Predicted Glycerol 3-Phosphate Dehydrogenase Homologs and the Glycerol Kinase GlcA Coordinately Adapt to Various Carbon Sources and Osmotic Stress in Aspergillus fumigatus

Glycerol plays an important role in the adaptation of fungi to various microenvironments and stressors, including heat shock, anoxic conditions and osmotic stress. Glycerol 3-phosphate dehydrogenase (G3PDH) is able to catalyze dihydroxyacetone phosphate to glycerol 3-phosphate (G3P), which is subsequently dephosphorylated into glycerol. However, current knowledge about the functions of G3PDH homologs in glycerol biosynthesis in Aspergillus fumigatus is limited. Here, we show that the A. fumigatus G3PDH gene, gfdA, is crucial for normal colony growth in glucose media under both normoxic and hypoxic conditions. In addition, failure of the overexpression of the gfdA homolog, gfdB, to rescue the phenotype of a gfdA null mutant suggests that gfdA plays a predominant role in the synthesis of G3P and glycerol. However, in a wild-type background, overexpressing either gfdA or gfdB is able to significantly enhance biomass production of mycelia, suggesting that gfdA and gfdB have similar functions in promoting the use of glucose. Interestingly, overexpression of the gene encoding the predicted glycerol kinase, GlcA, which is capable of phosphorylating glycerol to form G3P, significantly rescues the growth defects of gfdA null mutants in glucose media, indicating that the growth defects of gfdA null mutants might be due to the absence of G3P rather than glycerol. Moreover, Western blotting analysis revealed that gfdA is inducibly expressed by osmotic mediators. However, in the absence of gfdA, osmotic stress can rescue colony growth defects and allow colonies to partially bypass the gfdA requirement in a high osmolarity glycerol pathway-dependent manner. Therefore, the findings of this study elucidate how saprophytic filamentous fungi have developed pathways distinct from those of budding yeasts to adapt to varied carbon sources and survive environmental stresses.

. In this pathway, the glycolytic intermediate dihydroxyacetone phosphate (DHAP) is first reduced to glycerol 3-phosphate (G3P), a reaction catalyzed by NAD-dependent glycerol 3-phosphate dehydrogenase (G3PDH), and then G3P is dephosphorylated into glycerol by specific phosphatases. Alternatively, glycerol can be synthesized by another pathway in which the glycolytic intermediate DHAP is dephosphorylated to dihydroxyacetone (DHA), which is followed by its reduction to glycerol (Siderius et al. 2000).
In different S. cerevisiae strain backgrounds, G3PDHs encoded by GPD1 and GPD2 belong to a group of rate-limiting enzymes that control glycerol synthesis under a variety of media and oxygenation conditions. Although lack of both GPD1 and GPD2 leads to the complete loss of glycerol formation under all tested conditions, loss of either GPD1 or GPD2 does not generate a noticeable change in the glycerol yield (Nissen et al. 2000;Hubmann et al. 2011). Previous studies have also shown that GPD1 plays a crucial role in osmotic adaptation, since the GPD1 mutant shows hypersensitivity to osmotic stress and the expression of GPD1 is induced by external osmotic stress (Albertyn et al. 1994). In contrast, the expression of GPD2 is not affected by elevated external osmotic stress, but is induced in response to anoxic stress. The GPD2 null mutant also grows poorly in an anoxic environment. A double GPD1, GPD2 mutant is highly sensitive to both osmolarity and anoxia (Guadalupe Medina et al. 2010). Further studies have shown that the increased transcript responses of GPD1 triggered by extracellular osmotic stress are regulated by the high osmolarity glycerol (HOG) pathway, which is the central signal transduction system the osmotic stress response (Albertyn et al. 1994). Unlike the activation of GPD1, the hypoxic activation of GPD2 transcription is reported to be independent of the HOG pathway. Instead, it is controlled by another oxygen-independent signaling pathway (Eriksson et al. 1995;Ansell et al. 1997). In comparison, the model fungus A. nidulans also possesses two G3PDH-encoding genes, gfdA and gfdB. The expression of gfdA from A. nidulans fully rescues the growth defects of the S. cerevisiae GPD1 null mutant under osmotic stress, suggesting that gfdA and GPD1 have conserved functions. The gfdA null mutant exhibits reduced intracellular G3P levels and osmoremediable defects on various carbon sources,with the exception of glycerol. In contrast, the functions of gfdB have not been explored in any species of Aspergillus (Fillinger et al. 2001;Furukawa et al. 2007).
As a saprophytic fungus with a large number of buoyant airborne conidia, Aspergillus fumigatus is ubiquitously present in the environment due to its rapid adaptation to different carbon and nitrogen sources, heat shock, anoxic conditions, osmotic stress and other environmental stressors (Brown and Goldman 2016;Bruder Nascimento et al. 2016;Ries et al. 2017). However, knowledge of how the glycerol pathway genes participate in stress responses and adaptation to varied carbon resources is limited in this medically important human pathogen. In this study, we show that gfdA of A. fumigatus is required for normal colony growth in glucose media under both normoxia and hypoxia and that the overexpression of a predicted glycerol kinase, GlcA, which phosphorylates glycerol to G3P, is able to significantly rescue the growth defects of a gfdA null mutant in glucose medium. Our findings indicate that, compared to the model organism S. cerevisiae, the opportunistic human fungal pathogen A. fumigatus has developed a unique glycerol biosynthesis network to adapt to various carbon sources and respond to osmotic stress.

MATERIALS AND METHODS
Strains, media, and culture conditions A list of A. fumigatus strains used in this study is provided in the supplementary data Table S1. Strains were grown in the following media: glucose media (minimal media) containing 1% glucose, 2% agar, 1 mL/L trace elements and 50 mL/L 20 · salt solution, as described previously (Zhang and Lu 2017); glycerol media containing 1% (10 mL/L) glycerol, 2% agar, 1 mL/L trace elements and 50 mL/L 20 · salt solution; Rich media (YAG) containing 0.5% yeast extract, 2% agar, 2% glucose and 1 mL/L trace elements; YUU (for uracil and uridine auxotrophic strains) media containing YAG, 5 mM uridine and 10 mM uracil. Liquid media were identical to the corresponding solid agar media, except for the omission of agar. All strains were cultured at 37°under the normoxic or hypoxic conditions. The hypoxic condition was established in 250 mL sealed bags containing an AnaeroPack-Anaero (Mitsubishi Gas Chemical Company), a disposable oxygenabsorbing and carbon dioxide-generating agent used in anaerobic pouches.

Constructions for GFP labeling and deletion strains
The gfdA-GFP strain was constructed using the MMEJ-CRISPR system as described previously (Zhang et al. 2016a;Zhang and Lu 2017). The sgRNA, which targets the terminator site of gfdA, was synthesized in vitro by MEGAscript T7 Kit (Life Technologies, cat. no. AM1333). The corresponding repair template, including a fragment of GFP-hph (hph is a hygromycin selectable marker) with microhomology arms, was amplified by PCR. Then, fusion products of GFP-hph and sgRNA were cotransformed into a Cas9-expressing A. fumigatus recipient strain. For gene deletions, a similar strategy was carried out, but using sgRNA targeted to the open reading frame (ORF) of the target gene. The primers and annotations for sgRNAs and repair templates are listed in Table S2.

Constructs for complementation and overexpression assays
The plasmid (p-zero-pyr4-gfdA) for gfdA complementation was generated as follows: the selectable marker pyr4 from Neurospora crassa was amplified by PCR using the primers pyr4-SpeI-F and pyr4-SpeI-R and then cloned into the the pEASY-Blunt vector (TransGen Biotech) generating a plasmid P-zero-pyr4. Primers gfdA-revertant-F and gfdArevertant-R were used to generate a fragment that includes the promoter sequence, the complete ORF, and the 39UTR of gfdA. This fragment was then cloned into the NotI site of plasmid P-zero-pyr4 to generate the p-zero-pry4-gfdA plasmid.
For construction of the overexpression strains, plasmids overexpressing gfdA, gfdB, or glcA were generated as follows: PCR, using the primers gpd-gfdA/gfdB/glcA-ATG-F and gfdA/gfdB/glcA-BamHI-R, was used to generate fragments that included the complete ORF and 39UTR of relative indicated genes. Next, the fusion fragments gpdA-gfdA, gpdA-gfdB, or gpdA-glcA, which included gpdA (promoter sequence amplified with gpd-BamHI-F and gpd-down) and its relative gene amplified with primers gpd-BamHI-F and gfdA/gfdB/glcA-BamHI-R, were generated by fusion PCR, respectively. The three fusion fragments were, respectively, subcloned into the BamHI site of the plasmid prg3-AMAI-NotI to generate overexpression plasmids pAMAI-gpd-gfdA, pAMAI-gpd-gfdB, and pAMAI-gpd-glcA, which contain the pyr4 marker (Aleksenko and Clutterbuck 1997). Transformation procedures were carried out as previously described (Zhang and Lu 2017). Transformants were selected in the medium without uridine and uracil or in the presence of 150 mg/mL hygromycin B (Sangon) or 0.1 mg/mL pyrithiamine (Sigma). In order to recycle the pyr4 selectable marker, 5-FOA resistance (1 mg/mL 5-FOA) was selected in the recipient strains. All primers used are listed in the supplementary data Table S2.
Strain verification by diagnostic PCR and Southern blotting All transformant isolates were verified by diagnostic PCR analysis using mycelia as the source of DNA. Primers were designed to hybridize upstream and downstream of the expected cleavage sites as labeled in Figure S1. For Southern blotting, genomic DNA was digested with BamHI, separated by electrophoresis, and transferred to a nylon membrane. A 0.7 kb fragment amplified with primers gfdA/gfdBprobeF and gfdA/gfdBprobeR was used as a probe. Labeling and visualization were performed using a DIG DNA labeling and detection kit (Roche Applied Science), according to the manufacturer's instructions. The details are given in the supplementary data Figure S1.

Western blotting
To extract proteins from A. fumigatus mycelia, 10 8 conidia were inoculated in liquid glucose media with or without 1M sorbitol/NaCl at 220 rpm on a rotary shaker at 37°for 24 h. Protein extraction was performed as previously described (Nandakumar et al. 2003). Western blotting was performed as routine procedures (Zhang et al. 2016b). GFP fusion protein was detected using an anti-GFP mouse monoclonal antibody (Roche) at 1:3,000 dilution. Actin was detected using an antiactin antibody (ICN Biomedicals Inc.) at a 1:50,000 dilution.

Microscopy
For microscopy, fresh conidia were inoculated onto sterile glass coverslips overlaid with 1 mL of liquid glucose media with or without 1 M sorbitol. Strains were cultivated on the coverslips at 37°for 14 h before observation. The coverslips with hyphae were gently washed with PBS buffer three times. Differential interference contrast (DIC) and green fluorescent images of the cells were collected with a Zeiss Axio Imager A1 microscope (Zeiss, Jena, Germany).

RNA extraction for qRT-PCR
The qRT-PCR analysis was performed after growth in liquid glucose or glycerol media for 24 h at 37°in a rotary shaker at the speed of 220 rpm. Total RNA was isolated from the mycelium with TRIzol (Roche) following the manufacturer's instructions. The genomic DNA digestion and the synthesis of cDNA were performed using HiScript R II Q RT SuperMix for qRT-PCR kit (Vazyme) following the manufacturer's instruction. qRT-PCR was executed by ABI One-step fast thermocycler (Applied Biosystems) with SYBR Premix Ex TaqTM (TaKaRa). Independent assays were performed with three replicates, and transcript levels were calculated by the comparative threshold cycle (DCT) and normalized against the expression of tubA mRNA level in A. fumigatus. The 2 -DDCT was used to determine the changes in mRNA expression. All the qRT-PCR primers and annotations are given in supplementary data (Table S2).

Statistics
Data are given as means 6 SD. The SD was from at least three biological replicates. Statistical significance was estimated with Origin8 using Student's t-test. P-values less than 0.05 were considered statistically significant.

Data availability
Strains generated in this study are available on request. The authors state that all data necessary for confirming the conclusions presented in article are fully represented within the article and the Supplemental Material. Supplemental material available at Figshare: https://doi.org/10.25387/ g3.6224552.

RESULTS
gfdA is crucial for colony growth in glucose media To identify the putative homologs in A. fumigatus, the S. cerevisiae GPD1 and GPD2 sequences were used as queries to perform a BLASTP analysis in the database of A. fumigatus. The BLASTP results showed the top homolog candidates are AFUB_002530 (46% identity to GPD1 and 45% identity to GPD2 in protein sequence) and AFUB_024230 (49% identity to GPD1, 48% identity to GPD2), hereafter gfdA and gfdB, respectively. Furthermore, according to BLASTP analysis, selected species Schizosaccharomyces pombe, Candida albicans, Candida glabrata, and A. nidulans also have two G3PDH homologs ( Figure S2). Moreover, according to a domain analysis via the SMART software (http://smart. embl-heidelberg.de/), these G3PDH homologs all have the NAD_ Gly3p_dh_N domain at their N-terminus and the NAD_ Gly3p_dh_C domain at their C-terminus ( Figure 1A). To investigate the functions of gfdA and gfdB in A. fumigatus, we constructed single and double deletions mutants of gfdA and gfdB. As shown in Figure 1B and 1C, the gfdA null mutant (DgfdA) displayed severe growth defects under both normoxic and hypoxic conditions in solid glucose media when glucose (1%) was used as sole carbon source. In contrast, the gfdB single deletion mutant (DgfdB) showed similar phenotypes to those of the parental wild-type strain in glucose media under the normoxic or hypoxic conditions. Moreover, the DgfdADgfdB double mutant showed a similar growth phenotype as that of DgfdA, suggesting that gfdA plays a predominant role, and that, despite the high levels of homology between their protein sequences, gfdA function cannot be replaced by gfdB for colony growth. To further characterize function of GfdA, the strain gfdA-GFP was generated and showed a similar colony phenotype to its parental wild-type strain in YAG (rich media), indicating that GfdA-GFP was functional ( Figure. S3). Florescence microscopy showed that the GfdA-GFP fusion protein was predominantly localized within the cytosol ( Figure 1D). Next, a Western blotting experiment was carried out to analyze the molecular mass of GfdA-GFP. Since the GFP protein is about for 27-kD, the relative molecular mass of GfdA was estimated approximately for 56.5 kD, which is consistent with the predicted size of the GfdA protein based on its protein coding sequence. This band was absent in the parental wild-type strain lacking the GFP tag ( Figure 1D).
Overexpressed gfdB is unable to rescue defects of DgfdA but has a similar growth enhanced-function as gfdA overexpression in background of wild type In order to further dissect the functions of gfdA and gfdB, we transformed the full-length ORF sequence of gfdA and gfdB under the control of the constitutive promoter gpdA (a strong promoter from A. nidulans) into the parental wild-type strain, separately, resulting in two overexpression strains, WT OE::gfdA and WT OE::gfdB . Taken together with the findings described above and the fact that GfdA and GfdB have the NAD_Gly3p_dh_N domain and the NAD_Gly3p_dh_C domain, we hypothesized that they may have complementary functions in growth and conidiation. Therefore, we overexpressed gfdB in a gfdA deletion strain (DgfdA OE::gfdB ) and confirmed the overexpression of gfdB by qRT-PCR (Figure 2A). Unexpectedly, the result showed that not only was gfdB overexpression unable to rescue the DgfdA defects, but that it even exacerbated the defects to some extent ( Figure 2B). However, both WT OE::gfdA and WT OE::gfdB displayed significantly enhanced production of mycelia biomass in glucose media ( Figure  2B and 2C), suggesting that gfdA and gfdB have a similar functions for promoting the utilization of glucose as carbon source. Taken together, these data suggested that, at least under the conditions tested, gfdA plays a dominant role over gfdB.
Glycerol rescues the defects of a gfdA null mutant Next, we assayed the phenotypes of DgfdA and DgfdADgfdB in A. fumigatus in glycerol media (glycerol (1%) as the sole carbon source). As shown in Figure 3A, the growth defects of DgfdA and DgfdADgfdB could be fully restored in glycerol media. Moreover, the three overexpression strains, WT OE::gfdA , WT OE::gfdB and DgfdA OE::gfdB , showed faster growth than that of the parental wild-type strain in glycerol media. Finally, a low dose of glycerol (0.1% glycerol) was able to rescue the growth defects of DgfdA and DgfdADgfdB to the levels of the parental wild-type colonies ( Figure 3B). Taken together, these results suggest that defects induced by the gfdA deletion may be due to a blockage of glycerol biosynthesis in A. fumigatus.
Overexpression of a predicted glycerol kinase GlcA bypasses the requirement of gfdA in glucose media for colony growth Next, we investigated whether the defects of growth induced by the loss of gfdA were due to lack of glycerol or G3P. G3P can be synthesized by two classical pathways ( Figure 4A) (Fillinger et al. 2001). The first pathway is catalyzed by an NAD-dependent glycerol 3-phosphate dehydrogenase converting dihydroxyacetone phosphate (DHAP) into G3P The second pathway is catalyzed by a glycerol kinase encoded by glcA (AFUB_068560) converting glycerol to G3P (David et al. 2006). To further dissect whether gfdA and glcA had overlapping functions during the glycerol biosynthesis, we knocked out glcA in the DgfdA and the parental wild-type strain backgrounds. The glcA null mutant showed defects of growth and conidiation in glycerol media but not in glucose media, suggesting glcA is required when glycerol is the sole carbon source, and gfdA is required when glucose is the sole carbon source ( Figure 4B). Notably, the DgfdADglcA double mutant showed an exacerbation of colony defects in both glucose and glycerol media ( Figure 4B). In glucose medium, DgfdADglcA colonies were very small, and colonies were nearly undetectable colonies in the glycerol medium ( Figure 4B and 4D). To further verify the functional relationship between gfdA and glcA, we transformed the full-length ORF sequence of glcA under the control of the constitutive promoter gpdA into the DgfdA and reference strains, resulting in two glcA-overexpression strains, DgfdA OE::glcA and WT OE::glcA . We also confirmed the overexpression of glcA by qRT-PCR ( Figure 4C). In glucose medium, DgfdA OE::glcA displayed rescued wild-type-like phenotypes for colony growth, but not for conidiation, suggesting that overexpressed glcA is able to rescue growth defects associated with loss of gfdA. In comparison, WT OE::glcA still showed wild-type like colony phenotypes ( Figure 4B, 4D and 4E). These data suggested that overexpression of glcA may result in the production of accumulated G3P, which allowed colonies to bypass the requirement of gfdA to produce glycerol use in colony growth, indicating that the growth defects of gfdA null mutant might be due to absence of G3P rather than glycerol. In contrast, when glycerol was used as the sole carbon source, DgfdA OE::glcA showed defective phenotypes similar to WT OE::glcA ( Figure 4B and 4F), indicating the overexpression of glcA may cause the production of accumulated G3P, which results in growth defects in glycerol media either in the presence or absence of gfdA.
Osmotic stress is capable of bypassing the gfdA requirement for the use of glucose To investigate whether gfdA is responsible for the response to osmotic stress, we analyzed the protein expression in the gfdA-GFP strain cultured under an osmotic stress condition (1 M sorbitol or NaCl). Our results showed that the expression of GfdA-GFP was clearly increased under the osmotic stress condition, especially by the addition of sorbitol in glucose media ( Figure 5A). In addition, microscopy consistently showed that the fluorescence intensity of GfdA-GFP in hyphae treated with sorbitol was greater than that of the non-osmotic stress treatment ( Figure 5B). Surprisingly, under osmotic stress, the DgfdA strain displayed near wild-type colony phenotypes in glucose media ( Figure 5C), indicating that the colonies under osmotic stress are capable of bypassing the gfdA requirement for the use of glucose. To further explore whether the above recovery of colony growth in DgfdA under osmotic stress requires the high osmolarity glycerol (HOG) pathway, we first deleted sakA, which encodes the protein kinase in the last step of the HOG response pathway (Bruder Nascimento et al. 2016), in the DgfdA and parental wild-type strains. As shown in Figure 5C, under the osmotic stress condition, the DsakA, DgfdA double mutant still showed a very severely sick colony phenotype compared to the recovered wildtype colony growth of a DgfdA mutant. These data indicated that sakA is required for bypassing the gfdA requirement and suggested that colonies under osmotic stress are capable of bypassing the gfdA requirement for the use of glucose in a SakA (HOG)-dependent pathway. In order to explore whether gfdB is responsible for the osmotically remediable growth, we tested the phenotypes of DgfdB and DgfdADgfdB in osmotic stresses. Both DgfdB and DgfdADgfdB displayed comparable phenotypes to the parental wild type and to the DgfdA strain ( Figure 5C), suggesting gfdB may not play a role in osmotic responses.

DISCUSSION
Using homology to the S. cerevisiae orthologs, we identified two genes, gfdA and gfdB, that encoding glycerol 3-phosphate dehydrogenase, which catalyzes dihydroxyacetone phosphate to glycerol 3-phosphate (the key substrate of glycerol biosynthesis), in the filamentous fungus A. fumigatus, gfdA, but not gfdB, is necessary for the normal growth of A. fumigatus in glucose media. In addition, growth-defective gfdA null mutants cannot be rescued by the overexpression of gfdB. One possible reason is that gfdB may have diverged sufficiently that it has no overlapping function with that of gfdA. Another possibility is that gfdB might have a different protein localization than that of gfdA. Interestingly, in a wild-type background, overexpressing either gfdA or gfdB is able to significantly enhance biomass production of mycelia. These data suggest that in the presence of gfdA, gfdB is able to play an additional role in colony growth in glucose media. However, understanding the relationship between gfdA and gfdB in more detail relationship will require additional work.
The osmolyte glycerol plays an important role in the cellular response to hyperosmositic stresses. Therefore, glycerol 3-phosphate dehydrogenase, a key enzyme in the glycerol biosynthesis pathway, may also play a role in this stress response. Indeed, deletion of the glycerol 3-phosphate dehydrogenase-encoding gene GPD1 in S. cerevisiae conferred hypersensitivity to osmotic stress (Albertyn et al. 1994;Ansell et al. 1997). However, deleting gfdA, the GPD1 ortholog in A. fumigatus, showed no detectable colony phenotype compared to the parental wild-type control strain under hyperosmotic conditions ( Figure 5C). This result suggests that, despite their high degree of protein homology, GPD1 and gfdA play different roles in budding yeasts S. cerevisiae and filamentous fungi Aspergillus. One possible explanation is that Aspergillus may use an alternative pathway (DHAP-DHP-Glycerol) to synthesize glycerol in response to hyperosmosis such that it bypasses the requirement of gfdA to the extent that the deletion of gfdA has no phenotype under osmotic stress (Redkar et al. 1995). In contrast, GPD1 in S. cerevisiae is required for the glycerol biosynthesis pathway and the osmotic stress could not bypass the requirement of GPD1 for colony growth, indicating there may be no alternative pathway to respond to external osmotic stress in the absence of GPD1. These data also suggest that Aspergillus has developed more a more complex adaptation system for producing glycerol for adaption to hyperosmosis.
In order to survive in a broad range of environmental niches, fungi also possess metabolic pathways that allow them to utilize diverse carbon and nitrogen sources. According to studies on the model fungus A. nidulans and the opportunistic pathogen A. fumigatus, the gfdA and glcA genes play crucial roles in the use of glucose and glycerol, respectively. When glucose is the sole carbon source, it is gfdA but not glcA that is necessary for normal colony growth and conidiation. In contrast, glcA is required for growth in when glycerol is the sole carbon source. Compared to single mutants of gfdA and glcA, the double mutant DgfdADglcA is very sick in both glucose and glycerol media. Previous studies in A. nidulans have demonstrated that the double mutant of gfdA and glcA produces very little G3P, but contains elevated glycerol content in a glucose media condition (Fillinger et al. 2001), which suggests that defects in DgfdA may result from the blockage of the synthesis of G3P, not glycerol. Our finding that glcA overexpression  was able to significantly rescue the growth defects of the gfdA deletion in glucose media, further supports the hypothesis that gfdA is involved in the synthesis of G3P. In addition, the aberrant colony phenotypes produced by the overexpression of glcA in a wild-type strain background imply that excessive G3P is toxic for cells. We therefore suggest that fungal cells coordinately produce G3P regulated by GfdA and GlcA to adapt varied carbon resources and survive in different environmental niches.
Therefore, our data demonstrate that G3P plays a crucial role in Aspergillus growth so that the reduced production of G3P induced by deletion of gfdA may result in blockage of glucose usage. In addition, the decrease of G3P of DgfdA in A. fumigatus may have impact on the integrity and biogenesis of cell wall for colony growth since it displayed the comparable decreased colony growth compared to that of wild type, which may be consistent with that of the gfdA deletion in A. nidulans (Fillinger et al. 2001). In contrast, the excess accumulation of G3P also causes sick colony growth in glycerol media.