The partial mitochondrial cytochrome b gene from 32 strains of 12 species belonging to Aspergillus section Nigri was amplified by the polymerase chain reaction and sequenced directly. Using 402 nucleotide characters, nucleotide-based and amino acid-based phylogenetic trees were inferred and the genetic divergence among the species was evaluated. Based on analyses of the 402-bp nucleotide and 133-amino acid sequences, strains were divided into 11 DNA types and five amino acid types. Aspergillus niger and Aspergillus awamori showed different amino acid sequences. A. niger clade included A. niger var. niger and Aspergillus ficuum. A. awamori clade included A. awamori, Aspergillus phoenicis, Aspergillus pulverulentus, Aspergillus tubingensis, Aspergillus foetidus, and two varieties of A. niger, var. nanus and var. intermedius. Two varieties of A. niger will be reclassified. One strain of A. phoenicis and one strain of Aspergillus carbonarius were reidentified.
The concept of black aspergilli has been classified as the Aspergillus niger group by Raper and Fennell  and Aspergillus section Nigri by Gams et al. . These species are very important because they are used in fermentation industries and they also are encountered as human and plant pathogens. In the past, the identification, classification and taxonomy of section Nigri had mainly been based on morphological characteristics. Although some of the species can be readily distinguished morphologically, results obtained in several attempts at classifying this section are debatable and identification of some species is still difficult . Recently, several methods were used in the studies of this section. The circumscription of species in the group varies among taxonomic treatments. For example, Raper and Fennell  described 12 species in the A. niger group, Al-Musallam  accepted seven species in section Nigri, and Kusters-van Someren et al. [5,6] accepted six species. Recently, molecular methods such as nuclear or mitochondrial (mt) DNA restriction fragment length polymorphism (RFLP) analysis , Western blotting and DNA hybridization with pectin lyase (pelD) gene , isoenzyme analysis, separation of chromosomal DNAs by contour-clamped homogeneous electric field (CHEF) and randomly amplified polymorphic DNA (RAPD) analysis  also have been used for the identification, classification and phylogenetic analysis of section Nigri.
Mt cytochrome b genes have proven to be valuable tools to examine evolutionary relationships among closely related species in animals [8–10]. Recently, relationships between fungal species have been examined by restriction enzyme digestion of mtDNA and the taxonomic advantages of this method have been reported [11–13]. In the fungi, systematic studies using the mt cytochrome b gene have not yet been used extensively. The currently available sequence data on mt cytochrome b from the fungi constitute a random sample of only a few species, such as Hansenula wingei, Podospora anserina, Neurospora crassa, Schizosaccharomyces pombe, Saccharomyces cerevisiae and Aspergillus nidulans. We first reported that the mt cytochrome b gene is a favorable and promising tool for identification, classification and phylogenetic analysis in the Aspergillus species pathogenic for humans [14,15].
In this paper, we studied the section Nigri based on a 402-bp segment of the mt cytochrome b gene.
2Materials and methods
Twelve species of Aspergillus section Nigri were studied (Table 1). Other Aspergillus species were included in the phylogenetic analysis for comparison.
Table 1. Strains used in this study
ATCC: American Type Culture Collection, Rockville, MD, USA; CBS: Centraalbureau voor Schimmelcultures, Baarn, The Netherlands; IAM, Institute of Applied Microbiology, the University of Tokyo, Tokyo, Japan; IFM: Institute for Food Microbiology (at present, the Research Center for Pathogenic Fungi and Microbial Toxicoses, Chiba University), Chiba, Japan; IFO: Institute for Fermentation, Osaka, Japan; IMI, CAB International Mycological Institute (formerly Imperial Mycological Institute) Eggham, UK; NRRL: ARS Culture Collection, Northern Regional Research Center, Peoria, IL, USA. AUT, authentic material; R, reference strain of IUMS (International Union of Microbiological Societies); T, ex type.
aWe thought these strains were misidentified.
(CBS 115.80; IFO 5330, T of Aspergillus yezoensis)
(CBS 101.14; IFO 4030)
(ATCC 16872; CBS 172.66)
(CBS 114.34; IFO 4062; IMI 312982; NRRL 4786, AUT of A. japonicus var. capillatus)
(CBS 114.29, T of Sterigmatocystis acini-uvae)
(var. niger f. hennebergii, CBS 117.80, T of Aspergillus kawachii)
(var. niger, ATCC 16888; CBS 554.65; IMI 050566; NRRL 326, NT, R)
(ATCC 10698; CBS 126.49; IFO 6648; NRRL 363)
(ATCC 16882; CBS 555.65; IMI 091881; NRRL 364, synonym of A. niger)
(ATCC 12074; CBS 107.55, T of A. awamori var. hominis)
(CBS 629.78; IMI 313487; NRRL 365)
(ATCC 16879; CBS 115.48; IMI 211396, T of A. elatior)
(var. nanus, CBS 306.80)
(CBS 161.79; NRRL 4700)
(var. kagoshimaensis, ATCC 11363; CBS 137.52 (A. phoenicis); IAM 2190; IMI 214827, T)
(ATCC 10550; CBS 115.29, T)
(ATCC 38855; IFO 5708)
(CBS 103.14; IFO 4338)
(var. intermedius, CBS 205.80; IFO 4281, T of A. luchuensis)
(ATCC 38854; IFO 4033)
(ATCC 16876; CBS 482.65; IMI 172283, T)
DNA template isolation was performed according to the method described by Wang et al. . In brief, the mycelium cultured in potato dextrose broth medium was used. Cells were washed with sorbitol buffer, then disrupted by zymolyase digestion and smashing with glass beads. Mitochondria were pelleted by centrifugation at 2000×g for 15 min, and mtDNA was extracted using protease K, PCI (phenol:chloroform:isoamyl alcohol – 25:24:1) and ethanol.
Primers used were E1m (5′-TGAGGTGCTACAGTTATTAC-3′) and E2 (5′-GGTATAGMTCTTAAWATAGC-3′) designed from the cytochrome b amino acid sequences of A. nidulans, N. crassa, S. cerevisiae, corn, potato and Rhodospirillum rubrum. In this paper, primers E1m4 (5′-TGRGGWGCWACWGTTATTACTA-3′) and rE2m4 (5′-GGWATAGMWSKTAAWAYAGCATA-3′) were designed from cytochrome b amino acids. E1m or E1m4 was used as a forward primer and E2 or rE2m4 as a reverse primer.
2.4Polymerase chain reaction (PCR) amplification and gel electrophoresis
A TaKaRa PCR Amplification kit (TaKaRa Shuzo Co., Ltd., Otsu, Shiga, Japan) was used for amplification. The volume of reaction mixtures was adjusted to 100 μl with sterilized water. Each PCR was done with 10 μl of 10× reaction buffer (100 mM Tris–HCl (pH 8.3), 500 mM KCl, 15 mM MgCl2), 2 μl of dNTP mixture (2.5 mM each of dATP, dCTP, dGTP and dTTP), 1.25 U Taq polymerase and 1 μl of each primer (approximately 130 ng μl−1). The PCR and electrophoresis were carried out as reported previously .
Both DNA strands of the fragments were sequenced using the Dye Terminator Cycle Sequencing Ready Reaction kit (Applied Biosystems Division of Perkin-Elmer Japan Co., Ltd.) and ABI Prism 377 DNA sequencer.
The sequences of the mt cytochrome b gene and derived amino acid sequences, using the yeast mt genetic code, were aligned with the computer program GENETYX-MAC Genetic Information Processing Software (Software Development Co., Ltd., Tokyo, Japan). Then using the unweighted pair group method with the arithmetic mean method (UPGMA), amino acid-based trees from the sequences were conducted and using the neighbor joining (NJ), maximum likelihood (ML), and maximum parsimony (MP) methods, nucleotide-based trees were conducted. PAUP program (version 4.0; β version) was used for the NJ, ML, MP methods. In the case of UPGMA, an estimation of phylogenetic relationship was conducted using standard errors for each branching point. Standard errors were calculated by the method of Nei .
The aligned 402-bp nucleotide characters were analyzed. The differences between the nucleotide sequences and amino acid sequences are shown in Fig. 1A,B.
Fig. 2 shows the phylogenetic trees inferred from the amino acid sequences by UPGMA and from nucleotide sequences by NJ, MP and ML methods. Based on these analyses, the 32 strains from 12 species were divided into 11 DNA types (from D-1 to D-11) and five amino acid types (from A-1 to A-5) (Table 1 and Fig. 2).
Strains identified as different species shared the same nucleotide sequence. For example, Aspergillus japonicus IFM 47640, Aspergillus aculeatus IFM 47724 and Aspergillus carbonarius IFM 47645 were included in D-1. Two strains of Aspergillus phoenicis, Aspergillus saitoi IFM 49646, two strains of Aspergillus pulverulentus, A. niger var. nanus IFM 47786 and Aspergillus tubingensis IFM 47787 in D-7. Three strains of Aspergillus foetidus, two strains of Aspergillus awamori and A. niger var. intermedius IFM 47726 in D-9. Eight strains of A. niger, A. phoenicis IFM 47725 and Aspergillus ficuum IFM 47779 in D-5. On the other hand, some strains identified as the same species showed different DNA and amino acid types. For example, strains of A. niger were placed in four different DNA types, D-5, D-6, D-7, and D-9, and grouped with two amino acid types, A-3 (D-5 and D-6) and A-4 (D-7 and D-9). Particularly, two varieties of A. niger, var. nanus and var. intermedius, showed an identical amino acid sequence with A. awamori but not with A. niger var. niger. The strains labelled as A. carbonarius were placed in two different amino acid types, A-1 and A-2. Also strains of A. phoenicis were placed in A-3 and A-4.
Aspergillus ellipticus, A. japonicus, A. aculeatus and A. carbonarius showed distinct divergences from other species of section Nigri. Except for these four species the other species showed a very close relationship (from D-5 to D-10) (Fig. 2).
Based on the identities of the amino acid sequence, the strains were divided into five types (Fig. 2), A-1 (two strains of A. japonicus, two strains of A. aculeatus and A. carbonarius IFM 47645), A-2 (A. carbonarius IFM 47662), A-3 (eight strains of A. niger, A. phoenicis IFM 47725 and A. ficuum IFM 47779), A-4 (three strains of A. awamori, A. niger var. nanus IFM 47786, A. niger var. intermedius IFM 47726, two strains of A. phoenicis, two strains of A. pulverulentus, two strains of A. tubingensis and three strains of A. foetidus), and A-5 (A. ellipticus IFM 47041). A. niger and A. awamori had different amino acid sequences and formed two clades, A. niger clade and A. awamori clade.
Because black aspergilli are of great importance in the fermentation industry, taxonomic studies and identification are very important . Phenotypic features have been in the past and are still used for classification of fungi. Most of the species of section Nigri are morphologically similar so their identification has been confused. For example, based on the mtDNA and rDNA types, the strains were divided into two large types rDNA, type I (mtDNA type 1) and rDNA type II (mtDNA type 2). Type I included some strains of A. niger, A. foetidus, A. awamori and A. phoenicis, but some other strains of the same species are also included in type II . Similar results were also found by using the SmaI-digested chromosomal DNA, pkiA, 28S, and pelA genes . The studied strains were divided into seven groups representing A. foetidus varieties, A. tubingensis, A. niger, A. carbonarius, A. japonicus, Aspergillus heteromorphus and A. ellipticus. Strains identified as A. awamori and A. phoenicis strains were placed in three groups, A. foetidus varieties, A. tubingensis and A. niger. Our data also showed strain misidentifications. For example, A. phoenicis IFM 47725 (D-5, A-3) was different from A. phoenicis IFM 47781 and IFM 48074 (D-7, A-4), but the same as A. niger (D-5, A-3). This result was similar to Parenicova's . They showed that this strain belonged to the A. niger group. We checked this strain and found it is very similar to A. niger morphologically. Furthermore, A. carbonarius IFM 47645 had the same nucleotide sequence as A. japonicus IFM 47640 and A. aculeatus IFM 47724 (D-1, A-1) and not with A. carbonarius IFM 47662 (D-4, A-2). Its morphological features were very similar to A. japonicus. Therefore, we suggest reidentifying A. phoenicis IFM 47725 as A. niger and A. carbonarius IFM 47645 as A. japonicus.
On the basis of the nuclear DNA RFLP analysis , isoenzyme analysis, separation of chromosomal DNAs by CHEF, RAPD analysis  and mtDNA RFLP analysis , the A. niger aggregate was divided into two groups, namely A. niger (including A. awamori and A. foetidus) and A. tubingensis. Because A. foetidus could not be distinguished from A. niger using RFLP analysis , these researchers treated A. foetidus as a synonym of A. niger, although Al-Musallam  classified these taxa as separate species. However, our data support the conclusion of Kozlowski and Stepien  who suggested that A. awamori and A. niger var. niger were different species. Strains of A. phoenicis, A. pulverulentus, A. tubingensis, A. foetidus and two varieties of A. niger (var. nanus and var. intermedius) shared the same cytochrome b amino sequences as A. awamori (A-4). A. niger var. nanus and var. intermedius were shown to have amino acid sequences identical to that of A. awamori but not to that of A. niger. We suggest that the taxonomic status of the two varieties needs to be reevaluated.
A. aculeatus is a close relative of A. japonicus, formerly considered as a variety of A. japonicus. Based on SmaI-digested rDNA patterns  and the results of Southern and Western blottings , A. japonicus and A. aculeatus were considered to be conspecific and could not be clearly distinguished from each other. Because these two species showed the same amino acid sequences, we also consider that they belong to a single species, A. japonicus, although they could be divided into three DNA types (D-1, D-2 and D-3).
A. carbonarius, A. japonicus and A. aculeatus are distinct members of section Nigri[1,2,4,5,17,19–22]. To our knowledge, at present, DNA or amino acid sequence data have not been used in a phylogenetic analysis of section Nigri. Our data based on the amino acid sequences of cytochrome b support the opinion of earlier researchers who considered A. carbonarius, A. japonicus and A. aculeatus as distinct species within this section.
In this study, 50 positions were polymorphic in nucleotide sequences (Fig. 1). In the case of cytochrome b gene, the substitution is in accordance with evolutionary time orderly, and one base pair substitution requires about 22.7 million years (Ma) . Therefore, a few differences also have significant implication in the evolutionary process. These differences should be phylogenetic signals.
Generally speaking, the structure and function of cytochrome b protein were remained. The rate of substitution is in proportion to evolutionary time in animals. On the other hand, the substitution rate of mtDNA is higher than that in chromosomal genes in animals . We compared the proportional differences (%) of the amino acid and nucleotide sequences between three species fungi, A. nidulans, S. cerevisiae and S. pombe in the complete region (approximately 386 amino acids) and the selected portion (133 amino acids) of cytochrome b. The similar percentages (%) were obtained (41.8–43.2%). We also compared the substitutions (%) from mt cytochrome b gene, β-tubulin gene, actin gene and 28S (or 26S) rRNA (D1–D2). The substitution rate of cytochrome b is the highest in these genes. Between A. nidulans and S. cerevisiae, the rates of cytochrome b, 26S rRNA (D1–D2), β-tubulin and actin were 38.1%, 24.1%, 21.8% and 8.8%, respectively. Between A. nidulans and S. pombe, and S. cerevisiae and S. pombe, the similar rates were obtained.
In this study, the amino acid-based UPGMA tree and DNA-based NJ, MP and ML trees were used for phylogenetic analysis. The amino acid-based UPGMA tree had similar topology with the DNA-based NJ, MP and ML trees (Fig. 2). Watson et al.  regarded that mtDNA can be used as a molecular clock from animals and plants which had fossil remains. However, fungi which have no fossil remains can be used. We assume that fungal mtDNA is similar to that of animals and plants. Generally, the evolutionary rate of amino acids is more constant than DNA. If the evolutionary rate is constant or mtDNA is used as a molecular clock, UPGMA gives the correct topology and correct branch lengths [26,27]. While NJ, ML and MP methods were suitable for no constancy in the evolutionary rate, Takezaki and Nei , Zhakikh and Li  have shown that the MP method can give an incorrect topology in the case of a constant rate of evolution. Although we could not test the molecular clock hypothesis in fungi, the similarities between amino acid-based UPGMA tree and DNA-based NJ, MP and ML trees will probably provide support for this hypothesis, indirectly. Therefore, we suggest, for phylogenetic analysis, UPGMA is suitable for amino acid sequences and NJ, MP and ML are suitable for nucleotide sequences, in the case of mt cytochrome b gene.
In summary, the cytochrome b sequences can be used in the analysis of phylogeny, classification and identification of black aspergilli.
We thank the Institute for Fermentation, Osaka, Japan (IFO), for generously providing some of the strains used in this study. We also thank Professor David Wood for assistance with the text. We thank the Honor Scholarship (Association of international Education, Tokyo, Japan), the Sumitomo Scholarship (Social Welfares Business Group of Sumitomo Life Assurance Company, Osaka, Japan), OSF (Okamoto Scholarship Foundation, Chiba, Japan) and Yonnmaru Scholarship (Yonnmaru alumni association, Chiba University, School of Medicine, Japan) for providing scholarships to L.W.