Discovery and Functional Analysis of a Salicylic Acid Hydroxylase from Aspergillus niger

Nonrenewable petroleum sources are being depleted, and therefore, alternative sources are needed. Plant biomass is one of the most abundant renewable sources on Earth and is efficiently degraded by fungi.

ABSTRACT

Salicylic acid plays an important role in the plant immune response, and its degradation is therefore important for plant-pathogenic fungi. However, many nonpathogenic microorganisms can also degrade salicylic acid. In the filamentous fungus Aspergillus niger, two salicylic acid metabolic pathways have been suggested. The first pathway converts salicylic acid to catechol by a salicylate hydroxylase (ShyA). In the second pathway, salicylic acid is 3-hydroxylated to 2,3-dihydroxybenzoic acid, followed by decarboxylation to catechol by 2,3-dihydroxybenzoate decarboxylase (DhbA). A. niger cleaves the aromatic ring of catechol catalyzed by catechol 1,2-dioxygenase (CrcA) to form cis,cis-muconic acid. However, the identification and role of the genes and characterization of the enzymes involved in these pathways are lacking. In this study, we used transcriptome data of A. niger grown on salicylic acid to identify genes (shyA and crcA) involved in salicylic acid metabolism. Heterologous production in Escherichia coli followed by biochemical characterization confirmed the function of ShyA and CrcA. The combination of ShyA and CrcA demonstrated that cis,cis-muconic acid can be produced from salicylic acid. In addition, the in vivo roles of shyA, dhbA, and crcA were studied by creating A. niger deletion mutants which revealed the role of these genes in the fungal metabolism of salicylic acid.

IMPORTANCE Nonrenewable petroleum sources are being depleted, and therefore, alternative sources are needed. Plant biomass is one of the most abundant renewable sources on Earth and is efficiently degraded by fungi. In order to utilize plant biomass efficiently, knowledge about the fungal metabolic pathways and the genes and enzymes involved is essential to create efficient strategies for producing valuable compounds such as cis,cis-muconic acid. cis,cis-Muconic acid is an important platform chemical that is used to synthesize nylon, polyethylene terephthalate (PET), polyurethane, resins, and lubricants. Currently, cis,cis-muconic acid is mainly produced through chemical synthesis from petroleum-based chemicals. Here, we show that two enzymes from fungi can be used to produce cis,cis-muconic acid from salicylic acid and contributes in creating alternative methods for the production of platform chemicals.

INTRODUCTION

Salicylic acid (2-hydroxybenzoic acid) plays an important role in the regulation of the plant defense response against pathogens (1, 2). Many pathogenic fungi are able to degrade salicylic acid to manipulate the plant defense response. However, nonpathogenic fungi also have the ability to degrade salicylic acid. In several Aspergillus species, two salicylic acid metabolic pathways resulting in the formation of catechol have been reported (3–6). In the first pathway, salicylic acid is decarboxylated to catechol by salicylate hydroxylase (decarboxylating) (ShyA). In the second pathway, salicylic acid is 3-hydroxylated to 2,3-dihydroxybenzoic acid, which is converted to catechol through decarboxylation by 2,3-dihydroxybenzoic acid decarboxylase (DhbA) (7). The aromatic ring of catechol is cleaved by a catechol 1,2-dioxygenase (CrcA), resulting in the formation of cis,cis-muconic acid. cis,cis-Muconic acid is further converted through the oxoadipate pathway to acetyl coenzyme A (acetyl-CoA) and succinate, which enter the tricarboxylic acid cycle. Current interest in this pathway consists of the fact that cis,cis-muconic acid can be used as a platform chemical that can serve as a precursor to adipic acid and terephthalic acid. These compounds are used for the synthesis of nylon and other valuable materials, such as polyethylene terephthalate (PET), polyurethane, resins, and lubricants (8–10). Currently, cis,cis-muconic acid is mainly produced through chemical synthesis from petroleum-based chemicals (11). However, several methods have been developed to produce biobased cis,cis-muconic acid using microorganisms (8, 12–15).

In the ascomycete fungi Fusarium graminearum and Epichloë festucae, salicylate hydroxylases (decarboxylating) have been identified (16–18), while in Aspergillus nidulans, four genes were suggested to encode salicylate hydroxylase (decarboxylating), DhbA, salicylate 3-hydroxylase, and CrcA, respectively, but their function has not been confirmed (5). In A. niger, DhbA has been enzymatically characterized, but its role in aromatic metabolism has not been studied (7). In addition, four intradiol dioxygenases have been characterized from A. niger, three of which are able to cleave catechol to cis,cis-muconic acid (19, 20). One of these intradiol dioxygenases was identified as hydroxyquinol 1,2-dioxygenase (HqdA) (20). Currently, the in vivo role of catechol 1,2-dioxygenase remains to be proven.

In this study, we identified the genes encoding ShyA, DhbA, and CrcA of A. niger and demonstrated that recombinant ShyA and CrcA together can efficiently convert salicylic acid into cis,cis-muconic acid through catechol as an intermediate. Deletion of shyA, dhbA, and crcA in A. niger resulted in reduced growth on salicylic acid, 2,3-dihydroxybenzoic acid, and catechol, respectively, confirming their in vivo roles.

RESULTS

Identification of the salicylic acid hydroxylase, 2,3-dihydroxybenzoic acid decarboxylase, and catechol 1,2-dioxygenase from A. niger.

To identify genes involved in the metabolism of salicylic acid, transcriptome data of A. niger grown on salicylic acid for 2 h was used as reported previously (21). In the A. niger NRRL3 genome, 32 genes are annotated as putative salicylate 1-monooxygenase, but only two genes, NRRL3_9723 and NRRL3_43, were highly induced (fold change ≥ 4) by salicylic acid compared to the control (see Table S1 in the supplemental material). NRRL3_9723 showed high amino acid similarity to Shy1 from F. graminearum (FgShy1) (16) (65.1%) and to a putative salicylate hydroxylase (AN2114) from A. nidulans (5) (77.3%), while NRRL3_43 is less similar to these two enzymes (38.8% and 31.6%, respectively). In addition, NRRL3_9723 has a 37.2% sequence similarity to Pseudomonas putida NahG, while NRRL3_43 is 25.3% similar. Therefore, NRRL3_9723 was selected as the putative ShyA of A. niger.

BLASTP analysis of a suggested salicylate 3-hydroxylase from A. nidulans (5) (AN7418) revealed 21 homologs in A. niger. While NRRL3_43 did not have the highest amino acid similarity to AN7418, it was the only gene that was induced by salicylic acid (Table 1 and Table S2). All three intradiol ring cleavage dioxygenases (HqdA, NRRL3_4787, and NRRL3_5330) were induced by salicylic acid compared to the control. The highest induction was observed for NRRL3_4787, which was therefore selected for further investigation as the putative CrcA of A. niger (Table 1).

TABLE 1.

Expression of candidate salicylic acid metabolic genes on salicylic acid compared to a no-carbon-source control^a

JGI database gene ID	Annotation according to JGI database	Described protein	FPKM		Deseq2 fold change	P valueb
			SA	NC
NRRL3_9723	FAD-dependent oxidoreductase		225.70	11.14	20.03	0.000
NRRL3_43	FAD-binding domain-containing protein		1350.53	134.09	9.84	0.000
NRRL3_4385	Amidohydrolase	2,3-Dihydroxybenzoate decarboxylase (DhbA)	2903.69	1406.88	2.03	0.041
NRRL3_2597	FAD-binding domain-containing protein		0.69	0.25	2.64	0.103
NRRL3_4787	Intradiol ring cleavage dioxygenase		2804.74	343.56	8.00	0.001
NRRL3_5330	Intradiol ring cleavage dioxygenase		46.85	15.05	2.98	0.000
NRRL3_2644	Intradiol ring cleavage dioxygenase	Hydroquinol 1,2-dioxygenase (HqdA)	77.16	25.25	3.01	0.000
NRRL3_8551	FAD-binding domain-containing protein	p-Hydroxybenzoate-m-hydroxylase (PhhA)	16.96	69.70	0.24	0.000
NRRL3_1405	Intradiol ring cleavage dioxygenase	Protocatechuic acid 3,4-dioxygenase (PrcA)	8.61	59.02	0.15	0.001

^aAbbreviations: ID, identifier; FPKM, fragments per kilobase per million; SA, salicylic acid; NC, no-carbon-source control.

^bP values were calculated using Deseq2 software (51).

Colorimetric assay revealed salicylate hydroxylase activity of ShyA on salicylic acid.

To confirm the salicylate hydroxylase activity of the candidate ShyA, Escherichia coli strains that produce recombinant ShyA were inoculated on plates containing 1 mM salicylic acid together with an E. coli strain containing the empty vector (negative control). After 2 days, plates containing the recombinant strains (ShyA.1 and ShyA.2) turned brown, while the medium of the empty vector control strain remained yellow (Fig. 1a). Degradation of salicylic acid was further determined by a colorimetric assay. This degradation resulted in the purple coloration of salicylic acid in the negative-control plate, while the plates with the recombinant ShyA E. coli strains turned black (Fig. 1a).

FIG 1.

Expression of A. niger shyA in two independent E. coli strains, ShyA.1 and ShyA.2. (a and b) LB plates (a) and liquid cultures (b) contained 1 mM salicylic acid and were incubated for 2 days. An E. coli strain containing the empty vector without an insert was used as a control. After incubation, the plates were stained with FeCl₃, coloring salicylic acid purple and catechol black. (c) Salicylic acid concentrations in culture medium were determined by HPLC. Error bars represent the standard deviations of three biological replicates. Asterisks indicate significant differences in salicylic acid concentration compared with the empty vector control (Student’s t test, P value ≤ 0.01).

A similar experiment was performed in liquid cultures, resulting in the brown color formation in the two strains expressing shyA (Fig. 1b), indicating conversion of salicylic acid to catechol (Fig. 1c). These results confirm that NRRL3_9723 encodes the salicylate hydroxylase ShyA.

Production of cis,cis-muconic acid using recombinant ShyA and CrcA.

To confirm the role of ShyA and CrcA in cis,cis-muconic acid production, both were produced in E. coli. The E. coli strain expressing hqdA was obtained from Lubbers et al. (20). DhbA was not included since it has been previously characterized (7). Cell-free extracts of E. coli ShyA.1, CrcA, HqdA, and the empty vector control strain were incubated with 0.25 mM salicylic acid at 30°C overnight, after which the reaction mixtures were analyzed by high-performance liquid chromatography (HPLC).

Salicylic acid (0.25 mM) was converted to catechol (0.12 ± 0.04 mM) by the cell extracts from the shyA expression strain, whereas for the empty vector control and cell extracts of crcA and hqdA expression strains, no catechol was detected (Fig. 2). The cell extracts from the crcA or hqdA expression strain were verified to convert catechol to cis,cis-muconic acid (Fig. S1).

FIG 2.

Conversion of salicylic acid by cell extracts of the E. coli strains producing ShyA, CrcA, and HqdA from A. niger. Black, light gray, and dark gray bars represent salicylic acid, catechol, and cis,cis-muconic acid, respectively. The concentration of cis,cis-muconic acid was slightly overestimated due to the impurity of the standard. Error bars represent the standard deviations of three biological replicates. Asterisks indicate significant differences in salicylic acid concentration compared with the empty-vector control (Student’s t test, P value ≤ 0.01).

Incubation of salicylic acid by a combination of cell extracts from the shyA with the crcA or the hqdA overexpression strains resulted in the formation of cis,cis-muconic acid (Fig. 2), demonstrating that the combination of these enzymes is sufficient for this biochemical conversion. Almost no salicylic acid or catechol was detected, indicating that the conversion was highly efficient.

Biochemical characterization of purified ShyA.

To investigate the biochemical properties of ShyA, the enzyme produced in E. coli was purified using immobilized-metal affinity chromatography. The calculated molecular mass of ShyA-HIS was 49.2 kDa and corresponded with the band observed by SDS-PAGE (Fig. S2). The optimal pH and temperature for activity were determined to be pH 5.6 to 6.0 (Fig. 3a) and 45 to 50°C (Fig. 3b), respectively. The kinetic profile of ShyA, obtained while only varying the salicylic acid concentration at a constant 1 mM concentration of the second substrate (NADH), did not follow Michaelis-Menten kinetics. It revealed an apparent V_max of 75.6 ± 5.3 μmol/min/mg protein and a K_0.5 of approximately 138.8 ± 12.8 μM (Fig. 3c).

FIG 3.

Effects of pH and temperature on ShyA activity with 1 mM salicylic acid. (a). ShyA activity at 30°C after 15 min under different pH conditions; (b) ShyA activity at pH 6.0 after 15 min of incubation at different temperatures. Error bars represent the standard deviations between three replicates. (c) Rate of reaction of ShyA with salicylic acid. The assay was performed at pH 6.0 and 30°C. Error bars represent the standard deviations between three experiments.

Salicylate hydroxylases are known to perform NADH oxidation reactions in the presence of pseudosubstrates, like benzoate, without hydroxylating such a compound (22, 23). In order to distinguish between the oxidase and hydroxylase activities, the reaction mixtures were analyzed by HPLC (Table 2). ShyA was active toward salicylic acid, 4-aminosalicylic acid, 2,3-dihydroxybenzoic acid, and gentisic acid, resulting into the formation of catechol, 4-aminocatechol, pyrogallol, and hydroxyquinol, respectively.

TABLE 2.

Conversion of benzoic acids by ShyA^a

^aThe reaction conditions using purified enzyme were as follows: 1 mM substrate, McIlvaine buffer (pH 6.0), 1 mM NADH, incubated at 30°C for 1 h. Conversion values were determined by HPLC. The structural differences of the substrates are indicated by the residues attached to the aromatic ring. A value of 0 indicates that there was no conversion after 16 h of incubation.

ShyA, DhbA, and CrcA are involved in the fungal salicylic acid pathway.

To verify that the candidate genes encode the enzymes involved in the in vivo salicylic acid metabolic pathway of A. niger, six deletion mutants, the ΔshyA, ΔdhbA, ΔcrcA, Δ43 (putative salicylate hydroxylase), Δ2597 (putative salicylate 3-hydroxylase), and Δ5330 (putative catechol 1,2-dioxygenase) mutants, were created. Cultivation on several aromatic compounds demonstrated that the growth of the ΔshyA mutant was reduced on salicylic acid but not on the other aromatic compounds tested (Fig. 4), confirming its role as the salicylate hydroxylase of A. niger. Growth of the ΔdhbA mutant was reduced on 2,3-dihydroxybenzoic acid but not on salicylic acid or catechol, confirming its metabolic role. Deletion of crcA resulted in growth similar to that of the no-carbon-source control on salicylic acid, 2,3-dihydroxybenzoic acid, and catechol, while deletion of NRRL3_5330 or hqdA did not result in a phenotype, indicating that crcA encodes the catechol 1,2-dioxygenase of A. niger. Growth of the Δ43, Δ5330, and Δ2597 mutants did not result in any phenotypes on the tested aromatic compounds.

FIG 4.

Growth profile of the A. niger ΔshyA, ΔdhbA, ΔcrcA, ΔhqdA, Δ43, Δ2597, and Δ5330 deletion mutants and the reference strain on aromatic compounds. Phenotypes were examined after 7 days at 30°C. Fructose and a no-carbon-source condition served as growth controls.

When the plates containing salicylic acid, 2,3-dihydroxybenzoic acid, and catechol were stained with FeCl₃ (Fig. 5), the plates containing salicylic acid turned purple and halos were observed for the reference, ΔdhbA, ΔhqdA, Δ43, Δ5330, and Δ2597 strains. No halos were detected for the ΔshyA and ΔcrcA mutants, indicating that salicylic acid is not converted in these strains. Plates containing 2,3-dihydroxybenzoic acid turned black after staining, and no halos were observed when dhbA and crcA were deleted. Catechol plates turned black and no halo was produced when crcA was deleted. These results confirmed the phenotypes of the growth profile (Fig. 4).

FIG 5.

Growth profile of the deletion strains and the reference on MM agar plates containing salicylic acid, 2,3-dihydroxybenzoic acid, and catechol after 7 days of growth and stained with FeCl₃. As a control, noninoculated plates were stained with FeCl₃.

Candidates of the oxoadipate pathway genes are highly induced by salicylic acid.

Further analysis of the A. niger transcriptome data revealed five additional genes that are highly induced (fold change ≥ 4) by salicylic acid (Table 3). Four of these genes are homologs of the oxoadipate pathway genes from A. nidulans (5). The final step of this pathway is the conversion of 3-oxoadipate-succinyl-CoA to acetyl-CoA and succinate by 3-oxoacyl CoA thiolase. Only one gene annotated as a 3-oxoacyl CoA thiolase (NRRL3_1526) was induced by salicylic acid and is possibly involved in the conversion of 3-oxoadipate-SCoA. In both the A. nidulans and A. niger genomes, the putative 3-oxoadipate enol-lactone hydrolase is located next to the putative catechol 1,2-dioxygenase. Interestingly, homologs of the putative muconate isomerase (AN3895 in A. nidulans, NRRL3_10508 in A. niger) and putative muconolactone isomerase (AN4061 in A. nidulans, NRRL3_10507 in A. niger) are also clustered in the A. niger genome but not in the genome of A. nidulans.

TABLE 3.

Expression of candidate oxoadipate pathway genes on salicylic acid compared to a no-carbon-source control

JGI database gene ID	Annotation according to JGI database	Suggested function	Homolog of A. nidulansa	Amino acid similarity (%)	FPKM		Deseq2 fold change	P valueb
					SA	NC
NRRL3_10508	Predicted l-carnitine dehydratase/alpha-methylacyl-CoA racemase	Muconate isomerase	AN3895	80.0	1933.05	132.57	14.20	0.00
NRRL3_10507	Hypothetical protein	Muconolactone isomerase	AN4061	72.0	1888.93	158.93	11.58	0.00
NRRL3_4788	Predicted alpha/beta hydrolase	3-Oxoadipate enol-lactone hydrolase	AN4531	70.1	1069.90	242.36	4.32	0.00
NRRL3_1886	3-Oxoacid CoA-transferase	Succinyl-CoA:3-oxoadipate CoA-transferase	AN10495	86.5	131.90	8.00	16.13	0.00
NRRL3_1526	3-Oxoacyl CoA thiolase	3-Oxoacyl CoA thiolase	AN5698	91.2	306.30	47.39	6.38	0.00

^aProtein IDs correspond to those in the AspGD database (http://aspgd.org/).

^bThe P values were calculated using Deseq2 software (51).

Phylogenetic diversity of ShyA in filamentous fungi.

To study the presence of ShyA in other fungi, a phylogenetic analysis was performed on selected ascomycete and basidiomycete genomes (Table S3). A BLASTP search using ShyA as a query did not retrieve NRRL3_43 of A. niger or the characterized FgShyC (16) and FgNahG (18) of F. graminearum, and these were therefore added manually to the phylogenetic analysis. A. niger ShyA and F. graminearum FgShy1 were both located in the same clade (Fig. S3), in which homologs of Aspergillus fumigatus, Aspergillus japonicus, A. nidulans, A. oryzae, Mycosphaerella graminicola, Podospora anserina, Talaromyces stipitatus, and Zymoseptoria pseudotritici were also found. FgNahG was clustered with NRRL3_43, while ShyI of the basidiomycete fungus Ustilago maydis (24) was clustered with two uncharacterized salicylate hydroxylase-like proteins from A. niger (NRRL3_2207) and F. graminearum (11948). In addition, a salicylate 1-monooxygenase (SalA, 8711) of A. nidulans, involved in terbinafine resistance (25), showed homology with ShyA but is located in a different clade (Fig. S3).

DISCUSSION

The salicylic acid metabolic pathway in microorganisms has been studied for decades, but only a few fungal enzymes have been described (6), and the pathway is mainly studied for fungi in the context of plant pathogenicity (16, 24, 26, 27). In this study, we identified the genes encoding ShyA, DhbA, and CrcA using transcriptome data of A. niger grown on salicylic acid and showed that cell extracts containing ShyA and CrcA or purified ShyA and CrcA can be used to produce cis,cis-muconic acid from salicylic acid (Fig. 2 and Fig. S4). The optimal temperature (45 to 50°C) and optimal pH (5.6 to 6.0) of ShyA differ from those of the previously described salicylate hydroxylase NahG of P. putida (30°C and pH of 7.0 to 7.5) (23). In addition, we observed that ShyA shows activity on o-hydroxylated benzoic acids, such as 4-aminosalicylic acid, 2,3-dihydroxybenzoic acid, and gentisic acid, but not on benzoic acid derivatives that are not o-hydroxylated. Activities on o-hydroxylated substrates have also been observed for NahG (23). Surprisingly, characterization of ShyA revealed a sigmoidal curve. This may be caused by inhibition or inactivation due to low salicylate concentrations and relatively high NADH concentrations; the enzyme also exhibits NADH oxidase activity besides the more rapid hydroxylase activity (28).

The first step of the salicylic acid metabolic pathway of A. niger converting salicylic acid to catechol is catalyzed by ShyA. The deletion of shyA results in severely reduced growth on salicylic acid, to a size similar to that of the no-carbon-source control. Based on the growth reduction, induction by salicylic acid, and the activity of the recombinant ShyA on salicylic acid, we conclude that NRRL3_9723 encodes the salicylate hydroxylase of A. niger. In contrast, in F. graminearum, FgShy1, which has activity on salicylic acid, and its corresponding gene, which is induced by salicylic acid, are considered not essential for growth on salicylic acid (16, 27, 29). Therefore, an additional salicylate hydroxylase was suggested. Despite the high induction of NRRL3_43 by salicylic acid, deletion of this gene did not result in a phenotype on salicylic acid. No other putative salicylate hydroxylases were induced by salicylic acid, indicating that no additional salicylate hydroxylases are present in A. niger. Therefore, we suggest that NRRL3_43 is a salicylic acid hydroxylase-like enzyme. Interestingly, FgNahG of F. graminearum also possesses salicylate hydroxylase activity and is, unlike FgShy1, essential for growth on salicylic acid (16, 18). However, the closest FgNahG homolog in A. niger is NRRL3_43 (58.3%), which is less similar to ShyA (40.5%) and FgShy1 (34.0%). In addition, FgNahG was not found as a homolog in the phylogenetic analysis. A recent review revealed that NRRL3_43 is part of a biosynthetic gene cluster in A. niger, but it is unknown which secondary metabolite is formed by this cluster (30). Further investigation of the transcriptome data showed that NRRL3_43 and 10 neighboring genes (NRRL3_35 to NRRL3_45) in the A. niger genome were induced by salicylic acid (Table S4). In the Fusarium genome, FgNahG appears not to be part of a secondary metabolic cluster, which indicates that no homolog of FgNahG is present in the A. niger genome.

The second salicylic acid metabolic pathway is the 3-hydroxylation of salicylic acid to 2,3-dihydroxybenzoic acid, which has been reported for A. niger, A. nidulans, and other fungi (26, 31), but the enzyme responsible for this conversion had not been demonstrated. In Arabidopsis thaliana, a salicylate 3-hydroxylase has been characterized (32). BLASTP analysis of this plant protein sequence against the A. niger NRRL3 genome resulted in 13 hits, none of which were induced by salicylic acid (Table S5). Therefore, this did not provide any leads for the identification of the gene encoding this enzyme. The second metabolic step in this pathway is the nonoxidative decarboxylation of 2,3-dihydroxybenzoic acid to catechol, which is catalyzed by DhbA. Deletion of dhbA in A. niger resulted in reduced growth on 2,3-dihydroxybenzoic acid, confirming the previously reported characterization (7). However, the deletion of dhbA did not result in reduced growth on salicylic acid, even though dhbA was induced by salicylic acid. This has also been observed in the fungal species Trichosporon cutaneum (33). In F. graminearum, deletion of the 2,3-dihydroxybenzoic acid decarboxylase resulted in delayed salicylic acid uptake, but no growth effect was detected on salicylic acid (27). If the in vitro activity of ShyA on 2,3-dihydroxybenzoic acid, resulting in the formation of pyrogallol, is also present in vivo, this would compete with the conversion of 2,3-dihydroxybenzoic acid to catechol mediated by DhbA. Therefore, we have strong indications that the 3-hydroxylation of salicylic acid plays either a minor or no role in the metabolism of salicylic acid in A. niger (Fig. 6).

FIG 6.

The salicylic acid metabolic pathway and oxoadipate pathway in A. niger. Confirmed pathways are shown with black arrows, and the suggested m-hydroxylation pathway is shown with a gray dashed arrow. Boxed in gray are the gene expression fold change values on salicylic acid compared to the no-carbon-source control for the corresponding gene. NRRL3_10508, putative muconate isomerase; NRRL3_10507, putative muconolactone isomerase; NRRL3_4788, putative 3-oxoadipate enol-lactone hydrolase; NRRL3_1886, putative 3-oxoadipate CoA transferase; NRRL3_1526, putative 3-oxoacyl CoA thiolase.

Recently, three putative intradiol dioxygenases have been shown to convert catechol to cis,cis-muconic acid (19, 20), but HqdA was not involved in degradation of catechol in vivo (20). However, it remained unknown which of the remaining two enzymes corresponds to the true catechol 1,2-dioxygenase of A. niger, as both were induced on salicylic acid. Deletion of crcA resulted in severely reduced growth on catechol, similar to the no-carbon-source control, while deletion of NRRL3_5330 or hqdA did not result in any phenotype. In addition, the FeCl₃ assay of the catechol plates with the ΔcrcA mutant did not produce a halo, which indicated that NRRL3_5330 or hqdA does not have an in vivo function as a catechol 1,2-dioxygenase-encoding gene in A. niger. Based on the deletion mutant, biochemical characterization, transcriptome data, and the previously characterized CrcA (19), we can conclude that NRRL3_4787 encodes this enzyme in A. niger. The biological function of NRRL3_5330 remains unknown, but it has been shown to convert several catechol derivatives, including hydroxyquinol (19).

The ability to degrade salicylic acid has been suggested to be important for the pathogenicity of fungi on plants (16, 17, 24, 27). A connection between the pathogenicity and salicylic acid degradation by salicylate hydroxylase was demonstrated in F. graminearum by deleting FgNahG (18). In our phylogenetic study, we observed that many fungi, including non-plant-pathogenic species, are equipped with putative salicylic acid hydroxylases. Therefore, the ability to degrade salicylic acid appears to also be important for biological processes other than pathogenicity. Studies have shown that salicylic acid is released from plant residues and is found in soil (34–37). Salicylic acid has antifungal properties (38), which is a possible explanation for the presence of salicylic acid metabolism in non-plant-pathogenic fungi. Another possible explanation is that this pathway is part of the degradation of polycyclic aromatic hydrocarbons such as naphthalene, phenanthrene, and carbaryl (39, 40).

The identification and characterization of the A. niger enzymes ShyA and CrcA contribute to a better understanding of fungal aromatic metabolic pathways. We demonstrated that whole-genome transcriptome analysis is a powerful tool for studying metabolic pathways and identifying genes. The combination of recombinant ShyA and CrcA is an example of production of valuable compounds from this pathway, such as cis,cis-muconic acid, which is an important platform chemical for synthesis of plastic and resins. Understanding the metabolic pathway and the corresponding enzymes is essential for the creation of efficient fungal cell factories for conversion of renewable and sustainable resources, such as lignin, to valuable chemical building blocks.

MATERIALS AND METHODS

Candidate gene identification.

To identify candidate genes involved in salicylic acid metabolism, transcriptome data of A. niger N593 ΔkusA grown on 2 mM salicylic acid for 2 h and on minimal medium (MM) (41) without a carbon source (as a control) were obtained from Lubbers et al. (21). Genes were considered induced by using a cutoff of fragments per kilobase per million (FPKM) of ≥10, fold change of ≥4, and P value of ≤0.01.

To identify the salicylate hydroxylase of A. niger, all genes annotated as salicylate 1-monooxygenase in the A. niger NRRL3 genome (https://mycocosm.jgi.doe.gov/Aspni_NRRL3_1) were obtained from the MycoCosm genome database. To reduce the number of candidates, BLASTP analysis was performed using the amino acid sequence of FgShy1 from F. graminearum (16), putative salicylate hydroxylase (AN2114 [http://www.aspergillusgenome.org/cgi-bin/locus.pl?locus=AN2114&organism=A_nidulans_FGSC_A4]) from A. nidulans, and the salicylate hydroxylase (NahG) (UniProt accession number P23262) of the bacterium Pseudomonas putida. To identify dhbA, the amino acid sequence of DhbA (7) was used as query in a BLASTP search against the A. niger NRRL3 genome. This resulted in the identification of a single gene (NRRL3_4385), which has been annotated as an amidohydrolase. For the identification of salicylate 3-hydroxylases, the amino acid sequence of a suggested salicylate 3-hydroxylase from A. nidulans (5) (AN7418 [http://www.aspergillusgenome.org/cgi-bin/locus.pl?locus=AN7418&organism=A_nidulans_FGSC_A4]) was used as a query in a BLASTP search against the A. niger NRRL3 genome.

Construction of expression plasmids.

Full-length shyA was synthesized based on the reference sequence (NRRL3_9723) in pET23b containing a C-terminal hexa-His tag (Genscript Biotech, Leiden, the Netherlands). The expression plasmid for crcA containing an N-terminal hexa-His tag was created by inserting the gene in pET28a through homologous recombination as described previously (42). The plasmid was amplified from pET28a flanking the multiple-cloning site. The CrcA-encoding gene was amplified from cDNA of A. niger grown in salicylic acid. Purification of the PCR fragments and transformation of E. coli TOP10 were performed as described previously (20). The plasmid was isolated and verified for correctness by sequencing. Both plasmids were used to transform the E. coli protein production strain BL21(DE3) (New England BioLabs, Ipswich, MA).

Protein production and purification.

Production and purification of the recombinant ShyA and CrcA proteins were performed as previously described (20). E. coli BL21(DE3) harboring pET23b-shyA or pET28a-crcA was grown in 10 ml of LB medium supplemented with 50 μg/ml of ampicillin for pET23b-shyA or 50 μg/ml of kanamycin for pET28a-crcA overnight at 37°C and 160 rpm. One milliliter of inoculum was transferred to a 1-liter Erlenmeyer flask containing 400 ml of LB medium supplemented with 50 μg/ml of ampicillin or kanamycin and grown to an optical density at 600 nm (OD₆₀₀) of 0.4 to 0.8. To induce recombinant protein production, isopropyl-β-d-thiogalactopyranoside (IPTG) with a final concentration of 100 μM was added to the cultures, which were further incubated overnight at 12°C and 160 rpm. Proteins were isolated and purified with a HisTrap FF 1-ml column coupled with the ÄKTA start system (GE Healthcare Life Sciences, Uppsala, Sweden) using the setup described previously (20). After purification, 0.5 mM flavin adenine dinucleotide (FAD) was added to the fractions containing ShyA.

The molecular mass of ShyA and CrcA was calculated in silico (https://www.bioinformatics.org/sms/prot_mw.html) and estimated by SDS-PAGE (12% [wt/vol] polyacrylamide gel, Mini-PROTEAN Tetra Cell; Bio-Rad, Hercules, CA) using the Precision Plus prestained protein standards (Bio-Rad) as a marker. Protein concentrations were determined from the SDS-PAGE gel, using bovine serum albumin as standard, and quantified by ImageJ software (43).

Enzyme activity assay of ShyA.

Activity assays were performed using cell extracts of E. coli BL21 expressing shyA, crcA, and hqdA, respectively. The E. coli hqdA expressing strain was obtained from Lubbers et al. (20). Cell-free extracts of E. coli BL21 harboring an empty vector were used as a negative control. The reaction mixture contained 100 μM phosphate buffer (pH 6.4), 200 μM substrate, and 40 μl of cell extract. Reaction mixtures were incubated for 1 h at 30°C, after which the reaction was stopped by incubation at 80°C for 10 min and analyzed with an HPLC (Dionex ICS-5000+ chromatography system; Thermo Scientific, Sunnyvale, CA) equipped with an Acclaim Mixed-Mode WAX-1 LC column (3 by 150 mm; Thermo Scientific) and a UV detector (225 nm; Thermo Scientific). The chromatographic separation was carried out using isocratic elution of 30°C at a flow rate of 0.425 ml/min with 25 mM potassium monophosphate and 0.8 mM pyrophosphate (pH 6.0) in 50% acetonitrile (44). Salicylic acid, catechol, and cis,cis-muconic acid standards were used as a reference.

Activity assay of purified ShyA with 1 mM concentrations of substrates were performed in 200-μl reaction mixtures that contained 1 mM NADH and McIlvaine buffer (pH 6.0) (45). After 3 h of incubation at 30°C, 100 μl of reaction mixture was diluted 10 times with acetonitrile and analyzed by HPLC.

A preliminary kinetic analysis of purified ShyA was carried out by varying the salicylic acid concentration between 25 and 500 μM at a fixed concentration of 1 mM NADH. A 96-well plate containing a fixed amount of McIlvaine buffer (pH 6.0) and various amounts of salicylic acid and demi water up to a total volume of 176 μl per well was preheated at 30°C. Then, 4 μl (total concentration of approximately 0.7 μg) of purified ShyA enzyme was added to the wells. The extract from the empty-vector-containing E. coli strain was used as a control. To start the reactions, 20 μl of 10 mM NADH was automatically injected with a FLUOstar Optima microplate reader (BMG Labtech, Ortenberg, Germany) with a preheated chamber at 30°C. Absorbance was measured every 2 s at 340 nm for 500 s. The K_0.5 (substrate concentration for half-saturation) and V_max were determined by fitting the data to the Hill equation using GraphPad Prism 8 software.

Growth conditions of Aspergillus niger strains.

A. niger strains used in this study (Table 4) were grown on complete medium (CM) (41) agar (1.5% [wt/vol]) plates at 30°C for 4 days. Spores were harvested with 10 ml of N-(2-acetamido)-2-aminoethanesulfonic acid buffer, and 2 μl of 10³ freshly isolated spores was inoculated on MM agar plates. Due to the toxicity of the aromatic compounds, different concentrations were used for the growth profile, i.e., 2 mM for ferulic acid and cinnamic acid and 5 mM for the other compounds. All chemicals were purchased from Sigma-Aldrich.

TABLE 4.

A. niger strains used in this study

Strain	CBS no.	Genotype	Reference
N593 ΔkusA	CBS 138852	cspA1 pyrG ΔkusA::amdS	47
Reference	CBS 145984	cspA1 pyrG ΔkusA::amdS ΔpyrG::pyrG	This study
ΔhqdA	CBS 145839	cspA1 pyrG ΔkusA::amdS ΔhqdA::pyrG	20
ΔshyA (9723)	CBS 146179	cspA1 pyrG ΔkusA::amdS ΔshyA::pyrG	This study
ΔdhbA (4385)	CBS 146180	cspA1 pyrG ΔkusA::amdS ΔdhbA::pyrG	This study
ΔcrcA (4787)	CBS 146181	cspA1 pyrG ΔkusA::amdS ΔcrcA::pyrG	This study
Δ43	CBS 146183	cspA1 pyrG ΔkusA::amdS Δ43::pyrG	This study
Δ2597	CBS 146182	cspA1 pyrG ΔkusA::amdS Δ2597::pyrG	This study
Δ5330	CBS 146184	cspA1 pyrG ΔkusA::amdS Δ5330::pyrG	This study

Construction of gene deletion cassettes and transformation of A. niger.

The gene deletion cassettes were constructed as described previously (46). The flanking regions contained 900 to 1,000 bp upstream and downstream of the gene of interest, including an overlap of the selection marker orotidine 5′-phosphate decarboxylase (pyrG) that was amplified from Aspergillus oryzae RIB40 (Table 5). These three fragments were combined in a fusion PCR using the GoTaq long PCR master mix (Promega, Madison, WI) as described previously (20). Protoplast-mediated transformation of A. niger N593 ΔkusA (47) and purification of the different transformants were performed as described before (46). To verify that the phenotypes are not caused by random mutations, three independent deletion mutants were screened for phenotypes (Fig. S5).

TABLE 5.

Primers used in this study

Primer	Sequence, 5′a	Target
ShyA-5′F	GGTTCGCTTTGCTTTGCTAGC	5′ shyA-pyrG-flanks
ShyA-5′R-pyrG	gcacttaccttcgcattttctggtatattAGGGTTCAATGATGGTTTGAGGT
ShyA-3′F-pyrG	cctgtgttggttctcaggaactgcgaatatGAGTAGATTTCAGGGAGGAGAGC	3′ shyA-pyrG-flanks
ShyA-3′R	CTGTTTGTTGCTGAAGCTGTCC
ShyA-Check-F	CCAAGAAAGACTTCCACGTCG	Deletion verification shyA
ShyA-Check-R	CTTTGCCAGTCGAAAGAATTGG
DbhA-5′F	GCCTTTGGCCGTGATCAG	5′ dbhA-pyrG-flanks
DbhA-5′R-pyrG	gcacttaccttcgcattttctggtatattCAGGGATGTGTGAGTAGAGAAGG
DbhA-3′F-pyrG	cctgtgttggttctcaggaactgcgaatatGGCTGACGAGCTAGGATGG	3′ dbhA-pyrG-flanks
DbhA-3′R	CAAACAGCCACATGAACAAGCC
DbhA-Check-F	CGAAGAGAAGACGCGCTG	Deletion verification dhbA
DbhA-Check-R	CTTCTTGCGGTCCTCGAAC
CrcA-5′F	GCAAGGCCACAGCATCAAG	5′ crcA-pyrG-flanks
CrcA-5′R-pyrG	gcacttaccttcgcattttctggtatattGTCGAAGCGCGGCATTTTG
CrcA-3′F-pyrG	cctgtgttggttctcaggaactgcgaatatGCGATTGTCTATTGACGTTTGGG	3′ crcA-pyrG-flanks
CrcA-3′R	GTCACCGTCAACAACCACC
4787-Check-F	GCATGGGTCCCAAAACGA	Deletion verification crcA
4787-Check-R	CAACTCGAGCAGCTTTCCAG
43-5′F	CACTACCACAGACCCAGGC	5′ 43-pyrG-flanks
43-5′R-pyrG	gcacttaccttcgcattttctggtatattAGAGAAGTGCTTGCTATGCTGT
43-3′F-pyrG	cctgtgttggttctcaggaactgcgaatatCATGGTTGTGGGACTGGC	3′ 43-pyrG-flanks
43-3′R	CAAGCGAGCTTCACGTGTC
43-Check-F	CAACAATCCATCCCCTGCA	Deletion verification 43
43-Check-R	CTTATTCCATCCCTCGCGGT
5330-5′F	GCCGTCCATGCCTTTCTCAAG	5′ 5330-pyrG-flanks
5330-5′R-pyrG	gcacttaccttcgcattttctggtatattCCGTTCGATTGTTGCTCAGACAT
5330-3′F-pyrG	cctgtgttggttctcaggaactgcgaatatCCTCGCAAGACTGAGGATTTGG	3′ 5330-pyrG-flanks
5330-3′R	GCTATCCATCCCGATAAGCATGC
5330-Check-F	ATCTTGAAGTTGATCCAGGCC	Deletion verification 5330
5330-Check-R	TAGGAAAGGGCGTCGACG
2597-5′F	CCAACCAAAACGTGGTCACG	5′ 2597-pyrG-flanks
2597-5′R-pyrG	gcacttaccttcgcattttctggtatattGTACTGAAATGGCCTGACTGGTC
2597-3′F-pyrG	cctgtgttggttctcaggaactgcgaatatCGTGCTTTTCCAGTTGAGAATGC	3′ 2597-pyrG-flanks
2597-3′R	GTCCAATCCAACCCAGGTGAG
2597-Check-F	GTCGACGTCCTGATCTGTG	Deletion verification 2597
2597-Check-R	CCGGATACTGGATTTCTTCCG
pET28-amplification F	CCACTGAGATCCGGCTGCTA	pET28 backbone amplification
pET28-amplification R	GCTAGCCATATGGCTGCC
CrcA-amplification F	CTGGTGCCGCGCGGCAGCCATATGGCTAGCCCGCGCTTCGACCCCAAC	crcA-6×HIS with pET28 overlap
CrcA-amplification R	CGGGCTTTGTTAGCAGCCGGATCTCAGTGGTCAGTTCGCCGGCGCCC
pET28-sequencing primer F	GTGAGCGGATAACAATTCCCCTCTAGAAATA	Confirmation of insert in pET28 vector
pET28-sequencing primer R	CCTTTCAGCAAAAAACCCCTCAAGACC

^aIn lowercase sequences are regions flanking pyrG.

Colorimetric assay for the visualization of phenolic acid degradation in solid media using FeCl₃.

FeCl₃ reacts with many phenolic compounds and is commonly used as a visualization reagent for thin-layer chromatography (48). To visualize if active ShyA was produced by the shyA expression strain, LB plates supplemented with 1 mM salicylic acid, 100 μM IPTG, and 50 μg/ml of ampicillin were inoculated with 100 μl of preculture and incubated overnight at 30°C. MM agar plates supplemented with 5 mM salicylic acid, 2,3-dihydroxybenzoic acid, or catechol were inoculated with 10³ freshly isolated spores of the ΔshyA, ΔdhbA or ΔcrcA mutant and incubated at 30°C for 7 days. After incubation, 5 ml of a 20% FeCl₃ solution was added on the plates for 5 min. Thereafter, the FeCl₃ solution was removed and pictures were taken.

Phylogenetic analysis.

The amino acid sequence of ShyA was used for BLASTP analysis on selected ascomycete and basidiomycete genomes (Table S3). Mainly industrially relevant species, fungal model species, or fungi that are potentially interesting for new applications were selected. To reduce the amount of insignificant hits, a cutoff E value of 10⁻⁴⁰ was used. The amino acid sequences of NahG from P. putida was included in the phylogenetic analysis as the outgroup (49). The putative salicylate hydroxylase FgShyC and FgNagH from F. graminearum were not obtained as homologs through BLASTP and were manually included in the analysis (16, 18). Amino acid sequences were aligned using the MUSCLE algorithm implemented in MEGA7 with the default settings (50). The maximum likelihood, neighbor joining and minimum evolution trees were constructed using MEGA7 with 500 bootstraps.