Collection and characterization of oncom samples from traditional producers

Oncom samples were collected from 10 red oncom and 6 black oncom producers throughout Western Java, Indonesia. Before sampling, a survey was carried out on the location of the oncom industry and observations related to detailed information on raw materials, the starter culture used and the production process method used. Samples were taken aseptically, placed in sterile plastic, brought to the laboratory using an ice box and immediately used for analysis. Two replicates were collected. One sample was lyophilised for subsequent sequencing analysis, the other one was used fresh to analyse pH and texture. For pH measurements, a total of 50 g of wet sample was aseptically put into 450 ml of buffer, then crushed and homogenized to obtain a sample suspension of 10% dilution. The pH value was measured using a standardized pH meter. A total of 50 ml of the derived suspension was measured for pH and read when the value was stable. Measurements were made twice, then the average was taken. Texture analysis of oncom samples was carried out by visual observation and by using a penetrometer. A previously described framework13 was used to score the appearance and texture of the oncom by observing the growth of mould covering the surface of oncom and the texture density of oncom with the following scale: + (poor mould growth, non-compact texture); ++ (decent growth of mould, texture is quite compact and dense); +++ (good mould growth, compact and dense texture); ++++ (very good mould growth, very compact and dense texture). The hardness level was measured using the Precision Scientific Penetrometer (73501) based on the level of penetration of the penetrometer needle into the sample for a certain time58. The oncom sample was placed in the space provided, then the needle was inserted vertically above the surface of the sample at five different points. The measurement was carried out for 5 s, so the hardness was expressed as mm per 5 s.

16S and ITS amplicon sequencing of oncom samplesDNA extraction, PCR, sequencing and sequence processing

Lyophilised oncom samples were ground into a fine powder using a mortar and pestle. Powder was placed into a DNA Isolation Bead Plate. DNA was extracted following Qiagen’s instructions (MagAttract PowerSoil DNA kit, 27100-4-EP) on a KingFisher robot. Bacterial 16S rRNA genes were PCR amplified with dual-barcoded primers targeting the V4 region (515F 5’-GTGCCAGCMGCCGCGGTAA-3’ and 806R 5’-GGACTACHVGGGTWTCTAAT-3’) ITS: (ITSF 5-CCTCCGCTTATTGATATGC-3 ITS2-R CCGTGARTCATCGAATCTTTG), following the protocol in ref. 59. Input DNA concentration was normalized before sequencing. Amplicons were sequenced with an Illumina MiSeq system using the 250 bp paired-end kit (v.3). Sequences were analysed using DADA2 (ref. 60) and Phyloseq61 following the recommended procedures.

Metagenome sequencing and analysis of oncom samplesDNA extraction and library preparation

Lyophilised oncom samples were ground into a fine powder using a mortar and pestle. DNA from the lyophilised material was extracted using the Qiagen MagAttract PowerSoil DNA KF kit (27100-4-EP) using a KingFisher robot. DNA quality was evaluated visually via gel electrophoresis and quantified using a Qubit 3.0 fluorometer (Thermo Fisher). Libraries were prepared using an Illumina Nextera library preparation kit (Illumina).

Sequencing, data curation and sequence processing

Paired-end sequencing (150 bp ×2) was done on a NextSeq 500 system. Shotgun metagenomic sequence reads were processed with the Sunbeam pipeline. Initial quality evaluation was done using FastQC v.0.11.5 (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Processing occurred in four steps: adapter removal, read trimming, low-complexity-reads removal and host-sequence removals. Adapter removal was done using cutadapt (v.2.6)62. Trimming was done with Trimmomatic (v.0.36)63 using custom parameters (LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:36). Low-complexity sequences were detected with Komplexity (v.0.3.6)64. For metagenomics analysis, the concatenated paired-end-read fastq files were analysed using kaiju (v.1.9.0)24,25. The NCBI BLAST nr+euk database, which contains bacteria, yeasts, viruses and microbial eukaryotes, was used to perform taxonomic analysis. To reduce noise caused by false positives in identification, a 1% abundance filter was applied to define the presence and absence of each taxon65. To evaluate how much of the N. intermedia genome was recovered in the sequencing of the oncom samples, all fastq files were concatenated and mapped to the N. intermedia genome using HiSAT2 (https://daehwankimlab.github.io/hisat2/).

General growth conditions and husbandry of N. intermedia strains

N. intermedia was grown on Vogels minimal medium (VMM)66 unless otherwise indicated. The standard medium (1 l) was made by combining the 50X salt solution with 15 g sucrose (unless another carbon source was used) and 1 l distilled water. Agar (15 g) were added for solid medium. To generate spore suspensions, which were used for inoculation in liquid and solid-state cultures, N. intermedia was grown from glycerol stocks on VMM agar slants for at least 72 h at 30 °C, or until robust pigmented conidial growth was established. Conidial suspensions were generated by adding sterile water to the slants followed by vortexing. Conidial concentrations were established using a haemocytometer. Liquid and solid-state cultures were maintained in the dark unless otherwise indicated.

Preparation of SSF okara, sensory analysis and GC–MS analysis of volatile aromasFermented okara preparation

To make okara, 250 g of raw soybeans were soaked with 4 l of water at 4 °C overnight and rinsed with water before processing. It was then mixed with three parts water (1:3 w/w), processed in the Thermomix (Thermomix TM6) (high speed; 20 s) and drained. Immediately, the solid (okara) was separated from the liquid and steamed in the oven for 30 min at 100 °C. Okara was cooled to 30 °C before inoculating with Neurospora intermedia #2613. For inoculation, N. intermedia was grown on two separate slants on VMM for 1 week at 30 °C. A 12 ml conidial suspension was created by adding distilled water to each slant and the resulting 12 ml solution was added to 300 g of substrate. The oncom was covered loosely with a sterile cloth and incubated for 24 h at 29 °C and 60% relative humidity (RH). After 24 h the spores germinated and the oncom was fermented without cloth for 24 h. After an additional 24 h, the aerial orange mycelium covered all the surface. The samples were kept at −80 °C until analysis.

Volatile aroma extraction

To 5 g of each sample (wet weight, okara, oncom, cooked oncom), 11 g of water (containing 0.9% NaCl) was added. The samples were vortexed, placed in stomacher bags (BagPage, Interscience) and homogenized in a laboratory blender (Stomacher 400, Seward) for 2 min at maximum speed. Next, 8 ml were transferred to 20 ml gas-tight vials, 2.88 g of NaCl was added and the samples stored at −80 °C. Ten microlitres of 2-methyl-3-heptanone was used as an internal standard (diluted to 27.2 mg l−1 in liquid chromatography–mass spectrometry (LC–MS)-grade methanol). The negative control consisted of distilled water with the same concentration of salt and internal standard67. The volatile composition of the samples was determined using headspace solid phase micro-extraction (HS-SPME) in combination with gas chromatography–mass spectrometry (GC–MS), as described below.

SPME–GC/MS analysis

We used the method in ref. 67 with minor modifications. After the extraction, the samples were incubated at 45 °C for 15 min before inserting into the GC injection port. The GC–MS analysis was performed on a Thermo Scientific TRACE 1310 gas chromatograph equipped with a Thermo Scientific Q Exactive Orbitrap mass spectrometry system with a Thermo fused-silica capillary column of cross-linked TG-5SILMS (30 m × 0.25 mm × 0.25 µm) (Thermo Fisher). The GC conditions were as follows: inlet and transfer line temperatures, 250 °C; oven temperature programme, 40 °C for 2 min, 5 °C min−1 to 120 °C for 2 min, 7 °C min−1 to 220 °C for 5 min, 50 °C min−1 to 325 °C for 3 min; inlet helium carrier gas flow rate, 1 ml min−1; split ratio, 20:1. The electron impact (EI)-MS conditions were as follows: ionization energy, 70 eV; ion source temperature, 250 °C; full-scan m/z range, 30–350 Da; resolution, 60,000; AGC target, 1 × 106; maximum IT, 200 ms. Method of identification: (1) by comparison of the MS spectra with the NIST library and (2) by comparison of RI (Kovat indices). The areas were normalized to the ISTD 2-methyl-3-heptanone. Retention index was based on an Agilent DB-5MS column using C7-C27 as external references. Data were acquired and analysed with Thermo TraceFinder 4.1 software package (Thermo Fisher).

Sensory analysis

Sensory analysis was conducted using a hedonic test with a total of 61 consumers from Copenhagen, Denmark. Sensory analysis was designed in compliance with the Declaration of Helsinki and the 2016/679 EU Regulation on the Protection of Natural Persons Regarding the Processing of Personal Data. The study protocol was reviewed and approved by the Ethics Committee at Mondragon Unibertsitatea before analysis. The experimental procedure was explained to the participants who then completed a written consent form indicating voluntary participation in novel food sensory analysis and gave permission for data processing. Okara oncom was prepared to conduct the sensory analysis, following the method above. Approximately 10 g of each sample was cut in squares, fried and served. The test was conducted in the same room, under controlled temperature and relative humidity (20 ± 2 °C; 95 ± 5% RH) and natural illumination. The 9-point hedonic scale (1 = ‘dislike extremely to 9 = ‘like extremely’) was used to evaluate the participants’ level of liking, as well as three different characteristics (flavour, texture and appearance). The CATA attributes used for the sensory attributes were selected on the basis of trials with Tempeh, a similar fungal fermented Indonesian food from the island of Java68. All consumers were instructed to rinse their mouths with water between samples.

Nutritional analysis of raw and fermented okara and other by-products

Nutritional analyses were performed by Cumberland Valley Analytical Services. All samples analysed were lyophilised and ground into a fine powder before analysis. Fibre (crude) was analysed according to the protocol in ‘Fibre (crude) in animal feed and pet food (978.10)’ described in ref. 69. Protein (crude) was analysed using the combustion method, following the method of ‘Protein (crude) in animal feed (990.03)’ described in ref. 69. The combustion was performed using a Leco FP-528 Nitrogen Combustion Analyzer (Leco). Crude protein was calculated as Nitrogen × 6.25. Fat (crude) was analysed using the protocol described in ‘Crude fat in feeds, cereal grains and forages (2003.05)’ described in ref. 70. A Tecator Soxtec System HT 1043 Extraction unit was used for lipid extraction (Tecator). The amino acid composition of all amino acids except tryptophan was analysed using standard acid hydrolysis of the material and following AOAC Official Methods 994.12 and 982.30 (refs. 70,71). Tryptophan was analysed by the standard alkaline hydrolysis (Official Method 988) in AOAC Official Methods70,72. High-performance liquid chromatography (HPLC) was used for detection of amino acids.

Extraction and LC–MS analysis of ergothioneine and free amino acids in raw and fermented okaraExtraction

All samples were lyophilised before analysis. For extraction, samples were ground into a fine powder using a mortar and pestle, ~30 mg was transferred to bead beating tubes for homogenization (Lysing Matrix Z, MP Biomedicals, 116961050-CF), 1 ml of 20% methanol with 0.1% formic acid was added and samples were subjected to bead beating for 2× 1 min using a Biospec Mini beadbeater. Following bead beating, samples were spun down at 12,000 g for 10 min to separate the solids and 500 µl supernatant was transferred to a centrifugal spin filter (Amicon Ultra, Sigma-Aldrich, UFC500324) to remove any particulates and larger molecules (3 kDa cut-off). The flow-through was collected and subjected to analysis by LC–MS.

LC–MS analysis

For LC–MS, analytes were chromatographically separated with a Kinetex HILIC column (100 mm length, 4.6 mm internal diameter, 2.6 µm particle size; Phenomenex) at 20 °C using a 1260 Infinity HPLC system (Agilent Technologies). The injection volume for each measurement was 2 µl. The mobile phase was composed of 10 mM ammonium formate (prepared from a pre-made solution; Sigma-Aldrich) and 0.2% formic acid (from an original stock at ≥98% chemical purity; Sigma-Aldrich) in water (as mobile phase A) and 10 mM ammonium formate and 0.2% formic acid in 90% acetonitrile with the remaining solvent being water (as mobile phase B). The solvents used were of LC–MS grade, purchased from Honeywell–Burdick & Jackson. Analytes were separated via the following gradient: linearly decreased from 90% B to 70% B in 4 min, held at 70% B for 1.5 min, linearly decreased from 70% B to 40% B in 0.5 min, held at 40% B for 2.5 min, linearly increased from 40% B to 90% B in 0.5 min, held at 90% B for 2 min. The flow rate was changed as follows: 0.6 ml min−1 for 6.5 min, linearly increased from 0.6 ml min−1 to 1 ml min−1 for 0.5 min, held at 1 ml min−1 for 4 min. The total run time was 11 min. The HPLC system was coupled to an Agilent Technologies 6520 quadrupole time-of-flight mass spectrometer (QTOF–MS) with a 1:4 post-column split. Nitrogen gas was used as both the nebulizing and drying gas to facilitate the production of gas-phase ions. Drying and nebulizing gases were set to 12 l min−1 and 25 psi, respectively, and a drying gas temperature of 350 °C was used throughout. Fragmentor, skimmer and OCT1 RF voltages were set to 100 V, 50 V and 250 V, respectively. Electrospray ionization (ESI) was conducted in the positive-ion mode with a capillary voltage of 3.5 kV. MS experiments were carried out in the full-scan mode (m/z 70–1,100) at 0.86 spectra per second for the detection of [M + H]+ ions. The instrument was tuned for a range of m/z 50–1,700. Before LC-ESI-TOF MS analysis, the TOF MS was calibrated with the Agilent ESI-Low TOF tuning mix. Mass accuracy was maintained via reference ion mass correction, which was performed with purine and HP-0921 (Agilent Technologies). Data acquisition was carried out using MassHunter Workstation Software v.B.08.00 (Agilent Technologies). Data processing was carried out using MassHunter Workstation Qualitative Analysis v.B.06.00 and MassHunter Quantitative Analysis v.10.00. External calibration curves were used to quantify the analytes.

High-performance anion exchange chromatography (HPAEC) for sugar profilingExtraction of free sugars

All samples were lyophilised before analysis. For extraction from solid, samples were ground into a fine powder using a mortar and pestle, ~30 mg was transferred to tubes for homogenization (Lysing Matrix Z, MP Biomedicals, 116961050-CF), 1 ml of 20% methanol with 0.1% formic acid was added and samples were subjected to bead beating for 2× 1 min. Following bead beating, samples were spun down at 12,000 RCF for 10 min to separate the solids and 500 µl supernatant was transferred to a centrifugal spin filter to remove any particulates and larger molecules (3 kDa cut-off) (Amicon Ultra, Sigma-Aldrich, UFC500324). The flow-through was collected and subjected to analysis by HPAEC. For extraction from culture supernatants, 1 ml culture supernatant was collected and subjected to lyophilisation, 1 ml of 20% methanol with 0.1% formic acid was added and samples were extracted by continuous vortexing for 10 min. Samples were then spun down at 12,000 g for 10 min and 500 µl supernatant was transferred to a centrifugal spin filter to remove any particulates and larger molecules (3 kDa cut-off) (Amicon Ultra, Sigma-Aldrich, UFC500324). The flow-through was collected and subjected to analysis by HPAEC.

Extraction of pectin-bound sugars

The method of extraction was adapted from ref. 73. Lyophilised material was ground to a fine powder before boiling in 96% ethanol for 30 min. After a centrifugation step of 5 min at 20,000 g, the supernatant was discarded. The resultant pellet was washed with 70% ethanol until the supernatant was clear. Last, the pellet was washed with 100% acetone and dried in a vacuum concentrator. Samples were hydrolysed in 2 N trifluoroacetic acid for 1 h at 120 °C.

HPAEC for sugar profiling

The method was adapted from ref. 73 with minor modifications. HPAEC with pulsed amperometric detection was performed on an ICS-6000 (Dionex) using a CarboPac PA20 (3 150 mm, Dionex) anion exchange column at a flow rate of 0.4 ml min−1. Before sample injection, the column was equilibrated with 5 mM NaOH for 5 min. The elution programme involved two isocratic elution steps with 5 mM NaOH from 0 to 23 min to separate the neutral sugars, followed by a ramp step to 450 mM NaOH from 23.1 to 41 min, which allowed separation of uronic acids and washing of the column. Monosaccharide standards comprised l-Fuc, l-Rha, l-Ara, CGal, d-Glc, d-Xyl, d-GalA and d-GlcA, as well as GlcNac and d-Man. A run of a standard mixture containing the eight monosaccharides was performed with each sample set to enable sample quantitation by linear regression.

Library preparation and sequencing to generate the high-quality genome of oncom-derived N. intermedia FGSC #2613Extraction of high-quality genomic DNA from N. intermedia

N. intermedia was grown in VMM with sucrose as the sole carbon source at 25 °C and 120 r.p.m. for 72 h. Mycelia were collected by vacuum filtration and immediately ground in a mortar and pestle with liquid nitrogen to generate a fine powder. Approximately 12 g of finely ground mycelium was subjected to further DNA extraction. The mycelium was transferred to a 50 ml falcon tube containing 10 ml lysis buffer (0.15 M NaCl, 0.1 M EDTA, 2% SDS at pH 9.5). Protease K (Thermo Fisher, EO0491) was then added at a final amount of 1 mg. The tube was left overnight at 37 °C with shaking at 90 r.p.m. Samples were then centrifuged at 6,000 g for 10 min to pellet the cellular debris. After centrifugation, avoiding the bottom pellet of debris, 15 ml of the supernatant was transferred to a new tube. Distilled water (10 ml) was then added to the supernatant, as well as 25 ml of phenol:chloroform:isoamyl alcohol (25:24:1) reagent (Sigma-Aldrich, P3803). The sample was rotated and shaken to precipitate proteins, spun down at 7,000 RCF for 15 min to separate the layers, and 23 ml of the top aqueous layer was moved to a new 50 ml falcon tube. Nucleic acids present in this fraction were precipitated with 0.6 volumes (14 ml) of ice-cold isopropanol, which had been frozen at −80 °C for 15 min. The precipitated solution was spun down at 7,000 g for 10 min to pellet the DNA. Then 6 ml TE buffer (10 mM Tris, 1 mM EDTA pH 8.0) was added and the sample was resuspended. RNA was digested by overnight incubation at 37 °C with RNase A (300 μg, Sigma-Aldrich, 10109142001). Then, an additional protein digestion step was performed by adding 300 μg protease K (Thermo Fisher, EO0491) and incubating at 37 °C for 2 h. A volume of 6 ml of the solution was extracted once with 6 ml of phenol:chloroform:isoamyl alcohol (25:24:1) reagent (Sigma-Aldrich, P3803). DNA in the top aqueous layer was precipitated with 0.6 volumes (3.5 ml) of ice-cold isopropanol, and following centrifugation, the pellet was dried at room temperature, followed by resuspension in TE buffer. To further purify the DNA for PacBio sequencing, we performed an additional cleanup step, which removed contaminating carbohydrates and organic reagents. In this protocol, 150 µl DNA suspension was mixed with 150 µl chloroform and 50 µl TE buffer. Samples were inverted several times to mix, 150 µl of the top layer was removed to a new Eppendorf tube, and 15 µl sodium acetate (pH 5.2) and 450 µl 100% ethanol (ice cold) were added. Following centrifugation to pellet the precipitated DNA, the supernatant was removed and the pellet was washed twice with 800 µl 70% ethanol, followed by air drying for 30 min at room temperature. The final pellet was resuspended in 150 µl TE buffer. DNA quality was assessed by agarose gel electrophoresis and nanodrop. DNA concentration was assessed using the Qubit DNA BR assay quantification kit (Q32850). The DNA was deemed of sufficiently high quality for reference genome sequencing using PacBio and Illumina.

Library preparation and genome sequencing of N. intermedia #2613

For the PacBio sequencing (multiplexed at >10 kb with Blue Pippin Size Selection, Tubes), an input of 1.5 µg of high-quality genomic DNA (extracted according to the protocol above) was sheared around 10 kb using Megaruptor 3 (Diagenode) or g-TUBE (Covaris). The sheared DNA was treated with exonuclease to remove single-stranded ends, DNA damage repair enzyme mix, end-repair/A-tailing mix and ligated with barcoded overhang adapters using SMRTbell Express Template Prep kit 2.0 (PacBio). Up to 16 libraries were pooled in equimolar concentrations and purified with AMPure PB beads (PacBio). Pooled libraries were size selected using the 0.75% agarose gel cassettes with Marker S1 and High-Pass protocol on the Blue Pippin (Sage Science). PacBio sequencing primer was then annealed to the SMRTbell template library and sequencing polymerase was bound to them using Sequel II Binding kit 2.0. The prepared SMRTbell template libraries were then sequenced on a PacBio Sequel IIe sequencer using sequencing primer, 8M v.1 SMRT cells and v.2.0 sequencing chemistry with 1 × 1,800 sequencing movie run times.

Genome assembly and annotation

Filtered subread data were processed with the JGI QC pipeline to remove artefacts. The mitochondria-filtered CCS reads were then assembled with Flye v.2.8.1-b1676 (https://github.com/fenderglass/Flye) [-g 40 M–asm-coverage 50 -t 32–pacbio-hifi] and subsequently polished with two rounds of RACON v.1.4.13 racon [-u -t 36] (https://github.com/lbcb-sci/racon). Small scaffolds (<1 kb) were removed from the assembly. The cleaned nuclear genome assembly was annotated using the JGI Annotation Pipeline74. The annotation pipeline uses a combination of ab initio, homology-based and transcriptome-based gene predictors. The best representative gene model at each locus was selected through automated filtering based on homology and transcriptome support to produce the gene model sets available in MycoCosm74.

Growth of N. intermedia FGSC #2613 across carbon sources and RNA extraction for sequencingGrowth experiment across carbon sources

We adapted the approach used for Neurospora crassa in ref. 35, with minor modifications. For all RNA-seq experiments, we first grew N. intermedia #2613 from a glycerol stock for 7 days on VMM slants harbouring sucrose as the sole carbon source. Conidia were then collected by addition of 4 ml sterile water to the slant, followed by vortexing. Conidia were counted using a haemocytometer and added at a final concentration of 5 × 105 conidia per ml into 3 ml VMM medium harbouring 2% (w/v) sucrose in 24-well Whatman Uniplates. Each well harboured 3 ml medium. Importantly, the bottoms of the wells for each Uniplate were initially scratched with a sharp syringe needle to allow adherence and formation of a mycelial mat. The plates were placed in a shaker (light on) for 2 h at 30 °C to enable conidial germination and initial mat formation. This step was necessary to establish a mat that could be manipulated in subsequent steps. Then, the shaker was turned on at 30 °C (200 r.p.m., light on). After 15 h of growth at 30 °C, mycelial mats were removed from the bottom of the plate where they had attached and were then washed three times in 3 ml of VMM without any added carbon to remove the sucrose medium. The mycelial mats were finally transferred to 1x VMM with the indicated carbon source for induction. All conditions were done in triplicate. In addition to the carbon sources, a set of samples were ‘induced’ in a no-carbon control, which facilitated downstream analysis to identify genes that were uniquely induced on the carbon source of interest. For induction conditions with carbon sources, 2 mM mono and disaccharides were used, while for complex carbon sources, including complex polysaccharides and plant biomass, 1% (w/v) was used. The following complex carbon sources were used in the experiments at 1% (w/v): Okara flour (Renewal Mill), soluble soy polysaccharides (Creative Enzymes Inc, NATE-1284), Avicel PH-101 (Sigma-Aldrich, 11365), rhamnogalacturonan from soy pectic fibre (Megazyme, P-RHAGN) and stachyose hydrate (Sigma-Aldrich, S4001-100MG). The following carbon sources were used in the experiment at 2 mM: d-(+)-raffinose pentahydrate (TCI, R0002), d-(+)-galactose (Sigma-Aldrich, G0750-10G), l-(+)-arabinose (Sigma-Aldrich, A3256-25G), d-(+)-xylose (Sigma-Aldrich, X3877-25G), d-(+)-galacturonic acid monohydrate (Sigma-Aldrich, 48280-25G-F), l-rhamnose monohydrate (Sigma-Aldrich, 83650-50 G), d-(+)-cellobiose (Sigma-Aldrich, C7252-100G) and d-(+)-mannose (MP Biomedicals, 02102250-CF). After 4 h of induction, mycelia were collected over Miracloth filter paper and flash frozen in liquid nitrogen for storage at −80 °C in Lysing Matrix Z tubes (MP Biomedicals, 116961050-CF), which allowed subsequent bead beating upon defrosting for RNA extraction. Three biological replicates were used for each condition.

RNA extractions

RNA extractions were performed on −80 °C stored biomass using TRIzol (Invitrogen, 15596026) and chloroform. Half of the mycelial biomass from a 3 ml culture was used for RNA extractions, and 1 ml of TRIzol reagent was added to the bead-beating tube. Tubes containing biomass, TRIzol and beads were bead beaten for 1 min, then allowed to incubate at room temperature for 5 min on a rocker. Chloroform (200 µl) was then added to each tube, which was vortexed and centrifuged for phase separation. Of the aqueous phase from each sample, 400 µl was combined with 400 µl isopropanol and incubated at room temperature for 10 min on a rocker for RNA precipitation. Samples were centrifuged at 4 °C for 10 min. RNA pellets that formed were washed with 75% ethanol and centrifuged at 4 °C for 2 min. Ethanol was removed via pipette, and the RNA pellet was allowed to dry for several minutes with the cap open. The RNA pellet was resuspended in 40 µl water and treated with 2 µl of Turbo DNAse (Thermo Fisher, AM2238) in a 50 μl reaction. After 20 min incubation at 37 °C, RNA was cleaned up using Qiagen RNeasy mini kit (Qiagen, 74104), eluting at the final step in 30 µl RNAse-free water. RNA was tested for quality using agarose gel electrophoresis and nanodrop. RNA was quantified using the Qubit RNA quantification kit (Thermo Fisher, Q10210).

Library preparation and sequencing of N. intermedia FGSC #2613 transcriptomes

Illumina sequencing was used to capture the transcriptomic profile of N. intermedia #2613 grown across carbon sources. RNA was extracted according to the protocol above. Plate-based RNA sample preparation was performed on the PerkinElmer Sciclone NGS robotic liquid handling system using Illumina’s TruSeq Stranded mRNA HT sample prep kit utilizing poly-A selection of mRNA following the protocol outlined in the Illumina user guide, and with the following conditions: total RNA starting material was 1 µg per sample and 8 cycles of PCR was used for library amplification. The prepared libraries were then quantified using the KAPA Illumina library quantification kit (Roche, NC2242092) and run on a LightCycler 480 real-time PCR instrument (Roche). The quantified libraries were then multiplexed and the pool of libraries was then prepared for sequencing on the Illumina NovaSeq 6000 sequencing platform using NovaSeq XP v1.5 reagent kits (Illumina, 20028401) and S4 flow cell, following a 2 × 150 indexed run recipe.

Analysis of transcriptome data from N. intermedia FGSC #2613 grown across carbon sourcesRNA-seq data processing

Raw fastq file reads were filtered and trimmed using the JGI QC pipeline resulting in the filtered fastq file (*.filter-RNA.gz files). Using BBDuk75, raw reads were evaluated for artefact sequence by k-mer matching (k-mer = 25), allowing 1 mismatch, and detected artefact was trimmed from the 3’ end of the reads. RNA spike-in reads, PhiX reads and reads containing any Ns were removed. Quality trimming was performed using the phred trimming method set at Q6. Finally, the reads under the length threshold were removed. Filtered reads from each library were aligned to the reference genome using HISAT2 (v.2.2.0)76. Strand-specific coverage bigWig files (fwd and rev) were generated using deepTools (v.3.1)77. featureCounts78 was used to generate the raw gene counts (counts.txt) file using gff3 annotations. Only primary hits assigned to the reverse strand were included in the raw gene counts (-s 2 -p–primary options). Raw gene counts were used to evaluate the level of correlation between biological replicates using Pearson’s correlation.

Differential expression analysis

To determine the genes that change significantly, we only considered as expressed those genes that have at least a sum of 10 counts per million in at least 15 libraries. All comparisons were paired and the profile in each carbon source was compared against the non-carbon-source condition; this was done with the edgeR package79. We used the generalized linear model (GLM) likelihood ratio test for this. To determine whether there is consistency between biological replicates, we performed a multidimensional scaling plot of distances between gene expression profiles. The parameters used to call a gene ‘differentially expressed’ between conditions were FDR < 0.05 and log2FC > |1|. A common dispersion between replicates of 0.01339982 was calculated and a tagwise dispersion was also calculated. Raw gene counts (counts.txt), not normalized counts, were used for differential gene expression analysis.

Annotation of the genes

To annotate genes, orthologous genes between N. intermedia and N. crassa were obtained. This was done with the OrthoFinder programme80. The curated annotation was inherited from N. crassa to N. intermedia. The general classification as ‘PlantDegradBio’ is based on the Neurospora crassa genes predicted to code for plant biomass degrading enzymes and transporters35.

Co-expression network

With the matrix of accounts, we built a co-expression network using the WGCNA software81. The transformation of the count table was performed with the VST (variance stabilizing transformation) function of DESeq2 (ref. 82). Subsequently, the Pearson correlation matrix and the weighted adjacency matrix with continuous values of 0 and 1 were obtained. To evaluate scale independence and mean connectivity, we utilized a gradient method by systematically adjusting the power value from 1 to 20. After identifying a degree of independence surpassing 0.90, the construction of a scale-free network was initiated using the blockwiseModules function, employing a power value of 8. To define the modules of the network, the minimum size was established at 30 genes per module and the threshold for merging similar modules was established at 0.15. Furthermore, the weighted adjacency matrix was transformed into a topological overlap measurement matrix (TOM) to estimate connectivity in the network. Finally, we exported the network using the function exportNetworkToCytoscape with a correlation threshold of 0.25. The network was visualized using the Cytoscape v.3.9.1 programme. Network metrics were obtained using Cytoscape v.3.9.1. To determine enriched metabolic pathways in the modules, we used the KEGG tool (https://www.genome.jp/kegg/).

Time-course profiling of galactose and arabinose in N. intermedia #2613 liquid cultures using okara as sole carbon source

Conidia (5 × 105) from N. intermedia #2613 were inoculated into 50 ml VMM medium harbouring okara flour as the sole carbon source (1% w/v) and 250 ml flasks were used. Flasks were then left to shake at 160 r.p.m. at 30 °C in the dark. Culture supernatants (1 ml) were removed at regular intervals, spun down to remove the solid material and mycelium, and immediately flash frozen at −80 °C. Clarified culture supernatants were then lyophilised and extracted for sugar analysis using HPAEC (see below).

Enzyme assays for cellulase detection

For all cellulase assays, we used the fluorescence-based cellulase kit from Abcam (Abcam, ab189817) and followed the standard protocol. All cellulase assays were performed in 96-well plates (Corning, Falcon Tissue Culture Plate, 353072) and fluorescence was measured at 550 nm excitation and 595 nm emission using a BIOTEK Synergy H1 microplate reader (Agilent). Readings were recorded every 3 min for 18 min. For enzyme assays from liquid cultures, 1 ml culture supernatants were removed from cultures by pipetting. These were then spun down to pellet the insoluble material and mycelium. Of the clarified supernatant, 20 or 25 µl was then used as the enzyme source in cellulase assays in the 100 µl reactions. For cellulase assays from solid-state samples, ~300 mg (wet weight) of raw or fermented okara was placed in an Eppendorf tube. Water (1 ml) was added followed by vortexing. To homogenize the sample, sonication was used (10 s on, 25 s off, 2 min total, 25% amplitude). The resulting slurry was spun down at maximum speed to clarify the supernatant, and 25 µl of the resulting supernatant was used as the enzyme source in the 100 µl enzyme reactions. Cellulase activity was calculated using the formula (B/(ΔT × V)) × D, where B is the amount of resorufin in the sample well calculated from the standard curve (µmol); ΔT is the linear phase reaction time T2 – T1 (min); V is the original sample volume added into the reaction well (µl); D is the sample dilution factor if the sample was diluted to fit within the standard curve range. Results from calculations were expressed as mU ml−1; 1 unit (U) was defined as the amount of enzyme that cleaves the substrate to generate 1.0 μmol of molecule per min.

Cellulase activity in liquid cultures of N. intermedia strainsConfirmation of cellulase activity in N. intermedia #2613 liquid cultures

To confirm the presence of cellulases suggested by the RNA-seq data, we replicated the RNA-seq culture setup in which the cellulase gene expression was first detected. We first grew N. intermedia #2613 from a glycerol stock for 7 days on VMM slants harbouring sucrose as the sole carbon source. Conidia were then collected by addition of 4 ml sterile water to the slant, followed by vortexing. Conidia were counted using a haemocytometer and added at a final concentration of 5 × 105 conidia per ml into 3 ml VMM medium harbouring 2% (w/v) sucrose in 24-well Whatman Uniplates. The bottoms of the wells for each Uniplate were initially scratched with a sharp syringe needle to allow adherence and formation of a mycelial mat. The plates were placed in a shaker (light on) for 2 h to enable conidial germination and initial mat formation. This step was necessary to establish a mat that could be manipulated in subsequent steps. Then, the shaker was turned on (200 r.p.m., light on). After 15 h of growth, mycelial mats were removed from the bottom of the plate where they had attached and were then washed three times in 3 ml of VMM without any added carbon to remove the sucrose medium. The mycelial mats were finally transferred to 1x VMM with either sucrose (1.5% w/v), avicel (1% w/v), okara flour (1% w/v) or rhamnogalacturonan (1% w/v). All conditions were done in triplicate. Cultures were left for 72 h at 30 °C and supernatants were then analysed for cellulase activity using the protocol above.

Screen for secreted cellulase activity in liquid cultures among N. intermedia strains grown on okara

We first grew N. intermedia strains from glycerol stocks for 7 days on VMM slants harbouring sucrose as the sole carbon source. Conidia were then collected by addition of 4 ml sterile water to the slant, followed by vortexing. The following strains were used from the by-product-associated clade: #2613, #2685, #2559, #5644, #1791, #5642, #5342. The following strains were used from the burn-associated clade: #1316, #8767, #8761, #1938, #1785, #7426, #7402, #3194. Conidia were counted using a haemocytometer and added at a final concentration of 5 × 105 conidia per ml into 250 ml flasks with 50 ml VMM medium harbouring 1% (w/v) okara flour as the sole carbon source. All strains were grown in triplicate and the experiment was repeated three times to verify the consistency of the results. Cultures were left for 72 h at 30 °C and supernatants were analysed for cellulase activity using the protocol above. Biomass was collected and dried at the end of the experiment. The biomass was dried for 7 days at 50 °C before the weight was recorded.

Genome sequencing of additional N. intermedia strains beyond the oncom-derived reference strain N. intermedia #2613Extraction and sequencing

N. intermedia strains subjected to sequencing (FGSC #2559, #2685, #5644, #2557, #1791, #5642, #5342) were grown in VMM-sucrose medium (50 ml medium in 250 ml flasks) for 72 h before collection by vacuum filtration and flash freezing in liquid nitrogen. Genomic DNA extraction, sample quality assessment, DNA library preparation, sequencing and bioinformatics analysis were conducted at Azenta Life Sciences. Genomic DNA was extracted using DNeasy Plant mini kit following manufacturer instructions (Qiagen, #69204). Genomic DNA was quantified using the Qubit 2.0 fluorometer (Thermo Fisher). NEBNext Ultra II DNA Library Prep kit for Illumina (New England Biolabs, #E7645L) clustering and sequencing reagents were used throughout the process following manufacturer recommendations. Briefly, the genomic DNA was fragmented by acoustic shearing with a Covaris S220 instrument. Fragmented DNA was cleaned up and end repaired. Adapters were ligated after adenylation of the 3’ ends, followed by enrichment by limited cycle PCR. DNA libraries were validated using a High Sensitivity D1000 ScreenTape on the Agilent TapeStation (Agilent Technologies) and quantified using the Qubit 2.0 fluorometer. The DNA libraries were also quantified by real-time PCR (Applied Biosystems). The sequencing library was clustered onto the lanes of an Illumina HiSeq 4000 (or equivalent) flow cell. After clustering, the flow cell was loaded onto the Illumina HiSeq instrument according to manufacturer instructions. The samples were sequenced using a 2 × 150 bp paired-end configuration. Image analysis and base calling were conducted using the HiSeq Control Software. Raw sequence data (.bcl files) generated from Illumina HiSeq were converted into fastq files and demultiplexed using Illumina bcl2fastq 2.17 software. One mismatch was allowed for index sequence identification.

Assembly and annotation

The reads were filtered with TrimmomaticPE (v.0.39)63 with the following parameters: LEADING:30 TRAILING:30 MINLEN:120. The filtered reads were used for de novo assembly using the SPAdes genome assembler (v.3.13.1-1)83 with the following parameters: –careful–cov-cut-off 100. The resulting assemblies were then processed with AUGUSTUS (v.3.4.0)84 to obtain coding sequences and protein predictions. Augustus was executed using a gene model for Neurospora crassa to identify start and stop codons, introns and exons.

Genomic and phylogenomic analysis of N. intermedia strainsCore/pan genome analysis for N. intermedia

In addition to the seven N. intermedia draft genomes obtained in our study using short reads and the genome of N. intermedia strain #2613 obtained using long and short reads, we retrieved the genomes of 29 Neurospora spp. strains from a previous study28 and 5 strains which were obtained from the public GenBank database. These genomes were annotated following the protocol described above and their protein sequences were used for core genome calculation using BPGA85 with a sequence identity cut-off of 0.5; this analysis led to a set of 6,416 conserved orthologous proteins. The functional analysis of the resulting sets of core and unique genes was done with blast2go86 implemented in omicsBox v.3.029 (https://www.biobam.com/omicsbox).

The taxonomic affiliation of the N. intermedia strains used in this study was defined by a multilocus phylogenetic tree constructed with the set of conserved orthologous proteins found in their core genome. For each set of orthologues, the amino acid sequences were aligned28 and trimmed87; after this process, 6,353 alignments were kept. They were then concatenated to form a matrix with 3,203,239 sites in 6,353 partitions. An evolutionary model was calculated for each partition. Then a phylogenetic tree was calculated with IQtree2 (v.2.0.7)