Introduction

Itaconic acid (methylenesuccinic acid, C5H6O4) is a soluble unsaturated dicarboxylic acid mainly produced from sugars by several fungi. In 1931, Kinoshita reported that the filamentous fungus Aspergillus itaconicus produced itaconic acid in a medium with high sugar content as a carbon source (Kinoshita 1931). Aspergillus terreus is also an itaconic acid producer (Calam et al. 1939), and has been used for the industrial production of itaconic acid because of its high production rate. Itaconic acid is used as a monomer to form polymers that are widely used as raw materials for latex, synthetic resins, adhesives, paint, additives for acrylic resin fibers and paper, etc. It is also used as an acidulant and for the pH adjustment of foods. However, applications of this useful material in other industries have been limited because it is rather expensive.

Several groups have attempted to clarify the biosynthetic pathway of itaconic acid in A. terreus, with the aim of obtaining knowledge for metabolic engineering. Bentley and Thiessen (1957a,b) showed that cis-aconitic acid, produced in a tricarboxylic acid (TCA) cycle, could be a substrate of cis-aconitic acid decarboxylase (CAD, EC 4.1.1.6), forming itaconic acid. However, the primary structure of CAD was still unknown when we achieved the first purification and characterization. We purified a 55-kDa protein that had CAD activity from A. terreus TN484-M1 to homogeneity and characterized the enzyme, concluding that the purified protein was essential for itaconic acid production in the fungus (Dwiarti et al. 2002). A gene coding for CAD, however, has never been cloned from any organism, despite the fact the enzyme directly participates in one of the major pathways of the TCA cycle, and the structure of the gene is still unknown. To date, some scientists have intended to clone the gene for CAD, but difficulties related to the low stability of the enzyme perturbed the enzyme purification, which is a necessary experimental step for its amino acid sequencing and its gene identification.

In this study, the CAD1 gene encoding CAD of A. terreus was cloned, and its function was confirmed by its functional expression in the yeast Saccharomyces cerevisiae.

Materials and methods

Strains, media, plasmids, and cultivation

A. terreus IFO6365 (wild-type strain) and A. terreus TN484-M1 (an high-itaconic acid-producing strain derived from the IFO6365) were used for the isolation of CAD1 (Dwiarti et al. 2002; Yahiro et al. 1995). Escherichia coli DH5α and the plasmid pT7Blue (Merck KGaA, Darmstadt, Germany) were used as the host and vector, respectively, for deoxyribonucleic acid (DNA) manipulation. S. cerevisiae INVSc1 (MATα, his3-Δ1, leu2, trp1–289, ura3–52; Invitrogen, California, USA) and plasmid pAUR123 containing the alcohol dehydrogenase gene promoter of S. cerevisiae (Takara Bio, Kyoto, Japan) were used as constructs for the expression of CAD1. Fungal genomic DNA was extracted using a DNeasy Plant Mini kit (Qiagen GmbH, Düsseldorf, Germany). YPD medium (1% w/v Bacto yeast extract, 2% w/v Bacto peptone, 2% w/v glucose) and YPD medium containing 1 μg/ml Aureobasidin A (Takara Bio) were used for the growth of S. cerevisiae cells and their transformants, respectively.

For the transcription analysis of CAD1, A. terreus strains were inoculated into a 500-ml Erlenmeyer flask containing 50 ml of the medium consisting of (per liter) 20 g of glucose, 35 g of corn steep liquor, and 20 g of NaCl (pH 3.0) and were grown at 30°C for 48 h using a rotary shaker set at 170 rpm. Mycelia of these strains were harvested on filter paper and washed twice with sterile water, and then a 1.0 g sample of wet mycelium was inoculated into a 200-ml Erlenmeyer flask containing 20 ml of the production medium consisting of (per liter) 140 g of glucose, 2.1 g of corn steep liquor, 2.9 g of NH4NO3, and 1.8 g of MgSO4·7H2O (pH 2.0). To investigate the effect of itaconic acid on CAD1 transcription, 100 g/l of itaconic acid was added to the production medium (pH 2.0). Cultivation was performed at 30°C for 12 h using a rotary shaker at 170 rpm.

Enzyme assay and measurement of protein concentration

The CAD activity was measured according to the method of Bentley (Bentley and Thiessen 1957c). Briefly, 0.1 ml of the enzyme solution was incubated with 0.4 ml of cis-aconitic acid solution (final concentration, 8.0 mM) and 2.5 ml of 0.2 M sodium phosphate buffer (pH 6.2) for 10 min at 37°C. The enzyme reaction was terminated by the addition of 0.1 ml of 12 M HCl. The released itaconic acid was measured by high-performance liquid chromatography (HPLC; JASCO, Tokyo). The samples were filtered through a 0.45-μm filter and injected onto a column (Capcell Pak C18 MG, Lot BS14 Shiseido, Shiseido, Tokyo) of 4.6 mm diameter and 150 mm length. Itaconic acid was identified using a UV detector (UV-970, JASCO) at 210 nm. A solution of 2.5% acetonitrile and 0.1% phosphoric acid was premixed, degassed, and used as the mobile phase for analysis at a flow rate of 1.0 ml/min and temperature of 45°C.

The protein concentration was measured using the protein assay CBB solution (Nacalai Tesque) based on the Bradford method (Bradford 1976) with bovine serum albumin as a standard. The results were expressed as mean ± standard deviation (SD) of three independent experiments.

Purification of CAD, amino acid sequence analysis, and isolation of a genomic clone encoding A. terreus CAD

Purified CAD was prepared from A. terreus mycelia as described previously (Dwiarti et al. 2002). The N-terminal and internal amino acid sequences of the protein were identified by Shimadzu (Kyoto, Japan). A draft A. terreus genome sequence and predicted protein data translated from the genome were obtained from the A. terreus genome database provided by the Broad Institute (http://www.broad.mit.edu). The protein data from the database were searched for a translated amino acid sequence datum that contained the analyzed amino acid sequences. To amplify a DNA fragment containing the positive locus from the A. terreus genome, oligo DNA primers targeted on 521 bp upstream (AT09971-F) and 213 bp downstream (AT09971-R) of the open reading frame (ORF) were synthesized (Table 1). Polymerase chain reaction (PCR) was done by KOD-plus-DNA polymerase (Toyobo, Osaka, Japan) under the following conditions: 95°C for 2 min, 30 cycles of 95°C for 30 s, 55°C for 40 s, 68°C for 2 min 20 s, and a final extension at 68°C for 5 min. The fragments amplified from genomic DNAs of wild-type and the itaconic-overproducing strains were ligated into the EcoRV site in pT7Blue and were named pwCAD1 and pmCAD1, respectively. DNA sequencing analysis of these plasmids was performed by Operon Biotechnologies (Tokyo, Japan).

Table 1 Primers used in this study
Full size table

Construction of the CAD1 expression plasmid for S. cerevisiae

An intron that existed in the genomic sequence in CAD1 was removed by two steps of PCR to express the gene in S. cerevisiae (Fig. 1). Information on the intron position was obtained from the A. terreus database. A primer set of Exon1-F and Exon1-R and a set of Exon2-F and Exon2-R were used to amplify the first and the second exons of CAD1, respectively (Table 1). Each first PCR was performed by KOD-plus-DNA polymerase under the following conditions: 95°C for 2 min, 30 cycles of 95°C for 30 s, 55°C for 40 s, 68°C for 2 min, and a final extension at 68°C for 5 min. The amplified fragments were mixed together and used as a template for the second PCR. The Exon1-F and Exon2-R primers were used for the second PCR, and the reaction was performed by KOD-plus-DNA polymerase under the following conditions: 95°C for 2 min, 30 cycles of 95°C for 30 s, 55°C for 40 s, 68°C for 2 min, and a final extension at 68°C for 5 min. The amplified fragment was ligated into the EcoRV site in pT7Blue, and the nucleotide sequence was confirmed using a DTCS Quick Start Master mix (Beckman Coulter, Fullerton, CA, USA) with M13-Forward and M13-Reverse primers (Invitrogen) according to the manufacturer’s protocol. The prepared sequencing samples were analyzed using a capillary DNA sequencer CEQ2000 (Beckman Coulter). The resulting plasmid and pAUR123 were digested with KpnI and XhoI and ligated to obtain a plasmid named pAUR-CAD1.

Fig. 1
figure 1

Construction of CAD1 excluding its intron. Two PCR steps were used to prepare the CAD1 nucleotide sequence without its intron, as described in “Materials and methods.” The resulting nucleotide sequence was used to construct the yeast expression plasmid pAUR-CAD1

Full size image

Functional expression of CAD1 in S. cerevisiae

The plasmid pAUR-CAD1 was introduced into S. cerevisiae INVSc1 as reported previously (Elble 1992). Transformants were screened on a YPD medium plate containing Aureobasidin A, and grown colonies were analyzed by colony-direct PCR using primers Exon1-F and Exon2-R. The reaction was performed by TaKaRa Ex Taq (Takara Bio) under the following conditions: 95°C for 2 min, 35 cycles of 95°C for 20 s, 55°C for 30 s, and 72°C for 2 min. A positive transformant was used for further analysis. A transformant that introduced pAUR123 was also prepared as a control strain. These transformants were inoculated into 500-ml flasks containing 50 ml of YPD medium supplemented with Aureobasidin A and cultured at 30°C on a rotary shaker at 200 rpm for 3 days. Grown cells were harvested by centrifugation and homogenized by an ultrasonic processor (Sonics & Materials, Newtown, CT, USA) to prepare crude enzyme solutions. The CAD activities of the transformants were analyzed as described above.

Estimation of the mRNA level of CAD1

The wild-type strain and the high itaconic acid producer strain of A. terreus were grown in the production medium with and without itaconic acid. After the incubation, cells were harvested on filter paper, and total ribonucleic acid (RNA) was extracted and purified using BioRobot EZ1 (Qiagen) and EZ1 RNA Tissue Mini Kit (Qiagen). A first-strand complementary DNA (cDNA) synthesis from the total RNA and estimation of the specific cDNA were performed using a real-time quantitative reverse transcription PCR (QRT-PCR) Mx3000P system (Stratagene, La Jolla, CA, USA) and FullVelocity SYBR Green QRT-PCR Master mix (Stratagene). The amount of the transcript level of CAD1 was measured using a primer set of Q-CAD1-F and Q-CAD1-R and an actin gene (ACT1; Table 1), respectively. ACT1 is a constitutively expressed gene and was used as an internal standard of the transcript level. The reaction and measurements were performed according to the protocol provided by the manufacturer.

Results

Cloning and sequencing analysis of CAD1 from a wild-type strain of A. terreus

The N-terminal and four internal amino acid sequences of the purified protein obtained by the amino acid sequencing are shown in Table 2. We searched for a gene that encoded these amino acid sequences from the database and found a positive translated polypeptide datum that contained these five amino acid sequences. The gene encoding this polypeptide was classified as ATEG_09971 in the database and is regarded as the CAD1 gene. A fragment including the 5′- and 3′-untranslated regions of CAD1 of the A. terreus wild-type strain was amplified by PCR, and the nucleotide sequence was analyzed (Fig. 2, DNA Data Bank of Japan accession number AB326105), and it showed 95.7% nucleotide sequence identity with that of ATEG_09971 ORF (data not shown). The difference between these sequences could be due to minor differences in the sources of A. terreus strains. The nucleotide sequences of CAD1 around its initiation codon, intron, and termination codon were highly identical with ATEG_09971, and we predicted a coding region of CAD1 based on the exon of ATEG_09971. The predicted CAD1 is interrupted by 56 bp of the intron, and the exon consists of 1,470 bp encoding a polypeptide of 490 amino acid residues (Fig. 2), whose calculated molecular mass is 52,721 Da. This result was mostly consistent with our previous report that a molecular weight of CAD purified from A. terreus was estimated to be 55 kDa via sodium dodecyl sulfate polyacrylamide gel electrophoresis (Dwiarti et al. 2002).

Fig. 2
figure 2

The nucleotide and deduced amino acid sequences of CAD1. The nucleotides that encode amino acids of CAD are given in capital letters. Initiation and termination (asterisk) codons are given in bold letters. Consensus-binding motifs for the HAP complex (CCAAT) are indicated by both underlined and italic type, and consensus-binding motifs for AREA (HGATAR, H = not G, R = A or G) in the complementary strand are indicated by underlined type

Full size image
Table 2 Amino acid sequence analysis of CAD
Full size table

Computing analyses of the primary structure of CAD1 and upstream and downstream untranslated regions of the gene

The amino acid sequence was analyzed by the BLASTP algorithm, and we found that the protein contained a conserved domain of the MmgE/PrpD family of proteins of bacteria and fungi, including a number of 2-methylcitrate dehydratases of bacteria that are involved in propionate catabolism (data not shown). The protein showing the highest identity (53%) with CAD was an unnamed protein product of A. oryzae that possessed a conserved region of the PrpD family (accession no. AP007175); nevertheless, this was not recognized as a high-yield organism for itaconic acid. The homology search algorithms found many proteins that had a high identity with CAD but none whose functions had been characterized. The WoLF PSORT (Horton et al. 2006) algorithms predicted this protein would be localized in the cytoplasm, suggesting that the CAD substrate of cis-aconitic acid, the TCA cycle intermediate, was transported from the mitochondria to the cytoplasm in A. terreus cells. The DIpro algorithm (Cheng et al. 2006) predicted that eight cysteine residues were involved in forming four disulfide bonds among the 12 cysteines in the protein, and the predicted disulfide bonds ordered by probability in descending order were as follows: Cys406 and Cys483, Cys19 and Cys48, Cys201 and Cys205, and Cys337 and Cys368.

In the 5′-untranslated region of CAD1, no typical sequence for the TATA box was detected, while consensus-binding motifs for the HAP complex (CCAAT), a global transcription activator identified in eukaryotes including filamentous fungi (Xing et al. 1993; Kato et al. 1998; Goda et al. 2005), were found in the regions from −60 to −56 and −400 to −395 of CAD1 (Fig. 2). In addition, the consensus-binding motifs for AREA (HGATAR, H = not G, R = A or G), a global transcription repressor involved in nitrogen metabolite repression in Aspergillus nidulans (Kudla et al. 1990), were found in the regions from −500 to −495, −69 to −64, and −39 to −34 in the complementary strand (Fig. 2). In contrast, no polyadenylation signal typical of eukaryotes could be found in the 3′-noncoding region of CAD1.

Enzyme activity of the CAD1 product

The transformation system for A. terreus has not been established; thus, we confirmed the function of the CAD protein by its gene expression in yeast. To investigate the actual activity of the CAD protein, the CAD1 gene excluding an intron as shown in Fig. 1 controlled by the ADH1 promoter was expressed in S. cerevisiae, and the CAD activity of the protein was measured. The plasmids pAUR-CAD1 and pAUR123 were introduced into S. cerevisiae, and the total soluble proteins of these cells were extracted. Each CAD activity was measured by directly detecting itaconic acid as a product from cis-aconitic acid as a substrate. As expected, we found that the strain introduced by pAUR-CAD1 possessed 276.1 ± 30.1 mU/mg protein of enzyme activity, while the strain introduced by pAUR123, as a control, showed little activity (37.9 ± 18.4 mU/mg protein; Fig. 3). These results confirm that the CAD1 gene encodes the CAD enzyme of A. terreus.

Fig. 3
figure 3

Recombinant CAD activity detected by HPLC analysis. Activities of CAD from the introduced yeast strains pAUR123 (a) and pAUR-CAD1 (b) were analyzed by HPLC as described in “Materials and methods.” Peaks for itaconic acid, a product of the enzyme, are indicated by arrows

Full size image

Regulation of CAD1 in A. terreus

The inhibitory effect of itaconic acid on its production by A. terreus was reported by Lockwood and Reeves (1945). To clarify whether feedback inhibition by itaconic acid existed at the transcription level of CAD1, a transcription analysis of the gene was performed. The A. terreus wild-type strain was grown in media with and without 10% (w/v) itaconic acid, and the amounts of CAD1 transcripts were measured by QRT-PCR. The transcripts of CAD1 that were prepared from the mycelia grown in each medium were present in similar quantities (Fig. 4). This fact suggests that the transcription of CAD1 is not affected by the presence of itaconic acid.

Fig. 4
figure 4

Transcription of CAD1 in A. terreus wild-type and the high-itaconic acid-producing strains. The amounts of transcripts of CAD1 in the wild-type strain (WT) and the high-itaconic acid-producing strain (MT) cultured in media with (Itaconate +) or without (Itaconate −) itaconic acid were measured using QRT-PCR as described in “Materials and methods”

Full size image

Structure and regulation of CAD1 in the high-itaconic acid-producing strain

A fragment including the 5′- and 3′-untranslated regions of CAD1 of the high-itaconic acid-producing strain was amplified, and the nucleotide sequence was analyzed using the same procedure as for the A. terreus wild-type strain, as mentioned above. A comparison of the nucleotide sequences from the wild-type strain and the high-producing strain showed they were exactly the same (data not shown).

A transcription assay of CAD1 in the high-producing strain indicated that there was no effect of itaconic acid on the transcription of CAD1, as seen in the wild-type strain. Moreover, the transcription of CAD1 was approximately fivefold stronger in the high-producing strain compared to that in the wild-type strain (Fig. 4). These facts suggest that high itaconic acid productivity was not caused by the substitution of the amino acid sequence of CAD but by higher expression levels of CAD1 in the high-producing strain compared to the wild-type strain.

Discussion

In this study, we succeeded in cloning CAD1 and confirming the CAD1 gene function for the first time and enhanced our understanding of the roles of the CAD protein in A. terreus. Generally, CAD is an enzyme of low stability so that it is difficult to obtain enough from the amino acid sequence analysis to perform gene cloning. In this study, we prepared a sufficient amount of purified CAD to analyze its amino acid sequence and succeeded in gene cloning and measurement of its functional expression.

It is interesting that the primary structure of CAD shared high identity with proteins possessing conserved regions of the MmgE/PrpD family, although the catalytic function of 2-methylcitrate dehydratase belonging to this family is different from that of CAD (Brock et al. 2002). This fact would be due to amino acid residue(s) in CAD being different from 2-methylcitrate dehydratase. This result also suggests that some uncharacterized proteins that have been classified into the MmgE/PrpD family according to homology might have CAD activity. Previously, CAD was found to be inactivated by reagents acting on cysteine residues including 5,5′-dithio-bis(2-nitrobenzoic acid), which suggested that the enzyme had cysteine residue(s) near the active site or that the residue(s) played an important role in the protein structure (Dwiarti et al. 2002). Computing the analysis on the deduced amino acid sequence of CAD indicated that 8 of the 12 cysteines were predicted to form four S–S bonds, which would be essential for stabilizing the protein structure. As far as the localization of CAD is concerned, there has been some discussion as to whether it exists in the mitochondria or the cytoplasm because cis-aconitic acid, the substrate of CAD, is produced in the TCA cycle, but itaconic acid is finally secreted into the cultured broth. The WoLF PSORT algorithm predicted that this protein was localized in the cytoplasm, and therefore it would be predicted that cis-aconitic acid was transported from the mitochondria to the cytoplasm in A. terreus. The existence of an element in the 5′-flanking region of CAD1, the consensus-binding sequences of the HAP complex (CCAAT; Xing et al. 1993; Kato et al. 1998; Goda et al. 2005), suggests that CAD1 is a highly transcribed gene.

In the industrial production of itaconic acid via A. terreus, its synthesis is inhibited by the itaconic acid produced. However, our findings clarified that the presence of itaconic acid was not the cause of gene repression.

The transcription of CAD1 in the high-itaconic acid producer was higher than that in the wild-type strain, while the alteration of nucleotide sequences including the 5′- and 3′-untranslated regions did not occur in the high-itaconic acid producer, suggesting that the higher itaconic acid production by this strain may be caused by higher expression levels of the gene. The reason for this phenomenon cannot be determined at present, and further analyses are needed, although the release from the repressor of the gene or unusual activation of the transcription factor is assumed.

In this study, we cloned and characterized the gene coding for CAD of A. terreus, which could be useful for molecular breeding of an itaconic acid-hyperproducing organism.