All the strains used in this study were listed in Table 1. Fungal strains were routinely grown on potato dextrose agar (PDA) or Vogel's medium agar (Vogel, 1956), which were supplemented or modified as indicated. For conidial production, PDA plates were incubated at 30°C for 5 days and aerial conidia were harvested by flooding the plate with sterile distilled H₂O containing 0.02% Tween 80. Conidial concentrations were determined by direct count using a hemocytometer.

Standard techniques were used for nucleic acid manipulations. An overlap PCR method (Davidson et al., 2002) was used to create the fragments for targeted gene deletion. P. oxalicum 114-2 genomic DNA was used as the template for the PCR amplification of the 5′- and 3′-flanking regions of the laeA. The plasmid pME2892 (Krappmann et al., 2005) was used as the template for the PCR amplification of resistant gene pyrithiamine (ptrA). Primers UlaeA-F/UlaeA-R, ptrA-F/ptrA-R, and DlaeA-F/DlaeA-R were usded to amplify the 5′ homologous flanking region (2296 bp), ptrA gene (2008 bp), and 3′ homologous flanking region (3072 bp) of laeA, respectively. The three resulting PCR fragments were fused and amplified by PCR using the nested primers ClaeA-F/ClaeA-R. The PCR products were directly transformed into OEclrB or OExlnR to obtain OEclrBΔlaeA or OExlnRΔlaeA. The primers hyg-F and hyg-R were used to amplify the hph gene (a hygromycin B phosphotransferase encoding gene (hphG, 2425 bp) from plasmid pSilent1 (Nakayashiki et al., 2005). The primers gpdA-F and gpdA-LaeA-R were used to amplify the glyceraldehydes-3-phosphate dehydrogenase promoter (gpdA) from A. nidulans genome (1256 bp), whereas the primers LaeA-F/LaeA-R were used to amplify the intact P. oxalicum laeA open reading frame with its terminator (2179 bp). The three resulting PCR fragments were fused and transformed into WT to obtain OEΔlaeA.

The protoplast preparation, transformation, and screening of recombinant clones were performed as previously described by Qin et al. (2013). The genomic DNA of the transformants was isolated and used as the template to verify the integration events. Primers YlaeA-1/ptrA-R and YlaeA-2/ptrA-F were used to certify the deletion of laeA and obtained 2840 bp and 2413 bp fragments, respectively. Primers Yhyg-gpdA-F/Yhyg-gpdA-R and YgpdA-F/YgpdA-laeA-R were used to certify the overexpression of laeA and obtained 2263 bp and 1299 bp fragments, respectively. Then, the mutants were further analyzed by Southern blot. Primers S-DlaeA-F/S-DlaeA-R and were used for PCR amplification to generate the probes for the examination of OEclrBΔlaeA and OExlnRΔlaeA. Primers S-OElaeA-F and S-OElaeA-R were used for PCR amplification to generate the probes for the examination of OElaeA. The probes were labeled via PCR using a PCR DIG Probe Synthesis kit (Roche). The genomics DNA were digested with the restriction endonucleases and separated from 0.75% agarose gel electrophoresis. Then the DNA was transferred to the Hybond-N + nylon membranes (Amersham Biosciences/GE Healthcare, USA). Blots were visualized using a DIG DNA labeling system following the manufacturer's protocol (Roche). The PCR verification and Southern blot results are shown in Figure S1. The primers used in this study were listed in Table S1.

One microliter conidial suspension (1 × 10⁶ conidia ml⁻¹) of each strain (1 μL hyphal fragment solution of ΔcreAΔlaeA strain was dropped onto the agar because it cannot generate conidia) were spotted onto various media on a 9-cm plate, including PDA or modified Vogel's salts agar medium (Vogel, 1956) with various carbon sources supplemented with 2% (W/V) glucose, 2% lactose, 2% glycerol, or 2% ball-milled microcrystalline cellulose at 30°C for 4 days. Canon EOS 600D (Canon, Japan) was used for photographing. To test for conidiation, 10⁶ conidia were spread on 9-cm Vogel's + glucose plates and were cultivated at 30°C for 4 days. Afterward, a 5-mm diameter colony agar plug was removed from the plate. Three replications were conducted for each treatment. Conidia were then harvested by gently rubbing them off in an equal volume of physiological salt (0.2% w/v Tween 80 and 0.8% w/v NaCl). Conidial concentrations were determined by direct count using a hemocytometer.

The strains were first grown in 100 mL liquid glucose minimal medium (GMM, per liter: glucose 20.0 g, KH₂PO₄ 3 g, (NH4)₂SO₄ 2.0 g, MgSO₄·7H₂O 0.56 g, CaCl₂ 0.56 g, FeSO₄·7H₂O 7.5 mg, MnSO₄·H₂O 2.5 mg, ZnSO₄·7H₂O 3.6 mg, CoCl₂·6H₂O 3.7 mg, CuSO₄ 3.2 mg) with initial pH 5.5 at 200 rpm for 24 h in 30°C. Then, 0.3 g vegetative mycelia was collected through vacuum filtration and was added to 100 mL of Vogel's salts medium supplemented with 20 g L⁻¹ inducing carbon sources (10.0 g L⁻¹ microcrystalline cellulose plus 10.0 g L⁻¹ wheat bran) at 30°C and 200 rpm. Samples were collected at the time points indicated in the text. Microcrystalline cellulose (CB0279) was purchased from Sangon (Shanghai, China). Culture supernatants were diluted with sodium acetate buffer solution (0.2 M, pH 4.8). The filter paper activities (FPA) of the culture supernatants were assayed against Whatman No. 1 filter paper using a DNS method (Sun et al., 2008). β-D-xylopyranoside (p-NPX) was used as the substrate for β-xylosidase activity assay. The experiment was conducted by incubating 50 μL of 0.1 mg/mL p-NPX in a 0.1 M sodium citrate buffer (pH 4.8) with 100 μL enzyme elute for 30 min at 50°C. Then, 150 μL of 10% (w/v) Na₂CO₃ was added to terminate the reaction. One enzyme activity unit was defined as the amount of enzymes required to produce 1 μmol glucose or pNP per minute under the assayed conditions. Three biological triplicates were performed in all analyses.

Unconcentrated supernatants were added to loading buffer, boiled for 5 min for degeneration, and loaded onto a 12% Tris-HCl polyacrylamide gel. Coomassie blue stain reagent was used for staining. The biomasses of all mutants were measured using the methods described by Yao et al. (2015). Briefly, the WT and mutants were pre-grown in 1 × Vogel's medium with glucose for at 30, 200 rpm for 24 h. The same quantity of dehydrated mycelia from each strain was transferred to the same freshly prepared media for another 6, 12, 24, 36, 48, 60, 72, and 96 h. All sampled mycelia were dried at 65°C to achieve constant weight, and dry cell weight was estimated.

The strains were cultivated following the method described in “Cellulase and xylosidase activity assay.” Total RNA was extracted from frozen lyophilized mycelia using the RNAiso Plus reagent (TaKaRa Biotechnology). Then, RNA was treated with DNase I, and first- and second-strand cDNA synthesis was performed using a PrimeScript RT Reagent kit with gDNA Eraser (TaKaRa Biotechnology). qRT-PCR was performed by SYBR Premix Ex Taq (Perfect Real Time, TaKaRa Biotechnology) using a LightCycler 480 system with software version 4.0 (Roche, Mannheim, Germany). The primers that were used to examine gene expression levels included act-F/act-R (for the actin gene, EPS26156.1), cel7A-F/cel7A-R (for the cel7A/cbh1 gene), cel7B-F/cel7B-R (for the cel7B/eg1 gene), cel3A-F/cel3A-R (for cel1A/bgl1 gene), amy15A-F/amy15A-R (for the amy15A gene), xln10A-F/xln10A-R (for xln10A gene), xyl3A-F/xyl3A-R (for xyl3A gene), laeA-F/laeA-R (for the laeA gene, EPS25650), creA-F/creA-R (for the creA gene, EPS28222), clrB-F/clrB-R (for the clrB gene, EPS31045), and xlnR-F/xlnR-R (for the xlnR gene, EPS32714). The primers used are shown in Table S1. Three biological triplicates were performed, and qRT-PCR of each gene was performed in three triplicates. The expression of actin was chosen as the reference gene for data normalization. The number of gene expression copies was calculated using the standard curves constructed for each gene, and the data were normalized to the expression levels of the actin gene. The relative expression level was defined as follows: Rel. expression level (gene X) = copy number of gene X/copy number of gene action. Statistical analysis was performed using Minitab, and P ≤ 0.05 were considered statistically significant. The primers used in this study were listed in Table S1.

The culture method was the same as that described in the “Cellulase and xylosidase activity assay” section. Digital gene expression profiling was performed according the methods described by Zhang et al. (2016). Briefly, after 24 h or 60 h of cultivation, total RNA was extracted from frozen mycelia using the RNAiso Plus reagent (TaKaRa Biotechnology). Oligo (dT) beads were employed to purify mRNA and guide double-stranded cDNA synthesis. Digital gene expression profiling based on Illumina sequencing was performed by the Beijing Genomics Institute in Shenzhen, China. Each tunnel generated millions of raw reads with a sequencing length of 35 bp. The 30 adaptor sequences were first removed from the sequencing reads to produce 21-nt long tags. Afterward, low quality tags, one copy tags, and tags which were not 21-nt long were removed to generate clean tags. All clean tags were obtained after raw data processing was mapped to the reference tag database (predicted transcripts plus downstream 300-nucleotide sequences) of the corresponding strain. The copy number of unambiguous tags (tags mapped to a single gene) for each gene was normalized to TPM for differential expression analysis. Genes with significantly different expression levels were identified through a significance test with combined thresholds (FDR ≤ 0.001 and fold change ≥ 2) (Audic and Claverie, 1997). Blast2GO was used for the function enrichment analysis of gene sets with the threshold at FDR ≤ 0.05 (Conesa et al., 2005).

Yeast two-hybrid assay was performed according the protocols of Clontech (Clontech Laboratories, Inc.). Briefly, pGBKT7 derived prey laeA and pGADT7 derived bait creA, clrB or xlnR were co-transformed into the Saccharomyces cerevisiae strain AH109, which has four reporter genes including lacZ, mel1, ade2 and his3. Transformants were selected on SD/-Trp/-Leu media. To further confirm the interaction between LaeA and transcription factos, the transformants were tested on the SD/-Trp/-Leu/ containing 10 mg/ml X-α-gal. And the transforms were spotted in the SD/-Trp/-Leu/-His media containing 2.5 mM 3AT.

The Whole Genome Shotgun projects were deposited in DDBJ/EMBL/GenBank under the accession number AGIH00000000. The raw data of expression profiling were deposited in NCBI's Gene Expression Omnibus (GEO) database under the accession number GSE71287.

The fungal strains used in this study are listed in Table 1. Equivalent fresh spores of the WT and mutants were inoculated on PDA plates or on 1 × Vogel's salts (Vogel, 1956) with 2% (w/v) glucose, 2% lactose, 2% glycerol, or 2% cellulose as the sole carbon source for 4 days at 30°C (Figure 1A). Then, the colony morphology and the conidiation were observed. The colonies of ΔlaeA, OEclrBΔlaeA, and OExlnRΔlaeA appeared greenish-brown compared with the dark-green colonies observed in WT, OEclrB, and OExlnR. ΔcreAΔlaeA even formed a white colony (Figure 1A). Conidiation was quantified by plating conidia on top of agar layers, and conidiospore production was subsequently analyzed. The conidiation levels in all laeA deletion mutants were dramatically reduced; the conidiation of ΔlaeA, OEclrBΔlaeA, and OExlnRΔlaeA was ~4.1, ~13.2, and ~9.0% of WT, respectively (Figure 1B). No conidium was observed in ΔcreAΔlaeA, consistent with previous report (Zhang et al., 2016). This observation indicated that LaeA positively regulated asexual development.

When the strains were grown on cellulose agar (1 × Vogel's salts plus with 2% cellulose as the sole carbon), the cellulolytic halo of all the four mutants ΔlaeA, OEclrBΔlaeA, OExlnRΔlaeA, and ΔcreAΔlaeA decreased remarkably compared with their parent strains WT, OEclrB, OExlnR, and ΔcreA, respectively. No cellulolytic halo was observed around the ΔlaeA colony, whereas a clear cellulolytic halo was found around the WT colony (Figure 2). ClrB and XlnR were transcription activators for the cellulase gene (Stricker et al., 2006; Coradetti et al., 2012); thus, it was expected their overexpression strain (OEclrB and OExlnR) showed increased cellulolytic ability with a more pronounced cellulolytic halo than that of the WT; the result was consistent with the previous report described by Li et al. (2015). When laeA was deleted, however, the overexpression of clrB (OEclrBΔlaeA) or xlnR (OExlnRΔlaeA) did not result in the formation of a cellulolytic halo (Figure 2). CreA suppressed the expression of most cellulase and hemicellulase genes through carbon catabolite repression mechanism (Nakari-Setälä et al., 2009). As a result, the deletion of creA (ΔcreA) was observed, exhibiting a large and clear cellulolytic halo around its colony compared with that of the WT. The result was consistent with the previous report described by Liu et al. (2013a). However, when laeA was absent, the diameter of the cellulolytic halo around ΔcreAΔlaeA was reduced (Figure 2).

For the agar, pure cellulose was used as the sole carbon source. Complex carbon sources from plant materials were more efficient than pure cellulose in promoting the expression of lignocellulose-degrading enzymes. Therefore, we cultivated the WT and the mutants in a submerged medium with wheat bran and microcrystalline cellulose. Then, the levels of FPA (representing overall cellulase activity) were assayed. To facilitate the comparison of FPA produced by different mutants, same scale of vertical ordinate was used in the Figure 3. It was expected that OEclrB, OExlnR, and ΔcreA showed increased FPA activities compared with WT. However, in the absence of LaeA, the FPA of all the four mutants ΔlaeA, OEclrBΔlaeA, OExlnRΔlaeA, and ΔcreAΔlaeA decreased remarkably compared with their parent strains WT, OEclrB, OExlnR, and ΔcreA, respectively, after 2 days cultivation. On the third day, the FPA of ΔlaeA, OEclrBΔlaeA, OExlnRΔlaeA, and ΔcreAΔlaeA was only 32.1, 16.4, 41.4, and 58.9%, respectively, compared with that of the their parent strain WT (Figure 3A), OEclrB (Figure 3B), OExlnR (Figure 3C), and ΔcreA (Figure 3D) strains. The FPA of OEclrBΔlaeA, OExlnRΔlaeA, and ΔcreAΔlaeA also decreased compared with that of 21.4, 56.8, and 73.2% WT, respectively. The difference between the mutants and WT increased along with prolonged culture time; the FPA of ΔlaeA, OEclrBΔlaeA, OExlnRΔlaeA, and ΔcreAΔlaeA was only 5.0%, 8.1%, 4.7%, and 8.4%, respectively, compared with that of WT on the fifth day (Figure 3).

Then, the supernatants from the WT and the mutants were profiled by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE) (Figure S2). Equal volumes of supernatants were loaded. Significantly lesser protein bands were detected in ΔlaeA, OEclrBΔlaeA, OExlnRΔlaeA, and ΔcreAΔlaeA than in WT, especially in the range of 40–116 kDa. Previous reports have shown that this range is an area for aggregation of glycoside hydrolases; in particular, most glucoamylases, cellobiohydrolases, endoglucanases, and β-glucosidases were located in the area (Liu et al., 2013b). Were the changes in cellulase synthesis in the mutants caused by their different biomass levels? The growth kinetics of WT and all the mutants were examined (Figure S3). Except that ΔcreA and ΔcreAΔlaeA exhibited lower biomass formation than WT, the similar growth kinetic curve of biomass level was observed among WT, ΔlaeA, OEclrB, OExlnR, OEclrBΔlaeA, and OExlnRΔlaeA, showing that the reduced cellulolytic activities in ΔlaeA, OEclrBΔlaeA, and OExlnRΔlaeA or increased cellulolytic enzyme activities in OEclrB and OExlnR did not correlate with the biomass level. The introduction of a wild type copy of laeA at the laeA locus (RlaeA) restored the growth defects and the cellulolytic enzyme synthesis defects of the ΔlaeA mutant (Figure S4). No evident difference was observed in the conidiation and cellulolytic enzyme synthesis in the mutant of overexpressed laeA (OElaeA) in P. oxalicum (Figure S5).

We noticed that the deletion of laeA resulted in a massive reduction of FPA, especially in prolonged batch cultures (Figure 3). To obtain a global view of the effect of laeA deletion on glycoside hydrolase expression, the expression profiles of WT and ΔlaeA were assayed. Glycoside hydrolases were characteristically produced after the shift from exponential growth to stationary phase. Therefore, two time points for gene expression profiling were selected. In particular, “24 h time point” indicates the exponential growth phase (earlier phases of cultures), whereas “60 h time point” indicates a stationary phase and exhibits a cellulase formation peak (later phases of prolonged batch cultures). WT and ΔlaeA were cultivated in cellulose and wheat bran medium, which is an optimum medium for cellulolytic enzyme formation. The mRNA from WT and ΔlaeA at the cultivation time of 24 h and 60 h, respectively, was collected and was subjected to high-throughput Illumina sequencing. More than 3 million tags were obtained for each sample. The copy number of unambiguous tags (tags mapped to one single gene) for each gene was normalized to the number of transcripts per million (TPM) clean tags. A genome-wide expression profiling analysis revealed the extensive expression of the entire P. oxalicum genome. Of the 10021 protein-coding genes in the genome database, 8192 (81.7%), 8292 (82.7%), 8351 (83.3%), and 8229 (82.1%) genes were expressed in WT-24 h, ΔlaeA-24 h, WT-60 h, and ΔlaeA-60 h, respectively. Genes of significantly differential expression levels were identified through a significance test with combined thresholds (FDR ≤ 0.001 and fold change ≥2). Blast2GO was used for function enrichment analysis of gene sets with a threshold of FDR ≤ 0.05.

The ΔlaeA mutant was compared with WT at 24 h, showing that the expression levels of 1657 genes exhibited significant differences (2-fold or greater, FDR < 0.001) between ΔlaeA and WT. Of these, 719 were regulated 4-fold or greater (FDR < 0.001), with 399 genes downregulated and 320 genes upregulated. GO enrichment analysis revealed that downregulated genes (4-fold or greater, FDR < 0.05) were mainly involved in heme binding, monooxygenase activity, catalytic activity, hydrolase activity, acting on glycosyl bonds, and binding of ACP phosphopantetheine attachment site binding involved in fatty acid biosynthetic process (Figure S6A). The upregulated genes were mainly involved in substrate-specific transmembrane transporter activity (Figure S6B). The downregulated and upregulated genes (GO category: molecular function), along with their predicted functions, are listed in Tables S2, S3.

For the ΔlaeA mutant compared with WT at 60 h, the expression levels of 2488 genes exhibited significant differences (2-fold or greater, FDR < 0.001) between ΔlaeA and WT. Of these, 1145 were regulated 4-fold or greater (FDR < 0.001), with 703 downregulated genes and 442 upregulated genes. GO enrichment analysis revealed that the downregulated genes (4-fold or greater, FDR < 0.05) in ΔlaeA compared with that in WT were mainly involved in cellulose binding, aspartic-type endopeptidase activity, cellulase activity, monooxygenase activity, heme binding, electron carrier activity, and cellulose 1,4-beta-cellobiosidase activity (Figure S6A). GO enrichment analysis revealed no statistically significant results for the upregulated genes with this cutoff. The downregulated genes (GO category: molecular function) and their predicted functions are listed in Table S4.

Of the downregulated genes, many secondary metabolism hallmark genes and glycoside hydrolase genes were present. As expected, many genes involved in secondary metabolism were downregulated by laeA, which agreed with previous reports (Palmer and Keller, 2010). We focused on the glycoside hydrolase gene expression. Among the 129 predicted glycosidase hydrolase genes expressed in the four samples, the expression levels of 61 genes (P < 0.001) were significantly different between ΔlaeA and WT at both 24 h and 60 h (Figure 4A). Moreover, the gene expression of 40 (65.6%) glycosidase hydrolase was downregulated in ΔlaeA compared with that in WT at the 24 h time point. Furthermore, the laeA exhibited a strong and wide-ranging effect on glycoside hydrolase transcriptional regulation observed at 60 h. The gene expression of 54 (88.5%) glycosidase hydrolase was significantly downregulated in ΔlaeA compared with that in WT (Figure 4A).

When P. oxalicum was cultivated on cellulose and wheat-bran-containing media, the secretome data revealed that the top 10 extracellular proteins were glycoside hydrolases, including Amy15A (PDE_09417, Genbank No. EPS34453.1), Amy13A (PDE_01201, Genbank No. EPS26265.1), Cel7A/CBHI (PDE_07945, Genbank No. EPS32984.1), Cel61A (PDE_05633, Genbank No. EPS30681.1), Chi18A (PDE_08122, Genbank No. EPS33160.1), Cel3A/BGLI (PDE_02736, Genbank No. EPS27792.1), Xyn10A (PDE_08094, Genbank No. EPS33132.1), Cel7B/EGI (PDE_07929, Genbank No. EPS32968.1), Cel5B/EGII (PDE_09226, Genbank No. EPS34262.1), and Cel6A/CBHII (PDE_07124, Genbank No. EPS32164.1). Their products account for 28.9, 11.0, 9.6, 5.6, 4.9, 3.4, 3.3, 2.5, 1.9, and 1.5% of the total extracellular protein of P. oxalicum, respectively (Liu et al., 2013b). All top 10 extracellular protein-encoding genes [Figure 4A; numbers (1) to (10) indicate the amount of secretion from highest to the lowest] were detected with significant regulation in ΔlaeA compared with that in WT. laeA exhibited a strong effect on the expression at 60 h. The transcripts of amy15A, amy13A, cel7A/cbh1, cel61A, chi18A, cel3A/bgl1, xyn10A, cel7B/eg1, cel5B/eg2, and cel6A/cbh2 detected in ΔlaeA were only 43.2, 34.3, 13.8, 6.8, 0, 28.3, 9.5, 2.2, 32.5, and 30.5% of WT at 60 h, respectively (Figure 4B).

Then, we detected the expression patterns of five out of the above 10 genes in WT and seven mutants at different culture times. These genes include the prominent glucoamylase gene amy15A, cellobiohydrolase gene cel7A/chb1, endoglucanase gene cel7B/eg1, β-glucosidase gene cel3A/bgl1, and xylanase gene xyn10A. The strains were pre-cultured on glucose medium for 20 h, starved for 4 h under no carbon source conditions, and transferred to cellulose and wheat bran medium for further cultivation. The transcript levels of the five genes were quantified in cells grown from 12, 24, 36, 48, and 60 h (Figure 5). The expression levels of amy15A (Figure 5A), cel7A/chb1 (Figure 5B), cel7B/eg1 (Figure 5C), and xyn10A (Figure 5E) in OEclrB, OExlnR, and ΔcreA dramatically increased from 12 to 60 h. For example, amy15A expression in the OEclrB, OExlnR, and ΔcreA increased respectively by 167, 232, and 173% at the 36 h time point compared with WT. The expression level of cel7A/chb1 in the OEclrB, OExlnR, and ΔcreA increased respectively by 37, 56, and 23% compared with that of WT. However, the expression of amy15A, cel7A/chb1, cel7B/eg1, and xyn10A in ΔlaeA, OEclrBΔlaeA, OExlnRΔlaeA, and ΔlaeAΔcreA decreased remarkably, especially in the later phases of prolonged batch cultures. At the 48 h time point, amy15A expression in the ΔlaeA, OEclrBΔlaeA, OExlnRΔlaeA, and ΔcreAΔlaeA decreased to 8.93, 23.2, 17.3, and 46.4% of WT, respectively; cel7A/chb1 expression in the ΔlaeA, OEclrBΔlaeA, OExlnRΔlaeA, and ΔcreAΔlaeA decreased to 44.0, 58.2, 91.3, and 64.7% of WT, respectively (Figure 5).

As transcription factors creA, clrB, and xlnR have been identified to play essential roles in the expression of cellulase, and Lae1 was reported controlling transcriptional activator xyr1 gene expression in T. reesei (Seiboth et al., 2012), does LaeA affect creA, clrB, or xlnR expression? Then, the transcription levels of genes creA, clrB, and xlnR were quantified in WT and the mutants grown from 12, 24, 36, 48, and 60 h (Figure 6). The expression levels of creA was considerably increased in ΔlaeA compared with that in WT (Figure 6A), suggesting LaeA negative roles in regulating creA expression. The expression of clrB or xlnR was higher than that of in WT before 24 h-point culture time, but lower than that of in WT along with prolonged culture time (Figures 6B,C). It was expected that overexpression of clrB in WT (OEclrB) upregulate clrB expression remarkably; and overexpression of xlnR in WT (OE xlnR) upregulate xlnR expression remarkably. However, when LaeA was absent, the clrB expression in OEclrBΔlaeA or xlnR expression in OExlnRΔlaeA decreased. The result was unexpected, because the over-expression of clrB or xlnR was driven from the gpdA promoter, but not the native promoters. Meanwhile, the relatively high expression of clrB or xlnR in OEclrBΔlaeA or OExlnRΔlaeA (compared with WT) cannot rescue the impairment of cellulolytic enzyme gene expression by the deletion of laeA (Figure 3A), suggesting LaeA was necessary in the expression of cellulolytic enzyme gene activated by ClrB and XlnR. To investigate why laeA overexpression did not affect the cellulolytic enzyme production, the expression level of the key transcription factor gene (clrB, xlnR, and creA) in the OElaeA was compared with that of WT (Figure S5D). In contrast with that the expression of transcription repressor gene creA increased remarkably in ΔlaeA, the expression levels of creA in OElaeA was identical with that of in WT. The expression of transcription activator gene xlnR was also identical with that of in WT. Interestingly, the expression of transcription activator gene clrB had a small increase (about 2-folds) compared with that of WT. Obviously, this insignificant change in clrB expression had no effect on cellulolytic enzyme production (Figure S5C). The results showed, normal expression of laeA was enough to trigger proper expression of the key transcription factor gene; excessive laeA expression was not needed.

Of the 61 glycosidase hydrolase genes, only seven genes were upregulated at 60 h (Figure 5A, blue dots). Of the 7 genes, only PDE_00049 (a β-xylosidase-encoding gene) product was detected in P. oxalicum secretome (Liu et al., 2013a). Thus, the extracellular β-xylosidase activity and PDE_00049 (xyl3A) expression levels were assayed. Although, most cellulase and hemicellulase formation activities were reduced by the deletion of laeA, this deletion activated β-xylosidase synthesis, especially when laeA deletion and xlnR overexpression was combined. At 48 h, the extracellular β-xylosidase activity in OExlnRΔlaeA was 1.5 IU/mL, which was 5.8-fold greater than that in WT (Figure 7A). The results of qPCR for xyl3A showed that the transcription level of OExlnRΔlaeA was 28.5-fold greater than that of WT at 48 h. This result suggests that β-xylosidase gene expression was regulated at the transcriptional level (Figure 7B).

Originally, we thought that transcription activators (ClrB and XlnR) might recruit LaeA playing a common regulating function, as their co-existence was required for proper cellulolytic enzyme gene expression. So, yeast two-hybrid assay was done to determine if LaeA had physical interaction with transcription factors. Yeast two-hybrid experiments did not reveal any physically interaction between LaeA and ClrB (or XlnR). On the SD/-Trp/-Leu media with X-α-gal, the strain with laeA and clrB (or xlnR) was not colored (Figure S7A). And They did not grown on the SD/-Trp/-Leu/-His media with 2.5 mM 3AT (Figure S7B). No interaction was observed between LaeA and CreA either.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.