Characterization of sulfhydryl oxidase from Aspergillus tubingensis

Production and purification of AtSOX

Aspergillus tubingensis strain D-85248 was obtained from the VTT Culture Collection [29] and cultivated as described in [28]. A. tubingensis was grown on PeptoneTM-D(+)-glucose media adapted from [3, 5] in liquid cultivation at 30 °C under shaking (250 rpm). After removal of the fungal biomass by filtration, AtSOX was purified from the cell-free extract with four chromatographic steps according to the procedure described below. The culture supernatant was concentrated and buffer exchanged to 20 mM Tris-HCl pH 7 with a PD10 column (no 17-0851-01, GE Healthcare, Uppsala, Sweden), and applied on a QSepharose fast flow column (volume = 20 mL, Amersham Biosciences, Piscataway, USA). Proteins were eluted with a linear gradient from 0 to 0.3 M NaCl in 20 mM Tris-HCl pH 7. The SOX activity of the fractions was detected using 5,5-dithio-bis(2-nitrobenzoic acid) (Ellman’s reagent, DTNB, no D 218200, Sigma-Aldrich, Helsinki, Finland) and oxygen consumption measurements. The SOX active fractions were pooled and applied to a Superdex 75 column (volume = 24 mL, Amersham Biosciences, USA) using 50 mM Tris-HCl pH 7 buffer containing 150 mM NaCl at a 0.1 mL/min flow rate. SOX-active fractions were pooled, concentrated, buffer exchanged to 20 mM Tris-HCl pH 7, and applied twice to a Resource Q Sepharose column (volume = 1 mL, Amersham Biosciences, USA). In the first separation using Resource Q Sepharose resin, the proteins were eluted with a linear gradient from 0 to 0.2 M NaCl in 20 mM Tris-HCl pH 7, whereas in the second separation a shallower gradient from 0 to 0.11 M NaCl was used. Prepacked anion exchange columns were connected to ÄKTA™ chromatography system and UNICORN Control Software (GE Healthcare, Uppsala, Sweden).

Protein concentration was determined using a Bio-Rad DC protein assay kit (Bio-Rad, Hercules, USA) and bovine serum albumin (BSA, no. A8022, Sigma, St. Louis, USA) as a standard. Proteins were analysed by ready-made 12% Tris-HCl SDS-PAGE gel (no 161-1156, Bio-Rad, Hercules, USA) to ensure purity of the isolated AtSOX enzyme.

Amino acid analysis of AtSOX and strain identification

The partial amino acid sequence of AtSOX was determined from the N-terminus and internal peptide fragments. SDS-PAGE protein bands were stained with Coomassie Brilliant Blue, and the protein band of interest was excised and subjected to N-terminal sequencing. Edman degradation was performed using a Procise 494A protein sequencer from Perkin Elmer, Applied Biosystems Division (Foster City, CA, USA). The Coomassie stained protein band, assumably containing AtSOX, was cut off from the SDS-PAGE and in gel digested essentially as described by [30] for obtaining internal sequences and for peptide mass fingerprinting (PMF). In gel digestion was done by reducing proteins with dithiothreitol (DTT), alkylating with iodoacetamide and digesting with trypsin (no V5111, Promega, Madison, USA). The enzymatic cleavage occurred during overnight incubation at 37 °C. The peptides produced by enzymatic cleavage were analysed by MALDI-TOF MS after desalting using mC18 ZipTip (no ZTC18M096, Millipore, Billerica, USA). MALDI-TOF mass spectra of peptide fragments for PMF were obtained using an Ultraflex TOF/TOF instrument (Bruker-Daltonik GmbH, Bremen, Germany) using α-cyano-4-hydroxycinnamic acid (CHCA) as matrix. The sample solution was pipetted onto the sample plate together with matrix and air-dried. An electrospray ionization quadrupole time-of-flight tandem mass spectra for de novo sequencing were acquired using a Q-TOF instrument (Micromass, Manchester, UK) as described by [31]. Protein identification with the generated data, the obtained peptide masses and the partial sequences, was performed using the Mascot Peptide Mass Fingerprint and MS/MS Ion Search programmes (http://www.matrixscience.com). Identification of the production strain (VTT D-85248) based on the morphology and DNA sequencing was performed at the Identification Services CBS (Utrecht, The Netherlands).

Peptide and protein substrate preparation

The reduced glutathione (GSH, no G4251), urinary trypsin inhibitor fragment (bikunin, no U-4751), and ribonuclease A (RNase A, no R5500) were purchased from Sigma (St. Louis, USA). Peptides, gliotoxin (no. A7665) and holomycin (no. sc-49,029), were purchased from PanReac AppliChem GmbH (Darmstadt, Germany) and Santa Cruz Biotechnology Inc. (Dallas, USA), respectively. The insulin B chain as Bunte salt derivative was kindly provided by Dr. Elisabeth Heine from Deutsches Wollforschungsinstitut (DWI, Aachen, Germany). Insulin chain B and RNaseA were reduced before use with 3% (v:v) 2-mercaptoethanol in the presence of 8 M urea according to the method used by [3]. For the reduction, insulin B was dissolved in 50 mM (NH₄)HCO₃ pH 8.3 containing 8 M urea and 3% (v:v) 2-mercaptoethanol was used as solvent. In the case of RNaseA, 200 mM Tris-HCl buffer (pH 7.4) with urea and 2-mercaptoethanol was used. The solutions were incubated overnight at room temperature. The final concentration of insulin chain B was 1 mM, and concentration of RNaseA was 0.5 mM. After reduction the free sulfhydryl groups were detected with 5,5-dithio-bis(2-nitrobenzoic acid). Gliotoxin and holomycin were reduced with two equivalence of Tris(2-carboxyethyl)phosphine hydrochloride (TCEP-HCl) under nitrogen flow for 1 h prior liquid chromatography mass spectrometry (LC-MS) analysis.

Assay of AtSOX activity and pH and temperature behaviour

Different methods were used to measure AtSOX activity. First DTNB [32] was used for the detection of free sulfhydryl groups in the peptides after the enzymatic reaction. The spectroscopic measurements were done with a Cary 100 Bio UV-vis spectrophotometer (Varian Inc., Houten, the Netherlands). Second, the oxygen consumption measurement with Fibox 3 PreSens fiber-optic oxygen meter (Presens GmbH, Regensburg, Germany) was used to measure the changes in the concentration of dissolved oxygen during enzymatic reaction as described by [28]. GSH (5 mM) was used as a substrate when determining AtSOX activity. The substrate was dissolved in phosphate buffered saline (PBS) containing 68 mM NaCl and 75 mM KH₂PO₄ pH 7.4. Activity of AtSOX was measured by the oxygen consumption assay in the presence of different GSH concentrations (0.25-10 mM) to determine the kinetic parameters. The Michaelis-Menten constant (K_m) and maximum velocity (V_max) were determined with graphing software GraphPad Prism (GraphPad Software Inc., San Diego, USA) using nonlinear curve fitting to the Michaelis-Menten equation.

The thermal stability of AtSOX was determined at 30, 40, 50, 60 and 70 °C. The enzyme preparation was incubated in McIlvaine buffer pH 6.5 in a 0.2 mg/mL protein concentration for 1, 2, 15.5 and 20 h at 30 - 70 °C and also for 15 min at 70 °C, and the residual activity was measured by the oxygen consumption assay. pH stability was determined for AtSOX in a pH range between 2.3 and 10, and the residual enzyme activity was analysed by the oxygen consumption assay after 1 and 20 h incubation. pH optimum was determined by measuring AtSOX activity with the oxygen consumption assay using glutathione in McIlvaine citrate/phosphate buffer (pH 2.3-7.5), 50 mM Tris-HCl (pH 7-9) and 50 mM Glycine-NaOH (pH 8.5-10).

Spectroscopy measurements

UV-vis absorption spectra were measured in 20 mM Tris-HCl pH 7.5 at 20 °C using a Cary 100/300 UV-vis spectrophotometer (Varian Inc., Houten, the Netherlands). In order to release a flavin cofactor, enzyme was thermally denaturated at 100 °C for 15 min followed by centrifugation (13,000 rpm, 10 min). Fluorescence was measured at 20 °C with a Cary Eclipse Fluorescence Spectrophotometer (Varian Inc., Houten, the Netherlands) using AtSOX solution in 20 mM Tris-HCl pH 7.5 in quartz cuvette with four optical faces. FAD fluorescence was recorded by exciting at 450 nm and monitoring emission between 450 and 600 nm.

Inhibition analysis of AtSOX

The AtSOX activity was analysed in a buffer solution and in the presence of different inhibitors. The effect of possible inhibitors on the AtSOX activity was determined using 5 mM GSH as substrate. The tested potential inhibitors were DTT, ethylenediaminetetraacetic acid (EDTA), potassium iodide, magnesium sulphate, manganese sulphate, sodium sulphate, sodium chloride, zinc sulphate, and sodium dodecyl sulphate (SDS). The inhibitors were tested at 10 mM concentration in 200 mM Tris-HCl buffer pH 7.5, and zinc sulphate was tested also at 1 mM concentration. The residual SOX activity was measured by oxygen consumption measurements (Fibox 3 PreSens fiber-optic oxygen meter, PreSens GmbH, Regensburg, Germany). Inhibition by zinc sulphate was confirmed by homovanillic acid (HVA) and peroxidase coupled assay as described in [33], and the assay was done according to [34] using GSH (5 mM) as a substrate. Chemicals, HVA (Cat. no. H1252) and peroxidase type II (Cat. no. P8250), were purchased from Sigma-Aldrich (St. Louis, USA). Fluorescence from the production of a HVA dimer was measured in a black 96-well microtiter plate at 320 nm excitation and 420 nm emission wavelengths using a Varioskan spectral scanning multimode microplate reader (Thermo Electron co., Vantaa, Finland).

Activity of AtSOX with reduced peptides

The ability of AtSOX to oxidise the peptides carrying free sulfhydryl groups was analysed by following the oxygen consumption. The activity of AtSOX (112 nkat) was analysed using reduced RNase A (0.25 mM solution in 200 mM Tris-HCl pH 7.4), and reduced GSH (5 mM solution in 75 mM KH₂PO₄ pH 7.4) as substrates in a total 1.86 mL reaction volume. The enzyme reactions were monitored by following the oxygen consumption of the co-substrate. The reactions were performed at room temperature and monitored for 5-20 min. For NMR spectroscopy, the enzymatically treated GSH was prepared by incubating the substrate with AtSOX for 20 min while the dissolved oxygen was totally consumed, as assessed by oxygen consumption measurements. Freshly prepared and 3 day old GSH solutions (5 mM) were used as a control samples to assess possible auto-oxidation. Reaction mixtures were analysed with ¹H NMR as well as with ¹³C heteronuclear single quantum correlation (¹³C-HSQC) spectroscopy. The enzyme treated sample and the control sample were prepared in 75 mM potassium phosphate pH 7.4 in Shigemi NMR tubes. The NMR spectra were recorded on Varian INOVA 600 MHz NMR spectrometer at 293 K. One dimensional ¹H spectra were recorded with water pre-saturation at 25 °C along with the gradient enhanced ¹³C-HSQC spectroscopy at 20 °C [35]. Homonuclear Total Correlation Spectroscopy (TOCSY) experiments [36] were recorded to confirm the assignment.

Besides oxygen consumption method and NMR spectroscopy, the selected peptides were analysed with a MALDI-TOF MS on an Autoflex II spectrometer (Bruker Daltonik GmbH, Bremen, Germany) using CHCA matrix for peptides and sinapic acid (SA) for proteins. The reduced substrates GSH and insulin B were dissolved in 50 mM (NH₄)HCO₃, bikunin in 75 mM KH₂PO₄, and RNase A in 200 mM Tris-HCl pH 7.4. Purified SOX was used in the experiments in a dosage of 5.6 nkat (4.1 μg protein) and in case of insulin B and RNase A also in 56 nkat (41 μg protein). The enzymatic reactions were performed at 40 °C, except for bikunin that was incubated at room temperature for 20 h. Matrix solutions were prepared by dissolving CHCA or SA in a 1:1 solution of 0.1% trifluoroacetic acid and 100% acetonitrile. The spots for MALDI plate were prepared by using 1:1 proportion of matrix and sample. Typically 1 μl matrix and 1 μl sample were used for MALDI spot, where matrix was spotted first and dried before adding sample.

Reduced gliotoxin and holomycin were used in 200 μM final concentration in 0.1 M phosphate buffer (pH 6.5) for spectroscopic measurements. Gliotoxin and holomycin were used in 440 and 680 μM final concentrations, respectively, for LC-MS analysis. UV-vis absorption spectra from 200 to 600 nm were recorded with Varioskan spectral scanning multimode microplate reader (Thermo Electron co., Vantaa, Finland). The UV-vis absorption spectra from 200 to 600 nm was recorded for 10 min before and after addition of AtSOX (100 μl reaction volume). Reduced peptides incubated (50 min at ambient temperature) with AtSOX or with denaturated AtSOX were analysed with a ultra performance LC (UPLC) combined with a photodiode array detector and SYNAPT G2-S High Definition Mass Spectrometry (Waters, Milford Massachusetts, USA). One microliter of the sample was injected to a LC pre-column. LC-MS system was using a C18 Acquity UPLC VanGuard pre-column (2.1 × 5 mm, 1.7 μm,) and a C18 Acquity UPLC column (2.1 × 100 mm, 1.7 μm). All solvents used were spectral grade. Eluents were 5 mM ammonium acetate 0.1% formic acid in H₂O (A) and in methanol (B). Elution was started with 10% B for 1 min, followed by a linear gradient from 10 to 100% B for 10 min and finally at 100% B for 2 min, with 0.4 mL min⁻¹ flow. In these conditions reduced gliotoxin eluted from the LC column after 5.23 min and gliotoxin standard after 6.02 min. Elution times for reduced holomycin and holomycin standard were 1.69 and 3.74 min, respectively.

Analysis of SOX-coding genes in fungal genomes

The search was carried out as described in [37]. In brief, scaffolds of 33 fungal genomes from [38] were divided in windows of 16 genes that overlapped with two genes. InterPro protein annotations of the genes were then used to look for windows that contained a flavin adenine dinucleotide (FAD)-dependent pyridine nucleotide-disulfide oxidoreductase (PNDR) that recognizes also SOX enzymes (InterPro: IPR000103), nonribosomal peptide synthetases (NRPS) (InterPro: IPR000873) or polyketide synthases (PKS) (InterPro: IPR001227) and cytochrome P450 monooxygenases (P450, InterPro: IPR001128) and/or Zn2Cys6 transcription factors (Zn2, InterPro: IPR001138). Searches were carried out and the results visualised with R essentially as described by [38].

Phylogenetic analysis

A phylogenetic analysis of the selected 25 proteins from the protein family pyridine nucleotide-disulphide oxidoreductase, class-II (InterPro: IPR000103) was carried out. The selected protein sequences were obtained from UniProtKB except AtSOX sequence was retrieved from the genome of A. tubingensis from Joint Genome Institute (JGI) genome portal [39]. The alignment of the sequences was done with MAFFT [40], and alignment was trimmed with trimAl [41]. The aligned sequences were from Ascomycetes species except two proteins were bacterial origin, namely, DepH from Chromobacterium violaceum (UniProtKB: A4ZPY8) and HlmI from Streptomyces clavuligerus (UniProtKB: E2PZ87). A phylogenetic tree was constructed with FastTree [42] and visualised by Geneious version 10.0 created by Biomatters.

Purification and biochemical characterization of AtSOX from Aspergillus tubingensis

Purification of AtSOX was performed in four chromatographic steps from the culture cell-free medium. Fractions containing SOX activity eluted at a 110-260 mM NaCl concentration from a Q Sepharose anion exchange column. Pooled active fractions were applied to a Superdex 75 column for separation by SEC. Further purification of AtSOX was achieved with high resolution anion exchange chromatography. The SOX containing fractions eluted at a 60-175 mM and 100 mM NaCl concentration from two sequential purification steps where Resource Q columns were used. The fraction with the highest specific SOX activity was applied on a Resource Q column once again. AtSOX containing fractions started to elute at a 100 mM NaCl concentration. The purification process was monitored by SDS-PAGE (Additional file 1), and the final yield of protein was 3% of the initial activity.

AtSOX preferred GSH (3 mM) as a substrate over cysteine and DTT. AtSOX had residual activity on L-cystein 14 ± 0.2%, on D-cystein 3 ± 1.2% and on DTT 9 ± 1.8% measured with oxygen consumption assay. Oxygen consumption in AtSOX-catalysed reaction using GSH (3 mM) as a substrate is shown in Additional file 2. Michaelis-Menten behaviour was observed for AtSOX using GSH as a substrate. Michaelis-Menten constant (K_m) was 0.80 ± 0.09 mM and maximum velocity (V_max) (33.15 ± 0.99) × 108 nkat/ml (Fig. 1a). AtSOX retained 89% of the initial activity after 20 h incubation at 40 °C (Fig. 1c). Under identical conditions at 50 and 60 °C, AtSOX showed 75% and 18% residual activity, respectively. At 70 °C, the enzyme was inactivated within 15 min. AtSOX was stable within a broad pH range as it retained 80-90% of the initial activity after 20 h incubation at pH range from 4 to 8.5 (Fig. 1d). The pH optimum of AtSOX was pH 6.5 (Fig. 1b). The enzyme activity was totally inhibited by 10 mM zinc sulphate in the assay conditions. The relative SOX activity was 4% with 1 mM zinc sulphate. The inhibition of zinc sulphate was confirmed with HVA-peroxidase coupled assay (Additional file 3). The other analysed compounds did not inhibit AtSOX activity, or had only minor effect on it, as residual activity was more than 90% in the presence of the analysed compound. For example the AtSOX retained 92% of its activity in 10 mM of potassium iodide solution and denaturant SDS did not inhibit enzyme activity. The flavoenzymatic nature of AtSOX was determined by spectral analyses using UV-vis and fluorescence spectrophotometry. The absorption spectra were recorded before and after thermal denaturation of AtSOX. The flavin cofactor was released by denaturation and detected in solution (Fig. 2). The UV-vis spectrum of AtSOX showed the characteristic peaks of flavoproteins with absorbance maxima at 275, 365 and 445 nm with a shoulder at 475 nm. AtSOX fluorescence emission spectrum had a peak at 525 nm when excited at 450 nm, which is characteristic for flavins.

The purified AtSOX was subjected to Edman degradation, and trypsin digestion to perform peptide mass fingerprinting (PMF), and to identify the protein. The PMF gave six peptides made of more than six amino acids (Fig. 3). Mining the available genome of A. tubingensis [39] with the sequences of the identified peptides allowed the identification of a single SOX-coding gene on scaffold 13. This gene coded for a 41.8 kDa secreted protein with a predicted N-terminus, after removal of the signal peptide, SSIPQ. This protein contained four of the six peptides identified from AtSOX, and the other two peptides were present but carry either an amino acid insertion or deletion. Due to the high sequence similarity with secreted SOX from A. niger (UniProtKB: A2QUK3), the strain was re-identified at the Identification Services CBS (Utrecht, the Netherlands). The identification results confirmed that the used VTT strain D-85248 was A. tubingensis.

Oxidation of peptides and proteins with AtSOX

The ability of AtSOX to oxidise peptide-bound cysteine was tested with reduced model peptides and protein namely GSH, bikunin, insulin B chain, gliotoxin, holomycin and RNase A as substrates. Reaction products were analysed by MALDI-TOF MS and NMR spectroscopy. The MALDI-TOF MS analysis of the reaction mixtures showed a dimer (613 Da) formation with GSH after overnight enzymatic treatment whereas only a peak correspond to the monomer (308 Da) was detected in the mass spectrum of the untreated substrate. The enzymatic disulphide bond formation between GSH peptides was observed also with the spectroscopic measurements at 250 nm (Additional file 4). No dimer formation was detected with bikunin (919 Da), insulin B chain (3496 Da) or RNaseA (13,700 Da) as substrates by MALDI-TOF analysis. Other analysed small peptides – gliotoxin (326 Da) and holomycin (214 Da) – were not enzymatically oxidised as observed in the LC-MS analysis. Holomycin was auto-oxidised as oxidation was observed also with the denaturated AtSOX enzyme after 50 min incubation.

Oxidation of GSH and RNase A by AtSOX was also followed. AtSOX could oxidise only the tripeptide GSH, which was the shortest tested peptide. The final product formed in the reaction of AtSOX was analysed using NMR spectroscopy. Hereby, it was possible to follow the change in the cysteine oxidation state at atomic resolution. The oxidation was confirmed by following the oxygen consumption for 20 min after addition of AtSOX to the tripeptide L-GSH. The fresh GSH and enzyme-treated sample were analysed with one dimensional¹H NMR. The one dimensional ¹H spectra of GSH with and without AtSOX enzyme are shown in Fig. 4. Upon addition of AtSOX to GSH solution, the two degenerate protons bonded to C_β of cysteine residue became non-degenerated due to the formation of covalent disulphide bond. In addition, the two dimensional ¹³C-HSQC spectrum, exhibiting one-bond ¹H-¹³C connectivities (Fig. 5a) showed a significant change in the ¹³C chemical shift of the Cys ¹³C _β -¹H correlations as common for disulfide bond formation. This confirmed the fast dimerization of the GSH tripeptide through the disulfide bond upon addition of AtSOX. The control GSH substrate was checked once more 3 days after preparation with the gradient-enhanced ¹³C-HSQC spectrum to analyse auto-oxidation. The ¹H, ¹³C cross peak, corresponding to the oxidised Cys C_ß form, started to appear which refers to dimerization (Fig. 5b). The dimerization of GSH occurred also without the SOX enzyme, although at a much slower rate. Measurement of the pure GSH sample 3 days after preparation showed an evidence of both monomeric and dimeric conformations in a ratio of about 5:1.

SOX-coding genes in fungal genomes and phylogenetics

Aspergilli are known for their expanded secondary metabolism in comparison to yeasts [38], while fungi in general are known for their metabolic gene clusters, which are involved in, for example, synthesis of secondary metabolites [43] and catabolism of nutrients [37]. In order to suggest physiological roles for fungal secreted SOXs, fungal genomes were analysed. In brief, 33 fungal genomes were searched with custom R scripts for short genomic regions that would contain a SOX-coding genes and type of fungal secondary metabolism genes i.e. nonribosomal peptide synthetases (NRPS), polyketide synthases (PKS) and associated genes, i.e. cytochrome P450 monooxygenases (P450), and/or Zn2Cys6 transcription factors (Zn2).

No genomic regions containing genes coding for PKS and SOX could be found, instead ten regions with SOX, NRPS and P450 or Zn2 coding genes were identified (Fig. 6). Two of these regions were the gliotoxin synthesis cluster of A. fumigatus [44] and the candidate gliotoxin cluster of Trichoderma reesei [45]. In addition, the A. tubingensis candidate gliotoxin cluster was detected (Fig. 6a). Based on protein clustering of SOX enzymes [38] the SOX enzymes found in the chromosomal gene clusters were divided in two groups: SOX enzymes of candidate gliotoxin gene clusters (Fig. 6a) and SOX enzyme of candidate secondary metabolism gene clusters (Fig. 6b).

AtSOX sequence retrieved from Aspergillus tubingensis genome project was aligned with selected 24 SOX and thioredoxin reductase sequences from IPR000103 InterPro family, and the phylogenetic tree was constructed based on the alignment (Fig. 7). The C-X-X-C-motif was found from the majority of the aligned sequences (Additional file 7). It was observed from the phylogenetic tree that Aspergillus SOXs were evolutionary closely related. Oxidoreductases with activity on nonribosomal peptides, namely GliT, DepH and HlmI, were also closely related to AtSOX.