Arxula adeninivorans recombinant adenine deaminase and its application in the production of food with low purine content




Construction of a transgenic Arxula adeninivorans strain that produces a high concentration of adenine deaminase and investigation into the application of the enzyme in the production of food with low purine content.

Methods and Results

The A. adeninivorans AADA gene, encoding adenine deaminase, was expressed in this yeast under the control of the strong inducible nitrite reductase promoter using the Xplor®2 transformation/expression platform. The recombinant enzyme was biochemically characterized and was found to have a pH range of 5·5–7·5 and temperature range of 34–46°C with medium thermostability. A beef broth was treated with the purified enzyme resulting in the concentration of adenine decreasing from 70·4 to 0·4 mg l−1.


It was shown that the production of adenine deaminase by A. adeninivorans can be increased and that the recombinant adenine deaminase can be used to lower the adenine content in the food.

Significance and Impact of the Study

Adenine deaminase is one component of an enzymatic system that can reduce the production of uric acid from food constituents. This study gives details on the expression, characterization and application of the enzyme and thus provides evidence that supports the further development of the system.


Dietary and cellular nucleic acids are constantly being degraded. The degradation process follows the route: DNA, RNA to mononucleotides, further to nucleotides and finally to free bases. The purine nucleosides and free bases are then deaminated to form hypoxanthine and xanthine by conversion via AMP, adenosine and inosine or by various specific deaminases such as adenine deaminase (Vogels and Van der Drift 1976). Further degradation leads to uric acid, which is the final product of purine degradation in primates. Underexcretion of urate by the kidneys or a disorder in purine metabolism results in the concentration of uric acid in body fluids being elevated, that is, hyperuricemia. Up to 10% of hyperuricemic individuals develop gout, an inflammatory joint disease. Gout is a painful disorder, characterized by deposition of sodium urate in joints, swelling and recurrent attacks of acute arthritis (Richette and Bardin 2010). Treatment of severe gout remains a challenge in medicine. Standard management of an acute attack is based on colchicine and nonsteroidal anti-inflammatory drugs that provide rapid relief of symptoms. The urate-lowering therapy maintains the monosodium urate concentration below its saturation point. The drugs for long-term therapy are xanthine oxidase inhibitors, which prevent the degradation of xanthine, and uricosurics agents, which increase the excretion of uric acid. The third way to deal with gout is by reducing dietary purines to maintain low serum urate levels. A reduction in dietary purines may reduce uric acid content by 15–20%, because approximately one-third of total body urate is produced from dietary purines (Wortmann 2002; Schlesinger 2005). Thus, we introduce an innovative approach to avoid uric acid accumulation and prevent the underlying cause of gout. This approach entails enzymatic reduction in purine content of the food during the production or preparation processes using all of the enzymes involved in the degradation pathway: purine nucleoside phosphorylase, adenine deaminase, guanine deaminase, xanthine oxidoreductase and urate oxidase. Adenine deaminase, the first enzyme in the purine degradation pathway, was overexpressed, characterized and successfully applied to lower adenine content of purine-rich foods and demonstrating the principle of this approach.

Adenine deaminase (ADase, adenase; EC catalyses the irreversible deamination of adenine to produce hypoxanthine and ammonia in the purine degradation pathway. To date, this enzyme has only been found in bacteria (e.g. Bacillus subtilis and Escherichia coli) and lower eukaryotes such as Saccharomyces cerevisiae and Crithidia fasciculata (Nygaard et al. 1996), but not in animals (Ribard et al. 2003; Pospisilova and Frebort 2007). Adenine deaminase has been reported as an intracellular enzyme in different yeast species (Kidder et al. 1977; Deeley 1992; Pospisilova et al. 2008). All structurally characterized deaminases belong to the subfamily of α/β barrel enzymes and bind one divalent cation in the active site (Ribard et al. 2003; Kamat et al. 2011). To date, the transcriptional regulation of adenine deaminase has not been thoroughly studied, except for the analysis of Aspergillus nidulans nadA gene, which gives initial information on the regulation of expression of this gene (Oestreicher et al. 2008).

Here, we report the generation of a transgenic A. adeninivorans strain that produces a high concentration of recombinant adenine deaminase and its application in the production of low-purine-content food. The AADA gene encoding adenine deaminase was isolated from the yeast A. adeninivorans, an ascomycetous, nonpathogenic organism that exhibits unusual biochemical properties (Wartmann et al. 1995, 2000). It assimilates a broad spectrum of substances such as purines, tannic acid, starch and n-alkanes as carbon and/or nitrogen sources (Middelhoven et al. 1984, 1991; Gienow et al. 1990; Van der Walt et al. 1990).

Böer et al. (2009a) introduced the Xplor®2 transformation/expression platform, an optimized platform for high expression of heterologous and homologous genes in A. adeninivorans, which allows multicopy integration of vector cassettes with the genes of interest, which results in the production of recombinant proteins at a high level.

This work is part of a programme to produce all of the enzymes involved in purine degradation so that they can be used to pretreat food and thus decrease uric acid production in the body.

Materials and methods

Strains, media and growth conditions

Escherichia coli TOP 10 [F-mcrA Δ(mrr-hsdRMS-mcrBC) Φ80lacZΔM15 ΔlacΧ74 recA1 araD139 Δ(araleu) 7697 galU galK rpsL (StrR) endA1 nupG] from Invitrogen (Carlsbad, CA, USA) was used as the host for DNA manipulation and plasmid isolation. The growth medium for E. coli was LB medium supplemented with ampicillin (100 μg ml−1; AppliChem, Darmstadt, Germany) or kanamycin (50 μg ml−1; Roth, Karlsruhe, Germany) for selection.

Arxula adeninivorans LS3 – wild-type strain, isolated from wood hydrolysates in Siberia (Russia) and deposited as A. adeninivorans SBUG 724 in the strain collection of the University of Greifswald – and its auxotrophic mutant G1212 [aleu2 atrp1::ALEU2] transformed with YIC102–AADA6H or YRC102–AADA6H cassette were used in this study (Kunze and Kunze 1994; Samsonova et al. 1996; Wartmann and Kunze 2000; Steinborn et al. 2007).

Wild-type and transgenic yeast strains were cultivated under nonselective conditions in a yeast extract–peptone–dextrose growth medium (YEPD) or under selective conditions in yeast minimal medium (YMM) with glucose (1% w/v) as a carbon source (Tanaka et al. 1967; Rose et al. 1990). Cultivation of yeast was carried out in two steps: a preculture with an initial OD600 = 0·1 in YMM–glucose with 43·5 mmol l−1 NH4H2PO4 as N source (YMM–glucose–NH4H2PO4) was cultured at 30°C, for 24 h (growth phase). Wild-type cells were then shifted to YMM–glucose–NH4H2PO4, YMM–glucose with 43·5 mmol l−1 NaNO3 or different purines as N sources and grown at 30°C for up to 72 h (production phase). Recombinant and control G1212 strains were shifted to YMM–glucose–NaNO3 at 30°C for up to 72 h.

The biomass (dry cell weight, dcw) of yeast samples was determined from 1 ml cell culture that was washed with H2OMQ and freeze-dried.

The water used in this work, defined here as H2OMQ, was purified from tap water using Milli-Q Advantage A10 Ultrapure Water Purification System (Merck Millipore, Billerica, MA, USA).

Transformation techniques and isolation of nucleic acids

Transformation of E. coli was performed according to Hanahan (1983), and yeast cells were transformed according to Rösel and Kunze (1998).

Plasmid DNA was purified from 2 ml of overnight culture using buffers P1, P2 and P3 (Qiagen, Hilden, Germany). After precipitation and washing steps, the isolated plasmid DNA was resuspended in H2OMQ. DNA bands were purified from agarose gels using NucleoSpin® Extract II (Macherey-Nagel, Düren, Germany). Total RNA was prepared from the cell pellet using Z6 extraction buffer (8 mol l −1 guanidine–HCl, 20 mmol l−1 MES, 20 mmol l−1 EDTA, 50 mmol l−1 β-mercaptoethanol, pH 7·5; Logemann et al. 1987). cDNA was synthesized according to the manual provided with RevertAid™ H Minus First-Strand cDNA Synthesis Kit (Fermentas, St Leon-Rot, Germany) using oligo(dT) primers.

Sequencing of all nucleic acids was performed using an automatic laser fluorescence DNA sequencer (Pharmacia, Uppsala, Sweden). The nucleotide sequence of isolated gene has been entered into the GenBank/EMBL data libraries (HF562995).

Real-time PCR

AADA transcript accumulation was quantified by the amplification of appropriate cDNA fragments by PCR in the presence of SYBR Green fluorophore (Power SYBR® Green PCR Master Mix, Applied Biosystems, Foster City, CA, USA) using ABI PRISM 7700 Sequence Detection System (Applied Biosystems). Primers for qPCR are listed in the Table 1. Reference (housekeeping) genes used for analysis were AHSB4, TEF1, ALG9 and TFIID (Rösel and Kunze 1995; Wartmann et al. 2003). Calculations were made using the Hellemans modification of the Pfaffl method (Pfaffl 2001; Hellemans et al. 2007).

Table 1. Oligonucleotides used in this study
NameNucleotide sequenceFeature
  1. Underlined letters – restriction enzymes site; bold type letters – 6xHis-tag encoding sequences.

ALG9-qPCR-fwCAT GGG CCA AGG TAT ACT GAmplification in qPCR
ALG9-qPCR-rvAGA ATG CGA CCG ATA CCAAmplification in qPCR
TEF-qPCR-fwCTC TTG GAC GAT TCG CCAmplification in qPCR

Enzyme assays and protein analysis

For measurement of enzyme activity in the crude extracts, 2 ml of yeast culture was harvested, washed with H2OMQ and suspended in 500 μl 0·1 mol l−1 potassium phosphate buffer pH 7·4. Cells were homogenized using 500 μl 0·5-mm silica beads (BioSpec Products, Bartlesville, OK, USA) in a Mixer Mill MM 400 (Retsch, Haan, Germany) for 3 min at 30 Hz, 4°C. The protein extract was separated from cell debris and silica beads by centrifugation at 17 000 g for 5 min, 4°C.

Adenine deaminase activity was assayed by adding the enzyme to a reaction mixture of 33 mmol l−1 potassium phosphate buffer (pH 7·4), 7 mmol l−1 DTT (only for tests with purified protein) and 1·65 mmol l−1 adenine (or water as a control) in a total volume 300 μl. The reaction was terminated by heating for 5 min at 95°C and the precipitate separated by centrifugation. 200 μl of supernatant was used for NH4+ determination at 690 nm according to the Merck Spectroquant Ammonium Test manual (Merck, Darmstadt, Germany). Activity units were calculated from the standard curve for ammonia. One unit of enzymatic activity is defined as that amount that catalyses the deamination of 1 μmol of adenine per minute at 40°C.

α-d-Mannosidase (EC was used as a marker for vacuolar enzymes and assayed according to Opheim (1978). Glucose-6-phosphate-dehydrogenase (EC served as a cytosolic marker and was assayed according to Gibon et al. (2002, 2004).

Protein concentrations were determined by the dye-binding method (Bradford 1976) using purified bovine serum albumin as standard.

Purified wild-type adenine deaminase was prepared from cells grown on 2·5 mmol l−1 hypoxanthine as nitrogen source as described in Materials and Methods. Cells from a 1 l culture were harvested after 6 h of cultivation (16·2 mg of wet cells) and resuspended in 5 ml of potassium phosphate buffer (50 mmol l−1, pH 7·4). The cells were then frozen with liquid nitrogen and ground with mortar and pestle and resuspended in buffer. The slurry was centrifuged, and ammonium sulfate was added to the crude extract to 65% saturation, stirred for 30 min and centrifuged. The ammonium sulfate concentration in the supernatant was then raised to 95%. The resulting precipitate was redissolved in potassium phosphate buffer and dialysed against potassium phosphate buffer (50 mmol l−1, pH 7·4) for 12 h. Adenine deaminase was purified using a DEAE-Sepharose ion exchange column equilibrated with 0·05 mol l−1 potassium phosphate buffer (pH 7·4). The enzyme was eluted with phosphate buffer containing a linear gradient of KCl. Active fractions were then loaded onto a gel filtration column (Superdex™ 200) equilibrated with 0·05 mol l−1 sodium phosphate buffer (pH 7·4) containing 0·15 mol l−1 sodium chloride. The results of gel filtration were used for molecular weight estimation. The calibration standards used were dextran blue (Mr 2 000 000), ferritin (Mr 440 000), catalase (Mr 232 000), alcohol dehydrogenase (Mr 150 000), bovine serum albumin (Mr 68 000) and ovalbumin (Mr 45 000).

Recombinant Aadap with a His-tag (Aada6hp) was purified by column chromatography on Ni2+-charged His-bind resin (Novagen, Madison, WI, USA). Eluted enzyme was dialysed against 50 mmol l−1 potassium phosphate buffer, pH 7·4, and concentrated using a Spin-X UF concentrator (Corning, Life Sciences, Tewksbury, MA, USA).

Aada6hp was analysed under denaturing conditions on 11% SDS polyacrylamide gel (SDS-PAGE). After electrophoresis, protein samples were blotted to a polyvinylidene difluoride (Roth) membrane (Kunze et al. 1998). The blots were treated with anti-His antibody as described by Böer et al. (2009b).

Subcellular fractionation

Protoplasts preparation

Yeast minimal medium–glucose–NaNO3, 50 ml, was inoculated with transgenic A. adeninivorans at OD600 c. 0·1 and grown to OD600 1·2. After harvesting, the cells were washed in 10 ml 0·9% NaCl, centrifuged 5 min at 3800 g and incubated for 10 min at 20°C in 10 ml solution VBM (10 mmol l−1 Tris–HCl, 5 mmol l−1 EDTA, pH 9·0, 1% β-mercaptoethanol). The cells were then washed twice with 10 ml 0·9% NaCl and once with 1 mol l−1 sorbitol and resuspended in 10 ml of PP medium (1 mol l−1 sorbitol, 30 mg lysing enzyme from Trichoderma harzianum, Sigma-Aldrich, St Louis, MO, USA) and incubated for 2 h at 20°C with gentle shaking. Protoplasts were examined microscopically and then washed twice with 5 ml 1 mol l−1 sorbitol.


Protoplasts were suspended in 0·4 ml 1·8 mol l−1 sorbitol and lysed by dropwise addition of 3·2 ml chilled 10 mmol l−1 Tris–HCl, pH 7·5. Cell debris was removed by centrifugation at 1100 g for 4 min, and the supernatant was loaded onto a 15–55% sucrose density gradient and centrifuged in an ultracentrifuge (Beckmann LE-70, Brea, CA, USA) for 2·5 h at 150 000 g, 4°C. 0·5-ml fractions were collected from a range of gradients and stored in 1·5-ml tubes on ice for up to 1 week.

Construction of AADA6H expression plasmid and generation of recombinant strain

To assess the expression of AADA6H in A. adeninivorans, the gene was amplified by PCR using A. adeninivorans chromosomal DNA as template. The primers (listed in Table 1) used for amplification flanked the gene sequence with EcoRI and BamHI restriction sites and incorporated the coding sequence for 6 histidines at 3′ end of the ORF. The amplified fragment, EcoRI–AADA6HBamHI, was inserted into pCR®4-TOPO® vector (TOPO TA Cloning® Kit for Sequencing, Invitrogen). It was released from the plasmids by digestion with EcoRI and BamHI and cloned into an appropriately linearized pBS–AYNI1–PHO5-SA vector between A. adeninivorans-derived inducible AYNI1 promoter and S. cerevisiae-derived PHO5 terminator. The promoter–gene–terminator sequence was excised using SalI-ApaI restriction sites and introduced to the Xplor2.2 vector (Böer et al. 2009a). This vector contains fragments of 25S rDNA, which are interrupted by the selection marker cassette (ALEU2 promoter, ATRP1 m gene, ATRP1 terminator; Steinborn et al. 2007) and a multicloning site. The resulting plasmid Xplor2.2–AYNI1–AADA6H consists of the AYNI1–AADA6H–PHO5 module and selection module flanked by 25S rDNA fragments and an E. coli section encoding the resistance marker. To prepare the cassettes for transformation, Xplor2.2–AYNI1–AADA6H and the control plasmid Xplor2.2 lacking AADA6H were digested with AscI (YRC) and/or with SbfI (YIC) to remove the E. coli sequences including the resistance marker. The resulting restriction products YRC102–AADA6H, YIC102–AADA6H and YRC102, YIC102 (controls) were used to transform A. adeninivorans G1212 (Fig. 1).

Figure 1.

Physical map of the yeast integrative expression cassettes (a) YRC102–AADA6H and (b) YIC102–AADA6H. Both cassettes contain the selection marker ATRP1m together with the ALEU2 promoter and one copy of the expression module: AYNI1 promoter–AADA6HPHO5 terminator. Only the YRC102–AADA6H cassette contains 25S rDNA sequences for targeting.

Yeast-positive transformants were selected by tryptophan auxotrophy in YMM–glucose–NH4H2PO4. The cells were then stabilized by passaging on selective (YMM–glucose–NH4H2PO4) and nonselective (YEPD) agar plates to attain a high level of protein production (Kuriyama et al. 1992; Klabunde et al. 2003).

HPLC analysis of adenine deaminase substrates and products

The adenine deaminase Aada6hp was used to hydrolyse adenine in beef broth. The progress of reaction was followed using HPLC (TSKgel Amide-80 analytical column; 4·6 mm ID × 25 cm, Tosoh Bioscience, Tokyo, Japan). 4·2 U of Aada6hp was added to 1 g of beef broth (REWE Bio Klare Rinderbrühe) dissolved in 1·5 ml 0·5 mol l−1 NaOH and 18·5 ml 0·1 mol l−1 potassium phosphate buffer, pH 3·0, and incubated at 40°C for up to 20 min. After the reaction, the samples were heated at 100°C for 5 min and centrifuged, and 20 μl was loaded on the column. The samples were fractionated with 0·1% trifluoroacetic acid/acetonitrile (15 : 85 v/v). The flow rate was 1·2 ml min−1, and the detector was set at 260 nm. The signals of undigested and digested beef broth were compared, and their retention times were compared with those of pure adenine, guanine, xanthine, hypoxanthine and uric acid.


Isolation of the AADA gene

The gene encoding adenine deaminase, AADA, was isolated from genomic DNA by PCR using gene-specific primers (Table 1). The gene was found by comparing annotated A. adeninivorans genomic sequences (manuscript in progress) with ADA sequences from other fungi (BLAST). A 1065-bp sequence (Gene ID: HF562995) was identified that contains an ORF and encodes a protein of 355 amino acids. It is located on smallest chromosome Arad0A from 1417700 to 1418764 bp and has no introns. The predicted sequence of Aadap showed a high degree of identity to the adenine deaminases of the yeast Dekkera bruxellensis (55%; ABP33170.1), Candida glabrata (54%; XP_447966.1), Scheffersomyces stipitis (54%; XP_0013 84132.1), S. cerevisiae (54%; NP_014258.1), Kluyveromyces lactis (53%; XP_454997.1), Schizosaccharomyces pombe (52%; NP_595071.1) and Candida albicans (53%; XP_714841.1).

Purine-dependent expression of the AADA gene

Arxula adeninivorans can use different purines as N source (Middelhoven et al. 1984). The LS3 strain was cultivated with adenine, hypoxanthine and uric acid, and the levels of AADA gene product were compared with control cells grown on YMM–glucose–NH4H2PO4.

Figure 2a shows that hypoxanthine is the best inducer, followed by adenine and xanthine. Uric acid seems to have no influence on induction of the enzyme (activity the same as in the control). A time course experiment with adenine and hypoxanthine at optimal concentrations also demonstrated a high level of inducer activity (Fig. 2b).

Figure 2.

The influence of different inducers on adenine deaminase activity and transcript level. (a) LS3 strain was cultivated as described in Materials and Methods with shift to YMM–glucose supplemented with purines in different concentrations. Samples were taken after 4 h of induction. (b) Time-dependent induction with adenine and hypoxanthine (both 5 mmol l−1). (c) The fold change in expression of the target gene at each time point and (d) at each inducer concentration. Legend: (□) adenine, (▲) hypoxanthine, (○) uric acid, (●) NH4Cl. Error bars represent standard error.

mRNA levels for Aadap synthesis followed the same patterns observed for the enzyme. Transcript accumulation was measured for each inducer concentration and period of induction. The wild-type strain A. adeninivorans LS3 cultivated on YMM–glucose–NaNO3 served as a control (nontreated sample). The level of AADA transcript was monitored for 24 h after cultivation with adenine and hypoxanthine as inducers. The transcript level reached maximum 2 h after the shift in both cases; however, it was twofold higher in the medium supplemented with hypoxanthine, which is in agreement with enzyme activity results. The mRNA levels then decreased over the next 10 h (Fig. 2c).

The transcript accumulation in relation to inducer concentration was measured 2 h after the shift where transcription was maximal. The medium was supplemented with 0·5, 1, 3, 4 and 5 mmol l−1 adenine, hypoxanthine and uric acid (treated samples) or with the same concentrations of NaNO3 (nontreated samples). The results confirm the outcome of enzymatic tests, where the hypoxanthine exhibited a twofold more efficient induction than adenine and uric acid having limited influence on AADA gene induction (Fig. 2d).

Expression of the recombinant Aada6hp in A. adeninivorans

The AADA gene, with a polyhistidine tag encoding sequence fused to the 3′ end of the ORF under the control of the inducible AYNI1 promoter, was overexpressed in the auxotrophic mutant strain A. adeninivorans G1212 [Δatrp1]. Cassettes with the AADA expression module (YRC102–AADA6H, YIC102–AADA6H; Fig. 1) and control (YIC-102 and YRC-102) were prepared as described in the Material and Methods section. After integration of the cassettes into the genome, a number of selected clones (YICs and YRCs) were passaged to establish high plasmid stability. The transformants were then cultivated in YMM–glucose–NH4H2PO4 for 24 h and shifted to YMM–glucose-NaNO3 for 4 h to induce the AYNI1 promoter. Cells were harvested and screened for adenine deaminase activity. Elevated activity was observed in 98% of transformants, when compared to the controls G1212/YIC102 and G1212/YRC102 cultured in identical conditions (results not shown). Moreover, the activity exhibited by the best transgenic strain (G1212/YIC102–AADA6H-82) was nearly 40 times higher than the control (G1212/YIC102). A time course test to determine the optimal induction time and to compare it with the control and the wild-type strain was undertaken (Fig. 3). The activity of transformants increased dramatically between 6 and 10 h after the shift to the inducer-containing medium. After 10 h, activity decreased to one-third of the highest value, but remained stable thereafter. Similar behaviour was observed for LS3 wild-type strain, although enzymatic activity is almost 20 times lower than in the transgenic strain 82, while the control strain G1212/YIC102 did not increase at all over the sampling period.

Figure 3.

Time course experiment for comparison of the best G1212/YIC102–AADA6H-82 transformant (white bars) with G1212/YIC102 (black bars) control and LS3 wild-type strain (striped bars) induced with 2·5 mmol l−1 hypoxanthine. All enzyme activities were measured in crude extracts. The recombinant strain and control strain were cultivated in YMM–glucose–NH4H2PO4 at 30°C for 24 h and shifted to YMM–glucose–NaNO3. Samples were taken after 0, 2, 4, 6, 8, 10, 12, 24, 36, 48, 60 and 72 h after the shift. LS3 strain was also cultivated in YMM–glucose–NH4H2PO4 at 30°C for 24 h, but shifted to YMM–glucose supplemented with 2·5 mmol l−1 hypoxanthine. Samples were taken in the same time points after the shift. Error bars represent standard error.

To improve the yield, the cultivation conditions of G1212/YIC102–AADA6H were optimized. The transgenic strain and the control strain were cultivated in different conditions: 30, 37 and 45°C and 1, 2 and 5% w/v glucose. Results (not presented) of the enzyme assay showed that the highest induction rate occurred at 37°C in 2% glucose. Figure 4 shows the enzyme yield (Y(P/X), enzyme formation P [U] per biomass X [g]) in optimal conditions for the transformant and the control over time. The yield reached its maximum at 5 h and then constantly decreased to be close to the initial concentration. The activity of crude extract in U ml−1 was highest after 10h cultivation, but decreased thereafter.

Figure 4.

High-efficiency production of Aada6hp in shake flasks. (a) Production strain G1212/YIC102–AADA6H, (b) control strain G1212/YIC102. Both strains were cultured in shake flasks up to 50 h at 37°C in 2% glucose YMM with shift from NH4H2PO4 to NaNO3 as N source. Dry cell weight was determined from 1 ml culture medium (▲, dcw [g l−1]), as well as adenine deaminase activity (■) [U ml−1] that resulted in calculation of adenine deaminase output (×) Y(P/X) [U g−1 dcw].

Purification and molecular weight estimation

The wild-type enzyme was partially purified (5·26 fold) as described in the Materials and Methods section and had a specific activity of 9·89 U mg−1. The purification steps are presented in Table 2. The calculated molecular mass for the enzyme after filtration on Superdex™ 200 was c. 40 kDa. A protein band of the same size can be identified on SDS-PAGE (Fig. 5a). The molecular weight of recombinant adenine deaminase was analysed using the same methods with an isolated protein of the same size, which was visible on the denaturing gel in the correct position (results not shown).

Table 2. Purification steps for wild-type adenine deaminase
StepTotal protein [mg]Total activity [U]Specific activity [U mg−1]Yield [%]Purification fold
Crude extract741·91396·11·881001
(NH4)2SO4 precipitation112·4131·11·1611·040·84
DEAE-Sepharose column16·8140·98·3410·094·43
Superdex™ 200 column4·237·49·892·685·26
Figure 5.

Purification steps of (a) wild-type enzyme visualized with Coomassie. M – protein marker, 1 – crude extract, 2 – 65–95% ammonium sulfate precipitation, 3 – DEAE-Sepharose chromatography column, 4 – gel filtration on Superdex™200, 5 – recombinant adenine deaminase as a control; (b) purification steps of recombinant Aada6hp on His-bind resin. M – protein marker, 1 – crude extract, 2 – flow through the column, 3 – washing step with binding buffer, 4 – washing step with washing buffer, 5 – eluted and concentrated protein, 6 – crude extract of noninduced, wild-type strain. Left side – Coomassie staining, right side – Western blot with anti-His-tag antibodies.

Recombinant adenine deaminase was used for characterization of the protein because of the low concentration and poor purity of the wild-type enzyme. The recombinant Aada6hp with the C-terminal His-tag was purified on affinity chromatography column. 2·8 mg of pure protein was obtained from 1 l of culture. The purity of eluted protein is shown on Fig. 5b.

Properties and subcellular localization of the recombinant Aada6hp

The recombinant adenine deaminase was subjected to basic biochemical analysis. It exhibits relatively broad pH range of 5·5–7·5 with the optimum between 6·5–7·0 in either phosphate or HEPES buffer. The highest activity was at 40°C, but the enzyme loses 35% of its catalytic function in 1 h, whereas at 30°C, the activity of the enzyme remains stable. At 50°C, the activity drops within 10 min–0 U mg−1. The enzyme had maximal stability at 30°C of 24 h in HEPES buffer at pHs between 5·5 and 7·5.

Investigation into substrate specificity included all purine nucleosides, selected purines and their analogues (Table 3). Adenine deaminase has only one natural substrate, adenine, but the enzyme can also hydrolyse 2,6-diaminopurine, 6-chloropurine and 2-amino-6-bromopurine. The Michaelis constant (KM) for adenine and 2,6-diaminopurine was calculated using nonlinear regression method (Excel Solver) and is as follows: adenine – 0·66 mmol l−1, 2,6-diaminopurine – 0·47 mmol l−1, with kcat of 20·8 1 s−1 and 12·3 1 s−1 respectively. The parameters for 6-chloropurine and 2-amino-6-bromopurine could not be measured (see the Discussion section for reasons).

Table 3. Substrate spectrum for adenine deaminase. Final concentration of all substrates was 0·33 mmol l−1
SubstrateRelative activity [%]

The influence of metal ions and other additives on Aada6hp activity was measured. Mg2+, Mo2+, Mn2+ and Ca2+ had either no effect or only a weak effect. Addition of Cd2+, Fe2+, Cu2+ and Zn2+ caused a major decrease in enzyme activity. The presence of EDTA in the reaction mixture was not inhibitory (Table 4).

Table 4. Effect of metal salts on adenine deaminase activity at 1 mmol l−1 final concentration
SupplementRelative activity [%]

The localization of adenine deaminase in the cell was investigated. Sequence analysis in HMMTOP transmembrane topology prediction server (Tusnady and Simon 2001) did not reveal transmembrane segments. Similarly, the SOSUI algorithm (Hirokawa et al. 1998) predicted that adenine deaminase is a soluble protein without a signal peptide.

Twenty three fractions from a sucrose density gradient centrifugation of an A. adeninivorans preparation were analysed for Aada6hp and marker enzyme activities as described in Materials and Methods. Vacuolar α-d-mannosidase was found in fractions 17–20. Glucose-6-phosphate-dehydrogenase (cytosolic) activity was measured in fractions with the least sucrose concentration (19–23). Another marker protein for mitochondria, Afrd1p, fumarate reductase, was detected in fractions 14–16 by Western blot analysis with anti-Afrd1p antibodies (results not shown). Adenine deaminase was found in fractions 18–23, which also suggests a cytosolic localization (Fig. 6).

Figure 6.

Subcellular localization of Aada6hp. Arxula adeninivorans G1212/YIC102-AADA6H cell walls were removed by enzymatic lysis; proto-plasts were disrupted and loaded onto a sucrose density gradient, and cell components were separated by centrifugation. Collected fractions were analysed for subcellular marker proteins; (●) adenine deaminase, (Δ) α-d-mannosidase, (▲) sucrose refractive index, (□) glucose-6-phosphate-dehydrogenase.

Use of adenine deaminase to reduce adenine in food

Initial experiments to test the effectiveness of recombinant adenine deaminase for food treatment were carried out using a mixture of purines and a yeast extract (Ohly® KAT GMP; Ohly, Hamburg, Germany). Both solutions were treated with the recombinant enzyme, and the products of reaction were analysed by HPLC. The analysis revealed the absence of adenine and an increase in hypoxanthine content. In the next experiment, the purine degradation in beef extract was examined. The beef broth, however, had an unexpectedly low level of adenine and had to be supplemented with 50 mg l−1 adenine to test our assumptions. The spiked sample was treated with 4·19 U adenine deaminase for up to 20 min at 40°C. Analysis showed that adenine concentration dropped immediately from 70·4 –0·4 mg l−1 (Fig. 7). Simultaneously, accumulation of the product, hypoxanthine, was observed (increasing from 30·4 to 83·7 mg l−1). The concentration of other purines remained constant over the incubation period.

Figure 7.

Chromatogram after HPLC analysis of beef broth (50 mg ml−1) and additional adenine (0·05 mg ml−1) treated with 4·19 U of Aada6hp. Dashed line – adenine (1), hypoxanthine (2), xanthine (3), guanine (4a,b) and uric acid (5) standards. Thick line – starting point, beef broth components before reaction. Thin line – end point, beef broth after 20-min incubation with adenine deaminase. Adenine peak vanishes and hypoxanthine peak appears when compared to the state before reaction.


Choices of diet and lifestyle, use of improved medical care and increased longevity are factors that are associated with a rise in the prevalence of hyperuricemia and gout (Saag and Choi 2006; Richette and Bardin 2010). As metabolic disorders, these diseases can often be controlled or eliminated by changes in lifestyle, diet and the use of medication. Current treatments for gout have two targets: to control the inflammation of a gout attack and to lower serum uric acid levels (Tausche et al. 2009). However, given the disadvantages of drug therapy, modifications to lifestyle and diet, which are relatively inexpensive and safe, are attractive alternatives and, moreover, may result in better control of this disorder (Saag and Choi 2006). Thus, we propose the production of low-purine-content food as a novel solution to the problem of elevated serum urate. The enzyme that is the subject of this work, adenine deaminase is one of five enzymes that will be required to decrease the uric acid potential of foods (Trautwein-Schult et al. 2013).

The adenine deaminase gene was identified by comparing A. adeninivorans genomic sequences with Adap sequences from other fungi (52–55% identity). The isolated gene was then used to generate a transgenic strain of A. adeninivorans that produces very high amounts of His-tag-modified enzyme. This was achieved using the Xplor®2 transformation/expression platform. This platform has also been successfully used for the construction of transgenic A. adeninivorans strains that had high expression levels of tannase, lipase and alcohol dehydrogenase (Böer et al. 2005, 2009b; Giersberg et al. 2012). The Xplor2 vector allows insertion of different expression modules such as promoters and terminators. Here, the inducible nitrite reductase (AYNI1) promoter was used, which has the advantage over constitutive promoters of separating the cell growth phase from the enzyme production phase, allowing better control over the cultivation process (Böer et al. 2009c).

The generation of transgenic A. adeninivorans strain was successful, giving a level of production that is among the highest of the enzymes produced with the Xplor®2 transformation/expression platform (Böer et al. 2009b; Giersberg et al. 2012). Furthermore, the recombinant enzyme could be highly purified because of the C-terminal hexa-His-tag. The best transgenic strain accumulated 2·8 mg Aada6hp in 1 l of culture, equivalent to 80 U ml−1 of culture, which is about 20 times more than the wild-type strain LS3 under inducing conditions and 40 times more than the G1212/YIC102 control. The reason for such a high yield may be that the gene was isolated from and expressed in the same organism (homologous expression) with the host cell providing all structural requirements (e.g. cofactors) for the target protein. Moreover, the best transformant, G1212/YIC102–AADA6H-82, consisted of a YIC cassette that integrates randomly to the host cell DNA and the expression level that depends on the integration position, which can have a positive or a negative effect.

Arxula adeninivorans can grow on different nitrogen sources including purines (Middelhoven et al. 1984); however, this study shows that induction of AADA is only triggered by adenine and hypoxanthine. Uric acid acted as a very weak inducer that could only be detected by qRT-PCR, and there was no induction with ammonium. The effect of adenosine was not investigated. The transcriptional regulation of adenine deaminase is not well understood with the only detailed study being on Asp. nidulans (Oestreicher et al. 2008). The Asp. nidulans nadA gene is induced by uric acid, adenosine and ammonium, but not by adenine or hypoxanthine. The S. cerevisiae adenine deaminase is not induced by adenine, but it is with ammonium (Deeley 1992), whereas the opposite can be observed in E. coli, where adenine acts as a trigger for yicP gene transcription.

Aada6hp was found to have properties similar to those of known fungal adenine deaminases (Pospisilova and Frebort 2007). The A. adeninivorans enzyme's thermostability is average, in contrast to the instability of the enzyme from Candida utilis (loses 50% of activity in 10 min at 37°C) and the thermostability of the eukaryotic animal cell, C. fasciculata (loses 10% of activity after 60 min at 55°C; Kidder and Nolan 1979). Metal ions are not essential for activity as they are in the adenine deaminase from B. subtilis (Nygaard et al. 1996) or from E. coli (Matsui et al. 2001), which require manganese ions. Conversely, Mn2+ strongly inhibits the hydrolysis of adenine, as is the case with the C. fasciculata and C. utilis enzymes (McElroy 1963; Kidder et al. 1977). Fe2+, Cd2+ and Ni2+ act as partial inhibitors.

The natural substrate of adenine deaminase is adenine, but this enzyme can also hydrolyse N6-substituted adenine and guanine derivatives such as 2-amino-6-bromopurine, 2,6-diaminopurine and 6-chloropurine, but not guanine itself. Hartenstein and Fridovich (1967) describe many structurally related cyclic compounds as substrates for adenine deaminase. Apart from the amino group, the enzyme has also been found to be capable of catalysing the displacement of chloro, iodo and hydrazino groups. In the kinetic study, 2-amino-6-bromopurine and 6-chloropurine do not reach the enzyme saturation level because of their limited solubility. Thus, the kinetic constants for those substrates could not be calculated without altering them to apparent values, which may be higher or lower than the true values measured when all substrates are saturated. Moreover, adenine deaminase seems to have a higher affinity to 2,6-diaminopurine; however, its activity is slightly lower than it is with adenine as the substrate. This is probably due to the non-natural leaving groups present in this molecule that affect the catalytic efficiency. This may also apply to 2-amino-6-bromopurine and 6-chloropurine. It is clear that a change in chemical properties of the leaving group may affect the efficiency of the enzyme reaction; thus, characterization should be generally based on assays of natural substrates.

Adenine deaminases possess different numbers of subunits, and their molecular masses vary from 37–120 kDa. A molecular mass of 37 kDa was calculated for Pseudomonas synxantha (Sakai and Jun 1978), whereas the E. coli enzyme is a homodimer with a subunit of 60 kDa (Matsui et al. 2001). The theoretical molecular mass of A. adeninivorans adenine deaminase based on its amino acid sequence is 39·6 kDa, which is consistent with the mass estimated by gel filtration and electrophoresis of the purified enzyme in SDS gel. Other yeasts also possess adenine deaminases of a very similar size: S. cerevisiae – 39·6 kDa and Schizosaccharomyces pombe – 41·2 kDa (Pospisilova et al. 2008).

Adenine deaminase was found in the soluble phase of C. fasciculata, S. cerevisiae and S. pombe cells (Jun and Sakai 1985; Pospisilova et al. 2008). The localization experiment of Jun and Sakai (1985), based on high-speed centrifugation, led the authors to conclude that the adenine deaminase of C. fasciculata is also located in cytosol. Aada6hp of A. adeninivorans was subjected to examination by sucrose density gradient centrifugation. The enzymatic activity could be detected in the fractions with the lowest sucrose concentration, along with glucose-6-phosphate-dehydrogenase, a standard cytosolic marker. However, the first two fractions of the adenine deaminase peak cover a peak for α-d-mannosidase, a vacuolar marker. This could be due to overloading the cell with recombinant protein. Aada6hp should not have any targeting sequences according to the analysis with the SOSUI algorithm, but transport of proteins to the vacuole in yeast can be independent of the secretory pathway. For example, the α-mannosidase of S. cerevisiae, which accumulates in the vacuole, has no obvious signal peptide sequence, and its transport to the vacuole does not occur through the normal secretory pathway (Yoshihisa and Anraku 1990; Nakamura and Matsuoka 1993). We thus conclude that Aadap of A. adeninivorans is an intracellular enzyme, probably located in cytosol, but further investigation using other methods is needed.

The ability of recombinant Aada6hp to degrade adenine in food was investigated. The enzyme was able to effectively reduce the adenine concentration in the supplemented beef broth leading to increase in hypoxanthine content, which indicates that the recombinant enzyme may be useful in the reduction in adenine in foods. Application for the reduction in all purines in food for people suffering from gout will, however, requires the inclusion of other purine degrading enzymes such as urate oxidase, guanine deaminase and xanthine dehydrogenase. To our knowledge, this is a first time it has been proposed that enzymes can be used to degrade purines in foods so that uric acid production in the body will be reduced. In contrast to the very efficient, but expensive treatment with purified uricase, which is used at the onset of symptoms, our approach will provide patients with a less expensive and safer method for prevention of the disease.


The authors are grateful to Dr. M. Giersberg and Dr. S. Worch for helpful discussions. The research work was supported by grant from the BMWi (Project No. KF2131601MD8).

Conflict of Interest

No conflict of interest declared.