Bringing up to date the toolkit for the catabolism of aromatic compounds in fungi: The unexpected 1,2,3,5-tetrahydroxybenzene central pathway.

Saprophytic fungi are able to catabolize many plant‐derived aromatics, including, for example, gallate. The catabolism of gallate in fungi is assumed to depend on the five main central pathways, i.e., of the central intermediates' catechol, protocatechuate, hydroxyquinol, homogentisate and gentisate, but a definitive demonstration is lacking. To shed light on this process, we analysed the transcriptional reprogramming of the growth of Aspergillus terreus on gallate compared with acetate as the control condition. Surprisingly, the results revealed that the five main central pathways did not exhibit significant positive regulation. Instead, an in‐depth analysis identified four highly expressed and upregulated genes that are part of a conserved gene cluster found in numerous species of fungi, though not in Aspergilli. The cluster comprises a monooxygenase gene and a fumarylacetoacetate hydrolase‐like gene, which are recognized as key components of catabolic pathways responsible for aromatic compound degradation. The other two genes encode proteins with no reported enzymatic activities. Through functional analyses of gene deletion mutants in Aspergillus nidulans, the conserved short protein with no known domains could be linked to the conversion of the novel metabolite 5‐hydroxydienelatone, whereas the DUF3500 gene likely encodes a ring‐cleavage enzyme for 1,2,3,5‐tetrahydroxybenzene. These significant findings establish the existence of a new 1,2,3,5‐tetrahydroxybenzene central pathway for the catabolism of gallate and related compounds (e.g. 2,4,6‐trihydroxybenzoate) in numerous fungi where this catabolic gene cluster was observed.

Our understanding of how fungi utilize gallate or other syringyl derivatives is, to date, scarce and mostly relies on unproven assumptions. This study demonstrates the existence of a new 1,2,3,5‐tetrahydroxybenzene central pathway for the catabolism of gallate and related compounds in numerous fungi. DUF3500 gene likely encodes a ring‐cleavage enzyme for the central intermediate. These findings have the potential to make a significant contribution to our understanding of key ecological and biotechnological processes in both fungi and bacteria.

Saprophytic fungi play an invaluable role in the turnover of the elements, including those present in highly abundant plant phenolic sources such as lignin and tannins (Mäkelä et al., [24]). Understanding the catabolism of aromatics, for example, gallate, a major intermediate of the metabolism of lignin constituents, esterified polysaccharides and hydrolysable tannins (Mäkelä et al., [24]), is crucial to decipher the ecological aspects of plant biomass degradation. This process ultimately occurs through the central pathways for the metabolism of key aromatic compounds. Countless peripheral pathways converge towards a small number of central intermediates that undergo ring cleavage in the central pathways. In fungi, there are five main described pathways of the central intermediates: catechol, protocatechuate, hydroxyquinol, homogentisate and gentisate (Martins et al., [27]), of which the corresponding catabolic genes are known (Gérecová et al., [8]; Holesova et al., [15]; Martins et al., [25]; Mingot et al., [30]; Semana & Powlowski, [37]). The five central pathways are present in the genomes of both Ascomycota and Basidiomycota, sharing high similarity, despite some noteworthy differences (Martins et al., [27]). A catabolic pathway for the central intermediate 3‐hydroxyanthranilate was recently discovered in fungi (Martins et al., [26]). The existence of other specific dioxygenases for gallate, hydroquinone, homoprotocatechuate and pyrogallol, as described for bacteria, remains to be verified in fungi, although mostly accepted that these intermediates are channelled into the five main central pathways (Ferrer‐Sevillano & Fernandez‐Canon, [6]; Gérecová et al., [8]; Patel et al., [32]).

The catabolism of gallate by Ascomycota is assumed to occur through three pathways but none is fully described (schematically represented in Figure 1). First, a path of initial reduction and metabolization to protocatechuate was proposed for Penicillium simplicissimum (Patel et al., [32]). Second, it was established for the yeast Blastobotrys adeninivorans an initial non‐oxidative decarboxylation of gallate to pyrogallol that undergoes intradiol ring cleavage, mediated by a catechol‐like dioxygenase, forming 2‐hydroxymuconate, which is then channelled to an undisclosed catabolic pathway for 4‐oxalocrotonate (Meier et al., [29]; Sietmann et al., [38]). The same pathway was suggested for Aspergillus and Penicillium species once they also accumulate pyrogallol (Gurujeyalakshmi & Mahadevan, [10]). Finally, direct ring cleavage was proposed for both A. niger and A. flavus, based on indirect pieces of evidence, including apparent gallate dioxygenase activity (Gurujeyalakshmi & Mahadevan, [10]; Watanabe, [40]). The catabolism of gallate in A. niger was further described to be independent of the protocatechuate branch of the 3‐oxoadipate pathway (Boschloo et al., [5]). Collectively, data on the catabolism of gallate in fungi suggest that different pathways may be used by closely related species, and possibly exhibit some redundancy.

To better understand the catabolism of gallate, in the present study, we resorted to A. terreus and A. nidulans as model fungi. Specifically, the transcriptional response of A. terreus growth on gallate was analysed by RNA‐seq (acetate was used as a control condition). This was complemented by targeted gene expression analyses in both fungi, functional studies of gene deletion mutants in A. nidulans, and analytical chemical identifications of specific intermediates. The data support the elucidation of the catabolism of gallate, and other similar compounds, through a new central pathway with initial hydroxylation (decarboxylating) and formation of the central intermediate 1,2,3,5‐tetrahydroxybenzene. Furthermore, genes essential for gallate utilization are organized in a cluster of genes present in several genomes, increasing the relevance of this novel central pathway for fungi in general.

Aspergillus terreus FGSC A1156 (equivalent NIH2624), Aspergillus nidulans FGSC A4, and additional strains (Table S1) asexual spores were harvested and maintained as frozen suspensions at −80°C (Hartmann et al., [11]). Liquid cultures were initiated with 106 spores/mL and incubated with orbital agitation (250 rpm) in the dark at 37°C. Batch cultivations were performed in 250‐mL Erlenmeyer flasks with a working volume of 50 mL. A low nitrogen minimal medium was used containing per litre 3 g NaNO3, 0.01 g ZnSO4·7H2O, 0.005 g CuSO4·5H2O, 0.5 g MgSO4·7H2O, 0.01 g FeSO4·7H2O and 0.5 g KCl. Filter sterilized salts were added to an autoclave sterilized 100‐mM potassium phosphate solution. The carbon sources were added either directly to the potassium phosphate solution for sterilization (60‐mM sodium acetate—control, pH 6.0) or into the mineral media after filter sterilization (e.g. 20 mM gallate, pH 4.0). Whenever relevant, the appropriate auxotroph nutrient requirements were added, and agar (10 g L−1) was used to gel the media. Cultures of gene replacement mutants for functional analysis in liquid medium were done in 100‐mM potassium phosphate buffer pH 5.0 incubated with orbital agitation (250 rpm) in the dark at 30°C using pre‐grown mycelium for 24 h in sodium acetate liquid minimal medium. Cultures in solid media (triplicates) were grown for 5 days at 30°C, in Ø 50‐mm plates.

Gallic acid monohydrate (ACS), 3,4,5‐trimethoxyphenol (>98.5%), ascorbic acid (99%) and sodium nitrate (extra pure) were purchased from Acros Organics; 3,4‐dihydroxybenzoic acid (97%) from Alfa Aesar; acetonitrile (>99.9%), ethyl acetate (>99.8%), diethyl ether (>99.5%) and methanol (>99.8%) from Fisher Chemical; sodium hydroxide (>98%) from José Manuel Gomes dos Santos; anhydrous sodium sulphate (99%), boron tribromide solution in anhydrous dichloromethane and anhydrous dichloromethane (99.7%) from Thermo Scientific; agar and resorcinol (99%) from Panreac; all other chemicals of the highest grade from Sigma‐Aldrich/Merck.

1,2,3,5‐tetrahydroxybenzene was synthesized as previously described (Paizs et al., [31]). A solution of BBr3 in anhydrous CH2Cl2 (1 M, 18.8 mL) was added, with stirring, to a solution of 3,4,5‐trimethoxyphenol in anhydrous CH2Cl2 (1.88 mmol, 10 mL) at −80°C under nitrogen during 1 h and the reaction was left to warm until room temperature. After overnight, the mixture was cooled to 0°C and distilled water (5 mL) was added. The CH2Cl2 was removed under nitrogen flow and the remaining mixture was extracted with ethyl acetate. The organic phase was dried with anhydrous NaSO4, filtered and concentrated under nitrogen flow. 1H NMR (800 MHz, DMSO‐d6): 5.72 (s, 2H); 13C NMR (201 MHz, DMSO‐d6): 94.34; 125.33; 146.77; 149.91.

The culture media or the respective ethyl acetate extracts were analysed by liquid chromatography (LC) for the identification and/or quantification of the aromatic compounds and their transformation products using a previously described method (Martins et al., [28]). Two other methods were used depending on the polarity of the compounds of interest. For alternative hydrophobic selectivity compared to the previously mentioned method (with a standard C18 column), a Synergi Polar‐RP column, 4 μm, 150 × 4.6 mm (Phenomenex), set at 26°C, was used. For the separation, the eluent (5‐mM potassium phosphate pH 3.05 with 2% methanol) was used for 6 min under isocratic conditions, followed by a 15‐min gradient of increasing percentage of methanol from 2% to 50%, at a flow rate of 0.8 mL·min−1 (HPLC method A). For increased retention of very hydrophobic compounds, a Synergi Hydro‐RP column, 4 μm, 250 × 4.6 mm (Phenomenex), set at 26°C, was used. The eluent, 20‐mM potassium phosphate pH 2.9, was used under isocratic conditions at a flow rate of 0.7 mL·min−1 (HPLC method B).

Samples were analysed by gas chromatography–mass spectrometry (GC–MS) with an Agilent GC (7820A) equipped with a single quadrupole Agilent (5977B) MS. The ethyl acetate extracts were dried under a soft nitrogen flow, and subsequently derivatized with N,O‐bis (trimethylsilyl)‐trifluoroacetamide containing 1% of trimethylchlorosilane and pyridine (5:1) for 30 min at 90°C. The separation of the analytes was carried out using an HP‐5MS column (30 m by 250 μm by 0.24 μm; Agilent) with the following ramp temperature: 80°C, 4°C/min until 290°C for 10 min. Full scan mode with a mass range of m/z 40–900, with a source at 230°C and electron impact ionization (EI+, 70 eV), was used for all samples, and acquisition delay was set at 240 s. Technical triplicates were analysed. Data acquisition was accomplished by MSD ChemStation (Agilent) and compounds were identified based on a spectral library (NIST 2017) and references relying on diagnostic ions distinctive of each derivative and its spectrum profile.

The extracts were also analysed by liquid chromatography–mass spectrometry (LC–MS) using a Q Exactive Focus Hybrid Quadrupole‐Orbitrap (Thermo Scientific). The separation was achieved in a Waters Acquity UPLC HSS T3 column (2.1 × 150 mm, 1.8 μm particle size), using a gradient of increasing percentage of 0.9% formic acid (FA) in acetonitrile (B) and 0.9% FA in ammonium formate 50 mM (pH 2.9) (C) and decreasing percentage of 0.9% FA in water (A). The flow rate was 0.2 mL·min−1 and the column was kept at 30°C. The sample injection volume was 1 μL. The data were acquired using Xcalibur software v.4.0.27.19 (Thermo Scientific). The method consisted of several cycles of full MS scans (R = 70,000; scan range = 75–1125 m/z) followed by three ddMS2 scans (R = 17,500; 20, 40, 60 NCE) in negative mode. External calibration was performed using LTQ ESI negative ion calibration solution (Thermo Scientific). The generated mass spectra were processed using Compound Discoverer 3.2 (Thermo Scientific) for small molecule identification.

Ethyl acetate extracts were dried under a soft nitrogen flow (approx. 5 mg) and solubilized in 500 μL of DMSO‐d6 for NMR analyses. All NMR spectra (1H, 1H−1H COSY, 1H−13C HSQC, 1H−13C HMBC) were acquired at 25°C using an Avance III 800 CRYO spectrometer (Bruker Biospin, Rheinstetten, Germany) in 5‐mm‐diameter NMR tubes. MestReNova, Version 11.04–18,998 (Mestrelab Research, S.L.) was used to process the raw data.

Total RNA extraction of liquid‐grown mycelia was performed essentially as previously described (Hartmann et al., [11]) using a Tissuelyser LT (Qiagen) for cell disruption. The quality and quantity of RNA were determined by capillary electrophoresis using an HS RNA kit and 5200 Fragment Analyser (Agilent). For single‐end RNA sequencing (RNA‐seq), libraries were generated using the Smart‐Seq2® mRNA assay (Illumina, Inc.) according to the manufacturer's instructions. Samples were indexed and sequenced on the Illumina NextSeq550 (20M reads per sample). Generated FastQ files were analysed with FastQC and any low‐quality reads were trimmed with Trimmomatic (Bolger et al., [4]). All libraries were aligned to A. terreus NIH2624 genome assembly (ASM14961.v1) with gene annotations (built 2010‐06) from Ensembl Fungi using HISAT2 v. 2.2.1 (Kim et al., [18]) with sensitive option and maximum intron length at 4000. Read counts were obtained using featureCounts v2.0.1 (Liao et al., [20]). All RNA‐seq experiments were carried out in three biological replicates. Differential expression analysis was performed using DESeq2 v.1.24.0 (Love et al., [21]). Transcript abundance was defined as transcripts per million (TPM). The genes that showed more than log2 1‐fold expression changes and above median TPM with adjusted p‐value <0.05 are defined as significantly differentially expressed genes in this analysis.

Two strategies were implemented to identify upregulated genes that were absent or incorrectly annotated in the current A. terreus NIH2624 genome assembly. To identify genes that could be incorrectly joined together in the annotation, the differentially expressed exons were determined as described above for gene features. Genes that showed simultaneously up‐ and downregulated exons were manually curated. To identify genes that were not in the current annotation, the entire genome was divided into 500 bp length sections, which were analysed as described above as if they were genes/exons. The upregulated differentially expressed sections (threshold set for upper quartile expression levels) not attributed to a currently annotated gene were manually curated. Automated gene structure prediction of new genes or poorly annotated genes was done using AUGUSTUS (Hoff & Stanke, [14]). Gene structure was manually curated according to RNA‐seq data and used in the transcriptome analysis (Table S2).

Protein domain entries were obtained from InterPro 90.0 either from predetermined data or using InterProScan (REST) service to search entries for the new gene predictions (Paysan‐Lafosse et al., [35]). The enrichment analyses were performed with the FunRich software v.3.1.3 (Pathan et al., [33]), with a p‐value <0.05 after Bonferroni correction. Functional characterization and metabolic pathway analysis was performed using the KEGG application KofamKoala (Aramaki et al., [1]). Metabolic gene clusters were identified within annotated genomes (GenBank or RefSeq) through sequence homology searches using cblaster v.1.3.12 (Gilchrist et al., [9]) using default filtering thresholds.

Total RNA was extracted as reported above, and cDNA synthesis followed a previously described protocol (Hartmann et al., [11]). Oligonucleotide pairs were designed using Primer‐BLAST (Ye et al., [42]) and supplied by STAB Vida (Oeiras, Portugal) (Table S3). The RT‐qPCR analysis was performed in a CFX96 Thermal Cycler (Bio‐Rad) using the SsoFast EvaGreen Supermix (Bio‐Rad), 250 nM of each oligonucleotide and the cDNA template equivalent to 10 ng of total RNA, at a final volume of 10 μL per well, in three biological replicates. The PCR conditions were: enzyme activation at 95°C for 30 s; 40 cycles of denaturation at 95°C for 5 s and annealing/extension at 60°C for 15 s and a melting curve obtained from 65°C to 95°C, consisting of 0.5°C increments for 5 s. Data analyses were performed using the CFX Manager software v.3.1 (Bio‐Rad). The expression of each gene was taken as the relative expression in pair‐wise comparisons of each condition relative to the acetate control. The expression of all target genes was normalized by the expression of the gamma‐actin gene ATEG_06973 or AN6542.

The genes AN0331, hmgA (AN1897), AN4576, AN10530 and AN12484 were selected for gene replacement with A. fumigatus pyrG in A. nidulans A1149 strain, performed according to a well‐established method (Szewczyk et al., [39]), essentially as previously described (Martins et al., [25]). Oligonucleotides were designed as described above (Table S3). Gene replacement was confirmed by PCR (Figure S1).

A comparative transcriptome analysis was performed here aiming to disclose the central pathways relevant to the catabolism of gallate in A. terreus. During gallate catabolism (approx. 30% consumed of 20 mM), 1545 encoding genes were found differentially upregulated and 1756 downregulated (see Figure S2 for principal component analysis and MA plots, and Table S4). Functional analysis of differentially expressed genes revealed that the global response to gallate in comparison with the control acetate differed in the regulation mechanisms (Figure S3A,B). Gallate‐induced response mainly occurs at the transcriptional regulation, and resorts more to fungi‐specific DNA binding transcription factors (IPR036864). On the contrary, acetate‐induced response apparently occurs more at the translational and post‐translational regulation and involves proteins containing RNA binding/modifying domains (IPR012340, IPR012677, IPR001412, IPR014729, IPR009000) and proteasome degradation domains (IPR001353, IPR029055, IPR023332). This may, however, be biased since more is unknown in the upregulated genes dataset. About 21% of upregulated genes do not have a known protein domain in comparison with only ca. 10% in the control (17.6% in all genes). Moreover, 72% of upregulated genes are currently not functionally annotated using KEGG, while only 36% of the downregulated ones are not. The major facilitator superfamily (MFS) of transporters (IPR036259) was 2‐fold enriched and represented 8% of all upregulated genes. The general response to gallate is relatively similar to that towards salicylate (e.g. enriched protein domain families; Martins et al., [26]), except for the secondary metabolism which is mostly unregulated in gallate but upregulated in salicylate.

The known central pathways, namely, the 3‐oxoadipate, the homogentisate and the 3‐hydroxyanthranilate pathways for the catabolism of aromatic hydrocarbons were either unregulated or showed upregulation with moderate transcripts abundance (Table S5). The 3‐oxoadipate pathway was globally the central pathway most responsive to gallate, especially the protocatechuate branch (Table S5). Quinate or shikimate catabolism (or hypothetically from an initial reduction of gallate) via protocatechuate is dependent on qutC (ATEG_00348), 3‐dehydroshikimate dehydratase, which mediates the reaction that connects the upper and the central catabolic pathways (e.g. protocatechuate branch), intersecting also with the anabolic shikimate pathway (Hawkins et al., [13]). The predicted quinate utilization cluster (ATEG_00348 to ATEG_00355), including qutC, was scarcely expressed. Similarly, a second quinate utilization‐like cluster (ATEG_07672_new to ATEG_07676), not present in, e.g., A. nidulans, was also scarcely expressed (Table S5). Other qutC homologues were unregulated, nearly excluding a path of initial reduction of gallate. Moreover, as previously reported for A. niger (Boschloo et al., [5]), we also observed that the protocatechuate branch is not essential for the catabolism of gallate in A. nidulans (see below).

The transcriptional reprogramming induced by gallate was confirmed by the upregulation and very high expression of several tannase genes: ATEG_02651 orthologue of AO090023000047 (Koseki et al., [19]), ATEG_03047 orthologue of An18g03570 (Ramírez‐Coronel et al., [36]) and ATEG_07937 orthologue of AO090103000074 (Hatamoto et al., [12]). This was anticipated since gallate, found in tannic acid comprising esters of polygalloylglucose or polygalloylquinic acid, is known as an inducer of tannase activity (Arentshorst et al., [2]; Mäkelä et al., [24]). The tannic acid transcriptional activator–repressor module composed of tanR (ATEG_10383) and tanX (ATEG_10381) (Arentshorst et al., [2]) was also found to be upregulated though more modestly expressed. Additional esterase genes (ATEG_08115b_new and ATEG_09944) and O‐methyltransferases (ATEG_07862, ATEG_09300 and ATEG_01928) were also highly expressed and upregulated, suggestive of a role in the metabolism of esters of gallate or their precursors (Table S5).

Some gene clusters of upregulated genes were identified but none contained the abovementioned tannases, esterases and O‐methyltransferases. The central pathways for the catabolism of aromatic compounds are every so often organized into clusters of genes in the genomes of fungi (Martins et al., [27]). The hypothesis that some gallate upregulated genes in A. terreus are in a gene cluster in other species was tested using the top 50 abundant transcripts with an FC > 2 (Figure 2; similar results were obtained using the top 100). Four hierarchical groups of gene clusters were identified; these comprise (i) sugar/inositol transporters genes (IPR003663), (ii) genes of the ethanol utilization pathway (discussed below), (iii) gallate metabolism auxiliary genes and (iv) four genes, including a putative monooxygenase and a fumarylacetoacetate hydrolase (FAH)‐like gene—hallmarks of a catabolic gene cluster of aromatic compounds. Further analysis of the last main hierarchical group (iv) revealed that three gene cluster's types, each having a distinct flavoprotein monooxygenase (CATH code 3.50.50.60, prototype reaction–aromatic hydroxylation, Paul et al., [34]) as a core gene, are present across species of Ascomycota (Figure 3) and only one in Basidiomycota (Figure S4). The genes of the three cluster types are in four loci in the genome of A. terreus, and four of them are highly expressed and upregulated (ATEG_04855, ATEG_04856, ATEG_05726b_new and ATEG_08221‐22_new; hereafter denominated the four catabolic genes of the gallate pathway) (Table S5). Two of the three flavoprotein monooxygenase genes were found upregulated, though ATEG_05726b_new was more highly expressed. The encoded protein of the A. niger orthologue (NRRL3_4659), can convert in vitro protocatechuate in 1,2,4‐trihydroxybenzene (hydroxyquinol), hence recently suggested to convert gallate to 1,2,3,5‐tetrahydroxybenzene (Lubbers et al., [22]). The other highly expressed genes present in the three types of clusters are a FAH‐like (ATEG_04856), a fungal conserved short protein gene (ATEG_04855) with no known domains (containing a conserved nucleophilic residue serine, a.a. 64, and only found in Dykaria) and a domain of unknown function DUF3500 gene (ATEG_08221‐22_new) that currently remains uncharacterized. The last two genes do not have any homologues in the genome, while the FAH‐like gene has eight homologues, although none of them upregulated. In bacteria, DUF3500 genes are also often located adjacent or in the vicinity of an FAH‐like gene not closely related to the characterized ones (data not shown). They are mostly only present in FCB, Proteobacteria, PVC and Terrabacteria groups, and Acidobacteriota phylum. Finally, in fungi, DUF3500 genes are mostly found in filamentous Ascomycota (i.e. Pezizomycotina subphylum) and Basidiomycota with macroscopic fruiting bodies (i.e. Agaricomycotina subphylum) (Paysan‐Lafosse et al., [35]).

The massive expression and upregulation of the ethanol utilization pathway (Table S5) constituted of alcohol dehydrogenase alcA (ATEG_09407 orthologue of alcA/AN8979), aldehyde dehydrogenase aldA (ATEG_05020 orthologue of aldA/AN0554) and transporter alcS1 (ATEG_03626‐27_new; with homology to alcS/AN8981 and AN8390; Flipphi et al., [7]) strongly indicate the production of their physiological inducer acetaldehyde (Flipphi et al., [7]). The regulator ATEG_03627 with homology to alcR/AN8978 was also positively regulated. Acetaldehyde can be produced from the catabolism of compounds as diverse as ethanol, amino acid threonine, catechol (meta‐cleavage) or phenylpropionic acids (aka hydroxycinnamic acids). Cytosolic pyruvate carboxylase pycA (ATEG_05433) was found to be upregulated, most likely contributing to the formation of acetaldehyde and induction of the ethanol utilization pathway. Additionally, acuA (acetyl‐CoA synthetase), acuC (transcriptional activator) and cpA (transport of acetate ions with homology to alcS) were upregulated despite the control condition acetate.

The four catabolic genes of the gallate pathway, but not the ethanol utilization pathway, were also found here upregulated in A. flavus in the presence of gallate, notwithstanding the co‐metabolic conditions used in this previous study (Zhao et al., [43]).

The gallate‐induced transcriptional response suggests that the main catabolic pathway for gallate in A. terreus starts with initial hydroxylation (decarboxylating) and formation of the central intermediate 1,2,3,5‐tetrahydroxybenzene. Processing of this intermediate through the central pathway results in the build‐up of end metabolites that trigger the induction of the ethanol utilization pathway and C2 metabolism. To confirm this hypothesis using gene deletion mutants, our option was to resort to A. nidulans for which a commercial strain capable of efficient recombination is readily available. Gene expression analysis first validated the results obtained in RNA‐seq of A. terreus (Figure 4A). Importantly, a similar gene regulation was observed for A. nidulans: the highly conserved orthologues of the four catabolic genes of the gallate pathway were upregulated (Figure 4B). None are in close vicinity to each other in the genome, despite synteny of the AN0331 loci with other aspergilli (Figure 3). These four genes showed a remarkable induction from the early culture timepoints in both aspergilli, in opposition to the other genes of the central pathways analysed. These results mostly verify that the same pathway is used in both species, furthering its relevance across aspergilli.

Based on these results, gene functional assays were performed using the four catabolic gene deletion mutants in A. nidulans. These mutants showed severely impaired growth in gallate and tannic acid solid media (Figure 5) demonstrating their essential role in the catabolism of gallate. Growth of deletion mutants in other aromatic compounds, namely, 2,3‐dihydroxybenzoate, phenylacetate, phloroglucinol, protocatechuate and resorcinol was unaffected (Figure 5 and Figure S5). This revealed that the four catabolic genes are most likely only essential to the catabolism of gallate and other compounds that share the same central pathway unrelated to those previously described. Growth in an isomer of gallate, 2,4,6‐trihydroxybenzoate, showed severe growth impairments like those imposed by gallate, except for the monooxygenase deletion mutant (ΔAN4576). This indicated that the central metabolite 1,2,3,5‐tetrahydroxybenzene was formed through hydroxylation (decarboxylating) of gallate and its isomer.

To further validate 1,2,3,5‐tetrahydroxybenzene as a new central intermediate, the metabolites accumulating after the addition of gallate or its isomer to acetate pre‐grown cultures of A. nidulans gene deletion mutants were analysed. Although the consumption rate of gallate seemed to be highly reduced, neither gallate nor its metabolites accumulated in incubations of the Δ gallate 1‐monooxygenase (ΔAN4576), suggestive of functional redundancy (Figures S6 and S7A,B). Transient accumulation of 1,2,3,5‐tetrahydroxybenzene in cultures of gallate and its isomer was observed for the Δ DUF3500 gene (ΔAN10530) only in the presence of supplemented ascorbic acid (Figure 6, see confirmation on the synthesis in Figure S8A,B), likely because it prevented autooxidation from occurring. 1,2,3,5‐tetrahydroxybenzene was also shown to be unstable in phosphate buffer even at low pH. The resulting autooxidation metabolites were mostly not identified by GC–MS except for 2‐furoic acid (Figure 6 and Figure S9). Accumulation of a single compound was detected in Δ no domains short protein (ΔAN0331); identified as 5‐hydroxydienelactone based on NMR and mass spectral data (Figures 6 and 7 and Figure S10). For the Δ FAH‐like gene (ΔAN12484), the transient accumulations of a major unidentified compound (calculated MW 223.04795 and putative formula C10H9NO5 by LC–MS; non‐aromatic by NMR; resolved as three analytes with the same molecular ion by GC after derivatization) and some other metabolites including acetopyruvate were noticed (Figure 6, Figures S11 and S12 and Table S7). FAH is a diversified superfamily of proteins (IPR011234) with catalytic functions of hydrolase, lyase and isomerase (Weiss et al., [41]). The nature of the molecule that transiently accumulated in cultures of this gene deletion mutant remains unresolved. It is also not possible to state whether it is a direct or indirect product, the latter being the result of redundant enzymatic activities. However, it may be hydrolysed to form, at least non‐specifically, acetopyruvate or a reduced form (putative 2‐hydroxy‐4‐oxopentanoic acid) as these were only identified in Δ FAH‐like cultivations (Figure S11).

In summary, the proposed pathway starts with an initial hydroxylation (decarboxylating) step, which is followed by ring‐cleavage (di)oxygenation and immediate (spontaneous) cyclization to create 5‐hydroxydienelactone. This compound is then activated to form a resulting nitrogenous adduct and ultimately hydrolysed to produce acetopyruvate or its reduced form (Figure 8).

In this study, we report the discovery of a novel central pathway in fungi responsible for the catabolism of gallate and related aromatic compounds using the central intermediate 1,2,3,5‐tetrahydroxybenzene (Figure 8). Surprisingly, this pathway differs from those described and/or proposed before for the catabolism of gallate in both bacteria and fungi (Lubbers et al., [23]). Based on this discovery, it is wrong to refer to gallate as a central metabolite for fungi; instead, gallate is one of several compounds channelled towards this novel central pathway. The pathway contains two genes encoding proteins having yet no known activities (i.e. DUF3500 and no domains short protein), which necessitates further investigation into their unusual mechanisms of action. However, based on the data, we hypothesize that the DUF3500 protein functions as a newly discovered ring‐cleavage enzyme, deserving focused analysis soon.

The intermediate 1,2,3,5‐tetrahydroxybenzene was previously reported in the catabolism of aromatic compounds in bacteria, specifically in the catabolism of gallate or phloroglucinol. It functions as a cocatalyst in the anaerobic conversion of pyrogallol to phloroglucinol, i.e., after initial non‐oxidative decarboxylation of gallate (Paizs et al., [31]), or as a hydroxylation intermediate in the catabolism of phloroglucinol (Armstrong & Patel, [3]). Only the last partially resembles the central pathway reported here, as acetopyruvate was also identified as a near‐end‐metabolite. They, however, may differ, because in the bacteria phloroglucinol catabolism proceeds through this pathway but this dependence was not observed here in A. nidulans (Figure 5). In a recent study, the derivative 6‐methoxy‐1,2,4‐trihydroxybenzene (6‐MeOTHB) of the central intermediate 1,2,3,5‐tetrahydroxybenzene was identified during the catabolism of syringate (also a methoxy derivative of gallate) in the Basidiomycota Phanerochaete chrysosporium (Kato et al., [17]). 6‐MeOTHB was converted in vitro, by hydroxyquinol 1,2‐dioxygenase, to 4‐hydroxy‐2‐methoxy‐cis, cis‐muconate (Kato et al., [17]) but the subsequent catabolic steps were unlooked. Syringate can be metabolized into 5‐hydroxyvanillate (Kato et al., [17]; Lubbers et al., [23]), which likely undergoes prompt conversion to either 6‐MeOTHB or 1,2,3,5‐tetrahydroxybenzene, thereby effectively excluding the formation of gallate. The DUF3500 and the no domains short protein genes are both absent in Phanerochaete genomes, despite being present in genomes of various species in the subphylum Agaricomycotina. These findings suggest that additional pathways for the metabolism of 1,2,3,5‐tetrahydroxybenzene or its methoxy derivatives await discovery. Finally, the DUF3500 gene is well represented in certain bacteria genomes, including, for example, Sphingomonas, where it may explain the redundancy observed in this bacteria; deletion of the known key gallate catabolic genes—LigAB (protocatechuate 4,5‐dioxygenase), DesZ (3‐O‐methylgallate 3,4‐dioxygenase) and DesB (gallate dioxygenase)—did not block gallate utilization (Kasai et al., [16]).

The newly discovered 1,2,3,5‐tetrahydroxybenzene central pathway is common in certain subphyla of fungi (i.e. those comprising the DUF3500 gene) and likely plays a significant role in the breakdown of aromatics, particularly those derived from syringyl‐rich lignin, including gallate and related compounds. These findings surely expand the current understanding of the toolkit of central pathways for the catabolism of aromatic compounds in fungi. Besides, it questions the accuracy of the generalized (yet unproven) assumption that gallate catabolism depends on the previously known five main central pathways. This novel central pathway has great ecological and biotechnological importance, not only for fungi but also potentially for bacteria. The lignin found in various economically significant plants, such as major grains like rice, wheat and maize, comprises a substantial portion of syringyl units (up to 60%). As a result, the future utilization of residues from these plants in biorefineries will yield significant quantities of syringyl derivatives. Therefore, the information on the new central pathway is crucial for comprehensive studies on the degradation of complex polymers like lignin and surely may impact the development of technologies aimed at deconstructing and converting them into added‐value chemicals.

Tiago M. Martins: Conceptualization (lead); formal analysis (equal); investigation (lead); methodology (lead); writing – original draft (lead). Artur Bento: Investigation (equal). Celso Martins: Investigation (equal). Ana S. Tomé: Investigation (equal). Carlos J. S. Moreira: Investigation (equal). Cristina Silva Pereira: Conceptualization (equal); project administration (equal); resources (equal); supervision (equal); writing – review and editing (equal).

The authors are grateful to Maria C. Leitão (ITQB‐NOVA Small Molecules Analysis) for the implementation and support in the chromatographic analyses. Mass spectrometry data were generated by the Mass Spectrometry Unit (UniMS), ITQB/iBET, Oeiras, Portugal. All members of the Silva Pereira lab are also thanked for useful discussions.

This work was financially supported by National Funds from Portugal2020 and FEDER under the project "BIOPINUS" (Ref. 13/SI/2020‐072630) and through Fundação para a Ciência e a Tecnologia (FCT) through MOSTMICRO‐ITQB R&D Unit (UIDB/04612/2020, UIDP/04612/2020) and LS4FUTURE Associated Laboratory (LA/P/0087/2020). TM and CM are grateful to FCT for the working contract financed by National Funds under Norma Transitória D.L. n.° 57/2016 and the Fellowship SFRH/BD/118377/2016 respectively. The NMR data were acquired at CERMAX, ITQB‐NOVA, Oeiras, Portugal with equipment funded by FCT.

The authors declare that they have no competing interests.

All relevant data are available in the manuscript. Supplementary figures and tables are provided to support the results presented in this manuscript. Gene transcript TPM values are contained in Table S4, and raw data with metadata were deposited in the Sequence Read Archive under Bioproject accession number PRJNA612036.

Not applicable.

GRAPH: Data S1.