formation. All these indicated that GpaB is involved in the regulation of A. flavus asexual development.
Many studies have shown that G proteins engage in fungal virulence [10,28]. In A. fumigatus, GapB was shown to engage in regulation of fungal infection [19,20]. The deletion of gpaB was almost avirulent in A. fumigatus [19]. Here, we found that deletion of gpaB causes a considerable reduction in the infection of crop seeds. In A. flavus, the pathogenesis is considered to be related to multiple factors, like sporulation, mycotoxins and adaptability to stress environments [1,2]. Here, we demonstrated that the ∆gpaB mutant grew less vigorously on maize kernels, which also severely impaired its sporulation on crop seeds. On the other hand, we found that the ∆gpaB mutant was defective in cell wall integrity, which might affect the colonization of A. flavus on crop seeds. In Valsa mali, the deletion of the Gα coding gene gvm2/gvm3 showed a reduction in fungal virulence, which had a close relationship with the decreased transcriptional levels of cell wall-degrading enzyme genes [10]. The roles of group III Gα proteins on fungal virulence seem to be conserved in numerous plant pathogens, like C. neoformans (Gpa1), Fusarium oxysporum (Fga2) and Botrytis cinerea (Bcg3), for which the encoding gene is involved in the regulation of their pathogenesis [5,22,29].
In this study, we found that the deletion of gpaB led to a dramatic drop in the intracellular cAMP level, which was consistent with an A. fumigatus study showing that deletion of gpaB reduced the cAMP level in A. fumigatus [20], indicating that GpaB is important for the regulation of cAMP signaling in A. flavus. Previous studies in A. nidulans/parasiticus have demonstrated that the Gα protein FadA and its downstream cAMP signaling played a negative role in AF/ST biosynthesis [30,31]. Here, we found that the inactivation of GpaB blocked AF biosynthesis and its related genes' expression, which might be caused by the decreasing level of intracellular cAMP in A. flavus. It is interesting to wonder why a downregulation of the intracellular cAMP levels by deleting gpaB from A. flavus results in a drop in AF biosynthesis and its regulated genes' transcription levels. Of interest, we found that the full deletion of the gpaB gene showed a higher PKA activity compared to the WT and gpaB C strains, which might inhibit the activity of AF global transcription factor AflR and block AF production. On the other hand, we found that deletion of gpaB led to a prominent decrease in asexual development, which has a close relationship with AF biosynthesis. However, it is still hard to explicate why a decreased level of cAMP in the ∆gpaB mutant displays an increased activity of PKA. Our former study also indicated that abnormally high internal cAMP levels caused by the deletion of the phosphodiesterase encoding gene pdes decreased the PKA activities, and promoted AF production in A. flavus [24]. The dysfunction of GpaB might relieve GpaB-mediated repression of events leading to PKA activation. What also makes sense is that the ∆gpaB mutant showed a significant decrease in asexual sporulation, which is repressed by the activation of PKA in A. flavus.

Conclusions
In conclusion, four heterotrimeric Gα subunits were identified in A. flavus, and GpaB is important for asexual sporulation, AF biosynthesis and virulence by regulating cAMP signaling in A. flavus.
These findings raise the possibility of designing specific strategies to prevent AF contamination and A. flavus invading important crops.

Strains and Culture Conditions
All strains utilized in this study are listed in Table 2. Potato dextrose agar (PDA, BD Difco, Franklin Lakes, NJ, USA) was used for the growth and conidiation assays, supplemented with the appropriate amounts of uridine (1 g/L), uracil (1 g/L) or pyrithiamine (0.1 µg/mL) when necessary. To analyze sclerotia production, the modified Wickerham medium (WKM) was used [32], and after being grown for 7 days, the sclerotium was visualized by using 70% ethanol to wash and kill conidia on the WKM plates. PDA agar supplemented with 100 µg/mL Calcofluor White (CFW), 200 µg/mL Congo-Red (CR) or 3 mM hydrogen peroxide (H 2 O 2 ) was used to determine sensitivities to multiple stresses. The experiments were conducted with four replicates and repeated three times.

Strain Construction
The target gene deletion and transformation were conducted according to the previously described protocols [23]. For disruption of gpaB and ganA, a homologous recombination strategy was used to replace each gene with Aspergillus fumigatus pyrG in the parental strain A. flavus CA14 PTs protoplasts. The double-joint fusion PCR was performed to create the deletion constructs [34]. In brief, the flanking regions upstream and downstream of gpaB or ganA were amplified using primers P1 with P3 and P4 with P6 (Table 3), respectively. A. fumigatus pyrG was amplified from genomic DNA using primers AfpyrG/F and AfpyrG/R. The nested primers P2 and P5 were used to create entire disruption constructs ( Table 3). The purified fusion PCR constructs were co-transformed into A. flavus CA14 PTs strain protoplasts.  To generate a gpaB C complemented strain, the ptrA selective marker was used. A 2.5-kb PCR product (1.1-kb gpaB coding sequence, 1.4-kb upstream sequence) was amplified from A. flavus wildtype genomic DNA using primers gpaB/P1 and gpaB/R (Table 3). A 0.7-kb gfp coding sequence and 0.7-kb A. nidulans trpC terminator region were amplified using primers gfp/F with gfp/R and trpC/F with trpC/R (Table 3), respectively, and a 1.0-kb 5 flanking fragment of A. fumigatus pyrG was amplified using primers AfpyrG-cm/F and AfpyrG-cm/R. The ptrA marker was amplified from vector pPTRI (Takara, Tokyo, Japan) with primers prtA/F and ptrA/R. The primers gpaB/P2 and AfpyrG-cm-1049/R were used to generate the gpaB complemented construct containing the 2.5-kb gpaB complemented PCR product, gfp tag, trpC terminator, the ptrA selection marker and 5 flanking region of the AfpyrG (Figure 2A). The purified fusion PCR construct was co-transformed into ∆gpaB strain protoplasts. The screened transformants were then confirmed by PCR and RT-PCR analysis.

Phylogenetic Analysis
For phylogenetic analysis, the G protein α subunit sequences of interest from the NCBI database (https://www.ncbi.nlm.nih.gov/protein/) were aligned together with A. flavus protein sequences (www.aspergillus.org) using MEGA5.1 software by the neighbor-joining method. Bootstrap analysis was performed with 1000 replicates.

AF Analysis
A 2.5 × 10 6 spore suspension of A. flavus conidia was incubated into 25 mL YES medium in the dark at 29 • C for 5 days for AF production. AF extraction was performed according to the previously described protocol [35]. Then, thin layer chromatography (TLC) was used for AF production analysis. For quantitative analysis of AF production, Gene Tools image analysis software was used.

Fungal Virulence Assays on Corn
The fungal virulence was carried out following the previously described protocol [24,36,37]. The treated corn kernels were incubated with 200 µL 10 6 conidia/mL of indicated strains in a 29 • C incubator under dark conditions for 5 d. The filter paper in the incubator was moistened every day. Three replicates were conducted for each strain. After the incubation, the infected kernels were collected in 50 mL Falcon tubes with 15 mL of sterile 0.05% Tween 80 water solution. One hundred microliters of spore suspension were removed for spore quantification. AF production from the infected kernels was extracted and analyzed following the previously described protocol [24,36,37].

Real-Time Quantitative Reverse Transcription PCR
Expression levels of genes involved in asexual development and AF biosynthesis were measured by RT-qPCR. The 48 h-old liquid shaken mycelium were harvested from PDA medium and lyophilized for the preparation of total RNA extraction. RNA was extracted from 100 mg of indicated mycelium using the Eastep Total RNA Extraction Kit (Promega, Madison, WI, USA) and treated with RNase-free DNase I (Thermo Fisher Scientific, Waltham, MA, USA). First-strand cDNA synthesis was performed using the RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, Waltham, MA, USA). To do RT-qPCR reaction, SYBR Green Supermix (Takara) was used and detected with the PikoReal 96 Real-time PCR system (Thermo Fisher Scientific, Waltham, MA, USA) using the program of an initial denaturing step at 95 • C for 5 min followed by 40 cycles, each consisting of denaturing at 95 • C for 5 s and extension at 60 • C for 30 s. The primers used for RT-qPCR are listed in Table 4. The efficiency of all the primers was between 90% and 110%.
States Department of Agriculture, New Orleans, USA), Yang Liu (Institute of Food Science and Technology CAAS) and Kong Qing for their kindness in providing the strains.