RCO-3 and COL-26 form an external-to-internal module that regulates the dual-affinity glucose transport system in Neurospora crassa

Background Low- and high-affinity glucose transport system is a conserved strategy of microorganism to cope with environmental glucose fluctuation for their growth and competitiveness. In Neurospora crassa, the dual-affinity glucose transport system consists of a low-affinity glucose transporter GLT-1 and two high-affinity glucose transporters HGT-1/HGT-2, which play diverse roles in glucose transport, carbon metabolism, and cellulase expression regulation. However, the regulation of this dual-transporter system in response to environmental glucose fluctuation is not yet clear. Results In this study, we report that a regulation module consisting of a downstream transcription factor COL-26 and an upstream non-transporting glucose sensor RCO-3 regulates the dual-affinity glucose transport system in N. crassa. COL-26 directly binds to the promoter regions of glt-1, hgt-1, and hgt-2, whereas RCO-3 is an upstream factor of the module whose deletion mutant resembles the Δcol-26 mutant phenotypically. Transcriptional profiling analysis revealed that Δcol-26 and Δrco-3 mutants had similar transcriptional profiles, and both mutants had impaired response to a glucose gradient. We also showed that the AMP-activated protein kinase (AMPK) complex is involved in regulation of the glucose transporters. AMPK is required for repression of glt-1 expression in starvation conditions by inhibiting the activity of RCO-3. Conclusions RCO-3 and COL-26 form an external-to-internal module that regulates the glucose dual-affinity transport system. Transcription factor COL-26 was identified as the key regulator. AMPK was also involved in the regulation of the dual-transporter system. Our findings provide novel insight into the molecular basis of glucose uptake and signaling in filamentous fungi, which may aid in the rational design of fungal strains for industrial purposes.

The signaling pathway required to adapt to nutrient limitation and to use alternative carbon sources centers on the kinase Snf1p, a component of the S. cerevisiae SNF1 complex [22,23]. The SNF1 complex is homologous to AMP-activated protein kinase (AMPK) in higher eukaryotes, which acts as a regulator of cellular energy homeostasis [24]. Like AMPK, SNF1 is a heterotrimer consisting of a catalytic α-subunit (Snf1p), a regulatory γ-subunit (Snf4p), and one of three β-subunits (Sip1p, Sip2p, or Gal83p) [25,26]. In low glucose conditions, Snf1p is activated by the phosphorylation of residue Thr210 by one of three upstream kinases (Sak1p, Tos3p, or Elm1p) [27][28][29]. In response to high glucose concentrations, Snf1p is inactivated by dephosphorylation of Thr210 by phosphatase Glc7p [30]. ADP association with Snf4p protects Thr210 of Snf1p from dephosphorylation [31]. Active Snf1p phosphorylates the inhibitor Mig1p, the homolog of CreA/CRE1 in filamentous fungi, to relieve CCR through its translocation to the cytoplasm [32], promoting the use of alternative carbon sources such as ethanol and acetate [33]. Snf1p also phosphorylates Cyr1p and negatively regulates PKA-dependent transcription [34], whereas PKA phosphorylates the Snf1-activating kinase Sak1p [35] and the β-subunit Sip1p [36], suggesting a crosstalk between the Snf1p and PKA pathways. Snf1p/Mig1p, Rgt2p/Snf3p, and cAMP/ PKA pathways, as well as their crosstalk, enable yeast cells to sense a wide range of glucose concentrations and subsequently express corresponding transporters such as low-affinity glucose transporters Hxt1p and Hxt3p or high-affinity glucose transporters Hxt6p and Hxt7p [37,38].
However, in contrast to the comprehensive analysis of hexose transport in S. cerevisiae, much less is known about glucose sensing and the subsequent transcriptional regulation of glucose transporter-encoding genes in filamentous fungi. In N. crassa, the S. cerevisiae Gpr1p ortholog GPR-4 functions as a carbon source receptor. Ligand binding to GPR-4 activates the downstream Gα protein GNA-1, leading to an increase in cAMP level produced by the activated adenylate cyclase CR-1 [63]. Similarly, in A. nidulans, the G-protein-coupled receptor GprH and Gα subunit GanB have been shown to be involved in glucose sensing [64,65]. During early conidial germination events, GanB activates cAMP synthesis and subsequent PKA activity in the presence of glucose [64]. In Ustilago maydis, the S. cerevisiae Rgt2p/ Snf3p ortholog Hxt1 was characterized as a high-affinity glucose transporter and sensor involved in glucose signaling [66]. Similarly in N. crassa, the yeast Rgt2p/Snf3p ortholog RCO-3 was shown to be involved in glucose sensing [67]. Mutation of this non-transporting glucose sensor leads to complete dysfunction of the low-affinity glucose transport system and partial impairment of the high-affinity system [67]. However, orthologs of S. cerevisiae Mth1p, Sth1p, and Rgt1p have not been found in filamentous fungi, suggesting the presence of a different signaling pathway from the S. cerevisiae Rgt2p/Snf3p pathway.
In this study, we identified a module including transcription factor COL-26, the ortholog of the Zn(II) 2 Cys 6 transcription factor AmyR in Aspergillus [68,69], and RCO-3, a non-transporting glucose sensor, that regulate the dual-affinity glucose transport system in the model fungus N. crassa. COL-26 binds to the promoter regions of glt-1 and hgt-1/-2 to promote their expression, and RCO-3 is also essential for glt-1 expression in both glucose-rich and starvation conditions, and regulates the pathway, possibly by indirectly affecting the expression of COL-26. Transcriptomic analysis showed that the response of Δcol-26 and Δrco-3 mutants to a glucose gradient was greatly impaired. In addition, AMPK is also involved in the pathway by inhibiting the activity of RCO-3 in starvation conditions. Since the dual-affinity glucose transport system is widely conserved among fungal species, knowledge about its regulation will provide a foundation for further investigation into the molecular basis of nutrient transport and signaling as well as plant cell wall degradation in fungi.

Identification of COL-26 as a key transcription factor to promote glt-1 expression in response to external glucose
To search for transcription factors essential for expression of glt-1, deletion mutants of 36 transcription factors of N. crassa, based on their expression in adequate glucose conditions (Additional file 1: Table S1) [51], were chosen and screened through batch culture with glucose as the sole carbon source followed by qRT-PCR assay. Most of the mutants showed no significant change in glt-1 expression compared with the WT strain. However, the Δcol-26 mutant showed dramatically decreased expression of glt-1 (Fig. 1a). COL-26 is a zinc binuclear cluster [Zn(II) 2 Cys 6 ] DNA-binding protein that is essential for starch utilization [69]. The abnormal expression of hgt-1/-2 and glt-1 in Δcol-26 mutant compared with WT strain (Fig. 1b) suggested that COL-26 is involved in regulating the dual-affinity glucose transport system in Fig. 1 COL-26 is a key transcription factor that regulates the dual-affinity glucose transport system in N. crassa. a Relative expression level of glt-1 in 36 transcription factor mutants of N. crassa grown in Vogel's minimal medium (VMM) supplemented with glucose as the sole carbon source. Detailed information about these transcription factors is shown in Additional file 1: Table S1. b Relative expression of glt-1 and hgt-1/-2 in the presence of glucose and in carbon-free conditions in the wild-type (WT) strain and Δcol-26 mutant. c Subcellular location of COL-26 in N. crassa exposed to a gradient of glucose. Scale bar represents 10 µm N. crassa. The nuclear localization of COL-26 was independent of glucose concentration (Fig. 1c).
To further investigate whether COL-26 directly regulates glt-1 and hgt-1/-2, EMSAs were performed. GSTfused DNA-binding domain of COL-26 was expressed and purified from E. coli (Additional file 2: Figure S1). EMSA results showed that the recombinant COL-26 bound to the promoter regions of all three genes in a typical protein concentration-dependent manner (Fig. 2a-c).

Δcol-26 mutant phenotypically and transcriptionally resembles Δrco-3 mutant
In N. crassa, RCO-3 acts as a non-transporting glucose sensor [67]. Expression of glt-1 is dramatically downregulated in a Δrco-3 mutant, whereas the expression of hgt-1/-2 is significantly upregulated in glucose-rich conditions [51], which is similar to the Δcol-26 mutant (Fig. 1a, b). In addition, both mutants are defective in glucose uptake and biomass accumulation in the presence of high glucose levels, and are resistant to 2-deoxyglucose (2-DG, which cannot be catalyzed during glycolysis and is a drug often used for glucose repression analysis in filamentous fungi) [4,67]. Also, both mutants were sensitive to osmotic stress and H 2 O 2 -induced oxidative stress compared with the WT strain, as shown by plate growth assays (Additional file 3: Figure S2) and the corresponding mycelial diameter of the Δcol-26 and Δrco-3 mutants (Fig. 3a). We assumed that COL-26 and RCO-3 are probably in a common regulatory cascade, in which membrane-located RCO-3 transduces a glucose signal to nuclear-located COL-26 in the presence of glucose.
To test this hypothesis and also obtain a broad view of the mode of expression, we conducted high-throughput sequencing (RNA-Seq) of wild-type, Δcol-26, and Δrco-3 mycelia exposed to a gradient of glucose (0, 0.05, 0.5, 2.0%) for 1 h. Pearson and Spearman correlation analysis demonstrated that the biological replicates were reliable for all tested samples (Additional file 4: Figure S3a). RNA-Seq data (Additional file 1: Table S2) from the WT, Δcol-26, and Δrco-3 biological replicates were subjected to principal component analysis and data from the same strain grown in the same growth conditions clustered together. Compared with the WT strain, data from the Δcol-26 mutant and the Δrco-3 mutant exposed to glucose (0.05, 0.5 and 2.0%) clustered together (Fig. 3b). This indicated that both mutants had impaired transcriptomic responses to 0.05, 0.5, and 2% glucose and had similar expression profiles, which was in accordance with sample-to-sample clustering (Additional file 4: Figure S3b). Consistent with these observations, the number of differentially expressed genes (DEGs) in Δrco-3 vs. Δcol-26 was much lower than that in Δrco-3 vs. WT and Δcol-26 vs. WT exposed to glucose ( Fig. 3c and e, Additional file 1: Table S3). In addition, the number of DEGs in the Δrco-3 mutant and Δcol-26 mutant comparing 0.05% glucose with 0.5% glucose was dramatically lower than the number in the WT (Fig. 3d, Additional file 1: Table S3). Further investigation showed that in the presence of glucose, rco-3 and col-26 regulate a large proportion of DEGs in common as described above, whereas there were far fewer DEGs in Δrco-3 vs. WT than in Δcol-26 vs. WT in carbon-free conditions, or in Δrco-3 vs. WT in glucose conditions (Fig. 3e, Additional file 1: Table S3). This indicated that COL-26 functions in both glucose and carbonfree conditions, whereas RCO-3 mainly functions in the presence of glucose.
Next, the effect of COL-26 and RCO-3 on the sugar uptake system was investigated. Among the 39 putative sugar transporters present in the genome of N. crassa [70], 26 showed robust expression levels (FPKM > 20) in at least one condition (Additional file 5: Figure S4). In the WT strain, the transcriptional responses of these sugar transporters to a glucose gradient were in good accordance with previously published data [51]. Some genes displayed a strong or moderate response to external glucose changes, including glt-1, hgt-1/-2, xyt-1, cdt-1/-2, NCU05897, NCU00821, xat-1, lat-1, gat-1, clp-1, and NCU09287. However, the responses of these transporter genes to a glucose gradient (2, 0.5, and 0.05%) were impaired in Δrco-3 and Δcol-26 mutants (Additional file 5: Figure S4). This is consistent with the observations that both mutants had similar transcriptional profiles and impaired transcriptomic response to glucose fluctuation ( Fig. 3b-e). As for glt-1, hgt-1/-2, and xyt-1, which displayed the strongest response to external glucose changes in the WT strain, their changed expression was probably due to the absence of COL-26 in the Δcol-26 mutant or an inactivated form of COL-26 in the Δrco-3 mutant. Significantly downregulated expression of glt-1 was observed in both the Δcol-26 and Δrco-3 mutants at all glucose concentrations, even though the expression level of col-26 in the Δrco-3 mutant was almost three times that in the WT strain in the presence of 2 and 0.5% glucose (Additional file 6: Figure S5 and Additional file 1: Table S2), indicating that both genes are essential for glt-1 expression and that RCO-3 acts upstream of COL-26. Notably, hgt-1 and hgt-2 had very similar expression profiles (Additional file 5: Figure S4), indicating the synergetic regulation of the two major components of the high-affinity glucose transport system [51].

Phosphoproteome of the WT strain grown on glucose vs. starvation conditions
Given that expression of col-26 at the transcriptional level is not affected by a gradient of glucose (0-10% w/v) [51], we wondered if a post-translational modification, such as the phosphorylation level of COL-26, explains the significant differentially expression of glt-1 between carbon-rich (2% glucose) and starvation (no-carbon, NC) conditions. Phosphoproteome profiling of the WT strain grown on glucose (Glu) compared with starvation conditions was performed. The coefficient of variation showed that the phosphopeptide abundance correlated well between the three replicates in each condition (Additional file 7: Figure S6). We identified 11,992 phosphopeptides, mapped to 2508 proteins (Additional file 1: Table S4). Of these phosphopeptides, 661 (representing 360 proteins) increased in abundance, and 709 (representing 410 proteins) decreased in abundance in the NC vs. Glu comparison ( Fig. 4a and Additional file 1: Table S5).
Proteins highly dephosphorylated in the NC vs. Glu comparison were over-represented in various categories Fig. 4 Phosphoproteomic analysis of the WT strain of N. crassa grown on glucose or in no-carbon (NC) conditions for 1 h. a Differential phosphopeptides between NC and glucose (Glu) conditions. b Venn diagram of proteins whose peptides showed differential phosphorylation levels. c Gene Ontology (GO) enrichment analysis of the 317 proteins highly phosphorylated in the comparison NC vs. Glu. d GO enrichment analysis of the 367 proteins highly dephosphorylated in the comparison NC vs. Glu mainly associated with the membrane, transport, and ATP metabolism ( Fig. 4d and Additional file 1: Table S6), suggesting active metabolism in the presence of glucose. Pathway enrichment analysis of the highly dephosphorylated proteins using the Kyoto Encyclopedia of Genes and Genomes (KEGG) identified only one pathway--the MAPK signaling pathway-yeast (ko04011)-including protoperithecium-1 (PP-1), osmotic sensitive-4 (OS-4), MAPKK kinase NRC-1, an uncharacterized protein (NCU06252), WSC-1, osmotic sensitive-2 (OS-2), and mitogen-activated protein kinase-2 (MAK-2). NRC-1 (MAPKKK) and MAK-2 (MAPK) are core components of the conserved N. crassa MAK-2 pathway [72] that mediates cell fusion and activates transcription factor PP-1 required for the activation of genes that play a role during the cell fusion [73]. However, another core component, MEK-2 (MAPKK), was highly phosphorylated in the NC vs. Glu comparison (Additional file 1: Table S5). Other highly dephosphorylated proteins in dataset NC vs. Glu included a scaffold protein HAM-5 of the MAK-2 pathway, HAM-8, HAM-9, CSP-6, RCO-1, ADA-3, and PRK1, which all relate to the NRC-1/MEK-2/MAK-2 signaling pathway and are required for cell-to-cell fusion [74,75]. OS-4 (MAPKKK) and OS-2 (MAPK) are components of the hyperosmotic response (OS) MAP kinase pathway involved in carbon sensing [76]. Other highly dephosphorylated peptides in the NC vs. Glu comparison were from CK-1b, which is involved in growth and developmental processes [77]; ASD-4, which functions in ascus and ascospore development [78]; CEL-2, which is involved in fatty acid biosynthesis [79]; an actin-binding protein FIM [80]; and COL-26. Three phosphopeptides from COL-26 showed S79 and S83 phosphorylation decreased in abundance in the NC vs. Glu comparison and no other phosphorylated sites were found in COL-26 (Additional file 1: Table S5).

The function of COL-26 itself may be regulated at the protein level rather than the phosphorylation level
Previous studies have identified four phosphorylation sites (S79, S83, S674, and S676) in COL-26 [71,72,81], among which S79 and S83 were also identified in this study and showed decreased abundance in the NC vs. Glu comparison (Additional file 1: Table S5). To dissect potential functions of these phosphorylation sites in the expression of the dual-transporter system, we constructed plasmids harboring col-26-egfp without or with site-directed mutations under the control of the promoter of the glyceraldehyde-3-phosphate dehydrogenase-1 gene (gpd-1). Nuclear localization of WT COL-26 and protein with simultaneous mutations at S79 and S83 (S79A, S83A), S674 and S676 (S674A, S676A), or all four sites (S4A) (Fig. 5a), as well as their recovered biomass relative to Δcol-26 mutant on culture grown with sucrose (Fig. 5b), indicated the successful expression and correct function of these analogs. Expression of glt-1 in Δcol-26::Pgpd-col-26 (S79A, S83A), Δcol-26::Pgpdcol-26 (S674A, S676A), and Δcol-26::Pgpd-col-26 (S4A) in glucose and NC conditions was not different from that in Δcol-26::Pgpd-col-26 (WT) (Fig. 5c and 5d), suggesting that these phosphorylation sites of COL-26 are not involved in the regulation of glucose transporter expression in N. crassa.
Notably, glt-1 expression in the Δcol-26 mutant complemented with WT and mutated COL-26 grown in starvation conditions was much higher than that in the WT strain (Fig. 5d), probably because of the high expression level of col-26 driven by the gpd-1 promoter which leads to a high level of COL-26. So, we constructed the complemented strain Δcol-26::Pn-col-26 expressing col-26egfp under the control of its native promoter. The protein level of COL-26 in the presence of different concentrations of glucose was determined by western blotting using anti-GFP antibody. The protein level of COL-26 in the presence of adequate glucose (0.5 and 2%) was higher than that in starvation conditions (0.05%, or no glucose) (Fig. 6), suggesting that the function of COL-26 itself might be regulated at the protein level rather than by phosphorylation.

AMPK represses glt-1 expression, possibly by inhibiting RCO-3 activity, in starvation conditions
Previous study showed that the OS MAP kinase pathway is involved in carbon sensing [76]. Two of its components (OS-2 and OS-4) were highly phosphorylated in the WT strain in the Glu vs. NC comparison (Additional file 1: Table S5). Thus, the roles of this pathway in regulation of the dual-affinity glucose transport system were investigated by characterization of deletion mutants of os-1 and os-2, two essential components of this pathway. Expression of glt-1 and hgt-1 in Δos-1 and Δos-2 mutants showed no difference from that in the WT strain in either glucose-rich or starvation conditions. glt-1 and hgt-1/-2 expression in Δrco-3;Δos-1 and Δrco-3;Δos-2 double mutants was identical to that in the Δrco-3 mutant in both conditions (Additional file 8: Figure S7). These results indicate that os-1 and os-2 are not involved in regulation of the dual-affinity glucose transport system.
In S. cerevisiae, the SNF1 complex plays an important role in regulation of glucose transport [33,38]. AMP-activated protein kinase (AMPK) complex from higher eukaryotes is homologous to SNF1 complex and mainly functions in nutrient-limited condition to maintain cellular energy homeostasis. In N. crassa, prk-10 (the ortholog of snf1) and NCU01471 (here named as snf4) encode the α-subunit and γ-subunit of the AMPK complex, respectively. To test whether AMPK involves in glucose transport, we measured glucose uptake rates in WT, the Δprk-10, and the Δrco-3 mutants in low glucose concentration. Within the first 10 min, the Δprk-10 mutant consumed the same amount of glucose as WT strain. However, over the remaining 20 min, glucose uptake rates decreased dramatically in the Δprk-10 mutant (Fig. 7a). The Δrco-3 mutant, as expected, showed decreased glucose uptake rates all the time (Fig. 7a). To further investigate the role of AMPK in glucose transport, the effect of AMPK on expression of glt-1 was investigated. Expression of glt-1 in Δprk-10 and Δsnf4 mutants was the same as that in the WT strain in the presence of glucose, but significantly upregulated in starvation conditions. Expression of glt-1 in Δrco-3;Δprk-10 and Δrco-3;Δsnf4 double mutants was the same as that in the Δrco-3 mutant (Fig. 7b), indicating that AMPKrepressed expression of glt-1 might occur via inhibition of RCO-3 activity in starvation conditions. This conclusion was supported by transcriptomic data, which showed that rco-3 mainly functions in the presence of glucose (Fig. 3e). Notably, the lower number of DEGs in the Δrco-3 vs. WT comparison in starvation conditions compared with glucose condition (Fig. 3e) and the significantly reduced expression level of glt-1 in the Δrco-3 mutant (Fig. 7b) indicated that RCO-3 activity was not Mycelia were grown in VMM plus 2% sucrose for 16 h, then transferred to VMM with (c) or without (d) 2% glucose. After additional cultivation for 1 h, mycelia were harvested and mRNA was extracted, after which quantitative real-time qPCR was performed to determine glt-1 expression. n.s. not significant; ***P < 0.001 totally inhibited in starvation condition. Though deletion of prk-10 or snf4 had no effect on glt-1 expression in glucose-rich conditions, expression of hgt-1 was upregulated when the Δprk-10 and Δsnf4 mutants were grown on glucose (Fig. 7c). Besides, deletion of prk-10 or snf4 in the Δrco-3 background further decreased hgt-1 expression in starvation conditions (Fig. 7c), indicating that other regulatory component(s) are also involved in regulation of hgt-1 expression.

The glucose transport system shows conserved regulation by COL-26-like transcription factors in ascomycete species
Despite some minor differences in use of some kinds of sugars, conserved roles of COL-26/AmyR homologs have been reported in various fungi, including Magnaporthe oryzae [82], Fusarium graminearum and F. verticillioides [83], A. nidulans [84], A. oryzae [68], A. niger [85], T. reesei [86], Penicillium oxalicum [87], Talaromyces pinophilus [88], and Myceliophthora thermophila [89]. In P. oxalicum, the N. crassa HGT-1 ortholog PDE_03475 showed a high expression level on cellulose and a decreased expression level in a ΔamyR mutant compared with the WT strain [87]. In A. niger, the A. nidulans MstE ortholog An02g03540 showed a high expression level on glucose and maltose, and its expression in a ΔamyR mutant was significantly downregulated [90]. To test the hypothesis that the dual-affinity glucose transport system and its regulation by COL-26 homologs are conserved in filamentous fungi, the effect of deletion of M. thermophila AmyR, the closest homolog to N. crassa COL-26, on expression of the putative dual-affinity glucose transport system was investigated. The M. thermophila ΔamyR mutant exhibited significantly reduced growth on glucose, fructose, sucrose, maltose, trehalose, xylose, and soluble starch, but grew well on cellobiose and cellulose [89], which is similar to the N. crassa Δcol-26 mutant. Alignment showed that Mycth_112491 (named as MtGLT-1-1) is the closest ortholog of N. crassa GLT-1. However, MtGLT-1-1 is only 352 amino acids long with six transmembrane helices (TMHs), compared with the typical 12 TMHs for glucose transporters predicted by TMHMM Server v2.0. The second closest GLT-1 ortholog is Mycth_108924 (named as MtGLT-1-2), which has 12 predicted TMHs. Both MtGlt-1-1 and MtGlt-1-2 showed an elevated expression level on glucose compared with NC and cellulose conditions [89,91]. Mycth_2308157 (named as MtHGT-1) and Mycth_2295230 (named as MtHGT-2) are the closest orthologs of N. crassa HGT-1 and HGT-2, respectively. Both MtHgt-1 and MtHgt-2 showed increased expression on cellulose compared with glucose [91], while MtHgt-2 also showed significantly higher expression  AMPK complex involves in regulation of glucose dual-affinity transport system. a Glucose uptake of WT, Δprk-10, and Δrco-3 mutants in low glucose concentration. **P < 0.01. b Relative expression levels of glt-1 in WT, Δprk-10, Δsnf4, Δrco-3, Δrco-3;Δprk-10, and Δrco-3;Δsnf4 strains of N. crassa in glucose and no-carbon conditions. c Relative expression levels of hgt-1 in WT, Δprk-10, Δsnf4, Δrco-3, Δrco-3;Δprk-10, and Δrco-3;Δsnf4 strains in glucose and no-carbon conditions. Mycelia were grown in VMM supplemented with 2% sucrose for 16 h, then transferred to VMM with or without 2% glucose. After additional cultivation for 1 h, mycelia were harvested and the glt-1 expression level was determined by qRT-PCR in NC conditions than in glucose-rich conditions [89], consistent with the expression pattern of high-affinity glucose transporters. These data were supported by qRT-PCR (Additional file 9: Figure S8a). We determined the expression levels of these glucose transporter-encoding genes in a ΔamyR mutant of M. thermophila. In glucoserich conditions, both MtGlt-1-1 and MtGlt-1-2 showed significantly decreased expression, whereas MtHgt-2 was upregulated in the ΔamyR mutant, compared with the WT strain (Fig. 8a). In NC conditions, MtGlt-1-2 and MtHgt-1 showed decreased expression levels in the ΔamyR mutant compared with WT strain (Fig. 8b). This indicates that AmyR is a key transcription factor involved in regulation of the dual-affinity glucose transport system in M. thermophila, like COL-26 in N. crassa. In addition, like the Δcol-26 mutant of N. crassa, the ΔamyR mutant of M. thermophila was sensitive to osmotic stress and H 2 O 2 -induced oxidative stress compared with the WT strain, as shown by plate growth assays (Additional file 9: Figure S8b) and the corresponding mycelial diameter (Fig. 8c). Since COL-26 orthologs and the dual-affinity glucose transporter system are ubiquitous in ascomycete species based on phylogenetic analysis [51,69], the regulatory role of col-26 orthologs may also be conserved in many other filamentous fungal species.

Discussion
Glucose uptake is the first and rate-limiting step of glucose metabolism. To cope with environmental changes in glucose availability, fungi express low-affinity glucose transporters when glucose levels are high and high-affinity glucose transporters when glucose levels are low. The N. crassa dual-affinity glucose transport system is the best-studied example, and consists of a low-affinity glucose transporter GLT-1 and two high-affinity glucose transporters HGT-1 and HGT-2 [51]. This dual-transporter system is conserved in filamentous fungi and many glucose transporters have been characterized. However, the mechanism underpinning the regulation of the dualaffinity glucose transport system in response to environmental changes remains elusive. Here, using N. crassa as a model fungus, we identified a glucose signaling pathway consisting of multiple components, including the major transcription factor COL-26, non-transporting glucose sensor RCO-3, and a cellular energy sensor AMPK. COL-26 regulates the dual-affinity glucose transport system at the transcriptional level. Deletion of col-26 or rco-3 leads to a significantly reduced expression level of glt-1 in both glucose-rich and starvation conditions (Figs. 1a and 7b), indicating that the basal level of glt-1 in starvation conditions is also maintained by COL-26 and RCO-3. Interestingly, we found that COL-26-dependent regulation of the dual-transporter system might depend on the regulation of its protein homeostasis, rather than on posttranslational modification such as phosphorylation. The expression level of col-26 in the WT strain remains the same in glucose-rich and carbon-free conditions, and it is only induced by starch, trehalose and maltose [51,69] (Additional file 1: Table S2). This could reduce the transcriptional and translational investment when N. crassa growing in nutrition-deficient conditions encounters nutrient-rich conditions. Upon nutrition shift, the basal level glt-1 mRNA can be immediately translated, and there is no need to synthesize col-26 mRNA de novo for translation. Once COL-26 is synthesized from existing mRNA, it promotes expression of glt-1 in a COL-26 concentration-dependent manner at the protein level ( Fig. 6) rather than in a phosphorylation-dependent manner (Fig. 5). However, overexpression of col-26 did not further promote expression of glt-1 in the presence Expression of glucose transporter genes in no-carbon conditions. *P < 0.05; ***P < 0.001. c Mycelial diameter of the WT strain and ΔamyR mutant grown on VMM plus 2% cellobiose at 37 °C for 4 days with or without 0.5 M NaCl or 1 mM H 2 O 2 of glucose (Fig. 5c), indicating that COL-26 does not function independently. Other components of the pathway that control COL-26 synthesis/degradation or dissociation/association with other protein(s) are worth investigating in the future. Moreover, as only four phosphorylation sites (S79, S83, S674, and S676) have so far been identified in COL-26 by various studies [71,72,81] (Additional file 1: Table S4), whether other phosphorylation site(s) are involved in the regulation of the dualaffinity glucose transport system requires investigation.
The detailed mechanism of regulation of glucose dualaffinity transport system by the key transcription factor COL-26 under different glucose conditions needs to be further investigated. Some explanation could be drawn from a previous study which showed that the absence of GLT-1 leads to the upregulated expression of hgt-1/-2 in high glucose condition [51]. This indicates that the upregulated expression of hgt-1/-2 in Δglt-1 mutant was due to relief from CRE-1-mediated CCR which is caused by its inability of transporting high amount of glucose.
Since glt-1 expression was abolished in Δcol-26 and Δrco-3 mutant, the upregulated expression of hgt-1/-2 in Δcol-26 and Δrco-3 mutant in response to glucose was probably also caused by relief from CCR, similar to the Δglt-1 mutant.
The high expression level of hgt-1/-2 in glucose-limited conditions is dependent on COL-26 (Fig. 1b), however, expression of hgt-1/-2 in the presence of adequate glucose was repressed by CRE-1-mediated CCR [51], even though COL-26 remains at a relatively high level (Fig. 6), indicating that CRE-1 could be antagonistic to COL-26. Therefore, the high expression level of hgt-1/-2 in glucose-limited conditions is due to a combination of derepression from CRE-1 and activation by COL-26 (Fig. 9). Taken together, we proposed that COL-26 and CRE-1 probably binds to the same promoter regions of hgt-1/-2. In WT strain under glucose-rich condition, CRE-1 binds to promoter regions of hgt-1/-2 to repress their expression. Deletion of col-26 leads to abolishment of GLT-1 and subsequent impaired glucose transport, leading to Fig. 9 Model for the role of COL-26 in regulating the dual-affinity glucose transport system in filamentous fungi. Under high levels of glucose, RCO-3 transduces a glucose signal to COL-26 to promote expression of glt-1 for nutrient assimilation. Meanwhile, adequate glucose stimulates CRE-1-mediated carbon catabolite repression (CCR) to repress expression of hgt-1/-2. When external glucose is depleted or limited, hgt-1 and hgt-2 are rapidly derepressed by the lifting of CCR and their expression is activated by COL-26. This process synergistically activates AMP-activated protein kinase (AMPK), which inhibits activity of RCO-3, leading to low expression of glt-1 relief from CCR and disassociation of CRE-1 from their promoters, thus derepressing the expression of hgt-1/-2; whereas, under starvation condition, after disassociation of CRE-1 from hgt-1/-2 promoters, COL-26 might bind to the same region to activate their expression. When col-26 was deleted, this activation was absent, thus leading to the downregulation of hgt-1/-2 (Fig. 1b). The hypothesis needs to be investigated through further studies.
Another finding of this study is that RCO-3 is deeply involved in the regulation of the glucose transport system, since the expression of low-affinity glucose transporter gene glt-1 under starvation was repressed by AMPK through partially inhibiting the activity of RCO-3 (Fig. 7b). The similarity of multiple phenotypes between the Δcol-26 mutant and Δrco-3 mutant indicates that RCO-3 and COL-26 are in the same signaling pathway, in which RCO-3 acts as a glucose sensor and transduces a glucose signal to nuclear COL-26. Supporting this, the expression level of col-26 in the Δrco-3 mutant was almost three times that in the WT strain in the presence of a high glucose concentration (Additional file 6: Figure  S5 and Additional file 1: Table S2), but expression of glt-1 was still significantly downregulated due to the absence of RCO-3 (Fig. 7b). Besides, components downstream of RCO-3 and upstream of COL-26, but not yet identified, probably regulate not only the dual-affinity glucose transport system, but also other biological processes. Identification of the missing part(s) of the RCO-3/COL-26 signaling pathway will help us to understand the architecture of nutrient signaling regulation in filamentous fungi.

Conclusions
In this study, we identified the crucial transcription factor COL-26 that regulates the glucose dual-affinity transport system by directly binding their promoter region in N. crassa. COL-26 is downstream element of the non-transporting glucose sensor RCO-3. In addition, the AMPK complex is required to suppress the expression of lowaffinity glucose transport system in starvation condition by partially inhibiting the activity of RCO-3. Knowledge about regulatory mechanism of the dual-transporter system will not only deepen our understanding of the environmental adaptability of filamentous fungi, but also aid in the rational design of fungal strains for industrial purposes.

Culture conditions
E. coli was grown at 37 °C in Luria-Bertani medium supplemented with 100 µg ml −1 kanamycin or ampicillin when necessary. M. thermophila strains were cultured on Vogel's minimal medium (VMM) [93] supplemented with 2% sucrose at 45 °C for 7-10 days to obtain conidia. N. crassa strains were inoculated on slants containing 3 mL VMM with 2.0% (w/v) sucrose as the sole carbon source and grown at 28 °C in the dark for 2 days, then at room temperature in constant light for 6-10 days to stimulate conidia production. Conidia were inoculated into 100 mL liquid VMM with various carbon sources at 10 6 conidia·mL −1 and grown at 25 °C in constant light with shaking (200 rpm). For plate growth assays, 1 μL of conidia suspension (1 × 10 6 conidia·mL −1 ) was plated on VMM supplemented with 2% cellobiose and cultured at 28 °C for 30 h for N. crassa, or 37 °C for 4 days for M. thermophila. NaCl and H 2 O 2 were added to media to a final concentration of 0.5 M and 2 mM, respectively, for N. crassa. For M. thermophila, H 2 O 2 was added to a final concentration of 1 mM.

Medium shift experiments
Conidia were inoculated into 100 mL liquid VMM supplemented with 2.0% sucrose and grown at 28 °C and 200 rpm in constant light for 16 h. The mycelial biomass was washed with sterilized water at least five times and then transferred to 100 mL VMM with 2% glucose or with no carbon source added for 1 h before RNA extraction.

Plasmid construction and transformation
A fragment containing 5′-and 3′-flanking regions of csr-1 was amplified with pLC-5-F/pLC-3-R from pCSR1 [92]. The promoter of gpd-1 was amplified from genomic DNA using primers Pgpd-NC-F/Pgpd-NC-R. The open reading frame of col-26 was amplified using Col26-ORF-F/Col26-ORF-R from cDNA which was synthesized from total RNA using a ReverTra Ace qPCR RT Kit (Toyobo, Japan). The coding sequence of gfp was amplified using primers GFP-F/GFP-R from pMF272. These four fragments were assembled using the NEB Gibson Assembly Kit (New England Biolabs, USA) to give pCSR-COL-26-GFP. The variants col-26 (S79A, S83A), col-26 (S674A, S676A), and col-26 (S4A) were generated by site-directed mutagenesis using PCR with high-fidelity polymerase. Transformation by electroporation was performed as described previously [92]. Transformants resistant to cyclosporin A were further confirmed by PCR and green fluorescent protein (GFP) fluorescence.

Quantitative real-time qPCR (qRT-PCR)
qRT-PCR was performed as previously described [94]. The actin gene (act) was used as an endogenous control for N. crassa, and MtAct was used as an endogenous control for M. thermophila. All primers used in this study are listed in Additional file 1: Table S7. The transcript level of each gene was estimated using the 2 −ΔΔCt method [95]. The ratio of each gene transcript in each mutant to that in the WT strain was calculated as the relative transcript level.

RNA sequencing and transcription expression analysis
After harvesting via vacuum filtration, mycelia were immediately homogenized in liquid nitrogen for total RNA extraction. Total RNA was isolated from frozen samples with Trizol reagent (Invitrogen, Carlsbad, CA, USA), treated with DNase I, and purified using a Qiagen RNeasy Mini Kit (Qiagen, Hilden, Germany). RNA integrity was checked by agarose gel electrophoresis and using an Agilent Bioanalyzer 2100 system (Agilent Technologies, Santa Clara, CA, USA). Qualified RNA with OD 260 / OD 280 > 2.0 and RNA Integrity Number > 8.0 was used for RNA-Seq, which was performed using the BGISEQ-500 platform at Beijing Genomics Institute (BGI) (Shenzhen, China). All data were generated by sequencing two independent duplicate samples. Prior to read mapping, adaptors and low-quality reads were trimmed using Trimmomatic v0.36 [96]. Filtered clean reads were aligned against predicted transcripts from the N. crassa OR74A genome v12 [97] using Bowtie2 v2.2.5 [98]. The read counts were determined using RSEM v1.2.8 [99]. The abundance of each transcript was calculated from fragments per kilobase of transcript per million mapped reads (FPKM) values (Additional file 1: Table S2). We used fuzzy c-means clustering to group genes based on similarity between concentration-specific gene expression patterns. Fuzzy clustering was conducted using Mfuzz v2.34.0 [100]. Differential gene expression analysis was performed using the DESeq package (v1.5.1). Genes with fold-change > 2.0 (|log2 ratio|≥ 1) and DESeq P adj -value (Q-value) < 0.001 were considered significantly differentially expressed between different conditions or strains. To discover significantly up-and downregulated genes, only genes with relatively high transcript abundance (FPKM-value > 20 in at least one strain) were considered for further analysis. RNA-Seq data are available at the Gene Expression Omnibus under accession number GSE157186.

Protein gel electrophoresis
Culture supernatants were mixed with 4 × SDS loading buffer and boiled for 10 min before loading onto Criterion 4-15% Tris-HCl Precast Gels (Bio-Rad). GelCode Blue Stain Reagent (Thermo Scientific) was used for gel staining.

Expression and purification of DNA-binding domain of COL-26
A fragment encoding the DNA-binding domain of COL-26 was amplified with primers Ecol26-F/Ecol26-R using cDNA as template. After digestion with BamHI and XhoI, this fragment was ligated into the corresponding sites of pGEX-4T-1 (GE Healthcare) to give pGEX-col-26. The plasmid was subsequently introduced into E. coli BL21 (DE3) for protein expression. E. coli BL21 (DE3) harboring pGEX-col-26 was grown at 37 °C in 100 mL LB medium supplemented with 100 μg/mL ampicillin to an OD 600 of 0.6. Isopropyl β-D-1-thiogalactopyranoside (IPTG) was then added to a final concentration of 0.5 mM, and the cultures were incubated for an additional 3 h at 37 °C. The cells were harvested by centrifugation and resuspended in phosphate-buffered saline followed by sonication, after which the insoluble material was removed by centrifugation at 8000 g for 10 min. The glutathione S-transferase (GST)-fused protein was purified using a BeaverBeads ™ GST-tag Protein Purification Kit (Beaver, China) according to the manufacturer's manual. Protein purity was determined by Coomassie Blue staining after 12% SDS-PAGE, and protein concentration was measured by BCA assay (Thermo Scientific, Waltham, MA, US).  Table S7. The PCR products were purified by electrophoresis and quantified using a NanoDrop 2000c Spectrophotometer (Thermo Fisher Scientific). The subsequent binding experiments were performed using a modified gel mobility shift assay as described previously [101]. In each EMSA, different quantities of recombinant protein were incubated with a constant amount (10 ng) of the DNA probes individually at 25 °C for 30 min. The experiments were performed at least three times. The information of genes mentioned in this study are shown in Additional file 1: Table S8.

Phosphopeptide identification by mass-spectrometry (MS)-based analysis
Proteins were reduced with 10 mM dithiothreitol for 1 h at 56 °C and subsequently alkylated with 55 mM iodoacetamide. Samples were digested with trypsin at 1:20 enzyme-to-substrate ratio. Digested samples were desalted using C18 solid phase extraction tubes. The resulting peptide samples were concentrated and a BCA assay was performed to determine the peptide concentration and samples were diluted with nanopure water for MS analysis. Desalted peptides were labeled with 8-plex iTRAQ reagents (AB SCIEX) according to the manufacturer's instructions. Peptide aliquots for each sample (200 mg) were dried for TiO 2 enrichment, and used for phosphoproteome analysis; TiO 2 enrichment of phosphopeptides followed a previously established protocol [102]. Phosphopeptide samples were analyzed using a nanoESI system (Waters NanoAcquity LC, Waters Corporation) coupled to a Q Exactive HF mass spectrometer (Thermo Fisher Scientific). Proteomics data were analyzed using a combination of Proteome Discoverer (v1.4) and Mascot (v2.3) software. The MS results were filtered based on a 5% false discovery rate and phosphoRS probability ≥ 0.75. Phosphopeptide abundance changes >1.5fold were considered and subjected to downstream analysis.

Western blot analysis
Western blotting was performed as previously described [103]. Anti-GFP or anti-actin rabbit antibody and antirabbit IgG horseradish peroxidase-conjugated antibody at a dilution of 1:3000 were used as the primary and secondary antibodies (Abmart, Shanghai, China), respectively.

Microscopy and imaging
To localize GFP fusion proteins using microscopy, all strains were inoculated into liquid VMM supplemented with 2% sucrose and grown for 16 h at 25 °C. The hyphae were harvested, washed several times with Vogel's salts, transferred into media containing different concentrations of glucose, and cultured for another 1 h. Microscopic observations were performed using an Olympus BX51 fluorescence microscopy system and images were processed using ImageJ software.

Glucose uptake assays
Conidia from 10-day-old cultures were inoculated into 100 mL VMM with 2% sucrose as the carbon source. After grown at 25 °C in constant light with shaking (200 rpm) for 16 h, cultures were centrifuged at 2000 g for 10 min and washed in VMM without a carbon source, followed by 1-h growth in 100 mL VMM with 0.05% glucose as carbon source. Then 0.05 g glucose was added to each culture. Culture supernatants were sampled at 0, 10, and 30 min. Glucose levels were measured using HPLC with an e2695 instrument (Waters, Manchester, United Kingdom). mycelia were harvested and gene expression levels were determined by qRT-PCR.
Additional file 9: Figure S8. Expression levels of glucose transporter genes in M. thermophila. a Relative expression levels of glucose transporter genes MtGlt-1-1, MtGlt-1-2, MtHgt-2, and MtHgt-1 in glucose-rich and no-carbon conditions. Mycelia were grown in VMM supplemented with 2% sucrose for 16 h, then transferred to VMM with or without 2% glucose. After additional cultivation for 1 h, mycelia were harvested and gene expression levels were determined by qRT-PCR. **, P < 0.01; ***, P < 0.001. b Effect of stress on growth of WT and ΔamyR strains of M. thermophila. VMM medium with 2% (w/v) cellobiose was used. NaCl and H 2 O 2 were added to the medium to a final concentration of 0.5 M and 1 mM, respectively. Plates were incubated at 37 °C for 4 days before imaging.