W361R mutation in GaaR, the regulator of D‐galacturonic acid‐responsive genes, leads to constitutive production of pectinases in Aspergillus niger

Abstract Polysaccharides present in plant biomass, such as pectin, are the main carbon source for filamentous fungi. Aspergillus niger naturally secretes pectinases to degrade pectin and utilize the released monomers, mainly D‐galacturonic acid. The transcriptional activator GaaR, the repressor of D‐galacturonic acid utilization GaaX, and the physiological inducer 2‐keto‐3‐deoxy‐L‐galactonate play important roles in the transcriptional regulation of D‐galacturonic acid‐responsive genes, which include the genes encoding pectinases. In this study, we described the mutations found in gaaX and gaaR that enabled constitutive (i.e., inducer‐independent) expression of pectinases by A. niger. Using promoter‐reporter strains (PpgaX‐amdS) and polygalacturonic acid plate assays, we showed that W361R mutation in GaaR results in constitutive production of pectinases. Analysis of subcellular localization of C‐terminally eGFP‐tagged GaaR/GaaRW 361R revealed important differences in nuclear accumulation of N‐ versus C‐terminally eGFP‐tagged GaaR.

mutational analysis of this element (5′-CCNCCAA-3′) showed its requirement for induction in response to the presence of GA . The same GARE element was also identified in the promoter regions of GA-induced genes in B. cinerea (Zhang et al., 2016), indicating a conserved transcriptional control mechanism. A yeast one-hybrid screen was conducted in B. cinerea which resulted in the identification of a fungal GA-responsive Zn 2 Cys 6 type transcription factor named GaaR (Zhang et al., 2016). The A. niger GaaR orthologue was identified via its homology to the B. cinerea GaaR and was shown to be required for growth on GA and polygalacturonic acid (Alazi et al., 2016). GaaR was found to be essential for the GA-responsive expression of several pectinases, (putative) GAtransporters, and GA catabolic pathway enzymes (Alazi et al., 2016).
The gene encoding GaaR is present in the genomes of Ascomycetes belonging to the Pezizomycotina subdivision which includes members of Eurotiomycetes (Aspergillus, Penicillium, and Talaromyces spp.), Leotiomycetes (Botrytis and Oidiodendron spp.), Sordariomycetes (Neurospora, Myceliophthora, Trichoderma, and Fusarium spp.), and Dothideomycetes (Zymoseptoria and Aureobasidium spp.). The function of GaaR has not been studied yet in species apart from A. niger and B. cinerea. In N. crassa, the PDR-1 transcription factor is important for growth on PGA and pectin (Thieme et al., 2017).
Phylogenetically, PDR-1 is the ortholog of the A. niger rhamnose regulator RhaR. However, the ΔrhaR mutant of A. niger has no apparent growth defect on PGA or pectin (Ram et al., unpublished results; Gruben et al., 2014), indicating that expression of pectinases in A. niger and N. crassa is at least partially under control of different regulators.
Aspergillus niger mutants showing constitutive production of pectinases were previously obtained via a forward genetic screen using the promoter-reporter strain PpgaX-amdS ∆creA (JN29.2) (Niu et al., 2017). PpgaX-amdS ∆creA (JN29.2) is able to grow on acetamide as the sole nitrogen source only when the GA-responsive promoter of the pgaX (NRRL3_03144) gene is activated to express the amdS gene that encodes the acetamide catabolic enzyme acetamidase . In total, 64 mutants showing constitutive production of pectinases were isolated on plates containing minimal medium (MM) with glucose as the carbon source and acetamide as the sole nitrogen source. The genomes of five mutants were sequenced and it was found that they carry allelic mutations in a common gene, which was named the repressor of D-galacturonic acid utilization, gaaX (NRRL3_08194) (Niu et al., 2017). Transcriptome analysis revealed that deletion of gaaX results in the constitutive expression of pectinase genes (Niu et al., 2017). GaaX is suggested to inhibit GaaR function by stoichiometric binding to GaaR under noninducing conditions (Alazi et al., 2018;Niu et al., 2017). Interestingly, the genes encoding the repressor protein (GaaX) and the transcriptional activator (GaaR) are located side-by-side on the chromosome of A. niger as well as in other Pezizomycotina species (Niu et al., 2017). The clustering of genes encoding the GA-specific transcriptional activator (GaaR) and the repressor protein (GaaX) is similar to that of the activator-repressor modules qutA-qutR and qa- 1F-qa-1S involved in the utilization of the plant cell wall constituent quinic acid in Aspergillus nidulans and N. crassa, respectively. In both the quinic acid and the GA utilization systems, deletion of the repressor results in constitutive expression of target genes which requires the presence of the corresponding Zn 2 Cys 6 type transcriptional activator (Giles et al., 1985;Grant, Roberts, Lamb, Stout, & Hawkins, 1988;Lamb et al., 1996;Niu et al., 2017). Moreover, A. nidulans strains expressing multiple copies of qutA showed constitutive expression of the genes involved in quinic acid utilization (Lamb et al., 1996). Similarly, overexpression of gaaR in A. niger resulted in constitutive expression of the genes involved in the breakdown of pectin and utilization of GA (Alazi et al., 2018). These results indicate that the activity of GaaR is inhibited by GaaX under non-inducing conditions possibly via protein-protein interaction and that GAresponsive gene expression only requires the presence of active GaaR relieved/escaped from GaaX inhibition. This suggests that the inducer binds to GaaX resulting in a GaaX-unbound and active form of GaaR under inducing conditions. The catabolic intermediate 2-keto-3-deoxy-L-galactonate produced from GA was shown to induce the production of pectinases by A. niger which makes it reasonable to suggest that 2-keto-3-deoxy-L-galactonate binds to GaaX resulting in the dissociation of GaaX from GaaR, thereby releasing GaaR as an active transcription factor . Because overexpression of the activators involved in GA or quinic acid utilization has similar effects, it is likely that activation mechanism of both QutA/QutR and GaaR/GaaX is conserved and comparable.
Pectinases produced by A. niger are used industrially (Edwards & Doran-Peterson, 2012;Kashyap, Vohra, Chopra, & Tewari, 2001;Khan, Nakkeeran, & Umesh-Kumar, 2013;Toushik, Lee, Lee, & Kim, 2017), especially in food industry and in hydrolysis of plant biomass for the subsequent production of biofuel and high-value biopolymers. Inducer-independent production of pectinases by A. niger would enable the use of waste and residues from agriculture, forestry, and food industries as a cheap and sustainable feedstock for fungal cultivations. In this study, we describe mutations in gaaX and gaaR that result in constitutive production of pectinases by A. niger.
We show that a gain-of-function mutation in gaaR causing an amino acid change from tryptophan to arginine at position 361 (GaaR W361R ) leads to a constitutively active form of the GaaR transcription factor, leading to inducer-independent production of pectinases.

| Strains, media, and growth conditions
All strains used in this study are listed in Table S1. Media were prepared as described in Arentshorst, Ram, and Meyer (2012).
Filter-sterilized carbon source solutions were added after autoclaving MM containing agar. PGA and AP were autoclaved together with the medium. All growth experiments were performed in duplicate.
For enzymatic analysis, 300 ml shake flasks that include 100 ml MM (pH 5.8) containing 50 mM D-fructose and 0.003% yeast extract were inoculated with 7.5 × 10 7 freshly harvested spores and cultivated for 36 hr in a rotary shaker at 30°C and 250 rpm. Experiments were performed in duplicate.
For microscopic analysis of the localization of the eGFP-tagged GaaR or GaaR W361R proteins, 4 × 10 5 freshly harvested spores were inoculated on cover slips in Petri dishes that include 4 ml MM containing 0.003% yeast extract and 10 mM D-fructose or 2-keto-3-deoxy-L-galactonate, and grown at 30°C for approximately 21 hr. For each condition, two biological replicates were performed.

| Sequencing gaaX and gaaR genes from mutants displaying constitutive production of enzymes involved in PGA utilization
Sixty-four trans-acting mutants spontaneous or obtained after mild UV mutagenesis with constitutive production of enzymes involved in PGA utilization were obtained as previously described (Niu et al., 2017) on solid MM containing 50 mM glucose as the carbon source and 10 mM acetamide as the sole nitrogen source. Genomic DNA of 11 spontaneous and 53 UV mutants was extracted as described by Arentshorst, Ram, and Meyer (2012), and gaaX and/or gaaR genes were PCR-amplified using the primers gaaX_P5f and gaaX_GSP9r, or gaaRP7f and gaaRP8r, respectively (Table S2). The PCR fragments were sequenced in both directions using gaaX or gaaR sequencing primers, respectively (Table S2).

| Construction of the promoter-reporter strains expressing gaaR W361R
Protoplast-mediated transformation of A. niger and purification of the transformants were performed as described by Arentshorst, Ram, and Meyer (2012).
The gaaR W361R gene together with its 964-bp promoter and 992-bp terminator regions was amplified by PCR using the primers gaaRP5f and gaaRP6r (Table S2)  Transformants were selected on plates (Arentshorst et al., 2012) containing 10 mM acetamide as the sole nitrogen source. Correct gene replacements in strains JN130.4 and JN129.1 were verified by Southern blot and sequencing analyses. For sequence analysis, the gaaR locus was amplified using the primers gaaRP7f and gaaRP8r (Table S2) and JN130.4 and JN129.1 genomic DNA as template, and ligated into pJET1.2/blunt cloning vector (Thermo Fisher Scientific, Carlsbad, CA). The resulting plasmids were amplified in Escherichia coli and sequenced using gaaR sequencing primers. Integration of the gaaR W361R into the endogenous gaaR locus was confirmed via Southern blot analysis. Genomic DNA was digested overnight with NcoI or HindIII restriction enzymes. A 501-bp fragment containing the gaaR gene was PCR-amplified using the primer pairs listed in Table S2 with N402 genomic DNA as template and was used as a probe.

| Construction of strains expressing gaaR-eGFP or gaaR-(GA) 4 -eGFP
The plasmid pEA9 containing the PgaaR-gaaR-eGFP-TgaaR construct was created as follows: The eGFP gene and the 715-bp terminator of the gaaR gene were amplified by PCR using the primer pairs listed in Table S2 with the plasmid pFG029 (unpublished vector, containing PgpdA-eGFP-TtrpC) and N402 genomic DNA as template, respectively. eGFP and TgaaR were combined by fusion PCR using primers GaaR_GFP1F and GaaR_GFP3R. The gaaR gene without the stop codon and together with the 731-bp promoter region was PCR amplified using the primer pairs listed in Table S2. PgaaR-gaaR and eGFP-TgaaR were combined by fusion PCR using primers eGFP-gaaR-For and eGFP-gaaR-Rev, ligated into pJET1.2/blunt cloning vector (Thermo Fisher Scientific) and amplified in E. coli. Sequencing of pEA9 showed that a PCR error (T to C) has occurred in TgaaR at a distance of 519 bp downstream of the stop codon of the eGFP gene.
The plasmid pEA10 containing the PgaaR-gaaR-(GA) 4 -eGFP-TgaaR construct was created in a similar way to pEA9, except that PgaaR-gaaR-(GA) 4 was amplified using the primer GaaR_GFP6R adding a linker containing four repeats of glycine and alanine residues between GaaR and eGFP. Sequencing of pEA10 showed that a PCR error (C to T) has occurred in TgaaR at a distance of 430 bp downstream of the stop codon of the eGFP gene.
To create the strain EA29.14, pEA9 was co-transformed into strain JN36.1 (ΔgaaR) together with the plasmid pMA357 containing the A. nidulans amdS gene behind the A. nidulans gpdA promoter (Alazi et al., 2016). EA30.6 was created by co-transformation of pEA10 into strain JN36.1 (ΔgaaR) together with the plasmid pMA357. Transformants were selected on transformation plates containing acetamide as the sole nitrogen source.

| Construction of the ΔgaaR::AOpyrG deletion strain
SO1.1 (ΔgaaR::AOpyrG) was created using the split marker approach . 5′ and 3′ flanks of gaaR gene were PCR-amplified using the primer pairs listed in Table S2 with N402 genomic DNA as template. The Aspergillus oryzae pyrG gene was PCR-amplified as two fragments using the primer pairs listed in Table   S2 and the plasmid pAO4-13 (de Ruiter-Jacobs et al., 1989) as template. Split marker fragments with the AOpyrG selection marker were created by fusion PCR and used to transform the strain MA169.4 (Carvalho, Arentshorst, Kwon, Meyer, & Ram, 2010), resulting in the strain SO1.1. Proper deletion of gaaR was confirmed by diagnostic PCR (data not shown) and via Southern blot analysis.
pSO1.2 and pSO2.1 were sequenced to ensure no PCR errors have occurred and proper ligation and orientation of the fragments.
Repair fragments for CRISPR-Cas9 mediated targeted integration (see below) were obtained as follows: The PgaaR-gaaR W361R -TgaaR construct containing the 964-bp PgaaR and 992-bp TgaaR was excised from pSO1.2 using the restriction enzymes NotI and XbaI.
To create SO2.1, the DNA template for the in vitro synthesis of the guide RNA via the T7 promoter was amplified by PCR using the primers sgRNAP1 and OTL19 (Table S2), and the plasmid p426- et al., 2013) as template. The PCR consisted of the following reactions: Initial denaturation at 98°C for 30 s; 30 cycles of denaturation at 98°C for 5 s, annealing at 77°C for 5 s, extension at 72°C for 2 s; and final extension at 72°C for 30 s. The amplified DNA template was transcribed in vitro using the MEGAscript T7 Kit (Thermo Fisher Scientific). PgaaR-gaaR W361R -TgaaR repair fragment was co-transformed into strain SO1.1 (ΔgaaR::AOpyrG) together with the in vitro transcribed guide RNA and the plasmid pFC332 containing the cas9 gene and the hygromycin selection marker (Nødvig et al., 2015), yielding to SO2.1.
Transformants were selected on transformation plates containing hygromycin and uridine, and purified twice on MM containing GA, 5′-fluoroorotic acid, uridine, and hygromycin. pFC332 was cured by further purifying transformants twice without hygromycin selection on MM containing glucose and uridine. Transformants were not able to grow on MM containing glucose, uridine, and hygromycin after the first round of purification, indicating that pFC332 was successfully cured.
The DNA template for the in vivo synthesis of the guide RNA via the RNA polymerase III promoter was amplified in two parts by PCR using the primers Fw_LIC2 and Rev_P1, and Fw_P1 and Rev_LIC2 (Table S2) using the plasmid ANEp8_Cas9-gRNAalbA (manuscript in preparation Song et al., 2018) as template. The two PCR products were combined by fusion PCR using the primers Fw_LIC2 and Rev_ LIC2. The fusion PCR product was amplified by PCR using the primers for_pTE1 and rev_pTE1 to introduce PacI restriction sites at both ends, digested with PacI, and ligated into PacI cut opened pFC332, yielding to plasmid pTE1. pTE1 was co-transformed together with PgaaR-gaaR-eGFP-TgaaR or PgaaR-gaaR W361R -eGFP-TgaaR repair fragments into the strain SO1.1 (ΔgaaR::AOpyrG) to create EA31.1 or EA32.1, respectively. Transformants were selected on transformation plates containing hygromycin and uridine, and purified twice on MM containing GA, 5′-fluoroorotic acid and uridine. Transformants were not able to grow on MM containing GA, 5′-fluoroorotic acid, uridine, and hygromycin after the first round of purification, indicating that pTE1 was successfully cured.
The gaaR locus in SO2.1 and EA31.1 was amplified using the primers gaaRP7f and gaaRP8r, and the one in EA32.1 using the primers PgaaR_seq_P2 and TgaaR_seq_P2, and ligated into pJET1.2/blunt cloning vector (Thermo Fisher Scientific). The resulting plasmids were amplified in E. coli and sequenced using gaaR sequencing primers to confirm the presence of the mutated gaaR allele in SO2.1 and EA32.1, and the wild-type gaaR allele in EA31.1. Integration of the repair fragments into the endogenous gaaR locus in SO2.1, EA31.1, and EA32.1 was confirmed via Southern blot analysis ( Figure S3).

| Enzymatic analysis
Supernatants from shake flask cultures were obtained by filtration through glass microfiber filters (Whatman, Buckinghamshire, UK), and the filtrate was stored at −80°C. PGA plate assays were performed as described by Niu et al. (2017). Twenty-five microliters of supernatant from each culture were spotted on plates containing 0.2% PGA, and plates were incubated at 37°C for 24 hr before staining with 0.1% Congo Red (Sigma-Aldrich, Zwijndrecht) solution.

| Microscopy
The cover slips with adherent germlings were placed upside down on glass slides. The GFP fluorescence was excited using 488 nm laser line in a Zeiss Observer confocal laser scanning microscope (Zeiss, Jena, Germany). The images were analyzed with the ImageJ software (Abramoff, Magalhaes, & Ram, 2004). On each image, the exact same brightness and contrast adjustments were applied.

| Genetic characterization of A. niger mutants showing constitutive production of pectinases
In a previous study, we isolated 64 mutants showing constitutive production of pectinases (Niu et al., 2017). Genome sequencing of five mutants showed that all five mutants harbor a mutation in a putative repressor protein that represses the expression of pectinases under non-inducing conditions (Niu et al., 2017). We sequenced the gaaX gene in the remaining 59 mutants showing constitutive production of pectinases (Table S3, Figure S1). In total, 28 and nine mutants were found to carry nonsense and frameshift mutations, respectively. Twenty-three mutants were found to have missense mutations in the gaaX gene. These missense mutations were found throughout gaaX, indicating the importance of the different domains of GaaX for proper functioning. The nonsense or frameshift mutation closest to the C-terminus was the frameshift mutation V653G, indicating the importance of the last 45 amino acid residues for a functional GaaX, where the missense mutations L676P and V684K were also found ( Figure S1). Four mutants did not carry any mutations in gaaX. Sequencing the gaaR gene in these mutants F I G U R E 1 Schematic representation of the domains present in GaaR (a) and conservation of GaaR W361 in Aspergillus species (b) and other Ascomycetes (c). Domains in GaaR (NRRL3_08195) were identified using NCBI's CDD (Marchler-Bauer et al., 2015). Protein sequences homologous to the Aspergillus niger GaaR were retrieved using the blastp algorithm from NCBI (Altschul, Gish, Miller, Myers, & Lipman, 1990) against the nonredundant protein sequences database and were aligned using Clustal Omega ( The UV mutant PpgaX-amdS ∆creA gaaR W361R-UV (JN103.1) carries a missense mutation (W 361 to R) caused by codon change from TGG to CGG at a distance of 1285 bp from the start codon of gaaR.
The tryptophan361 residue lies in the fungal-specific transcription factor domain (cd12148) that spans the residues 139-518 in the 740 amino acid-long GaaR (Figure 1a), and is 100% conserved in the homologous GaaR sequences present in Aspergillus species (Figure 1b) and in other Ascomycetes analyzed belonging to the Pezizomycotina subdivision ( Figure 1c).

| Analyses of pectinase production in gaaR W361R
To show that the W361R mutation in GaaR is responsible for the constitutive production of pectinases, the GaaR gene containing the W361R mutation was introduced via a CRISPR/Cas9 approach into the endogenous gaaR locus in SO1.1, a strain carrying The expression of pgaX was previously shown to be induced only in the presence of an inducing carbon source, such as GA . Radial growth assays on plates containing acetamide showed that PpgaX-amdS gaaR W361R (JN130.4) and PpgaX-amdS ∆creA gaaR W361R (JN129.1) can grow on glucose or D-fructose, while their parental strains cannot (Figure 2d). This indicates that GaaR W361R can activate the expression of amdS via the pgaX promoter in an inducer-independent way, confirming the results obtained by PGA plate assays.
The expression of pgaX is known to be repressed strongly by glucose and mildly by D-fructose in a CreA-dependent way (Alazi et al., 2018;Niu et al., 2015). These previous results were confirmed as shown in Figure 2d, where growth of the promoter-reporter strains on plates containing acetamide and GA decreased when glucose was added to the growth media, and a short delay in growth and sporulation was observed when fructose was added as a repressing sugar.
The repressing effect of glucose or fructose was lost upon creA deletion, indicating that CreA is required for the repressing effect of glucose and fructose. To examine whether the CreA-mediated car-

| Subcellular localization of GaaR-eGFP and GaaR W361R -eGFP
We previously showed that eGFP-gaaR (EA19.2), a strain expressing the N-terminally eGFP-tagged gaaR, exhibits slightly reduced growth on GA compared to the reference strain (Alazi et al., 2018), indicating that N-terminal eGFP-tagging might affect the function of GaaR. To check whether C-terminal eGFP-tagging affects GaaR functionality, strains expressing C-terminally eGFP-tagged GaaR, with or without a linker between GaaR and eGFP, were created in a ∆gaaR background. While ∆gaaR (JN35.1) could not grow on GA, growth on GA was restored in both gaaR-eGFP (EA29.14) and gaaR-(GA) 4 -eGFP (EA30.6) ( Figure S4). This indicates that C-terminal eGFP-tagging does not affect the function of GaaR, independently of the presence of a linker between GaaR and eGFP.
We previously showed that the partially functional N-terminally eGFP-tagged GaaR in eGFP-gaaR (EA19.2) is localized mainly in the nucleus under both inducing (i.e., on GA) and non-inducing (i.e., on D-fructose) conditions (Alazi et al., 2018). Here, we repeated this experiment using 2-keto-3-deoxy-L-galactonate as an inducer instead of GA and confirmed a visual signal of eGFP-GaaR in the nucleus after induction by 2-keto-3-deoxy-L-galactonate or on the non-inducing carbon source D-fructose (Figure 3). C-terminally eGFP-tagged GaaR and GaaR W361R were also visualized mostly in the nucleus after growth in 2-keto-3-deoxy-L-galactonate. Surprisingly, this was not observed after growth in D-fructose. This result indicates that on D-fructose (under non-inducing and mildly repressing F I G U R E 3 Nuclear and cytoplasmic fluorescence intensities of the eGFP-tagged GaaR and GaaR W361R proteins. The eGFP-gaaR (EA19.2), gaaR-eGFP (EA36.1) and gaaR W361R -eGFP (EA39.1) strains were grown in MM containing 10 mM D-fructose or 2-keto-3-deoxy-L-galactonate for 21 hr. Example micrographs representing each condition are shown. Scale bar: 10 μm conditions) nuclear accumulation of C-terminally eGFP-tagged GaaR is restrained and suggests that the C-terminal tagging of GaaR with eGFP affects is localization under non-inducing conditions. The inability of the GaaR-eGFP protein to localize in the nucleus under non-inducing conditions is most likely also the reason why the GaaR W361R -eGFP protein does not exert its constitutive activity.

| D ISCUSS I ON
We previously isolated via a forward genetic screen A. niger strains displaying constitutive expression of pectinases (Niu et al., 2017).
The majority of the mutants in the collection (60 out of 64 strains) was found to contain mutations in GaaX and are probably all loss-offunction mutations as deletion of gaaX leads to the same phenotype (Niu et al., 2017). From the four mutants that did not carry mutations in gaaX, one mutant was found to possess a missense mutation leading to an amino acid change from tryptophan to arginine at position 361 in GaaR. The nature of the mutations in the three UV mutants with no mutations in gaaX or gaaR remains to be discovered. Introducing the GaaR W361R mutation in the wild-type genetic background and subsequent phenotypic analyses allowed us to conclude that this mutation results in constitutive activation of GaaR and therefore constitutive production of pectinases under non-inducing conditions.
To our knowledge, XlnR/Xyr1 is the only transcription factor involved in plant biomass degradation in which missense mutations leading to constitutive expression of enzymes are described.
The two mutations are the XlnR V756F mutation in A. niger, and the Xyr1 A824V mutation in Trichoderma reesei and orthologous mutations in N. crassa and Penicilium oxalicum, resulting in constitutive production of xylanases (Craig, Coradetti, Starr, & Glass, 2015;Derntl et al., 2013;Gao et al., 2017;Hasper, 2004;Hasper, Trindade, van der Veen, van Ooyen, & de Graaff, 2004). In protein sequence alignment, these two mutations are only three amino acids apart from each other ( Figure S6). Whereas the GaaR W361R mutation is located roughly half way of the 740 amino acid-long GaaR protein, the XlnR V756F mutation is located closer to the C-terminal in the 875 amino acid-long XlnR protein. The GaaR W361R mutation is predicted to be in the fungal-specific transcription factor domain (cd12148) which comprises amino acids 139-518 in GaaR. The cd12148 domain comprises amino acids 341-653 in A. niger XlnR and 368-717 in T. reesei Xyr1 (NCBI's CDD (Marchler-Bauer et al., 2015)), and the constitutive mutations in XlnR/Xyr1 are therefore outside of the predicted fungal-specific transcription factor domain in the Cterminal parts of XlnR and Xyr1, which has been suggested to be the activation domain. The XlnR/Xyr1 constitutive mutations are not located in a defined protein domain family, but the C-terminal parts of XlnR and Xyr1, including the amino acids V756 and the A824, respectively, have been suggested to be the activator domain (Derntl et al., 2013;Hasper et al., 2004). Alignment of protein sequences of XlnR and GaaR from A. niger, A nidulans, Aspergillus fumigatus and A. oryzae and Xyr1 of T. reesei using Clustal omega (Sievers et al., 2011) showed low sequence similarity in the amino acid residues surrounding the GaaR W361 residue between GaaR and XlnR/Xyr1 ( Figure S6). Whereas the W361 residue is fully conserved in GaaR in all four Aspergilli, only A. niger, A. oryzae, and T. reesei had a W residue positioned at the aligned protein sequence of XlnR/Xyr1 ( Figure S6).
Of note is that sequence similarity surrounding this region of XlnR/ Xyr1 is weak, making it possible that this conservation has no biological relevance. To test its biological relevance, it will be of interest to examine whether mutations in XlnR of A. niger at position W639 would also result in a constitutive phenotype. Protein sequence alignment revealed that the V756 and A824 residues are fully conserved in the XlnR/Xyr1 proteins of the four Aspergilli and T. reesei, but not strictly conserved between XlnR/Xyr1 and GaaR ( Figure S6).
Because of the low level of conservation between the regions and amino acids that result in constitutive activation of the respective transcription factors, it is reasonable to suggest that the mechanisms leading to the activation of GaaR and XlnR/Xyr1 are different. We identified in the A. niger genome four transcription factor/repressor modules (Niu et al., 2017). For two modules, we know that they are involved in either GA or quinic acid utilization (Niu et al., 2017;Ram et al., unpublished results). Analysis of the knockouts for the two remaining genes encoding the repressors, and subsequent phenotypic and transcriptomic analysis has shown that the two repressors are not involved in the expression of genes encoding xylanases (Alazi and Ram, unpublished results). Therefore, the activity of XlnR is likely not controlled by a GaaX-like repressor protein.
The activity of GaaR is proposed to be inhibited by GaaX under non-inducing conditions via protein-protein interactions (Niu et al., 2017). Pectinase genes are expressed under inducing conditions (i.e., in the presence of inducer), or under non-inducing conditions (i.e., in the absence of inducer) when gaaX is deleted or gaaR is overexpressed (Alazi et al., 2018;Martens-Uzunova & Schaap, 2008;Niu et al., 2017). These findings indicate that GA-responsive gene expression only requires the presence of active GaaR relieved/escaped from GaaX inhibition but does not require the presence of the inducer. This suggests that the inducer binds to GaaX under inducing conditions resulting in a GaaX-unbound and active form of GaaR.
Therefore, we propose that W361R in GaaR disrupts the interaction between GaaX and GaaR under non-inducing conditions. Both PpgaX-amdS gaaR W361R and PpgaX-amdS ∆gaaX mutants were found to produce pectinases under non-inducing conditions (Figure 2a and   b). The qualitative phenotypic similarity of these strains resulting from the W361R mutation in GaaR or the absence of GaaX, respectively, suggests that the GaaR-GaaX interaction is disturbed in the GaaR W361R mutant. In future experiments, the proposed proteinprotein interaction of GaaR and GaaX and mutant variants thereof will be studied further in detail. Initial attempts to express GaaR and GaaX in E. coli to perform interaction studies have been unsuccessful and will be subject for further research. In this respect, it would also be of interest to isolate and identify additional mutations in GaaR that lead to constitutive expression of pectinases to identify regions in GaaR that are necessary for the proposed interaction with GaaX.
GaaX of A. niger, QutR of A. nidulans, and qa-1S of N. crassa are all multidomain repressor proteins with sequence similarity to each other and to the three C-terminal domains of AROM, a large pentafunctional protein involved in the shikimate pathway. This suggests that these repressors share a common evolutionary origin. The last three domains of the AROM protein encode the shikimate kinase, 3-dehydroquinate dehydratase, and shikimate dehydrogenase enzymes (Lamb et al., 1996). The sequence similarity of the repressor domains to AROM is low (between 20% and 30% identity), suggesting that the enzymatic functions that are present in AROM are lost in the repressors. Loss-of-function mutations (nonsense and missense mutations) were found throughout all three domains of GaaX (shikimate kinase, 3-dehydroquinate dehydratase, and shikimate dehydrogenase domains) ( Figure S1). The 22 missense mutations could especially be valuable for structure/function relationships to predict which domains exert different function as they might affect the proposed protein-protein interaction between GaaX and GaaR. It is however difficult to predict whether a missense mutation affects the overall secondary/tertiary structure of the protein or whether a mutation affects the interaction without affecting its structure.
We identified a hot spot for mutations in the region 190-194 in which four mutants were found carrying mutations in three amino acid positions. This region could either be important for GaaR-GaaX interaction or could be important for the structure of GaaX.

Alignment of protein sequences homologous to GaaX present in 18
Pezizomycetes species using Clustal omega (Sievers et al., 2011) revealed that these four mutations were found in three residues that are highly conserved ( Figure S7). Several other missense mutations were found in conserved residues and are indicated in Figures S1 and S7. Relatively more missense mutations were found in the shikimate dehydrogenase domain (one mutation per 24 amino acids) ( Figure S1) compared to the shikimate kinase (one mutation per 76 amino acids), or 3-dehydroquinate dehydratase domains (one mutation per 37 amino acids), indicating that the shikimate dehydrogenase domain might be more important in relation to binding to GaaR than the other domains.
In N. crassa, dehydroquinic acid is suggested to be the physiological inducer of the quinic acid catabolic pathway genes, as mutants that accumulate or cannot produce dehydroquinic acid showed induced or diminished expression of these genes, respectively (Giles, Partridge, Ahmed, & Case, 1967 (Huiet & Giles, 1986). Hawkins, Lamb, Moore, Charles, and Roberts (1993) have proposed that the QutR domain homologous to the AROM 3-dehydroquinate dehydratase can recognize and bind to dehydroquinic acid produced from quinic acid, but does not react on it enzymatically in A. nidulans. For A. niger, accumulation of 2-keto-3-deoxy-L-galactonate in ΔgaaC results in hyper-induction of GA-responsive genes and the inducer 2-keto-3-deoxy-L-galactonate is likely to interact with the GaaX repressor protein . The GA catabolic pathway enzyme GaaC is an aldolase and converts 2-keto-3-deoxy-L-galactonate to pyruvate and L-glyceraldehyde.
Both the aldolase domain of GaaC and the 3-dehydroquinate dehydratase domain of GaaX belong to the same Aldolase Class I_superfamily (NCBI's CDD Marchler-Bauer et al., 2015). These findings suggest that 2-keto-3-deoxy-L-galactonate could bind to the GaaX domain homologous to the AROM 3-dehydroquinate dehydratase. However, direct protein-metabolite interaction studies are required to determine which part (or parts) of the repressor protein interacts with the inducer.
N-terminally eGFP-tagged GaaR was previously shown to mainly localize in the nucleus under both inducing and noninducing conditions but was unable to fully complement the growth of ∆gaaR when grown on inducing carbon sources (Alazi et al., 2018). C-terminally eGFP-tagged GaaR/GaaR W361R accumulated in the nucleus under inducing conditions and complemented the growth of ∆gaaR when grown on inducing carbon sources.
Based on these observations, we conclude that C-terminally eGFP-tagged GaaR/GaaR W361R is functional under inducing conditions. However, prominent nuclear eGFP-fluorescence signal was not observed in the strains producing the C-terminally eGFPtagged GaaR/GaaR W361R under non-inducing conditions. This result indicates that C-terminal eGFP-tagging restrained nuclear accumulation of GaaR/GaaR W361R under non-inducing conditions, explaining the inability of GaaR W361R -eGFP to constitutively activate the production of pectinases, as GaaR W361R can. In B. cinerea, expression of the C-terminally GFP-tagged BcgaaR via the strong oliC promoter resulted in higher nuclear GFP fluorescence intensity under both inducing and non-inducing conditions (Zhang et al., 2016). This indicates that nuclear accumulation of overexpressed and wild-type BcGaaR is not restrained by C-terminal GFP-tagging under non-inducing conditions, unlike the GaaR-eGFP/GaaR W361 -eGFP expressed from its endogenous promoter.
Constitutive production of plant biomass degrading enzymes like pectinases is interesting for the industrial production of these enzymes because it allows the production independently of the feedstock. By understanding the transcriptional regulatory mechanism of pectinases, several approaches have been conducted leading to inducer independent and abundant production of pectinases, such as deletion of GaaX (Niu et al., 2017), overexpression of GaaR from a constitutive promoter (Alazi et al., 2018), or accumulation of the inducer . Here, we show that a mutation in GaaR can also lead to constitutive production of pectinases. Combination of these approaches can be considered to further boost pectinase production in A. niger. The conservation of GaaR and GaaX in other fungi also allows improved pectinase production in other industrially important filamentous fungi.

ACK N OWLED G M ENTS
EA and JN were supported by grants from BE-Basic (Flagship 10) and the China Scholarship Council, respectively. This work was in part supported by the Natural Sciences and Engineering Research Council of Canada Strategic Industrial Biocatalysis Network.