Cytochrome P450 Monooxygenase-Mediated Metabolic Utilization of Benzo[a]Pyrene by Aspergillus Species

To test our initial hypothesis that Aspergillus species are able to degrade BaP, we employed A. nidulans, A. flavus, A. oryzae, and A. fumigatus. These fungal species are distantly related to each other, covering a broad range of Aspergillus (14). In all four species, the amount of BaP recovered from 7-day-old cultures was significantly lower in the living cells than that of dead cells (Fig. 2A), indicating that BaP was degraded or transformed in all species tested. The chromatogram showed no additional fluorescent peaks, suggesting that the degraded products were water soluble and/or not fluorescent at the same wavelengths as BaP (see Fig. S1 in the supplemental material). A. nidulans and A. oryzae were able to remove 92% ± 4.9% and 95% ± 3.5% of the added 200 μM BaP, respectively. As A. nidulans is a well-studied genetic model, we used it to further uncover the genetic and biochemical mechanisms of BaP degradation. Thin-layer chromatography (TLC) analyses of residual BaP in different concentrations of glucose in minimal medium (MM) have revealed that MM with 0.1% glucose led to the most effective degradation of BaP by A. nidulans (Fig. 2B).

To test whether A. nidulans uses BaP as a substrate for growth, we measured changes in cell viability. The alamarBlue assay showed that cells exposed to BaP had significantly higher viability than the dimethyl sulfoxide (DMSO) controls, including time points after the addition of fresh medium to supply cells with the additional nutrients for the continuous proliferation (Fig. 2C). As found in TLC data, the prolonged viability of the cells with BaP was observable with the presence of a small amount of glucose in the medium (0.1%), whereas BaP addition alone was able to increase the cell viability at 2 days postexposure (Fig. 2C). This suggests that BaP can be used as a carbon source, and such effects can be enhanced with the supplementation of a small amount of glucose, which may provide additional resources for energy and enzyme production.

To further test the hypothesis that the fungus uses BaP as a growth substrate, we investigated the genome-wide expression responses of A. nidulans to BaP via transcriptome sequencing (RNA-seq). Transcript levels were measured in BaP-treated cells relative to controls (DMSO) at 6, 24, and 48 h postexposure by KEGG analysis (Fig. 2D). The number of differentially expressed genes (DEGs) was low at 6 h (703 DEGs); this increased to 1,503 and 1,529 DEGs at 24 and 48 h, respectively (Fig. 2E). KEGG pathway analysis indicated that, as time progresses with BaP exposure, genes associated with cell growth, such as ribosome biogenesis, biosynthesis of amino acids, nucleotide metabolism, biosynthesis of unsaturated fatty acids, N-glycan biosynthesis (cell wall), and glycerophospholipid (membrane), were upregulated (Fig. 2E). Conversely, genes categorized into the pathways indicative of cell starvation and stress, including amino acid degradation, autophagy, aflatoxin (sterigmatocystin) biosynthesis, and starch metabolism, were downregulated in BaP-treated cells (Fig. 2E). These results indicate that BaP-treated cells are actively growing compared to controls and that the fungus is able to use BaP to sustain growth.

The transcriptomic response to carbon starvation in A. nidulans shows upregulation of genes involved in programmed cell death, macroautophagy, cell wall component degradation, asexual reproduction, and secondary metabolite production (15). Downregulation of genes involved in glycolysis and oxidative phosphorylation, cell wall component synthesis, and nitrogen and lipid anabolic pathways was also seen in the starving cells (15). We carried out an integrated analysis of the differentially expressed genes by BaP treatment with the carbon starvation stress response and found that the BaP-treated cells showed upregulation of the following genes: (i) citA, gsdA, acuE, acuD, and cycA, involved in the tricarboxylic acid (TCA) cycle, replenishment of TCA cycle intermediates, and oxidative phosphorylation; (ii) AN11161, involved in phospholipid biosynthesis; (iii) AN10779, involved in β-1,6-glucan biosynthesis; and (iv) several genes involved in amino acid biosynthesis pathways (Fig. 2F). BaP-treated cells show downregulation of aflR and other sterigmatocystin biosynthesis genes, abaA, involved in conidiation, and agnC, nagA, chiC, and AN4825, involved in cell wall component hydrolase enzymes (Fig. 2F). Taken together, the data demonstrate that BaP enables the cells to grow more actively than control cells and provide some evidence that BaP is metabolized and shuttled into carbon utilization pathways.

Finally, in an attempt to address whether BaP metabolism causes oxidative stress and/or DNA damage responses in A. nidulans as in mammalian cells, we compared our RNA-seq data with those representing responses to cells treated with other known oxidizing compounds. Some redox-balancing genes were upregulated in BaP-treated cells, including catA and sodB (Fig. 2F). Expression of some DNA repair genes was also induced by BaP, including AN0604 and AN0097 (Fig. 2F). These results suggest that BaP may cause DNA damage in A. nidulans as in mammalian cells, although it is shown that fungi have additional capacity to prevent extensive DNA damage (16).

Initial oxidation of BaP in mammalian cells is mediated by a CYP, adding a single molecular oxygen, leading to the formation of BaP epoxide intermediates, which can be further converted into hydroxylated products (4, 17–19).

With the hypothesis that A. nidulans employs a CYP to degrade BaP, we first examined our RNA-seq data to search for the CYP genes upregulated by BaP treatment and found that no specific CYPs were clearly induced by BaP. We then used the CYPs of P. chrysosporium, Pc-PAH1 and Pc-PAH3, which when expressed in the yeast cells, converted BaP to 3-hydroxybenzo[a]pyrene (13), to search for closely related CYPs in the A. nidulans genome. Despite the 723 million years of divergence between the two fungi (20), several CYPs similar to Pc-PAH1and Pc-PAH3 were identified (Fig. 3A; see Table S1 in the supplemental material). Among these, mRNA levels of AN1884 in glucose-limited medium and kinetics of BaP degradation were closely aligned, indicating that this CYP may be associated with metabolism of BaP (Fig. 3B; Fig. S2B).

To test its function in BaP metabolism in vivo, we generated independent deletion (Δ) strains of AN1884 in A. nidulans and found that all three independent null mutant strains lost the ability to degrade BaP, with no distinct growth and developmental phenotypic changes (Fig. S3). Reintroducing an AN1884 coding region into null mutant strains restored the BaP degradation ability to that of the wild type (WT) (Fig. 3C), supporting the hypothesis that AN1884 is responsible for the breakdown of BaP, so it was named bapA (benzo[a]pyrene metabolism locus A). Measuring the cell viability of the WT, the ΔbapA mutant cells, and complemented (C′bapA) cells treated with BaP relative to those treated with DMSO revealed that the ΔbapA mutant cells were not only less viable than WT and C′bapA cells treated with BaP, but less viable than ΔbapA mutant cells treated with DMSO (Fig. 3D). This indicates BaP may be toxic to cells unable to metabolize it. To our knowledge, this is the first report providing evidence of in vivo function of a CYP in cellular degradation of BaP.

We attempted to identify the BaP metabolite(s) formed by BapA. We isolated microsome-containing fractions from the WT, ΔbapA, and C′bapA, cells and incubated them with BaP. The WT and C′bapA, but not ΔbapA, chromatograms showed a small fluorescent peak with a shorter retention time than BaP (Fig. S4A). The retention time of this unknown metabolite was compared to those of known BaP metabolites and matched that of benzo[a]pyrene-3,6-dione (Fig. S4A). However, the fluorescence spectra revealed that sample peaks do not show the same fluorescing as benzo[a]pyrene-3,6-dione (Fig. S4B). No other known metabolite standards matched the retention time of this peak (Fig. S4C). Although we were not able to reveal the exact identity of the product formed within the microsome fraction of cells expressing BapA, we were able to rule out many BaP metabolites formed by other organisms. This may indicate that BapA is involved in forming a unique metabolite not previously reported in other organisms.

The availability of the ΔbapA mutant allows us to further investigate the downstream metabolomic outcomes of BaP. Since BaP treatment caused upregulation of genes associated with amino acid biosynthesis, which is an easily measured endpoint for carbon utilization, we quantified free amino acids in WT and ΔbapA cells treated with BaP relative to the control (DMSO). In agreement with alamarBlue data, only BaP-treated WT cells showed significant accumulation of glutamate, aspartate, glutamine, lysine, and the intermediate amino acid ornithine (Fig. 3E). The ΔbapA cells showed significant accumulation of alanine and α-ketoglutarate in control (DMSO) compared to BaP treatment (Fig. 3E). This provides additional evidence that loss of BaP metabolism causes lack of cell growth upon exposure to BaP and alludes to toxicity caused by the lack of BaP-degrading ability. While insignificant due to large deviations among samples, accumulation of several primary metabolites involved in energy metabolism, such as citrate and malate, and nucleotides involved DNA/RNA synthesis or signaling, such as cyclic AMP (cAMP), appeared to be affected by BaP (Fig. 3E).

The structural analysis showed that the four widely recognized consensus regions (a to d) (Fig. 4A) contributing to the core function of P450s (10) are highly conserved in the CYP617 family. Interestingly, the conserved motif a (AGHETT) of the CYP617 family is very specific and is highly similar to those of archaea and bacteria (10).

We performed a phylogenic analysis to determine how widely distributed BapA is in fungi. We found 64 CYP617 family members in the fungal kingdom, covering dothideomycetes, eurotiomycetes, leotiomycetes, and sordariomycetes, with the number ranging from one to several per species (see Fig. S5 and Table S3 in the supplemental material). CYP617 members are limited to ascomycetes. The BapA subfamily CYP617D1 was mostly distributed in the genus Aspergillus (Fig. 4B), suggesting a conserved role within the genus.

To examine a potential conservation of its function, we identified a likely orthologue of BapA in A. flavus (AFLA_036020) and generated three individual null mutant strains. All ΔAFLA_036020 strains lost the ability to degrade BaP, corroborating the conserved and essential role of BapA in degradation of BaP in Aspergillus species (Fig. 4C) under glucose-limiting conditions.

Identification of bapA allowed us to further investigate its upstream regulatory components. The velvet proteins are a family of global transcription factors (TFs) involved in diverse aspects of fungal biology (21). They contain a DNA binding domain structurally similar to that of the human TF complex nuclear factor kappaB (NF-κB) (22) p50, which governs cell survival upon exposure to BaP (23–26).

Due to this structural similarity and diverse regulatory functions between human NF-κB and the fungal velvet complex, we hypothesized that velvet proteins might play a role in BaP degradation. To address this hypothesis, we first tested mRNA levels of bapA in each velvet deletion mutant (ΔveA, ΔvelB, ΔvelC, and ΔvosA) and found that the two regulators VeA and VelB were necessary for proper expression of bapA (Fig. 5A). Interestingly, the other two regulators, VosA and VelC, seem to play a repressive role at 10 h post-glucose starvation and an activating role at later time points (Fig. 5A). To verify that expression of bapA translates into BaP degradation, we tested the degradation ability of each deletion mutant and found that the ΔveA and ΔvelB mutants were unable to degrade BaP. On the contrary, the ΔvosA and ΔvelC mutants were able to degrade the same amount of BaP, yet faster than the WT (Fig. 5B). These results indicate that VeA and VelB play a key role in metabolic utilization of BaP, in part by controlling proper expression of bapA, whereas VosA and VelC may indirectly play a role in coordinating the proper timing of BapA expression in response to glucose limitation.

The Aspergillus strains used in this study are listed in Table 1. Initially, 10⁶ spores/ml were added to 400 ml minimal medium (MM) (31) with 1% glucose in 2-liter flasks and incubated for 18 h at 220 rpm at 37°C. The mycelial aggregates were then collected on sterile Miracloth (Sigma-Aldrich), rinsed, and transferred to 100 ml MM with 0.1% glucose in 250-ml Erlenmeyer flasks. Control dead cells were autoclaved on a liquid cycle at 121°C for 20 min to account for any nonmetabolic sources of loss of BaP. A 100 mM stock solution of BaP (Sigma-Aldrich) in dimethyl sulfoxide (DMSO) was added to the cultures to a final concentration of 200 μM; the same volume of DMSO was added to controls. All flasks were further incubated at 220 rpm at 37°C for the designated time. Escherichia coli DH5α cells were grown in Luria-Bertani medium with ampicillin (100 μg/ml) for plasmid amplification.

Extraction of BaP was optimized to recover all BaP adhered to and taken up by cells, but not biotransformed. Individual fungal cell cultures were extracted using 100 ml 1:1 hexane-ethyl acetate with pyrene (Sigma-Aldrich [final concentration, 200 μM]) as an internal standard to correct for extraction efficiency. The entire mixture was sonicated using a Sonic dismembrator model 100 (Fisher Scientific) with a 1/2-in. probe on full power for 6 min to ensure disintegration of hyphal pellets. Solvent (1 ml) was removed and centrifuged to remove particulate matter and diluted 100-fold in 1:1 solvent A (30 mM acetate buffer at pH 4.7, 10% acetonitrile)-solvent B (acetonitrile). BaP and pyrene were quantified by high-performance liquid chromatography (HPLC) using an Agilent 1260 system equipped with a 3- by 50-mm Poroshell 120 EC-C₁₈ 2.7-μm column (Fig. S1). A linear gradient that ramped from 55% B to 90% B over 10 min at a flow rate of 0.75 ml/min was used, followed by fluorescence detection (FLD [λ_excitation = 248 nm and λ_emission = 465 nm]). All standard curves were linear, and the detection limits were ≤0.1 μM for pyrene and BaP.

To further verify the degradation of BaP by A. nidulans at different glucose concentrations (Fig. 2B), we also carried out TLC analyses 10 times and obtained a high degree of reproducibility. BaP extraction was performed by adding 0.5% (vol/vol) 6 N HCl (to stop all metabolic activity) to the fungal cultures (100 ml). The mycelium was collected through Miracloth and squeezed to maximize collection of the supernatant, which was transferred to a fresh 250-ml flask and mixed with 100 ml of ethyl acetate (1:1 ratio). Both liquids were then transferred to a new 250-ml separatory funnel. After shaking vigorously for 2 min, the organic phase was transferred in a new flask. Fresh solvent was added to the separatory funnel, and shaking and collecting were repeated two additional times. The resulting solvent was allowed to evaporate in the fume hood, and each dried sample was resuspended with ethyl acetate (1 ml) for TLC analysis. Ten microliters of each sample was applied to a TLC silica plate, including a fluorescence indicator (Kiesel gel 60, 0.25 mm thick; Merck). Authentic BaP standard was loaded as a control. The TLC plate was then developed with toluene-acetone-hexane (1:1:1 [vol/vol/vol]), where the R_f value of BaP was 0.9. The TLC plate was exposed to UV at A₃₂₀ for 30 s, and images were captured using a Canon EOS camera. To quantify the residual BaP shown in Fig. S3B, the density of each BaP spot on TLC was determined using ImageJ (NIH): the relative amount of BaP in live cells to the dead cell control (=100%) is presented.

Cell viability was determined by percentage of alamarBlue (Bio-Rad) reduction as described previously (32), with the following exceptions. Cells were prepared as described for BaP degradation with solvent (DMSO) only as a control, and 0.45 ml of cells was added to 0.45 ml fresh MM with 0.1% glucose and 100 μl alamarBlue and incubated for 2 h at 37°C.

Fungal cells from submerged cultures were collected at designated time points, squeeze-dried, flash frozen in liquid N₂, and stored at −80°C until subjected to RNA preparation. Total RNA isolation was done using TRIzol as described previously (33). cDNA was prepared using an avian myeloblastosis virus (AMV) reverse transcriptase kit (NewEngland Biolabs) with oligo(dT). Reverse transcriptase quantitative PCR (RT-qPCR) was performed with iTaq universal SYBR green supermix (Bio-Rad) on a Bio-Rad CFX96 real-time PCR detection system. mRNA was normalized using threshold cycle (2^−ΔΔCT) method (34). Levels of bapA mRNA were determined using 2^−ΔΔCT, in which bapA expression (C_T) was found relative to the reference gene actA (γ-actin) (ΔC_T) and then relative to time point 0 (ΔΔC_T). Time point 0 indicates exponential growth (18 h postgermination) in regular MM (1% glucose), and each time point represents hours after the switch to MM with 0.1% glucose. Each experiment was performed using technical triplicates for RT-qPCR accuracy, and three biological triplicates were used for each time point. The oligonucleotides used are listed in Table S1. Total RNA was extracted and submitted to ProteinCT Biotechnologies (Madison, WI) for library preparation and RNA sequencing.

RNA sequencing was done as described previously (35). The library was constructed and purified and sequenced (SE100bp) using the Illumina HiSeq2500, and over 20 million high-quality reads per sample were achieved.

Verification of the quality of reads (quality control [QC]), alignments, gene annotation, and differential expression analysis were performed as described previously (35).

The KEGG pathway database was used to search against A. nidulans KEGG pathway maps in order to identify A. nidulans metabolic pathways with the differentially expressed genes (DEGs) after exposure to BaP on 20 February 2018 (36).

Fungal cells prepared as described for BaP degradation with the DMSO control were subject to extraction of cellular components as described previously (37), with the following exceptions. Hyphal mats were filtered and squeeze-dried, noting the mass after removing liquid, 1 day after transfer to BaP-containing medium. Tissue was flash frozen in liquid N₂ and stored at −80°C. Two milliliters of extraction solvent (37) was added, and samples were sonicated using a 1/4-in. probe for 3 min and centrifuged to remove cell debris. Additional sample prep and analysis were performed as described previously (37).

CYP sequences similar to those of Pc-PAH1 and Pc-PAH3 in Aspergillus sp. were identified using blastp (38). Protein sequences were found using NCBI, and protein alignment was calculated using Clustal Omega at EMBL-EBI output ClustalW with character counts (39). A phylogenetic tree was created using Jalview nearest neighbor joining (40).

BapA (AN1884) was assigned to CYP617D1 (11) and analyzed according to the rules of the International P450 Nomenclature Committee (41). BapA protein sequence was used to query FungiDB (42). CYP617 members were aligned, and the phylogenetic tree was constructed as previously described (43).

Double-joint PCR was used to generate the deletion constructs of A. nidulans bapA (AN1884) and A. flavus bapA (AFLA_036020) (33). Briefly, the deletion construct containing the A. fumigatus pyrG marker with 5′ and 3′ flanking regions of bapA was introduced into the recipient strain RMJP1.59 (A. nidulans) or NRRL 3357.5 (A. flavus). Three independent ΔbapA strains each in A. nidulans (TMK6-1, -35, and -47) and A. flavus (TEO1 2, 8, and 9) were confirmed and analyzed. To generate complemented strains of the ΔbapA mutant in A. nidulans, a bapA⁺ gene region, including its upstream 2-kb region, was introduced to pHS13 (44) and introduced into E. coli DH5α for transformation. Upon sequence verification of the insert, the purified plasmid was introduced into the recipient ΔbapA A. nidulans strain (TMK6). Three independent complemented strains (C′bapA 12, 16, and 17) were verified and analyzed.

Cells were prepared as described without BaP treatment to capture peak bapA mRNA levels (Fig. S2A and B). After 1 day of incubation, cells were filtered, washed, squeeze-dried, and flash frozen in liquid N₂. Frozen tissue was ground to a fine powder in liquid N₂ with the addition of glass beads in a mortar and pestle. The powder was resuspended in 30 ml homogenization buffer (0.1 M KPO₄ at pH 7.25, 0.1 M KCl, 10 mM EDTA at pH 8, 0.25 mM phenylmethylsulfonyl fluoride [PMSF], 0.1 mM dithiothreitol [DTT]) and kept in ice. A sonication probe (1/8 in.) was used on full power for 30 s to homogenize cells and form microsomal structures. Large debris was filtered using Miracloth, and the supernatant was centrifuged for 20 min at 20,000 × g. The supernatant was then transferred and centrifuged for 60 min at 105,000 × g. The resulting supernatant was discarded, and the pellets were resuspended in 200 μl dilution buffer (0.25 M KPO₄ at pH 7.25, 20% [vol/vol] glycerol, 10 mM EDTA at pH 8.0, 0.25 mM PMSF, 0.1 mM DTT), flash frozen in liquid N₂, and stored at −80°C for a maximum of 1 week. The protein concentration was determined with the Coomassie protein assay (Thermo Fisher). BaP metabolism was measured by incubating 2 mg/ml microsomal protein, 1 μM BaP, and 1 μM NADPH in up to 1 ml 50 mM phosphate buffer at pH 7.5 at 37°C for 1 h. As a control, protein was denatured by boiling for 20 min prior to incubation. Metabolites were extracted by adding 2 ml of 2:1 acetone-ethyl acetate and vortexing for 2 min. Solvent was removed and centrifuged at 13,000 × g for 10 min and dried under N₂. Samples were resuspended in 50 μl 1:1 HPLC solvent A-solvent B and analyzed by HPLC.

Statistical significance was determined using Student's t test with a two-tailed distribution and two-sample unequal variance.

All RNA-seq data files are available from the NCBI Gene Expression Omnibus database (45) under accession no. GSE116804.