Fungal Genetics and Biology 45 (2008) 1081–1093

Contents lists available at ScienceDirect

Fungal Genetics and Biology

journal homepage: www.elsevier .com/locate /yfgbi

Involvement of the nadA gene in formation of G-group aflatoxins in Aspergillusparasiticus

Jingjing Cai a,1,4, Hongmei Zeng a,2,4, Yoko Shima a,d,4, Hidemi Hatabayashi a, Hiroyuki Nakagawa b,Yasuhiro Ito a, Yoshikazu Adachi d, Hiromitsu Nakajima c, Kimiko Yabe a,*,3

a Food Biotechnology Division, National Food Research Institute, Tsukuba, Ibaraki 305-8642, Japanb Food Safety Division, National Food Research Institute, Tsukuba, Ibaraki 305-8642, Japanc Faculty of Agriculture, Tottori University, Tottori 680-8553, Japand Laboratory of Animal Health, School of Agriculture, Ibaraki University, Ibaraki 300-0393, Japan

a r t i c l e i n f o

Article history:Received 24 November 2007Accepted 10 March 2008Available online 16 March 2008

Keywords:Aflatoxin biosynthesisAspergillus parasiticusnadA geneAflatoxin G1

NADA

1087-1845/$ - see front matter � 2008 Elsevier Inc. Adoi:10.1016/j.fgb.2008.03.003

* Corresponding author. Fax: +81 298 38 8122.E-mail address: yabek@affrc.go.jp (K. Yabe).

1 Present address: Biotechnology Research Institute,tural Sciences, Beijing 100081, China.

2 Present address: Institute of Plant Protection, ChiSciences. Beijing 100094, China.

3 Present address: National Agriculture and Food ReIbaraki 305-8517, Japan.

4 Equal contribution.

a b s t r a c t

The nadA gene is present at the end of the aflatoxin gene cluster in the genome of Aspergillus parasiticus aswell as in Aspergillus flavus. RT-PCR analyses showed that the nadA gene was expressed in an aflatoxin-inducible YES medium, but not in an aflatoxin-non-inducible YEP medium. The nadA gene was notexpressed in the aflR gene-deletion mutant, irrespective of the culture medium used. To clarify the nadAgene’s function, we disrupted the gene in aflatoxigenic A. parasiticus. The four nadA-deletion mutants thatwere isolated commonly accumulated a novel yellow-fluorescent pigment (named NADA) in mycelia aswell as in culture medium. When the mutants and the wild-type strain were cultured for 3 days in YESmedium, the mutants each produced about 50% of the amounts of G-group aflatoxins that the wild-typestrain produced. In contrast, the amounts of B-group aflatoxins did not significantly differ between themutants and the wild-type strain. The NADA pigment was so unstable that it could non-enzymaticallychange to aflatoxin G1 (AFG1). LC–MS measurement showed that the molecular mass of NADA was360, which is 32 higher than that of AFG1. We previously reported that at least one cytosol enzyme,together with two other microsome enzymes, is necessary for the formation of AFG1 from O-methylste-rigmatocystin (OMST) in the cell-free system of A. parasiticus. The present study confirmed that the cyto-sol fraction of the wild-type A. parasiticus strain significantly enhanced the AFG1 formation from OMST,whereas the cytosol fraction of the nadA-deletion mutant did not show the same activity. Furthermore,the cytosol fraction of the wild-type strain showed the enzyme activity catalyzing the reaction fromNADA to AFG1, which required NADPH or NADH, indicating that NADA is a precursor of AFG1; in contrast,the cytosol fraction of the nadA-deletion mutant did not show the same enzyme activity. These resultsdemonstrated that the NadA protein is the cytosol enzyme required for G-aflatoxin biosynthesis fromOMST, and that it catalyzes the reaction from NADA to AFG1, the last step in G-aflatoxin biosynthesis.

� 2008 Elsevier Inc. All rights reserved.

1. Introduction

Aflatoxins are highly toxic and carcinogenic secondary metabo-lites produced primarily by certain strains of Aspergillus flavus andAspergillus parasiticus (Bhatnagar et al., 1994; Payne and Brown,1998; Bennett and Klich, 2003). Recently, some other strains ofAspergillus nomius, Aspergillus pseudotamarii, Aspergillus bombycic,

ll rights reserved.

Chinese Academy of Agricul-

nese Academy of Agricultural

search Organization, Tsukuba,

and Aspergillus ochraceoroseus have also been reported to produceaflatoxins (Payne and Brown, 1998). The contamination of foodand feed crops such as wheat, corn, cotton, peanuts, and tree nutswith aflatoxins is not only a very serious health hazard to both ani-mals and humans but also causes economic problems all over theworld (Eaton and Groopman, 1994; Massey et al., 1995). Despitethis serious situation, there are currently no control approachesto prevent aflatoxin contamination (Jelinek et al., 1989). To obtaininformation useful for developing effective methods of preventingaflatoxin contamination, the biosynthetic pathways of aflatoxinhave been extensively studied. Recently, the majority of the en-zyme reactions in aflatoxin biosynthesis have been clarified (re-viewed in Minto and Townsend, 1997; Yu et al., 2002; Yabe andNakajima, 2004).

The genes involved in aflatoxin biosynthesis have also beenstudied. Most of the genes constitute a large gene cluster

1082 J. Cai et al. / Fungal Genetics and Biology 45 (2008) 1081–1093

encompassing 70 kb in the fungal genome (Fig. 1A), and theirexpression is positively regulated by the product of the regulatorygene, aflR (Payne et al., 1993; Woloshuk and Prieto, 1998; Yu et al.,2004a, 2004b; Price et al., 2006). The chromosomal location of thegenes in the cluster is also important in the regulation of geneexpression (Chiou et al., 2002). Although the cluster ends were firstsuggested to be norB and hypA, recently one terminus of the hypAgene was extended to the next gene, nadA, based on a microarraystudy (Price et al., 2006). Four other genes, hxtA, glcA, sugR, andorf, are adjacent to the nadA gene, and the first three of those geneswere suggested to constitute a sugar utilization gene cluster(Fig. 1A) (Yu et al., 2000). No similar sugar gene cluster has beenreported in any other related Aspergillus species such as A. flavusor A. oryzae (Yu et al., 2000). The function of the sugar gene clusterhas not been clarified.

More than 20 species, including A. nidulans, produce sterigmat-ocystin, one of the latter intermediates in aflatoxin biosynthesis, astheir final product (Brown et al., 1996; Keller and Hohn, 1997). Theenzyme pathway for sterigmatocystin biosynthesis in A. nidulans isthe same as that for aflatoxin biosynthesis, and the genes in thesterigmatocystin biosynthesis in A. nidulans are mostly found inthe aflatoxin gene cluster in the genomes of the aflatoxigenic fungi.However, several genes in the aflatoxin gene cluster are not pres-ent in the sterigmatocystin gene cluster and are thought to beexclusively involved in aflatoxin formation reactions after steri-matocystin formation (Yu et al., 2004a, 2004b). Among these genesis nadA, whose deduced amino acid sequence showed significantidentity (30–40%) with other NADH oxidases, though the functionof this gene has not been clarified (Yu et al., 2000).

Among the naturally occurring aflatoxins, the four major onesare afltoxin B1 (AFB1), AFB2, AFG1, and AFG2. The 1-group aflatoxins(AFB1 and AFG1), each of which contains a dihydrobisfuran ring inthose molecules, are produced from O-methylsterigmatocystin(OMST). The 2-group aflatoxins (AFB2 and AFG2), each containinga tetrahydrobisfuran ring, are produced from dihydro-O-methyl-sterigmatocystin (DHOMST) in the latter part of the biosyntheticpathway (Yabe et al., 1988a). The branching step for these two

Fig. 1. Genomic gene structures and expressions of the nadA gene. (A) The aflatoxin biosynadA gene was herein confirmed to be involved in the aflatoxin gene cluster. (B) Total RNhas been cultured in YEP (aflatoxin-non-conductive) (b) or YES (aflatoxin-conductive) (cpkaC (3) gene was analyzed using the resulting RNA sample and each primer set of nadA-template was also performed (a). (C) The RNA was prepared from the mycelia of the A. pacultured in YES medium for 72 h. Expression of the nadA (2 and 3), omtA (5 and 6), or cmomtA-F-2/R-2, and cmd-F/R (Table 1) in RT-PCR. The same PCR except using the A. para

groups is the desaturation step from versicolorin B to versicolorinA in aflatoxin biosynthesis (Yabe et al., 1991a). In biosynthesis of B-group aflatoxins, either the reaction from OMST to AFB1 or thatfrom DHOMST to AFB2 is commonly catalyzed by a P450 cyto-chrome monooxygenase enzyme, which is encoded by the ordAgene in the aflatoxin gene cluster (Yu et al., 1998; Yabe et al.,1999). The recombinant OrdA enzyme expressed in yeast showedthe conversion of OMST to AFB1 or that of DHOMST to AFB2, indi-cating that OrdA may be the sole enzyme required for these path-ways (Yu et al., 1998). The OrdA is thought to be a multi-functionalenzyme because either pathway from OMST to AFB1 and DHOMSTto AFB2 is composed of several types of complicated reactions, suchas oxidation, decarboxylation, and dehydration. Recently, 10-hy-droxy-OMST (HOMST) was suggested to function as an intermedi-ate between OMST and AFB1 (Udwary et al., 2002).

G-group aflatoxins, AFG1 and AFG2, are also produced fromOMST and DHOMST, respectively (Yabe et al., 1988a, 1999). Wepreviously reported that at least three enzymes, that are two mem-brane enzymes and one cytosol enzyme, are required for G-groupaflatoxins production from these precursors (Yabe et al., 1999).The activity of the OrdA monooxygenase enzyme was found inthe microsome fraction. Another monooxygenase gene, cypA, wasrecently reported to be involved in the formation of G-group afla-toxins by Ehrich et al. (Ehrlich et al., 2004), and the CypA enzymehas some transmembrane regions on its deduced amino acid se-quence, suggesting that it is a membrane enzyme. Therefore, theOrdA and CypA enzymes likely correspond to the two membraneenzymes required for G-group aflatoxins. In contrast, the geneencoding the cytosol (soluble) enzyme has not been determined.

We recently found that the hypA gene, which is a presumptiveend of the aflatoxin gene cluster, was involved in the step from ver-sicolorin A to demethylsterigmatocystin in aflatoxin biosynthesis,and Ehrlich et al. independently reported the same result recently(Ehrlich et al., 2005). The nadA gene exists next to and outside ofthe hypA gene, and shares the promoter region in the 866 bp inter-genic region with the hypA gene. A putative AflR-binding site wasrecently reported in this region (Price et al., 2006). Therefore, we

nthesis genes and assumptive sugar utilization genes are schematically shown. TheA was prepared from the mycelia of the wild-type strain A. parasiticus SYS-4, which

and d) medium for 48 h (b and c) or 72 h (d). Expression of the nadA (1), omtA (2) orF/R, omtA-F/R, and pkaC-F/R (Table 1) in RT-PCR. PCR using the genomic DNA as therasiticus SYS-4 (2, 5, and 8) or the aflR-gene deletion mutant (3, 6, and 9), which wasd (8 and 9) gene was analyzed using the RNA and each primer set of nadA-F-2/R-2,

siticus SYS-4 genomic DNA was also performed (1, 4, and 7).

J. Cai et al. / Fungal Genetics and Biology 45 (2008) 1081–1093 1083

supposed that the nadA gene is involved in the aflatoxin gene clus-ter. In this work, RT-PCR revealed that the nadA gene expressiondepended on the culture condition as well as on the presence ofthe aflR gene, a regulatory gene that positively controls expressionsof the aflatoxin biosynthesis enzyme genes. We also disrupted thenadA gene of A. parasiticus and then characterized the resulting dis-ruptant to clarify the gene’s function. Finally, we confirmed thatthe nadA gene encodes the cytosol enzyme required for the laststep in biosynthesis of G-group aflatoxins. A novel intermediateof AFG1 was also found in this work.

2. Materials and methods

2.1. Fungal strains

Aspergillus parasiticus SYS-4 (=NRRL2999) was a wild-type afla-toxin-producing strain, and the nadA-deletion mutants were newlyisolated in this work. A. parasiticus NIAH-26 was a UV-irradiatedmutant of A. parasiticus SYS-4 (Yabe et al., 1988b), and inducedall of the enzymes required to convert norsolorinic acid to aflatox-ins, although it produced no aflatoxins or pigmented precursors(Yabe et al., 1991b). This strain was suggested to be blocked inthe fas-2 gene (Suzuki R., personal communication).

Aspergillus parasiticus aflR-deletion mutant was also isolatedfrom A. parasiticus SYS-4 for RT-PCR analysis of nadA gene expres-sion. The nadA gene of SYS-4 strain was replaced with the pyrithi-amine (PT)-resistant marker (ptrA, PT-resistant gene, a selectablemarker (Kubodera et al., 2000) (Fig. 2A). The aflR gene disruptioncassettes were prepared by double-joint PCR (Yu et al., 2004a,2004b). The 50-flanking region (0.9 kb) and the 30-flanking region(0.9 kb) of coding region were, respectively, amplified using geno-mic DNA of A. parasiticus SYS-4 with primer pairs of P1 [50-CCGGCTGGTTCGTGGAAGTC-30] and P2 [50-GATGCAAGAGCGGCTCATCGTCACCCCAGAAAAGCCCCACCGCCAGAGCA-30] or P3 [50-CCAATGGGATCCCGTAATCAATTGCCCCGTGGAGGTGAGGAAGGAATTCA-30] andP4 [50-CATCGACCTTGTGGCCGACG-30], in which P2 and P3 primerscarried 24 bases of homologous sequence overlapping with theends of the ptrA gene (ptrA sequence is underlined). The ptrA gene(2.0 kb) was also amplified with primer pairs P5 [50-GGGCAATTGATTACGGGATCCCATTG-30] and P6 [50-GGGGTGACGATGAGCCGCTCT-30]. After second round PCR using the all fragments with-out any primers, third round PCR was performed with nested prim-ers P7 [50-CTGTGCAGGCCATGTGGGTG-30] and P8 [50-CCTCCACATGAGCCTTGAGCG-30]. The finally obtained DNA fragment was used asa disruption cassette to transform A. parasiticus SYS-4. After trans-formation of A. parasiticus SYS-4 wild strain, 25 PT-resistant trans-formants were isolated. One among them was suggested to be theaflR-deletion mutant when PCR analyses were done using the threesets of gene-specific primer pairs: P9 [50-GATCCATCGCGGATAGG-30] and P11 [50-GCAAGAGCGGCTCATCGTCA-30], P12 [50-TGGGA TCCCGTAATCAATTGCCC-30] and P10 [50- TCTCGTATCTCGCCCATG-30],or P13 [50-TCATTCTCGATGCAGGTAATC-30] and P14 [50-ATGGTTGACCATATCTCCCC-30] (Fig. 2B). The deletion of the aflR gene inthe mutant was further confirmed by Southern blot analysis, inwhich total genomic DNAs of the mutant and the wild strainSYS-4 were, respectively, digested with BglII enzyme and thenhybridized with the 0.76 kb aflR probe, which had been amplifiedfrom genomic DNA using primer pairs P15 [50-CAATTTGAGGGTTTACAGGG-30] and P14, and then using nested primers, P13 and P16[50-TGCCGATTTCTTGGCTGA-30]. The SYS-4 DNA gave the expected3.0-kb band together with another 6.0-kb band, in contrast, themutant DNA did not showed the 3.0-kb band, but showed onlythe 6.0-kb band (Fig. 2C). A. parasiticus contains a partially dupli-cated region of the aflatoxin gene cluster together with the wholegene cluster in the genome, and the additional aflR, that is calledaflR-2, is a non-functional gene in aflatoxin biosynthesis (Cary

et al., 2002; Chang and Yu, 2002). These results demonstrated thatthe resulting mutant was the aflR-deletion mutant. The aflatoxinproductivity was in fact lost in the aflR-deletion mutant whenanalyzed by thin-layer-chromatography (TLC) (Fig. 2D).

2.2. Media and growth conditions

GY (glucose 2%, yeast extract 0.5%), GY agar plates (GY supple-mented with 2% agar) and YES medium (20% sucrose, 2% yeast ex-tract) were used as aflatoxin-inducing media. YEP medium (20%peptone, 2% yeast extract) was a non-aflatoxin-inducingmedium.

For standard culture and TLC or HPLC analysis, tip culture meth-od was used (Yabe et al., 1988b). Spores (about 1 � 104 each) ofeach of the wild-type strain SYS-4 or the nadA-deletion mutantswere inoculated into 250 ll of YES medium in a tip (Yabe et al.,1988b). After stationary surface culture at 28 �C for 3 days, culturemedia and the mycelia were separated by centrifugation and thenused for further analyses. To prepare larger amount of the nadApigment (NADA) or other intermediate for physic-chemical analy-ses, the nadA-deletion mutant was cultured in 100 ml YES mediumin a bottle at 28 �C for 3–6 days (Wen et al., 2005).

2.3. Metabolites

The nadA pigment (NADA) was prepared as described below.The NADA concentration was determined by UV absorption spec-trum using the molecular coefficient of AFG1 (Cole and Cox,1981). OMST was prepared by the methylation of ST with methyliodide and sodium carbonate in acetone (Yabe et al., 1988a).

2.4. RNA preparation and RT-PCR

The total RNA was prepared from the mycelia of A. parasiticusSYS-4 or the aflR-deletion mutant, which had been cultured inYES or YEP medium for 48 or 72 h using TRI reagent (200 ll, Sig-ma–Aldrich, St. Louis, MO, USA) and FastPrep FP100A (Q-Biogene,Santa Ana, CA, USA) as previously described (Yan et al., 2004). Aftertreatment of the RNA with DNase I to degrade the slight amount ofremaining DNA, RT-PCR was performed using an RT-PCR kit (Rev-erTra Dash; Toyobo, Osaka, Japan) and the primers correspondingto the nadA, omtA, pkaC, or calmodulin (cmd gene) (Table 1) accord-ing to the manufacturer’s instructions. PCRs with the same primerswere also done using a genomic DNA of the wild-type SYS-4 strainas the template.

2.5. Construction of nadA gene disruption vector

To disrupt the nadA gene in A. parasiticus SYS-4, the nadA genedisruption plasmid, pNADA-L/R, was constructed in a two-stepprocedure (Fig. 3A). A 1.3-kb PCR fragment containing 50-flankingregion of the nadA gene in the genome of A. parasiticus SYS-4(AY371490) was amplified with KOD-plus enzyme (Toyobo) usingthe primers of nadA-L-F-SalI [50-AACGCGTCGACTCGAAGTTGCCTAGGCC-30 (SalI)] and nadA-L-R-PstI [50-AAACTGCAGTCCAGCTCGAGCCATTG-30 (PstI)]. The 1.4-kb PCR fragment containing 30-flank-ing region of the nadA gene was also amplified with the primersof nadA-R-F-(KpnI) [50-TGCAGGGGTACCGATCTC-30] and nadA-R-R[50-ATCCTAGCTGCTGCGGTG-30]. The PstI–SalI-digested PCR frag-ment containing 50-nadA and the KpnI-digested 30-nadA fragmentwere cloned sequentially into the PstI–XhoI site and then theKpnI–EcoRV site of the pSP72-ptrA (Wen et al., 2005) to give pNA-DA-L/R. The fragment containing the 50-flanking region, ptrA, and30-flanking region was used for fungal transformation followinglinearization of the plasmid pNADA-L/R by digestion with AlwNIand XcmI enzymes.

Fig. 2. Disruption of the aflR gene via double-crossover recombination. (A) Strategy for the disruption of the aflR gene is shown. The aflR gene disruption cassette wastransformed into wild-type strain SYS-4. A. parasiticus has two copies of aflR gene in its genome (Cary et al., 2002; Chang and Yu, 2002), and the functional aflR gene andanother non-functional aflR (aflR-2) are shown. The double-crossover recombination events resulted in the replacement of the target gene aflR with the selectable marker ptrAgene. Long arrows, gene direction; short arrows, positions of primers used for confirmation of gene disruptions; vertical arrow, gene replacement. (B) PCR analysis usingdifferent combinations of primers was done to confirm that the aflR gene was deleted in the aflR disruptant. The expected lengths of the PCR products are shown as a table. (C)Southern analysis of the aflR-deletion mutant. Genomic DNA of SYS-4 (1) or aflR-deleted mutant (2) were digested with BglII and analyzed by Southern hybridization usingaflR probe. kHindIII-digested markers were used as size standards. (D) Production of aflatoxin in the aflR disruptant. Aflatoxin produced by the aflR disruptant or SYS-4 wereanalyzed by TLC. Lane 1, culture medium (10 ll) of SYS-4; lane 2, culture medium of the aflR disruptant; lane 3, medium only.

1084 J. Cai et al. / Fungal Genetics and Biology 45 (2008) 1081–1093

2.6. Fungal transformation

Protoplasts were prepared from A. parasiticus SYS-4. Transfor-mation of fungi was done as previously described (Wen et al.,

2005). The PT-resistant transformants were screened on a CD-selective medium with 0.1 mg L�1 PT (Kubodera et al., 2000) andthen transferred to a GY agar plate to detect aflatoxin productionand the accumulation of precursors.

Table 1The oligonucleotide primers used in RT-PCR

Primera Sequence Positionb

nadA-F 50-CAATGGCTCGAGCTGGAC-30 70446–70463nadA-R 50-GACTGATAAGGGAGCCGC-30 71817–71834nadA-R-2 50-AAGTCCAATGCCGTCAAC-30 71523–71540omtA-F 50-CCTTCCTCGCCTTTGCG-30 54776–54792omtA-R 50-GGTGAGACGAAGAGCCC-30 56125–56141omtA-F-2 50-CCTTCCTCGCCTTTGCG-30 54776–54792omtA-R-2 50-GGTGAGACGAAGAGCCC-30 56125–56141pkaC-F 50-AGGTGGTCAAGATGAAGCAG -30 563–582c

pkaC-R 50-TGCTGCTATTTCTGTGGC-30 1708–1722c

cmd-F 50-GGTGATGGCCAGATCACCAC-30

cmd-R 50-CCGATGGAGGTCATGACGTG-30

a F, forward primer; R, reverse primer.b Positions correspond to the genome sequence of A. parasiticus (AY371490).c Sequence of A. parasiticus protein kinase A catalytic subunit (the pkaC, Cai J.,

personal communication), and the numbers for pkaC gene were based on its startsite (+1).

J. Cai et al. / Fungal Genetics and Biology 45 (2008) 1081–1093 1085

2.7. Confirmation of nadA deletion by PCR

The genomic DNAs of the resulting PT-resistant transformantswere prepared from the mycelia with the rapid extraction method(Wen et al., 2005), and then PCR analyses were performed usingthe three sets of gene-specific primer pairs: P17 (nadA-F) [50-CAATGGCTCGAGCTGGAC-30] and P18 (nadA-R) [50-GACTGATAAGGGAGCCGC-30], P19 (hypA-R2) [50-GCTAACAGATCCTCCGTCAACGT-30] and P11 (ptrA-F), P12 (ptrA-R) and P20 (hxtA-F) [50-TCATCCGCGGCATCGAG-30] (Fig. 2A).

2.8. Southern analysis

Genomic DNAs of nadA mutant and the SYS-4 were purified byNucleon PhytoPure (GE Healthcare) according to the manufac-turer’s instruction. Total genomic DNA of each strain was subjectedto restriction enzyme digestion of BglII or XhoI and separated byagarose gel electrophoresis, followed by vacuum blotting to a pos-itively charged nylon membrane, Hybond-N+ (GE Healthcare),using VacuGene XL Vacuum Blotting System (GE Healthcare). Thefilters were hybridized with the nadA and ptrA probes. The 1.1-kbnadA probe was amplified from genomic DNA using primer pairsnadA-F3 [50-CGTATCTCAGTTATGCAATGTCTC-30] and nadA-R3 [50-GTAACAGTATACCATGAAGGC-30], nadA-F [50-CAATGGCTCGAGCTG-GAC-30] and nadA-R2 [50-AAGTCCAATGCCGTCAAC-30] for nestedPCR. The 2.0-kb ptrA probe was amplified from pPTRI plasmid(Takara) using primer pairs ptrA-F2 [50-GGGCAATTGATTACGG-GATCCCA-30] and ptrA-R2 [50-TGACGATGAGCCGCTCTTGC-30].Hybridization and detection were performed using AlkPhos DirectLabelling and Detection System (GE Healthcare) according to thesuppliers’ manuals.

2.9. TIP culture and TLC analysis

The resulting nadA mutants or the SYS-4 were cultured in 250 llof YES medium in the tip culture for 3 days (Yabe et al., 1988b). Theresulting culture medium or mycelia extract of each fungus wasanalyzed by thin-layer-chromatography (TLC) using a silica gelKieselgel 60 TLC plate (No. 5721; Merck, Rahway, NJ, USA) andthe developing solution of chloroform–ethyl acetate–formic acid(90%) (6:3:1, vol/vol/vol). The substances on the plate were thenobserved under 365-nm UV light.

2.10. HPLC analysis

The aflatoxin concentration was determined by high perfor-mance chromatography (HPLC) (Yabe et al., 1991a). The tip culture

medium was extracted with chloroform, and the resulting chloro-form extract was analyzed using an HPLC apparatus (LC-10A, Shi-madzu, Kyoto, Japan) equipped with a silica gel HPLC column(0.46 � 15 cm), the mobile phase of toluene–ethyl acetate–formicacid (99%)–methanol (89:7.5:2:1.5, vol/vol/vol/vol) at a flow rateof 1.0 ml min�1 at 35 �C. The retention times of AFB1, AFB2, AFG1,and AFG2 were compared with those of standard samples (afla-toxin B–aflatoxin G mixture; Sigma).

2.11. Purification and characterization of the NADA pigment

NADA was extracted with acetone from the mycelia of thenadA-deletion mutant, which had been cultured in YES mediumfor 3 days, and purified with silica gel TLC using the developingsolution. The spot corresponding to the novel yellow pigmentabove AFB1 on the TLC plate was scraped off and then extractedwith ethyl acetate or acetone. Water was then added to the extractup to a final concentration of about 3% because the pigment ap-peared to be relatively stable in an aqueous condition. (1) Stability.The ethyl acetate solution containing NADA was spotted on a TLCplate, dried in air in the dark for various times, and then developedwith the same developing solution. The products were observedunder 365-nm UV light. (2) TFA assay. A trifluoroacetic acid (TFA)solution (1 ll) was mixed with 10 ll of the NADA ethyl acetatesolution or the methanol solution containing four aflatoxins(AFB1, AFB2, AFG1, and AFG2) at room temperature for 30 min. Afterdrying in the air, the residues were dissolved in 10 ll methanol andspotted onto a TLC plate, then developed. The product from NADAwas compared with those from aflatoxins. (3) LC–MS measurement.To determine the retention time in chromatogram, the molecularmass, and the absorption spectrum of the reaction products, thesubstances were analyzed with a liquid chromatography–massspectrum (LC–MS) apparatus (LCMS-2010, Shimadzu). The LC–MSsystem consisted of a LC-VP separation module equipped with aSPD-M10Avp photodiode array detector and a LC–MS 2010A singlequadrupole mass spectrometer with an atmospheric pressure pho-toionization (APPI) source probe. The probe was operated in thepositive/negative mode; the nebulizer temperature was 200 �C or250 �C. The moving solvents were: solution A, 0.1% acetic acid inwater; and solution B, 0.1% acetic acid in methanol. The ratio ofsolution B was changed according to the following linear gradientcycle: 0 min, 5%; 2 min, 5%; 17 min, 95%; 22 min, 95%; 23 min, 5%;and 33 min, 5%. An Inertsil column (150 mm � 2.1 mm, 5 lm,Superco, USA) was used, and the flow rate was 0.2 ml min�1.

2.12. Preparation of microsome and cytosol fractions

The microsome fraction was prepared from the mycelia of A.parasiticus NIAH-26 cultured in YES medium at 28 �C for 4 days(Yabe et al., 1999). The cytosol fractions were, respectively, pre-pared from the mycelia of the SYS-4 strain and the nadA-deletionmutant (Yabe et al., 1999). Further purification of the cytosol pro-teins was performed by successive gel-filtration and ion-exchangechromatography as follows. The cytosol fraction was desaltedthrough a Sephadex G-25M gel-filtration column (PD-10; GE Phar-macia LKB Biotechnology, Uppsalla, Sweden), which had beenequilibrated with Solution 1 (20 mM Tris–HCl [pH 7.5], 10% [vol/vol] glycerol, 5 mM MgCl2, 0.4 mM EDTA, and 1 mM 2-mercap-toethanol). The desalted cytosol fraction (0.4 ml) was then dilutedby adding the same volume of Solution 1 supplemented with 0.5%[weight/vol] Tween 80, and then applied onto the wet DEAE–Seph-acryl resin (GE Pharmacia LKB Biotechnology) in a filtration cup(UFC30 LH00, pore size 0.65 lm; or UFC3 OSV00, pore size 5 lm;Millipore, Bedford, MA, USA), which had been equilibrated withSolution 1 supplemented with 0.5% [weight/vol] Tween 80, andthen mixed. After the whole set containing the mixture was kept

Fig. 3. Disruption of the nadA gene via double-crossover recombination. (A) Strategy for the disruption of the nadA gene is shown. The gene disruption vector pNADA-L/R waslinearized and then transformed into wild-type strain SYS-4. The double-crossover recombination events resulted in the replacement of the target gene nadA with theselectable marker ptrA gene. Long arrows, gene direction; short arrows, positions of primers used for confirmation of gene disruptions; vertical arrow, gene replacement. Theexpected lengths of the PCR products are shown as a table. (B) PCR analysis using different combinations of primers was done to confirm that the nadA gene was deleted in thenadA disruptant. (C) Southern analysis of the nadA-deletion mutant. Genomic DNA of SYS-4 (a) or nadA-deleted mutant (b) were digested with BglII or XhoI and analyzed bySouthern hybridization using nadA gene coding region or ptrA selectable marker gene as probes. kHindIII-digested markers were used as size standards.

1086 J. Cai et al. / Fungal Genetics and Biology 45 (2008) 1081–1093

on ice for 5 min, it was centrifuged at 2000g using a mini-centri-fuge machine (PMC-060, AU Techno Services, Osaka, Japan). The re-sin in the cup was washed five times with 0.15 ml of the samesolution, and then five times with the same solution withoutTween 80, in order to remove the detergent which, if it contami-nated the mixture, would inhibit enzyme activity in the cell-freesystems. The proteins bound to the resin was eluted applying0.15 ml of Solution 1 supplemented with 0.5 M KCl to the resin fol-

lowed by centrifugation at 2000g. The resulting cytosol fractionwas stored at �80 �C until use.

2.13. Enzyme assays

For the enzyme reaction from OMST to aflatoxins, the pigment-removed cytosol of the SYS-4 or the nadA-deletion mutant (finalconcentration: 0.44 mg ml�1) was incubated in the reaction mix-

J. Cai et al. / Fungal Genetics and Biology 45 (2008) 1081–1093 1087

ture (total volume: 50 ll) containing 60 lM OMST and the micro-some fraction (1.8 mg ml�1) of NIAH-26 in solution A containing90 mM potassium phosphate [pH 7.5], 10% glycerol, and 4 mMNADPH as described previously (Yabe et al., 1999). The volume ofthe reaction mixture was 50 ll in a 0.5-ml microtube. After incuba-tion at 24 �C for 40 min, the reaction was terminated by adding70 ll of chloroform and mixing it with a Vortex mixer followedby centrifugation at 10,000g for 2 min. An aliquot (usually 20 ll)of the lower layer was analyzed by HPLC.

To detect the enzyme activity catalyzing the reaction fromNADA to AFG1, the same reaction mixture, except that it contained110 lM NADA instead of OMST was incubated in the presence orabsence of either 4 lM NAPDH or NADH as described above. Afterincubation at 24 �C for 40 min, aflatoxins produced were measuredby HPLC.

3. Results

3.1. Expression of the nadA gene

The wild-type A. parasiticus SYS-4 strain was cultured in YESmedium or in YEP medium. Expression of the nadA gene and theaflatoxin-related gene omtA was detected when the fungus wascultured in YES for 48 or 72 h. The RT-PCR product of 1328 bp usingthe primers encompassing the whole coding region of the nadAgene was slightly smaller than the PCR product (1389 bp) obtainedwhen the genomic DNA was used as the template, because theamplified region contained a splicing site (Fig. 1B, 1, lanes a, c,and d). In contrast, RT-PCR products were not detected when thefungus was cultured in YEP (Fig. 1B, 1, lane b). An aflatoxin-relatedgene, omtA, was also expressed in the YES medium, but not in theYEP medium (Fig. 1B, lane 2). Expression of the pkaC gene, a consti-tutively expressed protein kinase A catalytic unit gene, occurred inboth media (0.9-kb band, Fig. 1B, lane 3).

Furthermore, when the RNA was prepared from the aflR-dele-tion mutant (Fig. 2) for RT-PCR, expression of the nadA and omtAgenes was not detected even in the YES medium (Fig. 1C, lanes 3and 6). In contrast, the cmd gene, a calmodulin gene that is not re-lated to aflatoxin production, was expressed irrespective of themedium used (Fig. 1C, lane 9). These results indicated that regula-tion of nadA gene expression is similarly to that of other aflatoxin-related genes.

3.2. Disruption of the nadA gene

To investigate nadA gene function, we disrupted the nadA geneof the SYS-4 strain by gene replacement with a ptrA selectionmarker gene (Fig. 3A). Five independent transformation experi-ments yielded 470 pyrithiamine-resistant transformants on CDselection agar plates. PCR amplifications with several combina-tions of primer pairs were then carried out to select the transfor-mants lacking the nadA gene using their DNA samples as thetemplates. Among 90 randomly selected transformants, 4 showedthe predicted results in the PCR analysis. When we used P17 andP18 primers encompassing the nadA coding region, the recipientstrain SYS-4 gave a PCR product with a predicted size of 1.4 kb.However, the mutants did not show the band (Fig. 3B, lane 1).Also, only the mutants generated 1.45 kb (Fig. 3B, 2, lane a) and1.5 kb (Fig. 3B, 3, lane a) PCR products when the primer pairsP19–P11 and P12–P20 were used, respectively. We also didSouthern analysis to confirm the deletion of nadA gene in theresulting mutants. As the expected aflatoxin gene cluster restric-tion enzyme mapping data, the nadA gene coding region was pre-sented on an approximately 4.1-kb BglII fragment or 6.6-kb XhoIfragment of the wild-type SYS-4 genome. While nadA-deletionmutants DNA gave no hybridization band (Fig. 3C, left). When

using ptrA gene as a probe which is a thiazole synthase (thiA)gene with mutation point in A. oryzae (Kubodera et al., 2000), be-sides the band supposed to correspond to the putative homolo-gous nmt-1 gene in A. parasiticus genome (Kubodera et al.,2003), specific ptrA hybridization signals of 4.7 kb for BglII diges-tion or 7.2 kb for XhoI digestion were only detected in the geno-mic DNA of nadA-deletion mutant (Fig. 3C, right). These resultsindicated that the ptrA-selectable marker was replaced with thenadA gene in the genomes of the mutants. No transformantsshowed any apparent morphological differences from the wild-type SYS-4 strain and all of them made dark figures on the UVpictures (Yabe et al., 1987), suggesting that disruption of the nadAgene did not show a remarkable decrease on aflatoxin production.

3.3. A novel pigment produced by the nadA-deletion mutant

The nadA-deletion mutants and the recipient strain SYS-4 werecultured in YES medium by the tip culture method for 3 days.When the culture medium of the nadA-deletion mutant was ana-lyzed by TLC, a novel substance, which showed bright yellow fluo-rescence under 365 nm UV light, was remarkably detectedtogether with four kinds of aflatoxins (Fig. 4A, lanes 2–5). The samepigment was also found in the SYS-4 strain, but in much smallerquantities tan in the mutants (Fig. 4A, lane 6). The same pigmentwas also obviously observed in the mutants’ mycelia extracts (datanot shown). The pigment, which was named NADA, was not pro-duced when the mutants were cultured in an aflatoxin-non-induc-ible YEP medium instead of YES (data not shown), supporting theintermediacy of NADA in aflatoxin biosynthesis.

The amounts of aflatoxins accumulated in the culture media ofthe nadA-deletion mutants after 3-day tip culture were analyzedby HPLC (Fig. 4B). The nadA-deletion mutants did not show anyremarkable difference in production of B-group aflatoxin (AFB1

and AFB2) compared to the wild-type SYS-4 strain. In contrast, theyshowed significant decrease in production of G-group aflatoxins(AFG1 and AFG2), to about 50% of that by the wild-type strain.The nadA gene was revealed to be involved in G-group, but notB-group, aflatoxin formation.

3.4. NADA characterization

Interestingly, when the mutants and the wild-type strain werecultured in YES medium in tip culture for 6 days, there was muchless NADA in the medium than in that cultured for 3 days (Fig. 5A).Similar results were obtained when GY medium was used insteadof YES, indicating that medium was not related to the results(Fig. 5B). Since the fungus’ aflatoxin productivity peak at around3–4 days and then decreased in tip culture (Yabe et al., 1988b),NADA once produced appeared to gradually change to another sub-stance while in the medium. In contrast, aflatoxins in the mediumtended to increase for at least 6 days, since they are relatively sta-ble substances.

We first tried to purify NADA from the mycelia. However, repe-tition of TLC purification followed by extraction and drying causeda spontaneous change of NADA into another compound, whose Rf

value was the same as that of AFG1 (Fig. 5C). HPLC confirmed thatthe resulting substance was AFG1 (data not shown). Remarkably,when a fresh sample of the TLC-purified NADA was injected intothe HPLC apparatus, AFG1 also appeared with the remaining NADA.Furthermore, TFA treatment of the resulting AFG1-like substancedrastically increased the hydrophilicity of the substance, whoseRf value was the same as that of AFG2a produced from AFG1 stan-dard with TFA (Fig. 5D) (Cole and Cox, 1981). This demonstratedthat NADA could non-enzymatically changed to AFG1 in vitro.

LC–MS analyses revealed that NADA had a peak whose reten-tion time (18.0 min) was different from that of AFG1 (16.7 min)

Fig. 4. Production of a novel intermediate, NADA, and aflatoxins in the nadA disruptants. (A) Culture medium (10 ll) of each of the nadA-deletion mutants no.21 (lane 2), no.53 (3), no. 67 (4), no. 85 (5), the recipient SYS-4 strain (6) (indicated with D) and no fungus (7) after 3-day tip culture were analyzed by TLC. Lane 1, standard of aflatoxins B1,B2, G1 and G2 mixture. (B) Aflatoxins in the media of the nadA-deletion mutants (2–5) and the SYS-4 (1) (indicated with D) were, respectively, analyzed by HPLC after 3-day tipculture. Experiments were done in tetra-plicate. The total of B-group aflatoxins (AFB1 and AFB2) (left) and that of G-group aflatoxins (AFG1 and AFG2) (right) produced wereshown. Less than 5% of the total B- and G-group aflatoxins were AFB2 and AFG2, respectively. There were no significant differences among the mycelia weights of the nadA-deletion mutants (2–5) and the SYS-4 (1) (23.4 ± 3.0 mg 250 ll�1).

1088 J. Cai et al. / Fungal Genetics and Biology 45 (2008) 1081–1093

in LC chromatogram (Fig. 6A). NADA’s molecular mass peak wasmeasured using various nebulizer temperatures. NADA ionizationwas as follows ([nebulizer temperature] m/z (relative intensity)):temperature 200 �C, 361 [M+H]+ (100%), 329 (91.3%) (Fig. 6B,upper), 359 [M�H]� (100%), 327 (0%) (Fig. 6B, lower); tempera-ture 250 �C, 361 [M+H] + (20.9%), 329 (100%) (Fig. 6C, upper),359 [M�H]� (100%), 327 (91.3%) (Fig. 6C, lower). These resultsindicated that the NADA molecular mass was 360 and that thissubstance was partially changed to AFG1 even under the mildconditions for LC–MS measurement. The absorption spectra ofNADA and AFG1 by photodiode array (PDA) detection also dem-onstrated that NADA was a different substance from AFG1

(Fig. 6D).

3.5. Preparation of the cytosol fraction

We first fed the NIAH-26 with purified NADA to confirm the NA-DA’s intermediacy for aflatoxin biosynthesis. Although AFG1, butno other aflatoxins, was apparently produced from NADA after 4days’ culture (data not shown), the possibility of NADA spontane-ously changing to AFG1 in the culture medium could not beavoided. Therefore, we then tried to do cell-free experiments to de-tect the enzyme activity relating to NADA.

However, since the SYS-4 and nadA-deletion strains endoge-nously produced large amounts of aflatoxins together with smallamounts of their precursors, the cytosol fractions prepared fromthose fungi contained aflatoxins and the precursors, which ham-pering the sensitive detection of reaction products in the cell-freesystems. Therefore, we had to purify the cytosol proteins to removethem before the cell-free experiments, using the following two-step gel-filtration procedure using a PD10 column and then an-ion-exchange chromatography using DEAE–Sepharose resin. Theaddition of a non-ionic detergent, Tween 80, in the washing solu-tion for the DEAE resin was useful for removing the hydrophobiccontaminants such as aflatoxins and their precursors, because thenon-ionic detergent effectively captured them from the proteins,but did not bind to the ionic resin. A fraction containing solubleproteins without any aflatoxins or precursors was obtained andthen used for cell-free assays.

3.6. Involvement of the nadA gene in the pathway from OMST to AFG1

When the microsome fraction of A. parasiticus NIAH-26 wasincubated with OMST in the presence of NADPH, a small amountof AFG1 was produced, depending on the extent of cytosol con-tamination in the microsome fraction (Table 2). When the puri-

Fig. 5. Conversion of NADA to AFG1. The SYS-4 (lanes 2 and 3) or nadA-deletion mutant (4 and 5) was cultured for 3 days (2 and 4) or 6 days (3 and 5) in YES (A) or GY (B)medium by the tip culture, and then the media were analyzed by TLC. Either medium without fungus was also analyzed by TLC (A1 and B1), in which a spot in YES mediumwas an unknown substance different from NADA (A1). NADA is indicated with a closed triangle. (C) After analyzing the culture medium of the nadA-deletion mutant by TLC,the spot corresponding to the NADA pigment was extracted from the TLC plate and then spotted onto the TLC plate again (lanes 2–5). Each spot was air dried for 1 h (2), 2 h(3), 4 h (4) or overnight (5) and then developed by TLC. Lane 1, the culture medium of the nadA-deletion mutant. (D) NADA pigment and mixture of aflatoxins (AFB1, AFB2,AFG1, and AFG2) were, respectively, treated with trifluoroacetic acid (TFA), and analyzed by TLC. Lane 1, mixture of aflatoxins: AFB1, AFB2, AFG1, and AFG2 from the top down;2, the culture medium of the nadA-deletion mutant; 3, TLC-purified NADA pigment which was air dried on TLC plate for 1 h; 4, TFA-treated NADA, in which small amounts ofremaining AFG1 and newly formed AFG2a were detected; 5, TFA-treated aflatoxin mixture (5), in which AFB2, AFG2, AFB2a, and AFG2a were detected from the top down.

J. Cai et al. / Fungal Genetics and Biology 45 (2008) 1081–1093 1089

fied cytosol fraction of the SYS-4 recipient strain was furtheradded to the same reaction mixture, AFG1 production increasedremarkably, supporting the previous observation that the cytosolfraction was required for AFG1 formation from OMST (Yabe et al.,1999). The requirement of NADPH for the reaction was also con-firmed. However, when the cytosol fraction of the nadA-deletionmutant was used instead of that of the SYS-4 strain, the sameenhancement of AFG1 production was not detected. These resultsindicated that the cytosol fraction required for the pathway fromOMST to AFG1 corresponded to the protein encoded by the nadAgene.

Since OMST serves as the common precursor of AFB1 and AFG1

(Yabe et al., 1988a, 1999), AFB1 was also produced from OMSTdepending on the presence of NADPH although the AFB1 formationfrom OMST does not require the cytosol enzyme. In fact, theamount of AFB1 produced in the presence of the cytosol of thenadA-deletion mutant was almost same as that in the presence ofthe cytosol of the wild strain (Table 2).

3.7. AFG1 formation from NADA

Since NADA seemed to be the last precursor just before AFG1,we tried to determine the enzyme activity using NADA as the sub-strate. When the cytosol fraction of the SYS-4 strain was incubatedwith NADA in the presence of NADPH, AFG1 was significantly pro-duced, although a small amount of AFG1 was formed from NADA inthe absence of the cytosol because of spontaneous conversion ofNADA to AFG1 (Table 3). Almost the same amount of AFG1 was pro-duced when used NADH instead of NADPH (data not shown),

whereas a smaller amount of AFG1 was detected in the absenceof either cofactors. These results indicated that NADA is the precur-sor of AFG1 and that the cytosol fraction of the wild-type straincontained an enzyme that catalyzed the reaction from NADA toAFG1. In the same experiment, a small amount of AFG2 was alsoproduced in the presence of NADPH, likely due to contaminationof a dihydro-NADA in the NADA sample prepared through TLC.

When the cytosol fraction of the nadA-deletion mutant insteadof the wild-type strain was incubated with NADA, AFG1 formationwas not detected, irrespective of the presence of NADPH (Table 3).These results indicated that the NadA protein encoded by the nadAgene is the enzyme involved in the reaction from NADA to AFG1,which is the last step in G-aflatoxin biosynthesis.

4. Discussion

4.1. The nadA gene in aflatoxin biosynthesis

This work demonstrated that the nadA gene is involved in afla-toxin biosynthesis, and that the cytosol enzyme involved in G-afla-toxin formation from OMST corresponds to the NadA enzyme. ThenadA gene was expressed in aflatoxin inductive medium, but not inthe non-inductive medium in RT-PCR analysis. Expression of thenadA gene depended on the presence of aflR gene, which has beenalso observed in transcription profiling using DNA microarray anal-ysis (Price et al., 2006). Deletion of the nadA gene of A. parasiticuscaused decrease of G-group aflatoxins, but not changed B-groupaflatoxins in the culture (Fig. 4) as well as in cell-free experiment(Table 2). The nadA-deletion mutant accumulated a novel interme-

Fig. 6. Physicochemical properties of NADA pigment. (A) HPLC spectrum of theNADA pigment in LC–MS analysis. Peaks of NADA (retention time, 18.0 min) andAFG1 that was non-enzymatically produced from NADA (16.7 min) were detected.Monitored at 360 nm. (B) Mass signals of the NADA pigment at 200 �C (upper,positive; lower, negative). (C) Mass signals of the NADA pigment at 250 �C (upper,positive; lower, negative). The ratio of molecular mass of NADA (360) to AFG1 (328)decreased by increasing the nebulizer temperature. (D) Absorption spectra of eitherpeak of AFG1 or NADA in LC–MS analysis.

Table 3Formation of AFG1 from the NADA pigment by the NadA enzyme

Cytosol source NADPH (mM) Aflatoxins (ng/50 ll reaction mixture)a,b

AFG1c AFG2

c

1 — 4 — —2 SYS-4 4 37.6 ± 4.1 0.24 ± 0.033 nadA� 4 — —4 SYS-4 0 11.6 ± 1.7 0.06 ± 0.015 nadA� 0 — —

a The pigment-removed cytosol from SYS-4 strain or nadA-deletion mutant wasincubated with NADA intermediate in the presence or absence of NADPH for40 min. The aflatoxins produced were measured by HPLC. Values aremeans ± standard deviations.

b B-group aflatoxins (AFB1 and AFB2) were not significantly formed.c The amounts of AFG1 (16.5 ± 1.2 ng 50 ll �1) or AFG2 (0.19 ± 0.02 ng 50 ll �1),

which was non-enzymatically produced in the absence of the cytosol, was sub-tracted from each value.

1090 J. Cai et al. / Fungal Genetics and Biology 45 (2008) 1081–1093

diate, NADA, which was here confirmed to be an AFG1 precursor.Although we previously demonstrated that DHOMST, the dihydro-

Table 2Aflatoxin formation from OMST depending on the soluble NadA enzyme in cytosol

Cytosol Microsome Time (min)

1 — + 02 — + 403 SYS-4 + 404 nadA� + 405 — + 406 SYS-4 + 407 nadA� + 40

a The pigment-removed cytosol from SYS-4 strain or nadA-deletion mutant was incubaThe aflatoxins produced were measured by HPLC. Values are means ± standard deviatio

b Production of AFG1 was significantly enhanced.

derivative of OMST, is the precursor of AFG2, which is the dihydro-derivative of AFG1 (Yabe et al., 1988a), it was supposed that adihydroderivative of NADA (named DHNADA) serves as a precursorof AFG2. The reaction pathway proposed is shown in Fig. 7.

As for the pathway for B-group aflatoxins, Udwary et al. re-ported that HOMST is produced from OMST by the OrdA enzyme,and HOMST is further converted to AFB1 by only the OrdA enzymethrough formation of the 4-substitute 4-methoxy-3-butenoic acidintermediate (the intermediate (1)) (Udwary et al., 2002). Interme-diate 1 was suggested to be successively converted to intermediate2 by decarboxylation, to intermediate 3 by dehydration, and toAFB1 by demethylation. They also suggested an alternative schemein which tautomerization of the intermediate 1 would give a stabi-lized intermediate, and the resulting intermediate is changed toAFB1 by successive reactions of demethylation following by decar-boxylation (Udwary et al., 2002).

In contrast, the detailed pathway for G-group aflatoxins has notbeen clarified. G-group aflatoxins are produced from OMST by acertain branching step between B-group and G-group aflatoxins(Yabe et al., 1999; Ehrlich et al., 2004). We recently found thatG-group aflatoxins are also produced from HOMST (data notshown), suggesting that the branching step between B- and G-group aflatoxins is present after HOMST. A certain intermediateat the branching point would change to NADA by the OrdA andCypA enzymes. Ehrlich et al. showed that disruption of the cypAgene caused a complete absence of G-group aflatoxin formation,and that the lack of G-aflatoxin productivity in A. flavus was likelydue to the deletion of the promoter region of the cypA gene in thegenome of the same species (Ehrlich et al., 2004). They predictedthat CypA enzyme oxidizes the double bond of the intermediate1 (Fig. 7) generated from HOMST to give an epoxide, which maybe subsequently rearranged to AFG1.

NADPH (mM) Aflatoxinsa (ng/50 ll reaction mixture)

AFG1 AFB1

4 — —4 0.92 ± 0.13 21.5 ± 2.74 1.99 ± 0.16b 15.8 ± 1.04 0.95 ± 0.16 15.3 ± 2.40 0.01 ± 0.01 0 ± 0.00 0.07 ± 0.00 0.6 ± 0.00 0.05 ± 0.01 0.2 ± 0.0

ted with OMST in the presence of the microsome fraction of NIAH-26 and/or NADPH.ns.

Fig. 7. Postulated pathway from OMST to aflatoxin G1 in aflatoxin biosynthesis. The intermediates 1, 2, or 3 can work as a possible branching point for G-aflatoxin formation.Order of the decarboxylation, dehydration, and epoxydation can be changed depending on the branching intermediate. Epoxydation of the intermediate causes a novelintermediate containing an epoxide (intermediate 4), and then further hydroxylation of the resulting substance makes a novel intermediate (NADA) with a molecular mass is360. NADA shows equilibrium between open and closed forms. Demethylation of the NADA intermediate seemed to cause the formation of AFG1. NADA was proposed to bethe last intermediate in the conversion of OMST to AFG1. For biosynthesis of AFG2 from dihydro-O-methylsterigmatocystin (DHOMST), dihydroderivatives of the sameintermediates including dihydro-NADA (DHNADA) are likely involved in the same enzyme pathway. The NadA enzyme may catalyze both the reaction from NADA to AFG1

and that from DHNADA to AFG2.

J. Cai et al. / Fungal Genetics and Biology 45 (2008) 1081–1093 1091

We herein hypothesized a reaction scheme in which someintermediates (structures 1–6) were involved in G-group aflatoxinformation (Fig. 7). We hypothesized that each or all of the interme-diates 1, 2, and 3 might work as the intermediate or intermediatesfor epoxidation by the CypA enzyme to form intermediates 4, 5,and 6, respectively. The resulting intermediate 6 might be con-verted to intermediate 7 through oxidation, probably by the OrdAenzyme, because our preliminary study indicated that the OrdA en-zyme was also required for G-group formation from the branchingstep for G-aflatoxin formation (data not shown). The intermediates7 and 8 would connect to each other through tautomerism in thosestructures. The NadA enzyme may catalyze the subsequentdemethylation reaction from NADA to AFG1. The LC–MS analysisshowed that the molecular mass of the NADA was 360 Da, whichwas larger than AFG1 by 32 Da, suggesting deletion of a methanol.NADA pigment isolated is proposed to be the intermediate 7, whichis 3,4,7aa,10 aa-tetrahydro-1-hydroxy-1,5-dimethoxy-1H,12H-

furo[30,20:4,5]furo[2,3-h]pyrano[3,4-c][1]benzopyran-12-one. Thespontaneous conversion of NADA to AFG1 strongly suggested thatNADA is the last precursor just before AFG1. In fact, although thewild SYS-4 strain accumulated small amount of NADA in YES med-ium (Figs. 4 and 5), the cypA-deleted mutant, which was kindlygifted from Dr. Kenneth C. Ehrlich (Ehrlich et al., 2004), did notaccumulate any NADA (data not shown), suggesting that the nadAgene is involved in a step downstream of the cypA-gene step. How-ever, the detailed mechanism of G-aflatoxin formation remains tobe studied.

The instability of NADA as well as of the hypothesized DHNADAmay be why disruption of the nadA-deletion did not completely in-hibit G-aflatoxin production in the tip culture (Fig. 4). Since NADAas well as probably the hypothesized DHNADA were non-enzymat-ically changed to AFG1 and AFG2 in the culture, respectively(Fig. 5D), the large parts of AFG1 and AFG2 detected in the mutantslikely corresponded to artificial products from these substances.

1092 J. Cai et al. / Fungal Genetics and Biology 45 (2008) 1081–1093

However, the cell-free experiments showed that the disruption ofthe nadA gene almost completely inhibited the enzyme reactionfrom OMST to AFG1 (Table 2) as well as that from NADA to AFG1 (Ta-ble 3). These results suggested that the NadA enzyme plays a majorpart in the reaction from NADA to AFG1 or DHNADA to AFG2 in cells.Furthermore, The instability found in the cell-free systems does notdirectly indicate that this substance is in fact unstable in cells, be-cause we cannot reproduce the same environment in the cell-freesystem as that in cells. Also, NADA transiently formed in cell mayrapidly and efficiently be converted to AFG1, indicating that accu-mulation of relative amount of NADA would seldom occur. There-fore, the non-enzymatic conversion of NADA to AFG1 would notsignificantly contribute to the same pathway in cells.

4.2. The NadA enzyme

We previously reported that at least three enzymes are involvedin the formation of AFG1 from OMST or AFG2 from DHOMST (Yabeet al., 1999). Two of them were membrane proteins, probably OrdAand CypA. The nadA gene was suggested to encode a soluble en-zyme based on the amino acid sequence deduced from its gene se-quence using the SOSUI system (http://bp.nuap.nagoya-u.ac.jp/sosui/) (Hirokawa et al., 1998). This work demonstrated that thenadA gene in all probability encodes the third soluble enzyme inthe cytosol, which is involved in the last reaction in the formationof G-aflatoxins in aflatoxin biosynthesis. We previously reportedthat the native molecular mass of the soluble enzyme was about220 kDa based on gel-filtration analysis (Yabe et al., 1999); molec-ular mass of the NadA was calculated to be 43,720 Da according tothe deduced amino acid sequence. Therefore, this enzyme was sup-posed to be a homo-tetramer or larger complex composed of theNadA proteins.

Homologous research into the nadA gene revealed that the nadAhomologous genes are present in A. flavus (AY510452, AY510453,AY510453, and AY510451) and A. oryzae (AP007159, AB196490,AB072433, and AP007175), each of which has the homologous re-gion of the aflatoxin gene cluster. Interestingly, A. fumigatus(BX649605) and FvN Gibberella moniliformis EST1039750(DR633771, DR608592, and DR655104) also have the nadA genehomologs, indicating that the nadA gene was relatively conservedin many microorganisms. Although the prediction of the conserveddomains showed that the nadA gene may encode a NADH-oxidaseas previously suggested (Yu et al., 2000), NADH oxidase generallyrequires either NADH or NADPH. In this work, we found that NADHand NADPH equally enhanced AFG1 formation. A detailed investi-gation into this enzyme is now in progress in our laboratory.

Acknowledgments

We thank Dr. Kenneth C. Ehrlich for giving an A. parasiticuscypA-deletion mutant, and Yohei Ono and Akemi Koma for techni-cal assistance. DNA search of the GenBank database was performedwith the assistance of Computer Center of Agriculture, Forestry andFisheries Research, MAFF, Japan. This work was also supported byJSPS Postdoctoral Fellowship program of Japan Society for the Pro-motion of Science, by grant-in-aid BDP (Bio-Design Program) fromthe Ministry of Agriculture, Forestry and Fisheries, Japan, and agrant from National Agriculture and Food Research Organization(NARO), Japan.

References

Bennett, J.W., Klich, M., 2003. Mycotoxins Clin. Microbiol. Rev. 16, 497–516.Bhatnagar, D., Cleveland, T.E., Cotty, P.J., 1994. Mycological Aspects of aflatoxin

formation. In: Eaton, D.L., Groopman, J.D. (Eds.), The Toxicology of Aflatoxins:Human Health, Veterinary and Agricultural Significance. Academic Press Inc.,San Diego, CA, pp. 327–346.

Brown, D.W., Yu, J.H., Kelkar, H.S., Fernandes, M., Nesbitt, T.C., Keller, N.P., Adams,T.H., Leonard, T.J., 1996. Twenty-five coregulated transcripts define asterigmatocystin gene cluster in Aspergillus nidulans. Proc. Natl. Acad. Sci. USA93, 1418–1422.

Cary, J.W., Dyer, J.M., Ehrlich, K.C., Wright, M.S., Liang, S.H., Linz, J.E., 2002.Molecular and functional characterization of a second copy of the aflatoxinregulatory gene, aflR-2, from Aspergillus parasiticus. Biochim. Biophys. Acta1576, 316–323.

Chang, P.K., Yu, J., 2002. Characterization of a partial duplication of the aflatoxingene cluster in Aspergillus parasiticus ATCC 56775. Appl. Microbiol. Biotechnol.58, 632–636.

Chiou, C.-H., Miller, M., Wilson, D.L., Trail, F., Linz, J.E., 2002. Chromosomal locationplays a role in regulation of aflatoxin gene expression in Aspergillus parasiticus.Appl. Environ. Microbiol. 68, 316–325.

Cole, R.J., Cox, R.H., 1981. Handbook of Toxic Fungal Metabolites. Academic PressInc., New York, N.Y. pp. 1–127.

Eaton, D.L., Groopman, J.D., 1994. The toxicology of aflatoxins: human health,veterinary, and agricultural significance. Academic Press, San Diego, CA.

Ehrlich, K.C., Chang, P.-K., Yu, J., Cotty, P.J., 2004. Aflatoxin biosynthesis cluster genecypA is required for G aflatoxin formation. Appl. Environ. Microbiol. 70, 6518–6524.

Ehrlich, K.C., Montalbano, B., Boué, S.M., Bhatnagar, D., 2005. An aflatoxinbiosynthesis cluster gene encodes a novel oxidase required for conversion ofversicolorin A to sterigmatocystin. Appl. Environ. Microbiol. 71, 8963–8965.

Hirokawa, T., Boon-Chieng, S., Mitaku, S., 1998. SOSUI: classification and secondarystructure prediction system for membrane proteins. Bioinformatics 14, 378–379.

Jelinek, C.F., Pohland, A.E., Wood, G.E., 1989. Worldwide occurrence of mycotoxinsin foods and feeds—an update. J. Assoc. Off. Anal. Chem. 72, 223–230.

Keller, N.P., Hohn, T.M., 1997. Metabolic pathway gene clusters in filamentousfungi. Fungal. Genet. Biol. 21, 17–29.

Kubodera, T., Yamashita, N., Nishimura, A., 2000. Pyrithiamine resistance gene(ptrA) of Aspergillus oryzae: cloning, characterization and application as adominant selectable marker for transformation. Biosci. Biotechnol. Biochem. 64,1416–1421.

Kubodera, T., Watanabe, M., Yoshiuchi, K., Yamashita, N., Nishimura, A., Nakai, S.,Gomi, K., Hanamoto, H., 2003. Thiamine-regulated gene expression ofAspergillus oryzae thiA requires splicing of the intron containing a riboswitch-like domain in the 50-UTR. FEBS Lett. 555 (3), 516–520.

Massey, T.E., Stewart, R.K., Daniels, J.M., Liu, L., 1995. Biochemical and molecularaspects of mammalian susceptibility to aflatoxin B1 carcinogenicity. Proc. Soc.Exp. Biol. Med. 208, 213–227.

Minto, R.E., Townsend, C.A., 1997. Enzymology and molecular biology of aflatoxinbiosynthesis. Chem. Rev. 97, 2537–2556.

Payne, G.A., Brown, M.P., 1998. Genetics and physiology of aflatoxin biosynthesis.Annu. Rev. Phytopathol. 36, 329–362.

Payne, G.A., Nystrom, G.J., Bhatnagar, D., Cleveland, T.E., Woloshuk, C.P., 1993.Cloning of the afl-2 gene involved in aflatoxin biosynthesis from Aspergillusflavus. Appl. Environ. Microbiol. 59, 156–162.

Price, M.S., Yu, J., Nierman, W.C., Kim, H., Pritchard, B., Jacobus, C.A., Bhatnagar, D.,Cleveland, T.E., Payne, G.A., 2006. The aflatoxin pathway regulator AflR inducesgene transcription inside and ourside of the aflatoxin biosynthetic cluster. FEMSMicrobiol. Lett. 255, 275–279.

Udwary, D.W., Casillas, L.K., Townsend, C.A., 2002. Synthesis of 11-hydroxyl O-methylsterigmatocystin and the role of a cytochrome P-450 in the final step ofaflatoxin biosynthesis. J. Am. Chem. Soc. 124, 5294–5303.

Wen, Y., Hatabayashi, H., Arai, H., Kitamoto, H.K., Yabe, K., 2005. Function of thecypX and moxY genes in aflatoxin biosynthesis in Aspergillus parasiticus. Appl.Environ. Microbiol. 71, 3192–3198.

Woloshuk, C.P., Prieto, R., 1998. Genetic organization and function of the aflatoxinB1 biosynthetic genes. FEMS Microbiol. Lett. 160, 169–176.

Yabe, K., Ando, Y., Hamasaki, T., 1988a. Biosynthetic relationship among aflatoxinsB1, B2, G1, and G2. Appl. Environ. Microbiol. 54, 2101–2106.

Yabe, K., Ando, Y., Hamasaki, T., 1991a. Desaturase activity in the branching stepbetween aflatoxins B1 and G1 and aflatoxins B2 and G2. Agric. Biol. Chem. 55,1907–1911.

Yabe, K., Nakamura, Y., Nakajima, H., Ando, Y., Hamasaki, T., 1991b. Enzymaticconversion of norsolorinic acid to averufin in aflatoxin biosynthesis. Appl.Environ. Microbiol. 57, 1340–1345.

Yabe, K., Ando, Y., Ito, M., Terakado, N., 1987. Simple method for screeningaflatoxin-producing molds by UV photography. Appl. Environ. Microbiol. 53,230–234.

Yabe, K., Nakajima, H., 2004. Enzyme reactions and genes in aflatoxin biosynthesis.Appl. Microbiol. Biotechnol. 64, 745–755.

Yabe, K., Nakamura, H., Ando, Y., Terakado, N., Nakajima, H., Hamasaki, T., 1988b.Isolation and characterization of Aspergillus parasiticus mutants with impairedaflatoxin production by a novel tip culture method. Appl. Environ. Microbiol. 54,2096–2100.

Yabe, K., Nakamura, M., Hamasaki, T., 1999. Enzymatic formation of G-groupaflatoxins and biosynthetic relationship between G- and B-group aflatoxins.Appl. Environ. Microbiol. 65, 3867–3872.

Yan, P.-S., Song, Y., Sakuno, E., Nakajima, H., Nakagawa, H., Yabe, K., 2004. Cyclo(L-leucyl-L-prolyl) produced by Achromobacter xylosoxidans inhibits aflatoxinproduction by Aspergillus parasiticus. Appl. Environ. Microbiol. 70, 7466–7473.

Yu, J., Bhatnagar, D., Ehrlich, K.C., 2002. Aflatoxin biosynthesis. Rev. Iberonam.Micol. 19, 191–200.

J. Cai et al. / Fungal Genetics and Biology 45 (2008) 1081–1093 1093

Yu, J., Chang, P.-K., Bhatnagar, D., Cleveland, T.E., 2000. Cloning of sugar utilizationgene cluster in Aspergillus parasiticus. Biochim. Biophys. Acta 1493, 211–214.

Yu, J., Chang, P.-K., Ehrlich, K.C., Cary, J.W., Bhatnagar, D., Cleveland, T.E., Payne, G.A.,Linz, J.E., Woloshuk, C.P., Bennett, J.W., 2004a. Clustered pathway genes inaflatoxin biosynthesis. Appl. Environ. Microbiol. 70, 1253–1262.

Yu, J., Chang, P.-K., Ehrlich, K.C., Cary, J.W., Montalbano, B., Dyer, J.M., Bhatnagar, D.,Cleveland, T.E., 1998. Characterization of the critical amino acids of an

Aspergillus parasiticus cytochrome P450 monooxygenase encoded by ordAinvolved in aflatoxin B1, G1, B2 and G2 biosynthesis. Appl. Environ. Microbiol.64, 4834–4841.

Yu, J.H., Hamari, Z., Han, K.H., Seo, J.A., Reyes-Dominguez, Y., Scazzocchio, C.,2004b. Double-joint PCR: a PCR-based molecular tool for genemanipulations in filamentous fungi. Fungal. Genet. Biol. 41,973–981.