Role of Glycerol-3-Phosphate Dehydrogenase in the Development, Pathogenicity, and Glycerol Biosynthesis of Aspergillus flavus Authors: Liurong Zhang, Jiaru Zhao, Hongyi Lin, Shaoze Wu, Fei Wu, Hongtai Ning, Xiuna Wang, Jun Yuan, Yanling Yang Published: 2026 ABSTRACT Aspergillus flavus, a ubiquitous phytopathogen, produces mycotoxins, especially aflatoxin B1 (AFB1), and infects crops worldwide. Glycerol-3-phosphate dehydrogenase (G3PDH) is a key enzyme in the glycerol synthesis and metabolic pathway catalyzing the reversible conversion reaction between glycerol-3-phosphate (G3P) and dihydroxyacetone phosphate (DHAP). However, the biological function of G3PDH in A. flavus remains uncharacterized. In this study, the glycerol-3-phosphate dehydrogenase GfdA and GfdB recombinant proteins of A. flavus were expressed, and the enzymatic activity of the GfdA protein was successfully determined. Subsequently, single-gene knockout strains (ΔgfdA, ΔgfdB), double-gene knockout strains (ΔgfdAΔgfdB) and their corresponding complemented strains (gfdAC, gfdBC) were constructed by a homologous recombination method to explore the biological functions of these two genes in A. flavus. The phenotypic analyses revealed that although both gfdA and gfdB encoded glycerol-3-phosphate dehydrogenases, gfdA plays major roles in colony growth, conidiation, sclerotium formation, crop infection and osmotic stress tolerance in A. flavus. Notably, the performance of the ΔgfdAΔgfdB strains is almost similar to that of the ΔgfdA strain. Biochemical assays demonstrated that the intracellular glycerol content increased significantly in all mutants compared to the wild type (WT) under both normal and osmotic stress conditions. Furthermore, we found that exogenous glycerol supplementation rescued the growth defect of the ΔgfdA and ΔgfdAΔgfdB strains. Taken together, GfdA is important for glycerol synthesis, while GfdB is functionally redundant with respect to GfdA. This study preliminarily explores the main biological functions of GfdA and GfdB, providing a theoretical basis for the study of glycerol anabolic pathways of A. flavus and also offering novel insights into the development of strategies to control aflatoxin contamination. 1. Introduction Aspergillus flavus is a ubiquitous pathogenic filamentous fungus that is distributed worldwide. It is not only a phytopathogen, but also an opportunistic pathogen in humans and animals [1,2]. On one hand, A. flavus can infect corn, peanuts, nuts, soybeans and other crops, causing serious economic losses. On the other hand, the conidia produced by A. flavus can easily invade the bodies of humans and animals through multiple routes, such as the respiratory tract, oral cavity, ears, eyes, nasal passages, and skin, causing various diseases and posing a significant threat to public and animal health [3,4]. Furthermore, aflatoxin, the secondary metabolite of A. flavus, is recognized as one of the most chemically and physically stable mycotoxins found in nature [5]. To date, over 20 distinct aflatoxins have been identified, and all these variants share a basic structure characterized by the presence of a difuran ring and a coumarin moiety (oxaphthalene ketone) [6]. The International Agency for Research on Cancer (IARC) has classified aflatoxin as a Group I carcinogen [7]. Among them, aflatoxin B1 (AFB1) is notorious for its highest production, extreme toxicity, and potent carcinogenicity. Glycerol is the simplest trihydric alcohol, and there is a relatively complete and uncomplicated glycerol anabolic pathway in eukaryotic cells. The primary reason for the production of glycerol by eukaryotic cells is that glycerol can respond to external stress, especially hypertonic stress, oxidative stress and heat stress [8]. And also, glycerol can be used as an alternative carbon source. Previous studies have reported that in the model fungus Saccharomyces cerevisiae, glycerol synthesis begins with the reduction of DHAP to G3P by NADH-dependent glycerol-3-phosphate dehydrogenase (GPD) [9]. Subsequently, G3P is dephosphorylated to glycerol by glycerol-3-phosphate phosphatase (GPP). Nevertheless, in the model filamentous fungus Aspergillus nidulans, it has also been proposed that glycerol was produced by using dihydroxyacetone (DHA) as an intermediate. In this pathway, the glycolytic intermediate DHAP is dephosphorylated to DHA, which is then reduced to glycerol [10]. G3PDH is a key enzyme in the glycerol synthesis pathway and can catalyze the production of G3P from DHAP [9,11]. Most bacterial and mammalian genomes encode only GAPDH, whereas fungal and plant genomes encode many GAPDH isozymes. Arabidopsis thaliana has seven genes that encode cytosolic NAD+-dependent GAPDH and chloroplast NADP+-dependent GAPDH [12]. For S. cerevisiae cells, there are two isoenzymes of GPD (GPD1 and GPD2). Previous studies have shown that GPD1 is a target gene of the HOG pathway and its expression is induced by osmotic pressure [13,14]. However, GPD2 is induced when cells are in an anaerobic environment [15]. Some studies have shown that GPD1 is located in organelles in the cytoplasm, with a small part present in peroxisomes, while GPD2 is also mainly located in the cytoplasm, with a few in mitochondria [16]. Some studies have found that in S. cerevisiae, the gpd1 and gpd2 genes have different effects on the synthesis of glycerol in the strain. Compared to the wild type, the Gpd1 deletion mutant strain has significantly reduced or even completely absent glycerol synthesis ability [11,17,18]. Gpd1-overexpressing strain can significantly increase the intracellular glycerol content [19], but the Gpd2 deletion mutant strain showed no significant difference in intracellular glycerol content compared with the wild type [15]. Recent studies have shown that in S. cerevisiae, Hog1 regulates glycerol synthesis by relieving Ypk1-mediated inhibition of Gpd1 [20]. In A. nidulans, glycerol-3-phosphate dehydrogenase Gfd, which can catalyze the production of G3P from DHAP, is located in the cytoplasm and can bind to the coenzyme NAD+. The study confirmed that both Gfd isozymes contribute to the environmental stress defense of A. nidulans, but their functions are specific, with GfdA involved in osmotic stress defense and GfdB involved in oxidative stress defenses in A. nidulans [20,21]. GfdA had a limited effect on glycerol synthesis, and glycerol accumulation was not reduced during conidial germination, but intracellular glycerol levels were reduced to 60% of wild type during vegetative growth under osmotic stress conditions [21]. It was found that the introduction of the gfdB gene from A. nidulans into Aspergillus glaucus results in an enhanced ability of GfdB to improve multi-stress tolerance in the heterologous host [20]. In subsequent experiments, the introduction of gfdB into Aspergillus wentii conferred only a modest enhancement of stress tolerance and partially reversed the osmophilic phenotype of the recipient strain [22]. This also indicated that even between closely related species, the function of GfdB may differ. Studies also indicated that the gene gfdA encoding G3PDH of A. fumigatus is essential for normal colony growth in glucose medium [23]. Furthermore, overexpression of the GfdA homolog GfdB failed to rescue the phenotype of the GfdA deletion mutant, suggesting that GfdA plays an important role in the synthesis of G3P and glycerol. Overexpression of the gene encoding the predicted glycerol kinase GlcA, capable of phosphorylating glycerol to form G3P, was able to compensate for the growth defect of the GfdA deletion mutant, suggesting that the growth defect of GfdA deletion mutants may be due to the absence of G3P rather than glycerol [23]. However, the functional roles of the genes encoding glycerol-3-phosphate dehydrogenase in A. flavus remain largely unexplored. In this study, we therefore set out to investigate the enzymatic characteristics of GfdA and GfdB, their respective contributions to fungal growth, development, and stress adaptation, and their potential link to aflatoxin accumulation. Elucidation of these questions will not only deepen our understanding of glycerol metabolism in A. flavus, but may also inform novel strategies for aflatoxin contamination control. 2. Materials and Methods 2.1. Strains and Culture Conditions All strains studied are listed in Table 1. All strains were cultured on YES and PDA media for vegetative growth and conidiation, as well as on CM medium for sclerotia formation. YES liquid medium was used for aflatoxin production as described before [24]. 2.2. Phylogenetic Analysis The protein sequences of A. flavus GfdA (QRD87930.1) and GfdB (XP_041150538.1) were retrieved from the NCBI (National Center for Biotechnology Information, United States) database and compared with the corresponding sequences of A. nidulans, Aspergillus niger, Aspergillus oryzae, Aspergillus fumigatus, Saccharomyces pombe and S. cerevisiae via BLASTP 2.17.0+ alignment. A maximum likelihood tree was generated using MEGA-7.0 software with a JTT Matrix model and 1000 bootstrap replicates. Protein domains were predicted by SMART (http://smart.embl-heidelberg.de/, accessed on 1 April 2026) software and drawn by DOG 2.0 software. 2.3. Expression of Recombinant Protein and Assay of Enzyme Activity All E. coil strains and plasmid are listed in Table 2. The gfdA gene was cloned by PCR using special primers that included restriction sites for BamHI and XhoI. The pET32a(+) vector and gfdA gene fragment were digested with BamHI and XhoI and then ligated by ExoIII. The pET32a(+) vector containing the gfdA gene fragment was transformed into an expression strain of E. coli BL21 grown in liquid LB medium and further incubated at 37 °C until the OD value reached 0.5. Isopropyl-β-D-thiolactopyranoside (0.4 mM) was added to the medium to induce target protein expression. After 18 h of induction, cells were harvested and sonicated on ice, and then Ni2+-NTA column was used for purification. The expression of recombinant protein GfdB was carried out by pET28a(+) vector. The enzymatic activity was determined according to the method of Ghoshal et al. [25]. The absorbance at 340 nm was measured with a microplate reader and plotted as a graph. 2.4. Construction of Knockout and Complemented Mutants Construction of knockout and complemented mutants was carried out by homologous recombination [26]. Table 3 lists the primers required for this study. For the construction of knockout mutants, the fusion product (5′-UTR-pyrG-3′-UTR) was amplified by PCR and transformed into protoplasts of A. flavus CA14. Correct knockout transformants were identified by PCR: the ORF fragment of the target gene could not be amplified, whereas the 5′-UTR-pyrG (AP) and pyrG-3′-UTR (BP) fragments indicative of proper homologous recombination were successfully amplified. Furthermore, qRT-PCR was performed to confirm the loss of target gene expression in the knockout strains. Double-gene knockout mutants were constructed based on single knockout mutants, using the selectable marker gene ptrA with pyrithione resistance to replace a second target gene, with the same PCR and qRT-PCR confirmation strategy applied. For the construction of complementation strains, the 5′-UTR-target gene-ptrA-3′-UTR fragment was introduced into protoplasts of the corresponding single knockout strains to replace the pyrG gene, followed by selection with pyrithione. The 5′-UTR-ORF (AP) and ptrA-pyrG (BP) fragments indicative of proper homologous recombination were successfully amplified. Then qRT-PCR was used to confirm restoration of target gene expression. All primer pairs used for validation are listed in Table 3. 2.5. Analysis of Colony Growth, Conidiation and Sclerotia Formation To record colony growth, 1 μL of conidia suspensions (1 × 106 pcs/mL) of different strains were inoculated into YES and PDA media, and diameters were measured 3 d later. For conidial counting, 1 μL of the conidia suspension was inoculated on PDA medium and incubated at 37 °C for 3 d. Then the conidia were washed down with 3 mL of distilled water and then counted by a hemocytometer. To determine the production of sclerotia, 1 μL of the conidia suspensions of each strain were incubated at 37 °C for 7 d, and then the conidia were washed away with 75% ethanol to count the sclerotia [27]. 2.6. Aflatoxin Analysis The aflatoxins were extracted with dichloromethane and then transferred to a centrifuge tube and evaporated to dryness. Aflatoxin production was detected by thin-layer chromatography (TLC), and the results were observed under 365 nm UV light. The JD-801 computer-aided image analysis system (JEDA Co., Nanjing, China) was used for quantitative analysis of aflatoxin biosynthesis [28]. 2.7. The Infection of Peanuts and Corn Seeds To test the infectivity of the strains, conidia suspensions of different strains were inoculated onto the surface of peanuts and corn, which were pretreated according to the method of Yang et al. [27]. The inoculated seeds were then placed in a Petri dish covered with wet filter paper and incubated at 29 °C in the dark for 5 d. Then conidia were counted and aflatoxin was analyzed as described above. 2.8. Determination of Intracellular Glycerol Content of A. flavus The intracellular glycerol content of A. flavus was quantified using a Tissue Cell Glycerol Enzymatic Assay Kit (Applygen, Beijing, China). Briefly, conidial suspensions from different strains were inoculated into the culture medium at equal densities; after 48 h, mycelia were harvested, scraped, and freeze-dried with liquid nitrogen. Subsequently, the samples were weighed, lysed, and analyzed for glycerol content according to the manufacturer’s instructions, with absorbance measured at 550 nm [29]. 2.9. Quantitative RT-PCR The mycelia of the strains were grown on YES medium at 37 °C. Total RNA from mycelium was extracted with a RNA Extraction Kit (Tianmo Bio, Jiangsu, China), and reverse transcribed cDNA was obtained by TransScript one-step gDNA Removal and cDNA Synthesis SuperMix (TransGen Biotech, Beijing, China) [30]. SYBR Green qPCR Mix (Guangzhou Dongsheng Biotech, Guangzhou, China) was used for qRT-PCR on a Thermo Fisher Scientific real-time PCR system (Vantaa, Finland). The relative expressions of the target genes were calculated using the actin gene of A. flavus as the reference [31]. The primers for qRT-PCR are listed in Table 3. 2.10. Statistical Analysis Statistical analyses were performed using GraphPad Prism 6. A p-value < 0.05 was considered statistically significant. The T-test was applied for pairwise comparisons, and analysis of variance (ANOVA) was used for multiple group comparisons. 3. Results 3.1. Identification and Analysis of GfdA and GfdB in A. flavus In order to explore the evolutionary relationship of glycerol-3-phosphate dehydrogenase in different species, the protein sequences of GfdA and GfdB in A. flavus were retrieved from the NCBI database. And then Blastp searches were conducted to compare with other Aspergillus and yeast proteins, such as A. fumigatus, A. nidulans, and S. cerevisiae, etc. A phylogenetic tree was constructed by MEGA 7.0 software. The analysis revealed that GfdA and GfdB proteins in the Aspergillus genus are highly conserved, but there are still evolutionary differences from yeast (Figure 1A). Subsequently, protein domains were analyzed by DOG 2.0. The results showed that all examined proteins have two coenzyme-binding conserved domains: an N-terminal NAD-Gly3P-dh-N domain and a C-terminal NAD-Gly3P-dh-C domain (Figure 1B). To further elucidate the function of GfdA and GfdB, gene knockout strains (ΔgfdA, ΔgfdB), complementary strains (gfdAC, gfdBC) and double knockout strain (ΔgfdAΔgfdB) were constructed using a homology rearrangement strategy (Figure 1C). The positive strains were verified with a PCR method (Figure 1D,E) and qRT-PCR (Figure 1F,G), and the results showed that all mutant strains were successfully constructed for further function study. 3.2. Recombinant Expression and Enzyme Activity Assay of GfdA and GfdB Proteins To determine whether A. flavus GfdA and GfdB proteins have glycerol-3-phosphate dehydrogenase activity, the coding sequences of GfdA and GfdB proteins were cloned using the cDNA of A. flavus WT as a template (Figure 2A,B) and ligated into pET32a(+) and pET28a(+) vectors, respectively. After sequencing and secondary transformation, the recombinant expression strains were successfully constructed. Subsequently, recombinant protein expression was induced by IPTG, and the target proteins were purified using Ni2+-NTA affinity chromatography, yielding GfdA protein (Figure 2C,D). The same procedure was applied to obtain GfdB protein (Figure 2E,F). Finally, an in vitro enzymatic assay demonstrated that GfdA protein possessed glycerol-3-phosphate dehydrogenase activity (Figure 2G), whereas no such activity was detected for GfdB in this study. 3.3. Effects of gfdA and gfdB Genes on the Growth and Conidiation of A. flavus To investigate whether gfdA and gfdB genes play a role in the growth of A. flavus, quantitative conidia suspensions of different strains were inoculated on YES and PDA solid medium and cultured at 37 °C for 3 d in the dark (Figure 3A). It was observed that the colony diameter of the ΔgfdA and ΔgfdAΔgfdB strains decreased and the colony edge appeared irregular, while the ΔgfdB strain did not change, compared with the wild type (WT) (Figure 3B). The results indicated that GfdA exerts a regulatory effect on the growth of A. flavus, while GfdB has no obvious impact on its growth. After growing on PDA plates for 4 d, compared to WT, the conidia production of ΔgfdA and ΔgfdAΔgfdB strains was significantly reduced, while the ΔgfdB strain did not change (Figure 3C). The results revealed that the gfdA gene was involved in the regulation of conidiation in A. flavus, but the gfdB gene was not. In fungi, the formation of conidia is mainly regulated by abaA and brlA genes [32,33], so this study also detected the expression levels of these genes at the transcriptional level. It was discovered that the expression levels of abaA and brlA genes were significantly downregulated in the ΔgfdA and ΔgfdAΔgfdB strains, but did not change in ΔgfdB, when compared with WT (Figure 3D). Together, it can be seen that the ΔgfdA and ΔgfdAΔgfdB strains regulate the formation of conidia by regulating the abaA and brlA genes. 3.4. Involvement of gfdA and gfdB Genes in Sclerotium Formation of A. flavus To study the effects of the gfdA and gfdB genes on sclerotium formation of A. flavus, conidia suspensions of strains were inoculated into CM solid medium and cultured at 37 °C for 7 d in the dark (Figure 4A). The sclerotia were then counted, and it indicated that compared to WT, the sclerotium production of the ΔgfdA and ΔgfdAΔgfdB strains was significantly reduced, while the ΔgfdB strain did not change obviously (Figure 4B). As previously reported, sclerotium morphogenesis in fungi is mainly regulated by the nsdC and nsdD genes [34]. To further dissect how GfdA and GfdB regulate the sclerotium formation in A. flavus, the transcriptional level of the nsdC and nsdD genes was detected. Compared to WT, the expression of ΔgfdA and ΔgfdAΔgfdB was obviously downregulated, while that of ΔgfdB was not changed (Figure 4C), suggesting that GfdA affected the expression levels of nsdC and nsdD genes, thereby influencing the sclerotium production of A. flavus. 3.5. Effects of gfdA and gfdB Mutations on Aflatoxin Biosynthesis In an attempt to analyze the influences of gfdA and gfdB genes on the production of AFB1, conidia suspensions were inoculated into liquid YES medium and cultured at 29 °C for 7 d. Then aflatoxins were extracted with organic solvent dichloromethane and detected with a thin-layer chromatography (TLC) method. As shown in Figure 5, the results showed that compared with WT, none of the tested strains exhibited a significant alteration in toxin content. It suggested that gfdA and gfdB genes did not make a notable contribution in the aflatoxin biosynthesis. 3.6. Pathogenicity of gfdA and gfdB Mutants on Seeds A. flavus and its secondary metabolites are widely present in nature and easily invade crops such as peanuts and corn, causing serious economic losses [35,36]. In the pathogenicity assay, the growth of the strains on the surface of peanuts and corn was recorded after 5 d at 29 °C. Preliminary observation showed that compared to the WT and complemented strains, the hyphal growth of ΔgfdA and ΔgfdAΔgfdB strains was relatively sparse, while no significant difference was observed in ΔgfdB (Figure 6A). Then the conidia on the surface of peanuts and corn were washed and counted. The results showed that compared to the WT, the conidia yield of the ΔgfdA and ΔgfdAΔgfdB strains decreased visibly, but the ΔgfdB strain had no effect (Figure 6B). Moreover, aflatoxin was extracted from the infected peanuts and corn and detected by the TLC assays (Figure 6C). It was observed that the amount of secondary metabolite AFB1 produced by ΔgfdA and ΔgfdAΔgfdB double mutants was markedly reduced relative to the WT (Figure 6D). The above results showed that the deletion of gfdA gene can significantly reduce the ability of A. flavus to infect the seeds, meaning that gfdA has an important impact on the pathogenicity of A. flavus to the host, while gfdB gene does not. 3.7. Impacts of gfdA and gfdB Mutations on the Response of A. flavus to Osmotic Stress Fungi possess unique growth and development mechanisms to adapt to environmental changes. Previous research has shown that glycerol, as an important compatibilizer in cells, can balance the changes under osmotic stress, and proteins encoded by the gfdA and gfdB genes are involved in the synthesis of glycerol in different organisms [37]. Therefore, we studied the functions of gfdA and gfdB genes in the response of A. flavus to osmotic stress. For osmotic stress assays, YES medium was used as the basal medium, supplemented with 1.2 M NaCl, KCl, or Sorbitol, respectively. The inhibition rate was calculated after cultivation for 3 d. The results showed that under osmotic stress, the growth inhibition rates of ΔgfdA and ΔgfdAΔgfdB strains were noticeably decreased compared to WT and complemented strains, while that of ΔgfdB strains had no significant change (Figure 7). This indicated that the gfdA gene plays a role in the response of A. flavus to osmotic stress, but gfdB gene does not. 3.8. Effects of gfdA and gfdB Mutations on Glycerol Synthesis in A. flavus According to previous reports [9,10], filamentous fungi have two glycerol synthesis pathways. One is from DHAP to G3P and then to glycerol; the other is from DHAP to DHA and then to glycerol. Glycerol-3-phosphate dehydrogenase encoded by gfdA and gfdB exactly catalyzes the conversion of DHAP to G3P and participates in the synthesis of glycerol. Therefore, we investigated whether gfdA and gfdB genes exert an influence on the synthesis of glycerol in A. flavus. Glycerol detection assays were performed as described above, and the results showed that the glycerol content of the ΔgfdA, ΔgfdB and ΔgfdAΔgfdB strains did not change compared with WT, whether under normal conditions or under osmotic stress conditions (Figure 8A). It demonstrated that under normal and osmotic stress conditions, the absence of this synthetic pathway may be compensated by another. Meanwhile, it can also be found that the glycerol content of all tested strains under osmotic stress was obviously higher than that under normal conditions. And then, the expression levels of genes related to glycerol synthesis (gppA) and metabolism (glcA) were determined. qRT-PCR results showed that under normal conditions, gppA and glcA were significantly downregulated in all mutants (ΔgfdA, ΔgfdB, and ΔgfdAΔgfdB). Under high osmotic stress, however, gppA was significantly upregulated, whereas glcA was downregulated in all mutants. Additionally, gfdB was downregulated in ΔgfdA, and gfdA was downregulated in ΔgfdB under both normal and high osmotic conditions (Figure 8B,C). To further verify if glycerol could make up for the growth defect caused by the deletion of gfdA gene, 2% glycerol was added to the PDA medium, and conidia were inoculated and cultured at 37 °C for 4 d in the dark (Figure 8D). The results shown that the addition of glycerol could almost completely restore the growth of the ΔgfdA and ΔgfdAΔgfdB strains, indicating that glycerol indeed compensated for the growth defect caused by the deletion of the gfdA gene (Figure 8E). 4. Discussion G3PDH binds to coenzyme I to catalyze the conversion of DHAP to G3P, and there are two homologous proteins, GPD1 and GPD2, in S. cerevisiae [15]. Using bioinformatics methods, we identified the homologous proteins GfdA and GfdB of A. flavus. The first step was to express these two proteins in vitro and determine their enzyme activities. It was verified that GfdA protein could combine with coenzyme NADH to catalyze the conversion of DHAP to G3P and exhibited G3PDH activity in vitro (Figure 2G). However, the corresponding activity of GfdB protein was not detected under the same conditions. Therefore, we aligned the protein sequences of GfdA and GfdB in A. flavus to determine the differences in the key domains and specific substrate-binding sites between the two proteins. It has been reported that the amino acid residues His94, Asn150, and Ser153 in mouse NAD+-dependent glycerol-3-phosphate dehydrogenase are highly conserved and function as specific substrate-binding sites, playing important roles in the catalytic reaction [38]. In A. flavus, sequence alignment revealed that His94 and Asn150 are highly conserved, whereas Ser153 in GfdB is replaced by Pro186. Thus, we speculated that the inactivation of the GfdB protein may be caused by amino acid substitutions at specific sites. Admittedly, this interpretation remains speculative and requires further experimental substantiation. In fact, studies in A. fumigatus have shown that gfdA gene deletion causes growth defects due to lack of G3P [23], and our study also demonstrated that ΔgfdA and ΔgfdAΔgfdB strains have growth defects compared to WT, but the ΔgfdB strain has no growth inhibition. So we speculated that GfdA, rather than GfdB, is the key enzyme responsible for G3P synthesis. Combined with the obtained enzyme activity profiling data, we further hypothesized that the GfdB protein may be a partial functional redundancy with its paralog GfdA. Although gfdA and gfdB are homologous genes in A. flavus, gfdA plays an important role in colony morphogenesis, conidia production and sclerotium formation, while gfdB does not show an obvious effect in these aspects. Specifically, compared to WT, the ΔgfdA mutant exhibited significantly reduced colony diameter, irregular marginal protrusions on colonies, a sharp decline in conidiation, and a severe defect in sclerotium formation. Studies in A. nidulans have demonstrated that deletion of gfdA results in severe growth defects and aberrant hyphal morphology [21], while the gfdB deletion mutant in A. nidulans showed no effect on hyphal growth and development [20], which is consistent with our findings. Notably, the double deletion mutant ΔgfdAΔgfdB displayed nearly identical phenotypic traits to the gfdA single mutant. One possible reason is that GfdB protein may be functionally redundant compared to GfdA protein. An alternative explanation is that the GfdB protein exhibits distinct subcellular localization, which accounts for its functional divergence from GfdA. In S. cerevisiae, it has been reported that due to the different subcellular distribution of GPD1 and GPD2 proteins, they play distinct roles in oxidative stress [16]. It is worth noting that the ΔgfdB knockout mutant of A. nidulans exhibited pronounced sensitivity to oxidative stress (induced by hydrogen peroxide and tert-butyl hydroperoxide), cell wall integrity stress (induced by Congo red), and heavy metal stress [20]. Therefore, the potential role of GfdB in A. flavus under these additional environmental conditions merits further study. As one of the most toxic secondary metabolites in nature, aflatoxin contamination poses a serious threat to food safety and human and animal health [39]. In this study, no effect of gfdA and gfdB mutations on aflatoxin biosynthesis was observed. However, in the seed infection experiment, in the process of infecting peanuts and corn with ΔgfdA and ΔgfdAΔgfdB strains, the output of conidia and aflatoxin decreased obviously, compared to the WT. It is speculated that the reduction in conidia attached to the seed surface during infection may lead to the reduction in aflatoxin production. A similar infection experiment was carried out in Pyricularia oryzae, and it was discovered that the deletion of PoGpd1 or PoGpd2 would lead to reduced virulence of conidia or hyphae on rice, and the glycerol-3-phosphate shuttle is involved in virulence [40]. In fungi, glycerol biosynthesis has been proposed to play an important role in response to hyperosmotic stress. Indeed, in S. cerevisiae, GPD1 is responsible for most glycerol production under osmotic stimulus. gpd1 knockout mutants produce very little glycerol, and gpd1gpd2 double deletion strains did not produce glycerol anymore [17]. Nevertheless, in A. nidulans, inactivation of GpdA has a limited impact on glycerol synthesis. Its glycerol accumulation is not reduced during conidial germination, but intracellular glycerol levels are reduced to 60% of WT during growth under osmotic stress [21]. However, in A. flavus, the glycerol content of the ΔgfdA, ΔgfdB, ΔgfdAΔgfdB strains showed nearly no change compared to WT under both normal and stress conditions. This result suggests that despite their high degree of protein sequence homology (84%), gpd1 and gfdA fulfill distinct roles in the model yeast S. cerevisiae and the filamentous fungal genus Aspergillus, respectively. One possible explanation is that Aspergillus can use an alternative pathway (DHAP-DHP-glycerol) to synthesize glycerol, compared to S. cerevisiae which has only one main glycerol synthesis pathway (DHAP-G3P-glycerol). Interestingly, it was also found that the glycerol content in different strains under osmotic stress conditions was significantly higher than that under normal conditions, and the additional glycerol content under osmotic stress could restore the ΔgfdA growth defect. The possible reason is that the ΔgfdA strain cannot form sufficient G3P under normal conditions, but under the osmotic stress condition the ΔgfdA strain increases the glycerol content, and GlcA converts glycerol to G3P, thus restoring the growth defect. In fungi, osmotic stress responses are coordinated by a complex network of signal transduction pathways, including the cell wall integrity (CWI) pathway, the High-Osmolarity Glycerol (HOG) pathway, and the Target of Rapamycin (TOR) pathway, which exhibit extensive crosstalk to maintain cellular homeostasis [41]. In A. fumigatus, the TOR substrate SchA modulates Hog1 activity and affects glycerol accumulation under osmotic stress [42], supporting a link between TOR signaling and glycerol homeostasis. In S. cerevisiae, glycerol serves as a mediator of crosstalk between the HOG and CWI pathways [43]. Thus, the loss of Gfd proteins may be compensated via signaling-dependent regulation of alternative glycerol biosynthetic or catabolic enzymes, ensuring that glycerol levels remain stable under the conditions tested. 5. Conclusions Overall, we preliminarily explored the main biological functions of the gfdA and gfdB genes encoding glycerol-3-phosphate dehydrogenase in A. flavus and confirmed that the gfdA gene played important roles in the growth and development of A. flavus, such as conidiation, sclerotium formation, etc., but gfdB did not play an obvious role in these aspects. Therefore, a more detailed understanding of the physiological functions of GfdB and the relationship between GfdA and GfdB remains to be fully elucidated in future investigations. Also, these findings advance our current understanding of glycerol metabolic pathways in Aspergillus and filamentous fungi.