How an essential Zn2Cys6 transcription factor PoxCxrA regulates cellulase gene expression in ascomycete fungi?

Background Soil ascomycete fungi produce plant-biomass-degrading enzymes to facilitate nutrient and energy uptake in response to exogenous stress. This is controlled by a complex signal network, but the regulatory mechanisms are poorly understood. An essential Zn2Cys6 transcription factor (TF) PoxCxrA was identified to be required for cellulase and xylanase production in Penicillium oxalicum. The genome-wide regulon and DNA binding sequences of PoxCxrA were further identified through RNA-Sequencing, DNase I footprinting experiments and in vitro electrophoretic mobility shift assays. Moreover, a minimal DNA-binding domain in PoxCxrA was recognised. Results A PoxCxrA regulon of 1970 members was identified in P. oxalicum, and it was displayed that PoxCxrA regulated the expression of genes encoding major plant cell wall-degrading enzymes, as well as important cellodextrin and/or glucose transporters. Interestingly, PoxCxrA positively regulated the expression of a known important TF PoxClrB. DNase I footprinting experiments and in vitro electrophoretic mobility shift assays further revealed that PoxCxrA directly bound the promoter regions of PoxClrB and a cellobiohydrolase gene cbh1 (POX05587/Cel7A-2) at different nucleic acid sequences. Remarkably, PoxCxrA autoregulated its own PoxCxrA gene expression. Additionally, a minimal 42-amino-acid PoxCxrA DNA-binding domain was identified. Conclusion PoxCxrA could directly regulate the expression of cellulase genes and the regulatory gene PoxClrB via binding their promoters at different nucleic acid sequences. This work expands the diversity of DNA-binding motifs known to be recognised by Zn2Cys6 TFs, and demonstrates novel regulatory mechanisms of fungal cellulase gene expression. Electronic supplementary material The online version of this article (10.1186/s13068-019-1444-5) contains supplementary material, which is available to authorized users.

processes such as sporulation, autophagy, apoptosis or necrosis [4]. Fungal cells control CAZyme production by utilising complex nutrient-sensing pathways comprising numerous sensors and receptors such as kinases, transcription factors (TFs) and their targets [5]. However, the overall process remains poorly understood.
Penicillium oxalicum produces integrated cellulolytic enzymes that degrade insoluble cellulose, and displays a preference for particular carbon sources. When P. oxalicum grows in the presence of glucose, expression of cellulolytic enzyme-encoding genes is repressed via carbon catabolite repression (CCR). Among the TFs involved in CCR, the core zinc finger TF CreA/Cre-1 is the one best studied in filamentous fungi, including P. oxalicum. CreA directly or indirectly represses the expression of all major CAZyme genes and their regulatory genes that are involved in the degradation of plant cell walls in the presence of glucose [6,7].
When cellulose is present as a sole carbon source, induction of cellulolytic genes in P. oxalicum is dependent on a few essential TFs including the Zn2Cys6 TFs PoxCxrA and PoxClrB. Individual deletion of PoxCxrA and PoxClrB results in almost no cellulase production by P. oxalicum [2,8]. Clr2, a homolog of PoxClrB in Neurospora crassa, binds to a DNA sequence identical to that bound to by the Saccharomyces cerevisiae TF Gal4p (CGGN11CCG) [9]. Although PoxCxrA binds directly to the promotor regions of major cellulase and xylanase genes [8], the DNA element recognised by PoxCxrA remains unknown.
In the present study, we employed high-throughput sequencing of transcripts (RNA-seq) to analyse transcriptional levels of genes in P. oxalicum deletion mutant ∆PoxCxrA following exposure to Avicel in comparison with the parental strain ∆PoxKu70 to identify the regulon of PoxCxrA. Moreover, we identified two different DNA motifs recognised by PoxCxrA, as well as the minimal DNA-binding domain of PoxCxrA via in vitro DNase I footprinting and electrophoretic mobility shift assay (EMSA) experiments.

PoxCxrA positively regulates the expression of most genes encoding plant-cell-wall-degrading enzymes in P. oxalicum
In previous work, an essential TF PoxCxrA was found to be required for cellulase production in P. oxalicum, when subjected to cellulose as a carbon source [2,8]. However, its regulon is not yet fully understood. To comprehensively explore the regulatory roles of PoxCxrA in P. oxalicum, RNA-Seq was employed to analyse the transcriptomes of the P. oxalicum mutant strain ∆PoxCxrA and the parental strain ∆PoxKu70 cultured in medium containing Avicel as the sole carbon source for 24 h after a shift from glucose. A total of 21-23 million clean reads with a length of 100 bp were generated across all samples (Additional file 1: Table S1), > 90% of which were successfully mapped into the genome of P. oxalicum wild-type strain HP7-1 [2]. To evaluate the correlations among the three biological replicates for each strain, the Pearson's correlation coefficient (r) was calculated. The resulting high r value (> 0.85; Additional file 2: Figure S1) suggests that the transcriptomic data were reliable and suitable for further analysis.
Comparative transcriptomic profiling identified 1970 DEGs in deletion mutant ∆PoxCxrA, compared with the parental strain ∆PoxKu70, according to the |log2 fold change| ≥ 1 and p value ≤ 0.05 thresholds (Additional file 3: Table S2), which were defined as the Pox-CxrA regulon. The PoxCxrA regulon included 1010 genes down-regulated compared with ∆PoxKu70. Functional annotation based on Eukaryotic Orthologous Group (KOG) classification revealed that most of these DEGs were involved in primary and secondary metabolism (category E, amino acid transport and metabolism; category Q, secondary metabolite biosynthesis, transport and catabolism), and fell into the general function prediction only category (category R) (Fig. 1).

PoxCxrA and PoxClrB dynamically regulates the expression of one another
Interestingly, PoxCxrA regulated the expression of a key regulatory gene PoxClrB through RNA-Seq. To further elucidate this regulation, RT-qPCR was employed. When ΔPoxCxrA was exposed to Avicel, transcription of Pox-ClrB was down-regulated to some extent (2.7-to 8.2-fold) after 4 h (p ≤ 0.05, Student's t test) compared with that in ∆PoxKu70. In contrast, PoxCxrA expression increased by 1.5-to 2.1-fold in ∆PoxClrB during the latter stages of cultivation (24-48 h after induction; Fig. 4a).

Expression of both PoxCxrA and PoxClrB is induced by cellulose
When P. oxalicum strain ∆PoxKu70 was transferred into medium containing Avicel (induced state), PoxCxrA and PoxClrB exhibited similar transcriptional levels to those without a carbon source (de-repressed state) during the early induction stage (0-12 h), but only the transcriptional level of PoxClrB was higher than that without a carbon source during later stage (48 h). The expression of both PoxCxrA and PoxClrB on Avicel were higher than that on glucose (repressed state). Surprisingly, Pox-CxrA expression on Avicel increased by ~ 70% compared with its expression without a carbon source, but only at 24 h. Expression level of PoxClrB in ∆PoxKu70 was higher than those of PoxCxrA during all states (induced, repressed and de-repressed). PoxCxrA expression under Avicel induction increased before 24 h, but reduced after 24 h (Fig. 4b).

PoxCxrA directly binds to the promoter regions of PoxClrB and PoxCxrA, and genes encoding cellodextrin and glucose transporters
To further confirm whether PoxCxrA directly or indirectly regulates PoxClrB expression, in vitro EMSA experiments were employed. The putative DNAbinding domain PoxCxrA 17-150 was recombinantly expressed and purified by fusing with thioredoxin oxalicum strains cultivated on Avicel at four different time points (4, 12, 24 and 48 h) after a shift from glucose by real-time quantitative reverse-transcription PCR. Relative expression on the y-axis indicates differences in values for transcription between tested and reference genes, or between tested genes in deletion mutants and the parental strain ΔPoxKu70. *p ≤ 0.05 and **p ≤ 0.01 indicate differences among samples assessed by Student's t-tests. All experiments were carried out with at least three independent replicates (Trx), His and S-tags. A 300-bp DNA fragment from the promoter region of PoxClrB tagged with 6-carboxyfluorescein (6-FAM) was amplified using specific primers (Additional file 4: Table S3) and used as the probe for EMSA experiments. A DNA fragment from the promoter region of the β-tubulin gene POX05989 was used as a control. The results revealed that a complex was formed between PoxCxrA 17-150 [8] and the promoter region of PoxClrB, and its concentration increased with an increasing amount of fusion protein (1.0-2.0 µg). Bovine serum albumin (BSA) and Trx-His-S did not interact with the PoxClrB probe, and there was no interaction between PoxCxrA 17-150 and the control POX05989 promoter region. Competitive EMSA experiments revealed that the concentration of the complex decreased gradually with an increasing amount of protein-binding DNA fragment without 6-FAM (Fig. 5), suggesting that PoxCxrA 17-150 specifically bound the promoter region of PoxClrB.
Each EMSA experiment comprised different DNAbinding domains (PoxCxrA 17-150 , PoxClrB 1-120 or both; 0-2 µg) and 50 ng of the target probe labelled with 6-FAM. Probes without 6-FAM, and the β-tubulin gene POX05989, were used as competitive probes and negative controls, respectively. The fusion protein Trx-His-S purified from Escherichia coli cells harbouring the empty plasmid pET32a (+), and BSA alone, were used as protein controls.
The RT-qPCR data described above revealed that PoxClrB negatively regulated PoxCxrA expression during the latter stages of induction, but the regulatory mode was unclear. In vitro binding experiments showed that the putative DNA-binding domain of PoxClrB 1-120 was unable to bind to the promoter region of the PoxCxrA gene (Fig. 5), indicating that PoxClrB indirectly regulates transcription of PoxCxrA in P. oxalicum. EMSA experiments also showed that PoxCxrA  bound to the promoter region of its own gene, but PoxClrB 1-120 did not, suggesting the autoregulation of PoxCxrA. When a mixture of PoxCxrA  and PoxClrB 1-120 was treated with probe POX05587/ Cel7A-2, a band representing a larger protein-DNA complex than that formed by either PoxCxrA  or PoxClrB 1-120 individually was observed (Fig. 5), indicating that the binding motifs recognised by PoxCxrA 17-150 and PoxClrB 1-120 are distinct.

Different PoxCxrA-binding motifs are present in the promoter regions of target genes POX05587/Cel7A-2 and PoxClrB
Based on the above results and those of previous work [8], PoxCxrA appears to regulate the expression of cbh gene POX05587/Cel7A-2 and TF gene PoxClrB by directly binding to their promoters. In vitro EMSA and DNase I footprinting experiments were, therefore, performed to identify protein-binding motifs (PBMs) in the promoters of the target genes. An initial DNase I footprinting experiment was performed using a 100bp DNA fragment corresponding to the upstream flanking sequence (starting from the transcription initiation ATG codon) labelled with 6-FAM at the 3′-terminus to identify the PBM of PoxCxrA      (Fig. 7a). Further EMSA experiments confirmed that PoxCxrA 17-150 bound to probes containing PBM1 or PBM1 plus PBM2, but not PBM2 (Fig. 7b), suggesting that PBM1 contains the binding motif of PoxCxrA in the POX05587/Cel7A-2 promoter region (PBM_POX05587).

Identification of a minimal DNA-binding domain of PoxCxrA
The putative DNA-binding domain (DBD) of Pox-CxrA used above (residues 17-150) was serially truncated from the C-terminus to identify a minimal functional DBD. To facilitate the purification of the expressed recombinant protein in E. coli, we first fixed at 11th amino acid of N-terminus. Recombinant proteins PoxCxrA 11-150 , PoxCxrA 11-114 , PoxCxrA 11-87 , PoxCxrA 11-58 and PoxCxrA 11-31 were produced in E.   and 50 ng of truncated PBM_PoxClrB as probe labelled with 6-carboxyfluorescein. The promoter of β-tubulin gene POX05989 was used as a probe control. The fusion protein Trx-His-S purified from E. coli cells harbouring the empty plasmid pET32a (+), and BSA alone, were used as protein controls coli cells and purified (Fig. 9a, b), and in vitro EMSA experiments were carried out to investigate their ability to bind a 6-FAM-labelled 300-bp DNA fragment of the POX05587/Cel7A-2 promoter region comprising the PoxCxrA binding site as EMSA probe. The results showed that each truncated protein bound the probe to form a protein-DNA complex that retarded electrophoretic mobility in gels, except for PoxCxrA  . The concentration of the protein-DNA complexes gradually increased with increasing protein loading (1.0-3.0 µg). Competitive EMSA experiments were simultaneously performed, and the results indicated that the amount of complex tended to reduce with increasing amounts of competitive probe without 6-FAM. BSA and Trx-His-S control proteins did not bind the POX05587/Cel7A-2 promoter region (Fig. 9b). PoxCxrA 11-58 was confirmed to bind to PMB_POX05587 via in vitro ESMA experiments (Fig. 9c). Subsequently, the PoxCxrA 17-58 was also expressed and purified. EMSA binding experiments indicated a clear band comprised of PoxCxrA 17-58 and PMB_POX05587 on the gel (Fig. 9d). Thus, the minimal Subsequently, an alignment analysis of PoxCxrA 17-58 with the DBDs in known Zn2Cys6-containing TFs including PoxClrB, Pho7, XlnR, Clr1 and Gal4 that came from P. oxalicum HP7-1 [2], Schizosaccharomyces pombe [18], P. oxalicum strain 114-2 [7], N. crassa strain OR74A [19] and Saccharomyces cerevisiae S288C [9], was respectively performed. The results found three pairs of highly conserved zinc-coordinating cysteines that are essential for protein binding [18] and relatively conserved flanking amino acids such as arginine (R), lysine (K), aspartic acid (D) and proline (P; Fig. 9e). The retained amino acids showed high diversity. Several relatively conserved amino acids (18th R, 19th R, 27th Q, 30th K, 32th K and 38th P) were respectively replaced by alanine to generate a mutated PoxCxrA  . In vitro EMSA further displayed that all the mutated PoxCxrA 17-58 lost the ability to bind the PM_POX05587 (Fig. 9f ).

Discussion
PoxCxrA is known to be a critical transcriptional activator in P. oxalicum exposed to cellulose as a carbon source, but its regulatory mechanism is unclear. Herein, we found that PoxCxrA regulates cellulase production in P. oxalicum by controlling the expression of PoxClrB, and further elucidated their regulatory network. Pox-CxrA autoregulated the transcription of its own PoxCxrA gene, but PoxClrB did not. Moreover, the DNA-binding domain of PoxCxrA  bound to the promoter regions of PoxClrB and POX05587/Cel7A-2 at different binding sites (5′-ATC AGA TCC TCA AAGA-3′ and 5′-GCT GAG TCCTT-3′, respectively) according to in vitro DNase I footprinting and EMSA experiments. A minimal functional DBD (residues 17-58) of PoxCxrA was identified. These findings provide novel insights into the regulatory mechanism governing cellulase gene expression in P. oxalicum.
The PoxCxrA regulon was identified, which included a number of members involved in primary and secondary metabolism. To withstand starvation caused by insoluble cellulose as the sole carbon source, expression of major cellulase genes in P. oxalicum, including genes encoding CBH1 (GH7) and EGs (GH5 and GH12), was rapidly up-regulated, resulting in abundant intra-and extracellular cellodextrin production in fungal cells [20][21][22]. Accumulated intracellular cellodextrin triggered signalling cascades that activated the expression of PoxCxrA and PoxClrB, subsequently resulting in activation of the expression of cellulase genes such as POX05587/Cel7A-2.
Intriguingly, PoxCxrA also directly activated PoxClrB expression, whereas PoxClrB indirectly repressed the transcription of PoxCxrA in P. oxalicum. Regrettably, exactly how PoxCxrA expression is repressed by PoxClrB remains unknown (Fig. 10). In the early phase of P. oxalicum exposed to Avicel, both PoxCxrA and PoxClrB were transcribed at a background level. Expression of PoxCxrA was first up-regulated at the middle of the culture period, and then reduced in the latter stages due to autoinhibition or repression by PoxClrB (Fig. 10). In contrast, Pox-ClrB expression gradually increased in the middle and later stages.
Moreover, PoxCxrA up-regulated the expression of genes involved in cellodextrin transportation, such as POX06051/PoxCdtC and POX05915/PoxCdtD [10] and retarded the expression of genes involved in glucose transportation such as POX07576/GLT-1 and POX08783/HGT-2 [12] that caused CCR at high concentration. PoxCxrA stimulated cellulase gene expression via two pathways; direct binding, and through key TF mediators such as PoxClrB (Fig. 10), PoxCxrB and/or PoxNsdD [8].
Differences in regulatory networks occurred in different host cells. In N. crassa, cellobiose-activated CLR1 was necessary for the increased expression of clr2, a homolog of PoxClrB. CLR1 also induced most cellulase genes, thereby positively affecting enzyme production [19]. In the present study, we found that PoxCxrA was required for PoxClrB expression, but not POX03837, which encodes a homolog of CLR1 in P. oxalicum. Knockout of POX03837 had no effect on cellulase production in P. oxalicum HP7-1 cultured on Avicel (data not shown).
The PoxCxrA DBD resembles those of Gal4-like TFs (e.g. Gal4 in S. cerevisiae, CLR1 and XlnR in N. crassa, and PoxClrB). PoxCxrA 17-58 was found to be the minimal functional DBD, and it comprises three pairs of zinc-coordinating cysteines and several conserved amino acids that are essential for protein binding [18]. The amino acids flanking these cysteines display high diversity, which might determine the binding motifs. In the literature, many binding motifs of Gal4-like proteins have been characterised, including CGGN5CGGNCCG (CLR1), CGGN11CCG (Gal4 or CLR2), GGNTAAA (XlnR) [9], TCG(G/C)(A/T)NNTTNAA (Pho7) [18] and 5′-ATC AGA TCC TCA AAGA-3′ or 5′-GCT GAG TCCTT-3′ determined in this study. This suggests that the amino acids flanking the cysteines are also required for binding.
Moreover, screening PoxCxrA-binding sequences in other target genes confirmed by in vitro EMSA experiments using PBM_POX05587 and PBM_PoxClrB revealed that the binding sequences were highly diverse, which supports the enormous regulon of PoxCxrA that includes 1970 members, accounting for almost a quarter of the entire P. oxalicum strain HP7-1 genome. However, identification and analysis of all PoxCxrA-binding sequences in the genome requires further study.
In summary, the present study determined regulon for the essential TF PoxCxrA that is required for cellulase production of P. oxalicum. PoxCxrA regulates cellulase gene expression via two mechanisms: regulating key TF mediator PoxClrB, and directly binding cellulase genes with diverse binding motifs. This work provides novel insights into the regulatory mechanisms of fungal cellulase gene expression.

Strains and culture conditions
Penicillium oxalicum mutant strains ΔPoxCxrA (no. 12965) and ΔPoxClrB (no. 3.15649), and the parental strain ΔPoxKu70 (no. 3.15650) were obtained from the China General Microbiological Culture Collection (CGMCC) [8]. Spores were collected after maintaining strains on potato dextrose agar (PDA) plates at 28 °C for 6 days with buffer containing 0.1% Tween-80. In general, an aliquot of conidial suspension (1 × 10 8 conidia per milliliter) was inoculated into 100 mL of modified culture medium (MMM; pH 5.5) containing (g/L) (NH 4  For transfer cultivation, P. oxalicum strains were first pre-cultured for 24 h on MMM supplemented with 1.0% glucose as the carbon source at 28 °C with shaking at 180 rpm. Mycelia of pre-cultured P. oxalicum strains were transferred into MMM containing 2.0% Avicel and cultured for 4-24 h at 28 °C with shaking at 180 rpm. Mycelia were harvested and total RNA was extracted for RNA-Seq and real-time quantitative reverse transcription-PCR (RT-qPCR) assays.

RNA extraction
For total RNA extraction, mycelia harvested from three replicate independent cultures at each time point were frozen, ground under liquid nitrogen using a pestle and mortar, and RNA was purified using a TRIzol RNA Kit (Life Technologies, Carlsbad, CA, USA) according to the manufacturer's instructions. RNA concentration and quality were determined from the ratio of the absorbance at 260 nm/280 nm measured using a Nanodrop ND-2000 spectrophotometer (ThermoFisher Scientific, Waltham, MA, USA).

RNA-Seq
Total RNA was sequenced as described previously by Zhao et al. [2], and a cDNA library for each total RNA sample was constructed and assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA) and an ABI StepOnePlus Real-Time PCR System (Applied Biosystems, Foster City, CA, USA). After confirming eligibility, cDNA libraries were sequenced using an Illumina HiSeq 4000 system. After quality control, clean reads were mapped onto the genome of the wild-type P. oxalicum HP7-1 strain to assess sequence homology and functional annotation using BWA v0.7.10-r789 [23]

Expression and purification of truncated PoxCxrA constructs
Recombinant expression in Escherichia coli and protein purification were carried out as described previously by Yan et al. [8]. DNA sequences encoding a series of truncated PoxCxrA constructs (PoxCxrA 11-150 , PoxCxrA 11-114 , PoxCxrA 11-87 , PoxCxrA 11-58 , PoxCxrA 17-58 , PoxCxrA 11-50 and PoxCxrA 11-31 ) were amplified by PCR with specific primer pairs (Additional file 4: Table S3) and digested using appropriate restriction endonucleases. Digested DNA fragments were inserted into the expression vector pET32a (+) digested with the corresponding restriction endonucleases to generate recombinant plasmids that were subsequently introduced into competent E. coli cells by chemical recombination. After confirmation, cells harbouring constructs were pre-cultured for 12 h at 37 °C, then induced with 0.5 mM isopropyl-β-d-thiogalactopyranoside, with culturing being continued for 20 h at 28 °C to produce fusion proteins possessing thioredoxin (TRX), His and S tags. Fusion proteins were purified by affinity chromatography on TALON Metal Affinity Resin (Clontech, Palo Alto, CA, USA). Cells harbouring empty pET32a (+) vector were cultured as described above and the resulting Trx-His-S protein was used as a negative control.

In vitro EMSA assays
In vitro EMSA experiments were performed as described previously by He et al. [17]. Briefly, DNA fragments of different lengths harbouring the putative promoter regions of PoxClrB and POX05578/Cel7A-2, labelled with 6-carboxyfluorescein (6-FAM) at the 3′ terminus, were amplified by PCR using specific primer pairs (Additional file 4: Table S3), and used as probes for in vitro EMSA experiments. Meanwhile, the same DNA fragments without 6-FAM and a 300-bp DNA fragment from the promoter region of the β-tubulin gene POX05989 served as competitive and negative probes, respectively.

DNase I footprinting
DNase I footprinting experiments were carried out as reported by Wang et al. [27] with minor modifications. For each assay, 350 ng of each probe (100-bp DNA fragments corresponding to the region upstream from the start codon ATG of POX05587/Cel7A-2) were separately incubated with 6.6 µg of recombinant PoxCxrA 11-150 protein for 20 min at 30 °C. Subsequently, 0.015 U DNase I (Promega, Beijing, China) and 100 nM CaCl 2 were added and the reaction was incubated for 1 min at 30 °C. DNase I stop solution, containing 200 mM unbuffered sodium acetate, 30 mM EDTA and 0.15% sodium dodecyl sulphate (SDS), was used to stop the enzymatic reaction, and DNA was extracted and sequenced.

RT-qPCR assays
RT-qPCR assays were carried out according to a previously reported method [8]. Briefly, total RNA was extracted from P. oxalicum deletion mutant ∆Pox-CxrA grown on Avicel, and from the parental strain ∆PoxKu70 that served as a control. First-strand cDNA was synthesized from the extracted RNA as template using a PrimeScript RT Reagent Kit (TaKaRa Bio Inc., Dalian, China) according to the manufacturer's instructions. Each target gene was subjected to PCR amplification using the first-strand cDNA as template in a 20 µL