ARTP/EMS-combined multiple mutagenesis efficiently improved production of raw starch-degrading enzymes in Penicillium oxalicum and characterization of the enzyme-hyperproducing mutant

Background Application of raw starch-degrading enzymes (RSDEs) in starch processing for biofuel production can effectively reduce energy consumption and processing costs. RSDEs are generally produced by filamentous fungi, such as Penicillium oxalicum, but with very low yields, which seriously hampers industrialization of raw starch processing. Breeding assisted by random mutagenesis is an efficient way to improve fungal enzyme production. Results A total of 3532 P. oxalicum colonies were generated after multiple rounds of mutagenesis, by atmospheric and room-temperature plasma (ARTP) and/or ethyl methanesulfonate (EMS). Of these, one mutant A2-13 had the highest RSDE activity of 162.7 U/mL, using raw cassava flour as substrate, a yield increase of 61.1%, compared with that of the starting strain, OXPoxGA15A. RSDE activity of A2-13 further increased to 191.0 U/mL, through optimization of culture conditions. Increased expression of major amylase genes, including the raw starch-degrading glucoamylase gene, PoxGA15A, and its regulatory gene, PoxAmyR, as well as several single-nucleotide polymorphisms in the A2-13 genome, were detected by real-time reverse transcription quantitative PCR and genomic re-sequencing, respectively. In addition, crude RSDEs produced by A2-13, combined with commercial α-amylase, could efficiently digest raw corn flour and cassava flour at 40 °C. Conclusions Overall, ARTP/EMS-combined mutagenesis effectively improved fungal RSDE yield. An RSDE-hyperproducing mutant, A2-13, was obtained, and its RSDEs could efficiently hydrolyze raw starch, in combination with commercial α-amylase at low temperature, which provides a useful RSDE resource for future starch processing.


Background
Plant biomass biorefineries use renewable and relatively inexpensive raw materials as feedstocks for processing into value-added biofuels and chemicals, with potential benefits for industry and the environment. Biorefinery processes can help alleviate problems associated with industrial chemicals and fossil fuels, such as increasing

Open Access
Biotechnology for Biofuels  [1,2]. A major challenge, however, for biorefineries, is the high cost of enzymes used for biomass degradation into fermentable sugars, which results in poor profitability [1].
Bioethanol production by cassava starch biorefineries has attracted considerable interest because of the competitive advantages of cassava, such as abundant availability, low cost, and non-competition with direct food/ feed supplies [3]. Conventional enzymatic starch processing requires an initial high-temperature liquefaction step, using thermostable α-amylase, followed by saccharification with glucoamylase, after cooling below the starch gelatinization temperature, then fermentation to produce bioethanol [4]. Even these enzymatic processes require a large energy input and special equipment, which increases processing costs. The liquefaction step accounts for approximately 30-40% of the total cost of bioethanol production [5].
Remarkably, a few amylase proteins contain starchbinding domains (SBDs) that allow them to bind to the surface of raw starch granules [8], thereby efficiently and directly digesting granular, raw starch into glucose, even below the gelatinization temperature of starch. Those amylases are known as raw starch-degrading enzymes (RSDEs) and have potential applications in starch processing. In the natural environment, RSDEs are mainly secreted by filamentous fungi, but with low yields, which have not yet met the quantitative and cost requirements for large-scale industrialization of raw starch biorefining.
Classical physical or chemical mutagenesis is an efficient strategy for breeding of fungal strains, in particular, for improvement of enzyme yields. The main mutagens used are ethyl methanesulfonate (EMS), nitroglycerin, UV-radiation, and "atmospheric and room temperature plasma" (ARTP) [9][10][11][12]. The chemical agent EMS alkylates nucleotides at random positions, thereby resulting in transition mutations [13]. ARTP can change the structure and permeability of the cell wall and plasma membrane, by generation of plasma jets, which causes DNA damage; ARTP is a recently developed and effective mutation technique [14]. The use of EMS combined with UV treatment has been reported [9,10], whereas EMS combined with ARTP has not, to the best of our knowledge.
In this study, we employed multiple rounds of ARTP/ EMS-combined mutagenesis to improve RSDE production in P. oxalicum. The starting strain used was the engineered strain OXPoxGA15A, derived from the parental strain ∆PoxKu70 via over-expressing a rawstarch-degrading glucoamylase gene PoxGA15A [11,15,16]. Subsequently, culture conditions were optimized for the resulting mutant with the highest RSDE production and enzymatic hydrolysis efficiency against raw starch flour. Both genome re-sequencing and realtime reverse transcription quantitative PCR (RT-qPCR) were employed to analyze single-nucleotide polymorphisms and transcriptional levels of genes encoding major RSDEs in the isolated mutants.

Development of a two-layer agar gel diffusion method for rapidly screening for RSDE-hyperproducing mutants
To optimize screening efficiency for mutants with improved RSDE production, a high-throughput screening method using two-layer agar gels was developed. A two-layer agar gel was prepared, containing raw natural cassava flour (RNCF) and ball-milled Avicel, at a series of different ratios, as the top layer, and modified minimal medium (MMM), without carbon source as the bottom layer. RNCF flour was ground directly from freshly harvested cassava tubers after drying, without any other pretreatments, such as cellulose removal. RNCF is very similar to the cassava flour used in the starch industry and different from edible raw starch available from farmer's markets. As expected, Avicel stimulated the expression of cellulase and xylanase genes, as well as induction of the pPoxEgCel5B promoter, which controls over-expression of the RSDG gene, PoxGA15A in OXPoxGA15A [15], whereas RNCF induced the expression of amylase genes, including RSDE genes. The two-layer agar gel containing 0.5% Avicel and 1.0% RNCF exhibited clear zones around colonies of OXPoxGA15A and the parental strain ∆PoxKu70, the clear zones around the OXPoxGA15A colonies being slightly bigger (Fig. 1a). When inoculated into liquid MMM containing 0.5% Avicel and 1.0% RNCF, the RSDE activity of OXPoxGA15A was 362.6% higher than that of ∆PoxKu70 (p < 0.01; Fig. 1b). These data confirmed that the two-layer agar gel was an effective way to screen for RSDE hyperproducers, after random mutagenesis.
The lethality of ARTP against mutant E3-16 (see following description) was also determined, after 350, 400, 450, 500, and 550 s of treatment at a radio-frequency power input of 130 W and a flow rate of 10 L/min. The lethality increased with treatment time and treatment for 500 s resulted in 95.1% lethality (Additional file 2: Fig.  S2B).
When cultured for 8 days after treatment, a total of 3532 colonies were observed on the two-layer agar plates, including 1616 colonies from three rounds of EMS treatment, 1120 from ARTP treatment and 796 from the ARTP/EMS combination (Additional file 1: Fig. S1). Comparison of the diameter ratio between each colony and its clear zone selected the mutants designated as E1-1, E2-3, E3-16, A1-2, and A2-13 as having the largest ratios after each round of mutagenesis; their diameter ratios ranged from 1.51 to 2.0, compared with 1.48 for the original strain, OXPoxGA15A (Fig. 2a, b). Measurement of enzymatic activities determined the RSDE activity of each mutant, using RNCF as the substrate, ranged from 103.2 to 162.7 U/mL, when cultured in liquid MMM containing 4% wheat bran plus 1% Avicel, which was consistent with its diameter ratio (Fig. 2c). A2-13 had the highest RSDE activity and the largest diameter ratio, so it was selected for further study.
RSDE production by A2-13 was 61.0% higher than that of the starting strain OXPoxGA15A, after multiple rounds of ARTP/EMS mutagenesis. This appears to be attributable to mutant selection using the two-layer agar screening plates, containing both cassava starch and Avicel. The clear zones on the screening plates indicated degradation of both starch and Avicel.
To evaluate the genetic stability of mutant A2-13, regarding enzyme production, the RSDE activities, using RNCF as the substrate, were measured after each of six successive sub-cultures. No significant change in RSDE production was observed after any of the sub-cultures (Fig. 2d).

Optimization of liquid culture conditions
The secretion of plant biomass-degrading enzymes (i.e., amylase, cellulase, and xylanase) by filamentous fungi is closely related to the characteristics of the lignocellulosic substrates in media, such as physical properties and chemical composition, as well as culture-related factors, such as the culture set-up and substrate loading [1,17]. Therefore, to further improve RSDE production by mutant A2-13, the following liquid culture parameters were optimized: the initial pH of the medium, incubation temperature, composition of carbon source and composition ratio, nitrogen source, and concentration of spore inoculum. The optimal conditions found were as follows: initial pH of medium, 5.5; incubation temperature, 28 °C; composition of carbon source, wheat bran plus Avicel (2:3 w/w); nitrogen source, 5 g/L NH 4 NO 3 ; and concentration of spore inoculum, 1% (v/v) inoculum with 10 8 spores/mL (Additional file 3: Fig. S3). When cultured under these optimal culture conditions, mutant A2-13 produced 191.0 U/mL of RSDEs, which was 17.4% and 89.1% higher than those of A2-13 and OXPoxGA15A under non-optimal culture conditions (p < 0.01), respectively. This was also far higher RSDE production than by P. oxalicum strain GXU20 (20 U/mL), using the same substrate [17]. The production of RSDE by A2-13 against OXPoxGA15A, grown on a two-layer agar gel plate (a) and their RSDE activity in liquid medium (b). The two-layer agar gel plate consisted of an upper layer containing ball-milled Avicel (0.5% w/v) and natural raw cassava flour (RNCF, 1% w/v), and a lower layer of MMM without carbon source. Carbon sources used in liquid medium were the same as that in the top layer of the two-layer plate. RSDE activity was determined using RNCF as the substrate. ** (p ≤ 0.01) indicates significant difference between the OXPoxGA15A and ∆PoxKu70 by Student's t test. RSDE raw starch-degrading enzyme, MMM modified minimal medium RNCF, under the optimal conditions, is the highest reported to date. Other reports of RSDE activities using different raw starch substrates, such as processed cassava starch, potato starch, or uncooked soluble starch, are not directly comparable ( Table 1).

Properties of crude RSDE produced by the mutant A2-13
To investigate the effects of ARTP/EMS-combined mutagenesis on the properties of the crude enzymes produced, the optimal pH and temperature of the crude enzymes were measured in comparison with those of OXPoxGA15A. The optimum pH and temperature, with RNCF as substrate, were 4.5 and 65 °C, respectively, which were similar to those of crude enzymes from OXPoxGA15A (Fig. 3a, b). Moreover, pH and heat stability analyses revealed that crude RSDE from A2-13 was highly stable when degrading RNCF under acidic conditions, again similar to RSDE from OXPoxGA15A.
Interestingly, crude RSDE from strain A2-13 also had improved tolerance to alkaline conditions ( Fig. 3c) and was more stable at low temperature (< 40 °C), but less stable at medium-high temperature (45-65 °C; Fig. 3d), compared with RSDE from OXPoxGA15A. These changes in RSDE properties are potentially very beneficial for industrial applications, such as simultaneous saccharification and fermentation of raw starch to produce bio-ethanol, in which yeasts grow at 30-35 °C [4, 11, 23, 24], using as carbon source, glucose released from raw starch by simultaneous RSDE degradation.

Hydrolytic analysis of raw starch with crude RSDE produced by mutant A2-13
Corn starch has been widely used as a renewable feedstock for bioethanol production in the US [25]. Cassava starch is readily available in subtropical areas, such as China, Southern Asia, and South Africa, Fig. 2 Investigation of potential P. oxalicum mutants isolated from each round of mutagenesis. a Clear zones on the two-layer agar gel plates after 8 days. Red circle indicates the isolated colony used for the next round of mutagenesis. b Diameter ratio between clear zone and colony. c RSDE activity. d RSDE activity of the strain A2-13 sub-cultured six times. RSDE activity was determined using RNCF as the substrate. ** (p ≤ 0.01) indicates significant difference between the isolates and the starting strain OXPoxGA15A by Student's t test. RSDE raw starch-degrading enzyme, RNCF natural raw cassava flour and extensively employed for producing bioethanol and other bio-based products, because of its low cost and non-competition with direct food/feed supplies [3]. Therefore, RNCF and natural raw corn flour (RNCOF) were chosen to evaluate potential applications of A2-13, with OXPoxGA15A as control. Analysis of enzymatic hydrolysis reactions revealed that the hydrolytic efficiency of the crude A2-13 enzyme for converting RNCOF into glucose was slightly lower than the OXPoxGA15A enzyme (degree of hydrolysis (DH) 87.6% and 98.1% after 72 h, respectively Fig. 4a). In contrast, with RNCF as the substrate, the DHs were  (Fig. 4b).
When hydrolyzing different substrates using the same enzyme, the substrate composition can contribute to the hydrolytic efficiency. The composition of raw starch granules is mainly amylose and amylopectin, with minor non-starch components, such as cellulose fibers, proteins, and fats [26]. Amylose is made up of long linear polymer chains of glucopyranose units, whereas amylopectin has shorter, but highly branched chains, with high molecular weights [27]. Normally, the amylose content of cassava starch is lower than that of corn starch, but the fiber content is higher [28].
Different DHs from the same substrate, using crude enzymes from strains A2-13 and OXPoxGA15A, may result from the two strains producing different RSDE mixtures. Fast degradation of raw starch requires the synergistic action of various enzymes, including α-amylases, glucoamylases, α-glucosidases, and α-1,4glucan debranching enzymes [7]. For example, the Fig. 4 Hydrolysis efficiencies of RNCOF and RNCF by crude enzymes from P. oxalicum strains A2-13 and OXPoxGA15A. Hydrolysis of RNCOF (a) and RNCF (b) by crude enzymes from mutants A2-13 and OXPoxGA15A. Hydrolysis of RNCOF (c) and RNCF (d) by crude enzymes from mutant A2-13 combined with commercial α-amylase. Hydrolysis of RNCOF (e) and RNCF (f) by crude enzymes from OXPoxGA15A combined with commercial α-amylase. In panels a and b, 50 U crude enzyme was added per g of solid substrate. In panels c-f, the legends show the amounts of crude enzymes from P. oxalicum added. The ratio of crude enzymes from P. oxalicum and commercial α-amylase was 1:1. Values are presented as the mean ± SD of three replicates for each treatment. RNCOF natural raw corn flour, RNCF natural raw cassava flour recombinant rPoxGA15A combined with commercial α-amylase efficiently degraded raw corn starch and processed raw cassava flour, but rPoxGA15A alone did not [11].
Therefore, crude enzymes from A2-13 and OXPox-GA15A were combined with commercial α-amylase and used to hydrolyze RNCOF and RNCF. Hydrolysis of RNCOF by crude enzyme from A2-13 reached a DH of 96.0% at 72 h, with an enzyme loading of 50 U/g substrate, and a DH of 100% at 60, 48 and 36 h with 100, 150, and 200 U/g substrate, respectively (Fig. 4c). With RNCF as substrate, the DH reached 87.1% at 72 h with an enzyme loading of 50 U/g substrate, and 100% at 72, 60 and 48 h with 100, 150, and 200 U/g (Fig. 4d). The hydrolytic efficiencies against RNCOF (Fig. 4e) and RNCF (Fig. 4f ) with crude enzymes from OXPox-GA15A were similar with the enzymes from A2-13, when they were combined with commercial α-amylase.

Analysis of extracellular proteins and transcriptional levels of major amylase genes in the mutant A2-13
To better understand the findings from the hydrolysis experiments, the proteins secreted by A2-13 and OXPoxGA15A, cultured under their optimal conditions, were compared by SDS-PAGE analysis. The composition of the two protein mixtures was similar, but the relative concentrations of the different components were noticeably different. Specifically, the band corresponding to raw starch-degrading enzyme PoxGA15A [11,15] in A2-13 was denser than that in OXPoxGA15A (Fig. 5a). The PoxGA15A band (labeled by purple arrow in Fig. 5b) in A2-13 was 20% denser than that in OXPox-GA15A (p < 0.01), but the PoxAmy13A band was not significantly different between A2-13 and OXPoxGA15A (Fig. 5). A2-13 apparently produced more raw starchdegrading glucoamylase than OXPoxGA15A. Glucoamylase can release glucose from the non-reducing end of starch chains, and acts synergistically with α-amylase, which generates new non-reducing ends in starch chains, In panel c, expression levels of the tested genes in the strain A2-13 were normalized against those in the starting strain OXPoxGA15A. The actin gene was used as reference. **p ≤ 0.01 and *p ≤ 0.05 indicate significant differences between the A2-13 and OXPoxGA15A by Student's t test. Each experiment contained three biological replicates. Each data point represents mean ± SD. RT-qPCR real-time reverse transcription quantitative PCR, SDS-PAGE sodium dodecyl sulfate polyacrylamide gel electrophoresis, MMM modified minimal medium thereby improving the hydrolysis efficiency. Moreover, the crude enzymes secreted by A2-13 were a good balance between the different synergistic amylases, which can efficiently digest raw starch.

Phenotypic and growth analyses of the mutant A2-13
The phenotype and mycelial growth of A2-13 were compared with those of OXPoxGA15A. A2-13 was directly inoculated onto solid PDA medium plates containing different carbon sources: wheat bran plus Avicel, Avicel, soluble starch, glucose, and NRCA (RNCF plus Avicel) and cultured for 2-6 days. The A2-13 colonies on wheat bran plus Avicel, Avicel, PDA, and glucose plates were smaller than those of OXPoxGA15A to varying degrees, appeared dense, thick with mycelia, and darker in color. A2-13 colonies on the other carbon sources were darker in color, but not different in size (Additional file 4: Fig.  S4A). Sporulation of A2-13 varied on different carbon sources; sporulation was delayed on PDA, wheat bran plus Avicel, and glucose plates, but accelerated on the other carbon sources (Additional file 4: Fig. S4A).

Genome re-sequencing of the mutant A2-13
To further elucidate the reasons behind the observed changes in enzyme production and phenotypes, genome re-sequencing of A2-13 was carried out, with OXPox-GA15A as control. In total, 2881 Mbp of clean data for each P. oxalicum strain were generated, with an average sequencing depth of approximately 90-fold, which covered the genome of the wild-type strain HP7-1 with > 99% coverage.
Comparative genomic analyses detected 230 single-nucleotide variations (SNVs) and 131 insertion/ deletions (InDels) in the genome of A2-13, compared with that of OXPoxGA15A. These SNVs were localized in 193 intergenic regions and 35 coding sequences (CDSs). Of the 35 SNVs in CDSs, 25 non-synonymous SNVs were detected, including one stop, non-synonymous mutation in gene POX03359 that encodes a hypothetical protein (Additional file 5: Table S1). There were two genes, POX02883 and POX04725, encoding putative transcription factors. Gene POX09060 encoded the homologous protein of histone methyltransferase Set1, and shared 99.9% identity with PoSet1 (PDE_02489) in P. oxalicum strain 114-2 and 64.4% with Set1 in Saccharomyces cerevisiae. Set1 performs the methylation of histone H3 lysine 4 (H3K4) [29]. In S. cerevisiae, Set1 contributes to assembly of the elongation complex associated with RNA polymerase (Pol) II, by binding nascent RNA [30]. Recently, Li et al. [31] reported that PoSet1 positively regulated the production of cellulases and xylanases in P. oxalicum, by modulating H3K4me1 and H3K4me2 signals, as well as regulating colony diameter and spore number. POX01661 encoded the subunit Tho2 of the THO complex that is required for RNA transcript elongation, mRNA biogenesis, and export in S. cerevisiae [32,33]. POX09438 was annotated as formin Bni1p, which is involved in the polarized localization in S. cerevisiae, controlled by the Rho GTPase Cdc42, via the effector Gic2p [34]. Cdc42 is activated by the Bud site selection protein Bud3 [35] encoded by POX05164. To our surprise, a β-glucosidase gene, POX03641 was found with a unique SNV and another gene, POX02958-encoded calcium/calmodulin-dependent protein kinase type I, and two ribosomal protein-encoding genes, POX01113 and POX01133.
In addition to the SNVs in CDSs, of 193 SNVs located in intergenic regions, 38 and 10 were localized in 1500 bp upstream-and 300 bp downstream regions of the coding genes, respectively (Additional file 5: Table S1). These regions may contain gene promoters and terminators, in which SNVs may regulate the expression of the corresponding genes. A notable finding was a gene, POX02086, encoding a zinc finger protein of CCHC-type, two sugar/ inositol transporter encoding genes POX05515 and POX08241 and an α-glucosidase gene POX07319.
InDels in the mutant A2-13 included 65 deletions and 76 insertions. Most of them were located in intergenic regions and introns. Of particular interest, there were six InDels located in four CDSs (POX00063, POX02669, POX07393, and POX09551) that encoded endo-β-1,4xylanase Xyn11A, alkaline phosphatase, ABC transporter, and a hypothetical protein, respectively. InDels in POX00063 (G deleted at positions 197,194) and POX09551 (A and G inserted at positions 17,671 and 17,697) resulted in a frameshift mutation in A2-13, whereas deletion of nucleotides in POX02669 (CTC CCG deleted at position 2,276,066) and POX07393 (G and TAC TCC CC deleted at positions 877 and 878) resulted in removal of a few amino acids at the C-terminus (Additional file 6: Table S2). However, it appears that changes resulting from the InDels in POX00063, POX02669, and POX07393 do not contribute to increased amylase production in A2-13. Elucidation of the biological role of POX09551 requires further study in P. oxalicum.

Conclusions
In this study, we carried out ARTP/EMS-combined mutagenesis to enhance raw starch-degrading enzyme (RSDE) production in P. oxalicum. A mutant strain, A2-13, was isolated and produced RSDEs at a concentration of 191.0 U/mL, a yield increase of 89.1% compared with that of the parental strain, OXPoxGA15A. Crude RSDE enzymes from A2-13 showed improved tolerance to alkaline conditions and high hydrolysis efficiency against raw starch, when combined with commercial α-amylase. Furthermore, the factors contributing to the improved RSDE production by A2-13 were elucidated. This study confirmed that combined ARTP/EMS is an effective tool to enhance fungal RSDE yields and provided a potential new source of RSDEs for future industrial starch processing.

P. oxalicum strains and culture conditions
All P. oxalicum strains were cultured on potato-dextrose agar (PDA) plates for spore production. The starting strain, OXPoxGA15A, was from the China Center for Type Culture Collection, Wuhan, China (accession number M 2017794; [15]). Spores were collected from P. oxalicum cultured on PDA plates for 6 days.  [36], supplemented with 4% w/v wheat bran, plus 1% w/v Avicel, in an orbital shaker at 180 rpm and 28 °C, for 6 days. The secreted crude enzymes were collected by centrifugation at 11,300×g for 15 min and the supernatant used for measurement of enzymatic activity and hydrolysis of natural raw starch flour.
For RT-qPCR analyses, P. oxalicum strains (1.0 × 10 8 spores) were pre-cultured in MMM containing glucose as the sole carbon source for 24 h, then transferred into MMM containing 2% w/v wheat bran, plus 3% w/v Avicel in a shaker, at 180 rpm and 28 °C, for 12-48 h. The mycelia were harvested by filtering for RNA extraction every 12 h.
For genome re-sequencing, P. oxalicum strains OXPox-GA15A and A2-13 were inoculated into Completed Medium containing (in g/L) yeast extract 1.0, glucose 10.0, casein hydrolysate 1.0 g, NaNO 3  A two-layer agar gel was prepared, containing raw natural cassava flour (RNCF; from a local farmer's market in Nanning, China) and ball-milled Avicel (Sigma-Aldrich, Darmstadt, Germany), at a series of different ratios as the top layer, and MMM, without a carbon source as the bottom layer. RNCF was ground directly from freshly harvested cassava tubers after drying, without any other pretreatments, such as cellulose removal.

Extraction of total DNA and RNA
Total DNA and RNA were extracted from hyphae of P. oxalicum strains using chemical methods, as reported previously [36].

Mutagenesis
For EMS treatment, fresh spores were re-suspended in phosphate buffered saline (PBS) at 10 8 spores/mL. EMS (1.2%, w/v) was added to the suspension and incubated in a shaker at 180 rpm and 28 °C, for 1-12 h, to determine the optimum treatment time. An equal volume of Na 2 S 2 O 3 solution was used to stop reaction. The treated spores were separated by centrifugation at 11,300×g for 10 min at 4 °C and re-suspended at different concentrations (between 10 2 and 10 8 /mL) with sterile water.
For ARTP treatment, fresh spores were re-suspended in 5.0% glycerol at 10 6 spores/mL. Spore suspension (10 µL) was spread on carrier plates and treated for 0 to 550 s at a flow rate of ten liters per minute and radio-frequency power input of 130 W, in a Type M ARTP Mutagenesis Bio-breeding Machine (Wuxi Tmaxtree Biotechnology Co., Ltd., China), to determine the optimum treatment time. The treated spores were washed from the carrier plates using sterile water and adjected to concentrations of 10 2 to 10 8 spores/mL. ARTP/EMS-combined treatment was performed thereafter, as described above, with the optimum treatment times of 8 h by EMS and 500 s by ARTP.
Treated spore suspension (~ 100 μL) was spread on plates containing MMM supplemented with RNCF (1% w/v) plus Avicel (0.5% w/v). After incubation for 8 days at 28 °C, the diameter ratio between each colony and its clear zone was measured and calculated. Spores without EMS and/or ARTP treatment were used as controls.

Measurement of fungal enzymatic activity
The activity of RSDEs was determined as described previously [37], using RNCF as the substrate. Briefly, crude enzymes from P. oxalicum cultures were diluted with citrate buffer (100 mM; pH 4.5). Diluted enzyme solution (50 μL) was added to RNCF (450 μL; 1.0% w/v), incubated at 65 °C for 30 min; then the reaction mixtures were transferred to boiling water for 10 min. Enzymes inactivated by boiling were used as controls. Amounts of released reducing sugars and p-nitrophenol were determined using the 3,5-dinitrosalicylic acid method [38] and spectrometry at 410 nm, respectively. One unit of enzymatic activity (U) was defined as the amount of enzyme that produced 1 μmol of reducing sugars per min from each appropriate substrate.

Optimization of culture conditions
For further enhancement of RSDE production by the mutant A2-13, culture parameters, including the initial medium pH, incubation temperature, carbon source composition, and nitrogen source, were optimized, respectively, using MMM containing wheat bran (4% w/v) plus Avicel (1% w/v). The effect of the initial medium pH on RSDE production by A2-13 was determined at pH 2.5, 3.5, 4.5, 5.5, 6.5, and 7.5, at 28 °C. Subsequently, incubation temperatures of 20, 24, 28, 32, and 36 °C were evaluated at the previously determined optimum pH, to determine the optimum incubation temperature.
Finally, the effects of different-sized inoculations of spores (10 2 , 10 3 , 10 4 , 10 5 , 10 6 , and 10 7 ) on RSDE production were determined. All optimization experiments were carried out in a shaking incubator at 180 rpm. P. oxalicum was cultured for 6 days and the crude enzymes were collected for RSDE activity measurement.

Properties of crude RSDE produced by the mutant A2-13
RSDE activity against raw cassava flour (1% w/v) was measured in 0.1 M citrate-phosphate buffer at pH 3.0, 3.5, 4.0, 4.5, 5.0, 5.5, 6.0, 6.5, and 7.0, at 65 °C for 30 min, as described above. RSDE activity at each pH was expressed as relative to that at the optimum pH (designated as 100%). The optimum reaction temperature was then determined at the optimum pH, by measuring the activity at various temperatures (30, 35, 40, 45, 50, 55, 60, 65, 70, 75, and 85 °C). RSDE activity at each temperature was expressed as relative to that at the optimal temperature.
To evaluate pH stability, crude RSDE was incubated at 4 °C for 24 h in buffers at various pHs, i.e., 0.1 M citrate-Na 2 HPO 4 buffer (pH 3.0-7.0), Tris-HCl buffer (pH 7.0-9.0), and glycine-NaOH buffer (pH 9.0-11.0); then the residual RSDE activity was determined at the optimum pH and temperature for activity. Activity of untreated RSDE was defined as 100% and used as the positive control. Similarly, to ascertain temperature stability, crude RSDE was added to buffer at the optimum pH for stability, and the mixture was incubated at various temperatures from 30 to 65 °C, for 1 h. The residual RSDE activity was calculated as described above.

Enzymatic hydrolysis of raw starch
RNCF and natural raw corn flour (RNCOF, purchased from a farmer's market in Nanning, China) were used as substrates for enzymatic hydrolysis. For raw starch hydrolysis by RSDEs from P. oxalicum A2-13 and OXPoxGA15A, hydrolysis reactions with 10% (w/v) solid substrate and an enzyme loading of 50 U/g substrate were conducted in citrate-phosphate buffer (pH 4.5) at 40 °C for 72 h. The reducing sugars produced were determined using the DNS method [38].

Genome re-sequencing
Genomic DNAs of P. oxalicum strains A2-13 and OXPoxGA15A were extracted and subsequently used for construction of the read libraries with a length of 400 bp. Libraries were sequenced on a BGISEQ-500 platform at the Beijing Genomics Institute (BGI, Shenzhen, China). After removing low-quality reads and adaptors, the generated clean reads were mapped onto the genome of P. oxalicum strain HP7-1 [16] using Bowtie2 v 0.7.10 [39] to detect SNVs.

RT-qPCR assay
RT-qPCR was carried out as described previously [16]. The tested genes were detected using specific primers (Additional file 7: Table S3). The actin gene POX09428 was used as reference. The expression level of each tested gene was calculated relative to that of POX09428, and subsequently normalized against that for the parental strain ΔPoxKu70. Each experiment was repeated at least in triplicate.