A Conditionally Dispensable Chromosome Controls Host-Specific Pathogenicity in the Fungal Plant Pathogen Alternaria alternata

Corresponding author: Takashi Tsuge, Graduate School of Bioagricultural Sciences, Nagoya University, Chikusa, Nagoya 464-8601, Japan., ttsuge{at}agr.nagoya-u.ac.jp (E-mail)

Communicating editor: J. J. LOROS

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The filamentous fungus Alternaria alternata contains seven pathogenic variants (pathotypes), which produce host-specific toxins and cause diseases on different plants. Previously, the gene cluster involved in host-specific AK-toxin biosynthesis of the Japanese pear pathotype was isolated, and four genes, named AKT genes, were identified. The AKT homologs were also found in the strawberry and tangerine pathotypes, which produce AF-toxin and ACT-toxin, respectively. This result is consistent with the fact that the toxins of these pathotypes share a common 9,10-epoxy-8-hydroxy-9-methyl-decatrienoic acid structural moiety. In this study, three of the AKT homologs (AFT1-1, AFTR-1, and AFT3-1) were isolated on a single cosmid clone from strain NAF8 of the strawberry pathotype. In NAF8, all of the AKT homologs were present in multiple copies on a 1.05-Mb chromosome. Transformation-mediated targeting of AFT1-1 and AFT3-1 in NAF8 produced AF-toxin-minus, nonpathogenic mutants. All of the mutants lacked the 1.05-Mb chromosome encoding the AFT genes. This chromosome was not essential for saprophytic growth of this pathogen. Thus, we propose that a conditionally dispensable chromosome controls host-specific pathogenicity of this pathogen.

HOST-SPECIFIC toxins, which are produced by plant pathogenic fungi, are generally low-molecular-weight secondary metabolites (YODER 1980 ; NISHIMURA and KOHMOTO 1983 ; SCHEFFER and LIVINGSTON 1984 ; WALTON 1996 ). They are critical determinants of host-specific pathogenicity or virulence in several plant-pathogen interactions (YODER 1980 ; NISHIMURA and KOHMOTO 1983 ; SCHEFFER and LIVINGSTON 1984 ; WALTON 1996 ). The imperfect fungus Alternaria alternata contains seven variants, which produce host-specific toxins and cause necrotic diseases on different plants (NISHIMURA and KOHMOTO 1983 ; KOHMOTO et al. 1995 ). Since A. alternata is one of the most cosmopolitan fungal species and is generally saprophytic (ROTEM 1994 ), these host-specific forms have been designated as pathotypes of A. alternata (NISHIMURA and KOHMOTO 1983 ; KOHMOTO et al. 1995 ). A. alternata pathotypes are a fascinating case for studying intraspecific variation and evolution of pathogenicity in plant pathogenic fungi.

Host-specific toxins from A. alternata are diverse in structure (KOHMOTO et al. 1995 ). However, AF-toxin of the strawberry pathotype, AK-toxin of the Japanese pear pathotype, and ACT-toxin of the tangerine pathotype have a common 9,10-epoxy-8-hydroxy-9-methyl-decatrienoic acid structural moiety (Fig 1; NAKASHIMA et al. 1985 ; NAKATSUKA et al. 1986 , NAKATSUKA et al. 1990 ; FENG et al. 1990 ; KOHMOTO et al. 1993 ). Thus, these three pathotypes should share genes required for biosynthesis of this common moiety.

View larger version (35K): In this window In a new window Download PPT slide   Figure 1. Host-specific toxins produced by three pathotypes of A. alternata. AF-toxins of the strawberry pathotype (NAKATSUKA et al. 1986 ), AK-toxins of the Japanese pear pathotype (NAKASHIMA et al. 1985 ), and ACT-toxins of the tangerine pathotype (KOHMOTO et al. 1993 ) have a common moiety, 9,10-epoxy-8-hydroxy-9-methyl-decatrienoic acid, indicated in shaded boxes.

The strawberry pathotype causes Alternaria black spot of strawberry and affects only one Japanese strawberry cultivar, Morioka-16 (MAEKAWA et al. 1984 ). The Japanese pear pathotype causes black spot of Japanese pear on a narrow range of susceptible cultivars, including the commercially important cultivar Nijisseiki (TANAKA 1933 ; OTANI et al. 1985 ). The tangerine pathotype causes citrus brown spot on tangerine, grapefruit, and grapefruit x tangerine hybrids, and this disease has not occurred in Japan so far (PEGG 1966 ; KOHMOTO et al. 1991 , KOHMOTO et al. 1993 ; PEEVER et al. 2000 ). Interestingly, the strawberry and tangerine pathotypes were also found pathogenic to Japanese pear cultivars susceptible to the Japanese pear pathotype (MAEKAWA et al. 1984 ; KOHMOTO et al. 1993 ). The host ranges of these pathotypes can be explained by toxicity of their toxins. Each of these toxins consists of multiple related molecules (Fig 1). AK-toxins I and II of the Japanese pear pathotype are toxic only to susceptible cultivars of Japanese pear (OTANI et al. 1985 ). AF-toxin I of the strawberry pathotype is toxic to susceptible cultivars of both strawberry and pear; toxin II is toxic only to pear; and toxin III is toxic to strawberry (MAEKAWA et al. 1984 ). ACT-toxin I of the tangerine pathotype is toxic to susceptible cultivars of both citrus and pear; toxin II is more toxic to pear than to citrus (KOHMOTO et al. 1993 ).

We previously isolated AK-toxin-minus, nonpathogenic mutants of the Japanese pear pathotype by restriction enzyme-mediated integration (REMI) mutagenesis and selected a cosmid clone (pcAKT-1) of the wild-type strain 15A, which contains the tagged site in a mutant (TANAKA et al. 1999 ). Structural and expression analyses identified four genes (AKT1, AKT2, AKTR-1, and AKT3-1) within an 15-kb region in pcAKT-1 (Fig 2; TANAKA et al. 1999 ; TANAKA and TSUGE 2000 , TANAKA and TSUGE 2001 ). Transformation of the wild type with the targeting vectors of these genes produced toxin-minus, nonpathogenic mutants. DNA gel blot analysis demonstrated that AKT1 and AKT2 were targeted in the mutants (TANAKA et al. 1999 ). AKTR-1 and AKT3-1, however, were not disrupted in the mutants, and the targeting vectors were integrated into different regions (TANAKA and TSUGE 2000 ). Thus, we isolated a cosmid clone, pcAKT-2, containing the targeted DNA and found that pcAKT-2 carries two genes, AKTR-2 and AKT3-2, with strong similarity to AKTR-1 and AKT3-1, respectively (Fig 2; TANAKA and TSUGE 2000 ). DNA gel blot analysis demonstrated that the wild-type strain has multiple copies of nonfunctional homologs of all the AKT genes and that these genes and their homologs are on a 4.1-Mb chromosome (TANAKA et al. 1999 ; TANAKA and TSUGE 2000 ). These results imply the structural and functional complexity of the genomic region controlling AK-toxin biosynthesis.

View larger version (17K): In this window In a new window Download PPT slide   Figure 2. Partial structure of cosmid clones pcAKT-1, pcAKT-2, and pcAFT-1. The cosmid clones pcAKT-1 and pcAKT-2 were isolated from a genomic library of strain 15A of the Japanese pear pathotype (TANAKA et al. 1999 ; TANAKA and TSUGE 2000 ). The cosmid clone pcAFT-1 was isolated from a genomic library of strain NAF8 of the strawberry pathotype in this study. Arrows indicate the protein coding regions, and white segments indicate positions of introns. AFT1-1, AFTR-1, and AFT3-1 are the AKT1, AKTR, and AKT3 homologs, respectively. The plasmid clones p1EX, p2XS, pRS, and p3HB were made by cloning the indicated fragments in pBluescript KS+ or pGEM-T Easy (TANAKA et al. 1999 ; TANAKA and TSUGE 2000 ) and used as probes for hybridization experiments. The fragments from p1EX and p3HB were cloned in pSH75 (KIMURA and TSUGE 1993 ) to make the targeting vectors pGDT1 and pGDT3, respectively (TANAKA et al. 1999 ; TANAKA and TSUGE 2000 ). Arrowheads denote the orientation and location of oligonucleotide primers used in RT-PCR experiments. B, BamHI; C, ClaI; E, EcoRI; H, HindIII; K, KpnI; P, PstI; S, SphI; V, EcoRV; X, XhoI.

In DNA gel blot analysis, these genes are unique to three pathotypes of A. alternata: Japanese pear, strawberry, and tangerine (TANAKA et al. 1999 ; TANAKA and TSUGE 2000 ). In the tangerine pathotype, polymerase chain reaction (PCR) amplification and sequence analysis of the AKT1 and AKT2 homologs showed that the sequences are 90% identical to AKT1 and AKT2 (MASUNAKA et al. 2000 ). The AKT homologs in the strawberry and tangerine pathotypes are hypothesized to be the genes responsible for biosynthesis of 9,10-epoxy-8-hydroxy-9-methyl-decatrienoic acid.

Our objective in this study was to characterize the structure and function of the AKT homologs of the strawberry pathotype. We found that all of the AKT homologs are located on a single chromosome of 1.05 Mb in strain NAF8 of the strawberry pathotype. We identified a genomic cosmid clone of NAF8, which contains the AFT1-1, AFTR-1, and AFT3-1 genes with strong similarity to AKT1, AKTR, and AKT3, respectively. Transformation-mediated targeting of AFT1-1 and AFT3-1 produced AF-toxin-minus (Tox^-) mutants, which also lost pathogenicity completely. These mutants were found to lack the 1.05-Mb chromosome, which was dispensable for saprophytic growth. Thus, it appears that AF-toxin biosynthesis genes are clustered on a supernumerary chromosome.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Fungal strains:
Strains NAF1, NAF8, T-32, O-187, and M-30 of the strawberry pathotype of A. alternata (KUSABA and TSUGE 1994 ) were used in this study. NAF1 and NAF8 were collected from the same field in Aichi Prefecture, Japan (KUSABA and TSUGE 1994 ). T-32, O-187, and M-30 were isolated in Tottori, Iwate, and Tottori Prefectures, respectively (KUSABA and TSUGE 1994 ). NAF8 was used for isolation of AF-toxin biosynthesis genes. The others were used for analysis of chromosomal distribution of the AKT gene homologs. Strains were routinely maintained on potato dextrose agar (PDA; Difco, Detroit).

Plasmids and genomic libraries:
The plasmid clones p1EX, p2XS, pRS, and p3HB, which contain the internal fragments of AKT1, AKT2, AKTR-1, and AKT3-1, respectively, in pBluescript KS+ (Stratagene, La Jolla, CA) or pGEMT Easy (Promega, Madison, WI) (Fig 2; TANAKA et al. 1999 ; TANAKA and TSUGE 2000 ), were used as probes in hybridization experiments. The targeting vectors pGDT1 and pGDT3 were used for transformation-mediated disruption of AFT1-1 and AFT3-1, respectively, in NAF8. To make pGDT1 and pGDT3, the internal fragments of AKT1 and AKT3-1, respectively, were cloned in a transformation vector pSH75 (Fig 2; TANAKA et al. 1999 ; TANAKA and TSUGE 2000 ), which carries the hygromycin B resistance gene (hph) fused to the Aspergillus nidulans trpC promoter and terminator (MULLANEY et al. 1985 ; KIMURA and TSUGE 1993 ). The plasmid clone pGDB2-2, which contains the A. alternata BRM2 gene, was used as a probe for hybridization experiments. BRM2 encodes 1,3,8-trihydroxy-naphthalene reductase involved in melanin biosynthesis (KAWAMURA et al. 1999 ).

A genomic cosmid library of strain NAF8 was constructed with a cosmid vector pMLF2 (AN et al. 1996 ) using half-site fill-in reactions (SAMBROOK et al. 1989 ). Total DNA of NAF8 was partially digested with Sau3AI to generate fragments of 40 kb and partially filled with dATP and dGTP. The cosmid vector pMLF2 was completely digested with XhoI and partially filled with dCTP and dTTP. Partially filled Sau3AI genomic DNA fragments were cloned at the partially filled XhoI site of pMLF2 to construct a genomic library. Screening of the library by colony hybridization was conducted by the standard method (SAMBROOK et al. 1989 ).

Fungal transformation:
Protoplast preparation and transformation of A. alternata were performed by the methods previously described (TSUGE et al. 1990 ; SHIOTANI and TSUGE 1995 ). Colonies that appeared 5–10 days after plating on the selective regeneration media (TSUGE et al. 1990 ) were transferred to PDA containing hygromycin B (Wako Pure Chemicals, Osaka, Japan) at 100 µg/ml, and transformants were selected after incubation at 25° for 5 days.

Assay for AF-toxin production, pathogenicity, and vegetative growth:
The wild type and transformants were grown statically in 5 ml of potato dextrose broth (PDB, Difco) in test tubes at 25° for 7 days, and culture filtrates and mycelial mats were harvested. Culture filtrates were tested for toxicity to leaves of strawberry cultivar Morioke-16 and Japanese pear cultivar Nijisseiki as previously described (MAEKAWA et al. 1984 ; TANAKA et al. 1999 ). AF-toxin and 9,10-epoxy-8-hydroxy-9-methyl-decatrienoic acid in culture filtrates were extracted and quantified by reverse-phase high-performance liquid chromatography as previously described (FENG et al. 1990 ; HAYASHI et al. 1990 ; TANAKA et al. 1999 ).

Mycelial mats were used for preparation of conidia as previously described (HAYASHI et al. 1990 ). Pathogenicity was assayed by spray inoculation of conidial suspension (5 x 10⁵ conidia/ml) to leaves of cultivar Morioka-16 and cultivar Nijisseiki as previously described (MAEKAWA et al. 1984 ; HAYASHI et al. 1990 ; TANAKA et al. 1999 ).

The wild type and transformants were grown on PDA at 25° for 4 days. Agar blocks (3 mm in diameter) carrying mycelia were prepared from the resulting colonies and inoculated on PDA. After incubation at 25° for 4 days, colony growth and morphology were observed. Prototrophy of transformants was tested on a minimal agar medium (10 g KNO₃, 5 g KH₂PO₄, 2.5 g MgSO₄·7H₂O, 0.02 g FeCl₃, 10 g glucose, 20 g agar per liter).

DNA manipulations:
Isolation of total DNA and RNA from A. alternata, isolation of plasmid and cosmid DNA, agarose gel electrophoresis, and DNA gel blot hybridization were performed as previously described (TANAKA et al. 1999 ). Hybridized blots were washed twice in 2x SSPE [1x SSPE: 180 mM NaCl, 10 mM NaH₂PO₄ (pH 7.7) and 1 mM EDTA] plus 0.1% sodium dodecyl sulfate at 65° for 10 min and once in 1x SSPE plus 0.1% sodium dodecyl sulfate at 65° for 10 min.

Reverse transcription-polymerase chain reaction (RT-PCR) was performed using the RNA PCR kit, version 2.1 (Takara, Shiga, Japan), according to the manufacturer's instructions. The following primers were used: 1-5' (5'-CTACCGCCTGAGTACATGCGTC-3') and 1-3' (5'-AGCAACAGCACCCTGGGGTT-3') for AFT1-1; R-5' (5'-GCATGGGGACAAGATCCAG-3') and R-3' (5'-CCAACACAGATGCTGAACTT-3') for AFTR-1; and 3-5' (5'-CCTGCGAAACTCTACCTCTG-3') and 3-3' (5'-TCAGAGCTTTGGCTTGGAAG-3') for AFT3-1 (Fig 2).

For analysis of nucleotide sequences, DNA was cloned in pBluescript KS+. DNA sequences were determined using the PRISM dye termination cycle sequencing ready reaction kit (Applied Biosystems, Foster City, CA) and an automated fluorescent DNA sequencer (Model 373A, Applied Biosystems) according to the manufacturer's instructions. DNA sequences were analyzed with BLAST (ALTSCHUL et al. 1997 ). Alignment of nucleotide and amino acid sequences was made using the CLUSTAL W program (THOMPSON et al. 1994 ).

Pulsed-field gel electrophoresis (PFGE):
Chromosome-sized DNA molecules were prepared from fungal protoplasts as previously described (ADACHI et al. 1996 ). PFGE was carried out on a contour-clamped homogeneous electric field (CHU et al. 1986 ) apparatus (CHEF-DRII, Bio-Rad Laboratories, Hercules, CA) using 0.5x TBE (SAMBROOK et al. 1989 ) at 8° in 0.8% agarose gel (Seakem Gold agarose, BioWhittaker Molecular Applications, Rockland, ME). The following electrophoresis conditions (duration/voltage/linear gradient of switching time; MORALES et al. 1993 ) were used: for separating 1.0–6.0 Mb DNA, 115 hr/50 V/3600–1800 sec, 24 hr/50 V/1800–1300 sec, 28 hr/60 V/1300–800 sec, and 28 hr/80 V/800–600 sec, and, for separating <2.0 Mb DNA, 13 hr/180 V/120 sec and 13 hr/180 V/180 sec. The gels were stained with ethidium bromide for 30 min and destained in distilled water for 30 min. Size standards were chromosome preparations of Schizosaccharomyces pombe and Saccharomyces cerevisiae (BioWhittaker Molecular Applications).

DNA gel blotting and hybridization were performed as previously described (ADACHI et al. 1996 ). Chromosomal DNA separated by PFGE was excised from the gel, nicked by UV irradiation, and purified using the GeneClean II kit (Bio 101, Vista, CA). Recovered DNA was used as a probe in hybridization experiments.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Distribution of the AKT homologs in the strawberry pathotype strains:
We previously showed that the AKT1, AKT2, AKTR, and AKT3 probes hybridized to DNA of all strains of the strawberry pathotype collected from different locations in Japan (TANAKA et al. 1999 ; TANAKA and TSUGE 2000 ). Hybridization patterns were similar in these strains and suggested that the homologs are present in multiple copies (TANAKA et al. 1999 ; TANAKA and TSUGE 2000 ). In the present study, strain NAF8 was selected, because it showed the high level of pathogenicity and of AF-toxin production.

The DNA of strain NAF8 was digested separately with 12 restriction enzymes and probed with the AKT1, AKT2, AKTR-1, and AKT3-1 fragments from p1EX, p2XS, pRS, and p3HB, respectively (Fig 2 and Fig 3). Fig 3 shows blots of ClaI- or EcoRV-digested DNA. All probes hybridized to multiple bands (Fig 3). To confirm complete digestion of DNA, the blots were subsequently stripped of the probes and rehybridized with the BRM2 fragment from pGDB2-2 (KAWAMURA et al. 1999 ) as a single-copy gene probe. The BRM1 fragment has no ClaI and EcoRV sites. The BRM2 probe hybridized to a single band (6.6-kb ClaI fragment or 13.0-kb EcoRV fragment) in each lane (data not shown). These results suggest that NAF8 has multiple copies of each homolog of four AKT genes.

View larger version (70K): In this window In a new window Download PPT slide   Figure 3. Distribution of the AKT homologs in strain NAF8. Total DNA of NAF8 was digested with ClaI (lane 1) or EcoRV (lane 2) and separated in 0.8% agarose gel. The blots were probed with the AKT1, AKT2, AKTR-1, and AKT3-1 fragments from p1EX, p2XS, pRS, and p3HB, respectively (Fig 2). Sizes (in kilobases) of marker DNA fragments (HindIII-digested DNA) are indicated on the left.

To investigate the genomic distribution of the AKT homologs, the chromosome-sized DNA of NAF8 was separated by PFGE, and the blot was probed with the AKT1, AKT2, AKTR-1, and AKT3-1 fragments (Fig 2 and Fig 4). Under the electrophoresis condition for separating 1.0–6.0 Mb DNA, at least 10 chromosomal DNAs of 1.0–5.7 Mb were separated (Fig 4A). All probes hybridized to a single band of 1.0 Mb. Chromosome-sized DNA was also separated by PFGE under the condition for providing good resolution of <2.0 Mb DNA. Under this condition, 1.0 Mb DNA detected in Fig 4A was resolved into two chromosomal DNAs of 0.95 and 1.05 Mb (Fig 4B). All probes hybridized to the 1.05 Mb DNA, indicating that the small chromosome encodes all of the AKT homologs (Fig 4B).

View larger version (61K): In this window In a new window Download PPT slide   Figure 4. Chromosomal distribution of the AKT homologs in strain NAF8. Chromosome-sized DNA molecules of strain NAF8 were separated by pulsed-field gel electrophoresis under the condition for 1.0–6.0 Mb DNA (A) and for <2.0 Mb DNA (B). The blots were probed with the AKT1, AKT2, AKTR-1, and AKT3-1 fragments from p1EX, p2XS, pRS, and p3HB, respectively (Fig 2). Sizes (in kilobases) of chromosomes of S. pombe and S. cerevisiae (A) and of S. cerevisiae (B) are indicated on the left.

The chromosomal distribution of the AKT homologs in other strains of the strawberry pathotype was also examined. Strains NAF1 and NAF8 were collected from the same field in Aichi Prefecture, Japan, and the others were from different prefectures (KUSABA and TSUGE 1994 ). Chromosome-sized DNA of each strain was separated by PFGE under the condition for <2.0 Mb DNA. All strains had 1.05 Mb chromosomal DNA (Fig 5). NAF1, T-32, and M-30 had additional small DNAs of 0.95, 1.13, and 0.37 Mb, respectively (Fig 5). The gel blot was hybridized with the AKT1, AKT2, AKTR-1, and AKT3-1 probes. Fig 5 shows the blot hybridized with the AKT1 probe. All probes hybridized to the 1.05 Mb DNA in all strains (Fig 5). The 1.13 Mb DNA of T-32 was also hybridized with all probes (Fig 5). All of the AKT homologs are located on small chromosomes in the strawberry pathotype strains.

View larger version (102K): In this window In a new window Download PPT slide   Figure 5. Chromosomal distribution of the AKT homologs in the strawberry pathotype strains. Chromosome-sized DNA molecules of strains NAF1, NAF8, T-32, O-187, and M-30 were separated by pulsed-field gel electrophoresis under the condition for <2.0 Mb DNA (left). (Right) The blots were probed with the AKT1 fragment from p1EX (Fig 2). Sizes (in kilobases) of chromosomes of S. cerevisiae are indicated on left.

Isolation of cosmid clones containing the AKT homologs:
A cosmid genomic library of strain NAF8 was screened with the AKT2 or AKTR-1 probes (p2XS and pRS inserts, respectively; Fig 2), and 39 positive clones were isolated. These clones were examined for distribution of the AKT homologs by DNA gel blot analysis with the AKT1, AKT2, AKTR-1, and AKT3-1 probes. This analysis divided these clones into at least four groups (Groups 1–4), which contain AKT1-AKTR-AKT3, AKTR-AKT3, AKT1-AKT2, and AKT1-AKT2-AKTR homologs, respectively. No clones had homologs for all four AKT genes.

Preliminary mapping of the AKT homologs in the representative clones of four groups suggested that these clones did not correspond to overlapping fragments of a single cluster of the AKT homologs. The cosmid clones pcAFT-1 (Group 1) and pcAFT-4 (Group 4) contain three homologs in the order AKT1-AKTR-AKT3 homologs and AKT1-AKT2-AKTR homologs, respectively. The cosmid clone pcAFT-2 (Group 2) has the AKTR and AKT3 homologs with an 31-kb region upstream of the AKTR homolog. The cosmid clone pcAFT-3 (Group 3) contains the AKT1 and AKT2 homologs with an 29-kb region downstream of the AKT2 homolog. The 31-kb and 29-kb regions in pcAFT2 and pcAFT-3, respectively, are apparently longer in size than the intergenic region upstream of the AKTR homologs in pcAFT-1 and pcAFT-4 (data not shown). These results strongly suggest that NAF8 has multiple clusters of the AKT homologs, which are different in distribution patterns of the homologs. Here we analyzed detailed structure of pcAFT-1 (Fig 2).

Structure of pcAFT-1:
The nucleotide sequence of a 32-kb region in pcAFT-1 was determined, and complete open reading frames (ORFs) with high similarity to AKT1, AKTR-1, and AKT3-1 were identified within a 24-kb region (Fig 2). We designated the AKT1, AKTR-1, and AKT3-1 homologs in pcAFT-1 as AFT1-1, AFTR-1, and AFT3-1, respectively (GenBank accession nos. AB070711, AB070712, and AB070713; Fig 2). Fig 2 shows the map of the 25-kb region containing these genes. The remaining 7.0-kb region downstream of AFT3-1 contained no putative ORFs with significant size, although short sequences homologous to fungal transposons were detected.

AFT1-1, AFTR-1, and AFT3-1 were deduced to consist of six, one, and four exons, respectively, on the basis of alignments with AKT1, AKTR-1, and AKT3-1 sequences. Introns of AFT1-1 and AFT3-1 contain consensus sequences for 5' splice sites [GT(A/G/T)(A/C/T)G(T/C)] and 3' splice sites [(C/T)AG], which is typical of fungal genes (BRUCHEZ et al. 1993 ). The cosmid clone pcAKT-1 of the Japanese pear pathotype encodes AKT1, AKT2, AKTR-1, and AKT3-1 on the same strand within a 15-kb region (Fig 2). However, pcAFT-1 lacks the AKT2 homolog and encodes AFT1-1 on the different strand from AFTR-1 and AFT3-1 (Fig 2).

The intergenic region between AFT1-1 and AFTR-1 was found to encode an additional ORF, ORFS1, and a transposon-like sequence, TLS-S1 (transposon-like sequence of the strawberry pathotype 1; Fig 2). These sequences have not been detected in the AKT cluster of the Japanese pear pathotype. ORFS1 possibly encodes 366 amino acids after splicing three introns. TLS-S1 has significant similarity to transposase genes of restless of Tolypocladium inflatum (KEMPKEN and KUCK 1996 ) and Tfo1 of Fusarium oxysporum (OKUDA et al. 1998 ), which are members of the hAT transposon family. TLS-S1, however, contains several stop codons and is probably a pseudogene of transposase. Structural and functional characterization of ORFS1 and TLS-S1 will be reported elsewhere.

Structure of the AFT genes:
AFT1-1 and AKT1 are almost identical in sequence (Fig 6). AFT1-1 potentially encodes a 578-amino-acid protein, the same size as Akt1 (TANAKA et al. 1999 ). AFT1-1 and AKT1 both consist of six exons of the same size (Fig 6). Of five introns, only the fourth intron is different in size (AFT1-1 = 60 bp; AKT1 = 61 bp). As with Akt1, Aft1-1 reveals similarity to carboxyl-activating enzymes, such as 4-coumarate-CoA ligases of higher plants, long chain acyl-CoA synthetases of microorganisms, and luciferases of insects (PAVELA-VRANCIC et al. 1994 ; TANAKA et al. 1999 ). Aft1-1 and Akt1 terminate with Ser-Lys-Ile, a peroxisomal targeting signal type 1 (PTS1) tripeptide that conforms to the PTS1 consensus motif established for S. cerevisiae (ELGERSMA et al. 1996 ).

View larger version (30K): In this window In a new window Download PPT slide   Figure 6. Comparison of structure of the AFT and AKT genes. Identity of amino acid sequences of the gene products is shown in parentheses. The open box in the intergenic region between AFTR-1 and AFT3-1 indicates the specific insertion sequence of 185 bp.

AFTR-1 is a single exon encoding a 445-amino-acid protein. It is more similar to AKTR-1 than to AKTR-2 (Fig 6). As with AktR-1 and AktR-2, AftR-1 contains two recognizable domains: a zinc binuclear cluster DNA-binding domain (Cys-X₂-Cys-X₆-Cys-X₆-Cys-X₂-Cys-X₆-Cys) in the amino terminal region and an internal leucine zipper domain (Leu-X₆-Leu-X₆-Leu-X₆-Leu) (TODD and ANDRIANOPOULOS 1997 ; TANAKA and TSUGE 2000 ). Proteins with the zinc binuclear cluster DNA-binding domain are known to be typical regulatory factors in fungi and have been placed in a single protein family, designated Zn(II)2Cys6 (TODD and ANDRIANOPOULOS 1997 ).

AFT3-1 encodes a protein of 296 amino acids. AFT3-1, AKT3-1, and AKT3-2 all consist of four exons of the same size (Fig 6; TANAKA and TSUGE 2000 ). Of three introns, only the second intron is different in size (AFT3-1 = 69 bp; AKT3-1 = 56 bp; AKT3-2 = 60 bp). AFT3-1 is more similar to AKT3-1 than to AKT3-2 (Fig 6). As with Akt3-1 and Akt3-2, Aft3-1 has similarity to members of the hydratase/isomerase enzyme family, such as enoyl-CoA hydratases, crotonases, and naphthoate synthases of microorganisms (MULLER-NEWEN and STOFFEL 1993 ; TANAKA and TSUGE 2000 ). Aft3-1 terminates with the same PTS1 tripeptide, Pro-Lys-Leu, as Akt3-1 and Akt3-2 (ELGERSMA et al. 1996 ; TANAKA and TSUGE 2000 ).

The 5' and 3' flanking sequences of AFT3-1 and AKT3-1 ORFs are highly conserved (Fig 6). However, the flanking sequences of the AFT1-1 and AFTR-1 ORFs show lower homology with those of the AKT1 and AKTR ORFs (Fig 6). The intergenic region between AFTR-1 and AFT3-1 is longer than that between AKTR and AKT3 due to a specific insertion of 185 bp (Fig 6).

Expression of the AFT genes:
Transcription of AFT1-1, AFTR-1, and AFT3-1 in strain NAF8 was investigated by RT-PCR. The AFT1-1-, AFTR-1-, and AFT3-1-specific primer pairs (1-5'/1-3', R-5'/R-3', and 3-5'/3-3', respectively; Fig 2) produced DNA fragments of 1.58, 0.92, and 0.92 kb, respectively, from total RNA of strain NAF8 (data not shown). Control reactions without reverse transcriptase produced no DNA fragments (data not shown), indicating that the RT-PCR products were amplified from cDNA templates. The sizes of RT-PCR products for AFT1-1 and AFT3-1 corresponded to those expected, when five and three introns, respectively, were spliced. Thus, AFT1-1, AFTR-1, and AFT3-1 are transcribed in NAF8.

Mutation of AFT1 and AFT3 by gene targeting:
To determine the function of AFT1-1 and AFT3-1 in AF-toxin biosynthesis, homologous recombination was employed to disrupt these genes in strain NAF8 with pGDT1 and pGDT3, which contain internal sequences of AKT1 and AKT3-1, respectively, of the Japanese pear pathotype in a fungal transformation vector pSH75 (Fig 2; TANAKA et al. 1999 ; TANAKA and TSUGE 2000 ). The targeting vectors pGDT1 and pGDT3 were expected to be useful to disrupt AFT1-1 and AFT3-1, respectively, in the strawberry pathotype, because the AKT1 and AKT3-1 sequences in these vectors are 94% identical to the corresponding regions of AFT1-1 and AFT3-1, respectively. Strain NAF8 was transformed with pGDT1 and pGDT3, and 162 and 92 transformants, respectively, were selected. AF-toxin production of transformants was evaluated on the basis of toxicity of culture filtrates to leaves of strawberry cultivar Morioka-16 and Japanese pear cultivar Nijisseiki. Two pGDT1 transformants (GD1-1 and GD1-2) and one pGDT3 transformant (GD3-1) were found to lose toxin production. When culture filtrates were applied to three leaves of each plant, culture filtrate of the wild type showed toxicity to all leaves of strawberry and pear, but those of GD1-1, GD1-2, and GD3-1 showed no toxicity to both plants (Fig 7A). When conidial suspensions were spray inoculated on three leaves of each plant, the wild type caused a number of spots on all leaves, but the transformants caused no spots on the leaves (Fig 7B). We used three sets of culture filtrates and conidia separately prepared and confirmed loss of toxin production and pathogenicity in these transformants.

View larger version (57K): In this window In a new window Download PPT slide   Figure 7. AF-toxin production (A), pathogenicity (B), and colony growth (C) of the wild-type strain and Tox- transformants. Tox- transformants GD1-1 (1-1) and GD1-2 (1-2) were produced by transformation of the wild-type strain NAF8 (W) with the AFT1-targeting vector pGDT1; GD3-1 (3-1) was produced by transformation of NAF8 with the AFT3-targeting vector pGDT3 (Fig 2). (A) Leaves of strawberry cultivar Morioka-16 (left) and Japanese pear cultivar Nijisseiki (right) were wounded slightly, treated with culture filtrates, and incubated in a moist box at 25° for 20 hr. (B) Leaves were spray inoculated with conidial suspensions (5 x 105 conidia/ml) and incubated in a moist box at 25° for 20 hr. (C) The wild type and transformants were grown on PDA (left) and PDA containing 100 µg/ml hygromycin B (right) at 25° for 4 days.

AF-toxin has been characterized as three related molecular species: AF-toxins I, II, and III (Fig 1), with toxin I being the predominant species with respect both to yield and to biological activity (MAEKAWA et al. 1984 ; NAKATSUKA et al. 1986 ; HAYASHI et al. 1990 ). Thus, toxins I, II, and III were quantified by reverse-phase high-performance liquid chromatography. Toxins I and II were detected in culture filtrates of the wild-type strain (1.81 and 0.11 µg/ml, respectively, in the averages of three cultures), but not in those of GD1-1, GD1-2, and GD3-1. Toxin III was not detectable in the wild type and transformants. The level of a precursor of AF-toxin, 9,10-epoxy-8-hydroxy-9-methyl-decatrienoic acid (FENG et al. 1990 ; NAKATSUKA et al. 1990 ), in culture filtrates was also quantified by high-performance liquid chromatography. Culture filtrates of the wild type contained the precursor molecule (0.75 µg/ml in the average of three cultures). In contrast, the same analysis could not detect the precursor molecule in culture filtrates of the Tox^- transformants.

All the Tox^- transformants were prototrophic (data not shown). To compare vegetative growth between the wild type and Tox^- transformants, they were grown together on three PDA plates at 25° for 4 days. On all plates, colony diameter of GD1-1 was similar to that of the wild type, but its colony morphology was different (Fig 7C). Colony growth of GD1-2 was slightly slower than that of the wild type (Fig 7C). Colony growth and morphology of GD3-1 were indistinguishable from those of the wild type (Fig 7C). The wild type and transformants were also grown on three PDA plates containing hygromycin B at 100 µg/ml. On all plates, colony diameter of GD1-2 was smaller than that of GD1-1 or GD3-1 (Fig 7C). This seemed to be also due to differences in copy number of the integrated hph cassette as shown below. When conidia of the Tox^- transformants were suspended in water, placed on glass slides, and incubated at 25°, they normally germinated and formed appressoria within 20 hr. These results strongly suggest that mutations, which affect pathogenicity in these transformants, are attributable to loss of AF-toxin production.

DNA gel blot analysis of the AFT1- or AFT3-targeted mutants:
The mode of integration of the vector in a subset of toxin-producing (Tox⁺) and Tox^- transformants was analyzed by DNA gel blot hybridization (Fig 8). Total DNA of the wild type and pGDT1 transformants was digested with HindIII, which has no site in pGDT1 and a single site in AFT1-1, and the blot was probed with the AKT1 fragment inserted in pGDT1 (Fig 2 and Fig 8A). The HindIII site in AFT1-1 is outside the region hybridizing to the AKT1 probe (Fig 2). The probe hybridized to two bands in the wild type (Fig 8A), showing the presence of multiple copies of the AFT1 homolog. Although Tox⁺ transformants preserved these two bands, Tox^- transformants lost both bands (Fig 8A). Tox⁺ and Tox^- transformants both carried high-molecular-weight hybridizing bands (>23 kb), suggesting integration of multiple copies of plasmids as tandem repeats (Fig 8A).

View larger version (70K): In this window In a new window Download PPT slide   Figure 8. DNA gel blot analysis of pGDT1 transformants (A) and pGDT3 transformants (B). Total DNA from the wild-type strain NAF8 (W), toxin-producing (Tox+), or toxin-minus (Tox-) transformants was digested with HindIII (A) or KpnI (B) and fractionated in 0.8% agarose gel. The blots were probed with the AKT1 (A) and AKT3-1 (B) fragments from pGDT1 and pGDT3, respectively (Fig 2). Sizes (in kilobases) of marker DNA fragments (HindIII-digested DNA) are indicated on the left.

Total DNA of pGDT3 transformants was digested with KpnI, which does not digest within pGDT3 and AFT3-1, and probed with the AKT3-1 fragment inserted in pGDT3 (Fig 2 and Fig 8B). The probe hybridized to three bands in the wild type (Fig 8B), showing the presence of multiple copies of the AFT3 homolog. Tox⁺ transformants preserved these three bands, and Tox^- transformants lost all of these bands (Fig 8B). The pGDT3 transformants also carried high-molecular-weight hybridizing bands (>23 kb) resulting from the integration of multiple copies as tandem repeats (Fig 8B).

DNA gel blot analysis suggested that all copies of the AFT1 or AFT3 gene were targeted in the Tox^- transformants. To determine whether the DNA flanking AFT1 or AFT3 was altered in the Tox^- transformants, the HindIII-digested DNA of pGDT1 transformants and the KpnI-digested DNA of pGDT3 transformants were hybridized with other AKT gene probes. The AKT2, AKTR-1, and AKT3-1 probes did not hybridize to DNA of the AFT1-targeted, Tox^- transformants (data not shown). The AKT1, AKT2, and AKTR-1 probes did not hybridize to DNA of the AFT3-targeted, Tox^- transformant (data not shown). Thus, it appeared that these Tox^- transformants lost all of the AKT homologs through deletion of a large part of the region controlling AF-toxin biosynthesis.

Electrophoretic karyotypes of the AFT1- or AFT3-targeted mutants:
To gain additional information on the extent of the deletion, which occurred in the Tox^- mutants, the chromosome-sized DNA of the wild type and mutants were separated by PFGE under the condition for <2.0 Mb DNA (Fig 9A). All mutants were found to lack the 1.05-Mb chromosome, which carries all of the AKT homologs in the wild type (Fig 4 and Fig 9A). No deletion products that originated from the 1.05-Mb chromosome were found in the 0.2- to 0.9-Mb region on the gel (Fig 9A). The pGDT1 transformant GD1-1 also lacked the 0.95-Mb chromosome (Fig 9A). The chromosome-sized DNA of the wild type and mutants were also separated by PFGE under the condition for 1.0–6.0 Mb DNA (Fig 9B). No other alterations were observed in the karyotype of GD1-2 (Fig 9B). However, GD1-1 also lacked a chromosomal DNA of 2.9 Mb, and the ethidium bromide-stained signal of 3.3 Mb DNA increased (Fig 9B). This suggested that the 2.9-Mb chromosome was increased in size to 3.3 Mb in GD1-1. In the pGDT3 transformant GD3-1, the signal of 3.3 Mb DNA decreased, and a larger-sized DNA of 3.4 Mb newly appeared (Fig 9B). The wild type probably has two chromosomes of 3.3 Mb, one of which was increased in size in GD3-1. These results indicated the possibility that the translocation of part of the 1.05-Mb chromosome to other chromosomes occurred in GD1-1 and GD3-1.

View larger version (40K): In this window In a new window Download PPT slide   Figure 9. Electrophoretic karyotypes of Tox- transformants. Chromosome-sized DNA molecules of the wild-type strain NAF8 (W) and Tox- transformants GD1-1 (1-1), GD1-2 (1-2), and GD3-1 (3-1) were separated by pulsed-field gel electrophoresis under the condition for <2.0 Mb DNA (A) and for 1.0–6.0 Mb DNA (B). The blots were probed with the 1.05 Mb chromosomal DNA of the wild-type strain (1.05 Ch), the backbone vector pSH75 of pGDT1 and pGDT3, the AKT1 fragment from pGDT1, and the AKT3-1 fragment from pGDT3. Arrowheads indicate chromosomal DNA hybridizing to any of the probes. Sizes (in kilobases) of chromosomes of S. cerevisiae (A) and chromosomes of S. pombe and S. cerevisiae (B) are indicated on the left.

To detect such translocation in the mutants, the 1.05 Mb DNA of the wild type was recovered and used as a probe for hybridization of the gel blots. The probe hybridized to the 1.05 Mb DNA in the wild type, but apparently not to any chromosomes in the mutants (Fig 9). Thus, it seemed that the translocation of the 1.05-Mb chromosome did not occur in the mutants. Longer exposure of the hybridized blot detected the 3.3 Mb DNA in GD1-1 with very weak signal. Such weak signal was probably due to hybridization of the probe to multiple copies of the AKT1 fragment in pGDT1 introduced into the chromosome as shown below.

Examination of the chromosomal distribution of targeting vectors in the mutants showed that integration of multiple copies of targeting vectors into the 2.9- and 3.3-Mb chromosomes of the wild type produced the 3.3- and 3.4-Mb chromosomes of GD1-1 and GD3-1, respectively. The gel blot shown in Fig 9B was stripped of the 1.05-Mb DNA probe and rehybridized with the backbone plasmid pSH75 of pGDT1 and pGDT3. This probe hybridized to 3.3, 2.6, and 3.4 Mb DNA of GD1-1, GD1-2, and GD3-1, respectively (Fig 9B). The hybridization signals in GD1-1 and GD3-1 were markedly intense compared with that in GD1-2 (Fig 9B). This result revealed that the 2.6 Mb DNA of GD1-2 contains single or few copies of targeting vector and that the 3.3 and 3.4 Mb DNA of GD1-1 and GD3-1, respectively, contain many copies of targeting vectors. The integration of pGDT1 or pGDT3 in these chromosomes of transformants was confirmed by rehybridizing the blot with the AKT1 and AKT3-1 fragments from pGDT1 and pGDT3, respectively (Fig 9B). The 3.3 Mb DNA of GD1-1 possibly resulted from the integration of 60 copies of pGDT1 (6.6 kb) into the original 2.9-Mb chromosome. The 3.4 Mb DNA of GD3-1 could be originated through the integration of 15 copies of pGDT3 (6.5 kb) into the 3.3-Mb chromosome. Such multiple-copy integration of transformation vectors is usual in transformation of A. alternata (TSUGE et al. 1990 ; SHIOTANI and TSUGE 1995 ; TANAKA and TSUGE 2000 ). It appeared that these mutants originated through loss of the 1.05-Mb chromosome simultaneously with ectopic integration of pGDT1 or pGDT3 into other chromosomes.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Here we found that the AKT homologs are present in multiple copies on a single chromosome of 1.05 Mb in strain NAF8 of the strawberry pathotype. We identified three AKT homologs (AFT1-1, AFTR-1, and AFT3-1) on a cosmid clone pcAFT-1 of NAF8 and also isolated three other types of cosmid clones, which are different in the distribution pattern of the AKT homologs. These results imply the structural complexity of the genomic region controlling AF-toxin biosynthesis.

In strain 15A of the Japanese pear pathotype, multiple copies of the AKT genes are also present on a single chromosome of 4.1 Mb; however, only one of each is functional (TANAKA et al. 1999 ; TANAKA and TSUGE 2000 ). DNA gel blot analysis of AK-toxin-minus mutants produced by transformation-mediated gene disruption allowed us to identify functional copies: toxin-minus mutants resulted when particular versions of the AKT genes were disrupted (TANAKA et al. 1999 ; TANAKA and TSUGE 2000 ). Transformation-mediated targeting of AFT1 and AFT3 in strain NAF8 of the strawberry pathotype also produced Tox^-, nonpathogenic mutants. However, these mutants resulted from loss of the 1.05-Mb chromosome encoding the AFT genes and not from site-directed integration of the disruption vectors into certain copies. Thus, we could not address whether all or particular copies of each of the AFT genes were functional. If multiple copies of the genes are functional, it is probably impossible to disrupt all of the copies by site-directed integration of targeting vectors. In this case, Tox^- mutants could result only when a large deletion of the region containing the AFT genes or loss of the AFT chromosome would result.

All strains of the strawberry pathotype tested had 1.05-Mb chromosomes encoding the AFT genes. Small chromosomes of several fungi have been identified as supernumerary (dispensable) chromosomes (COVERT 1998 ). Such chromosomes are similar to B chromosomes in other eukaryotic organisms, which are described as small, dispensable for growth, and inherited in a non-Mendelian manner (JONES 1991 ). The function of supernumerary chromosomes in most species is still cryptic. However, in the pea pathogen Nectria haematococca, the supernumerary chromosomes have been characterized to encode functional genes, which have roles in virulence to host plants (MIAO et al. 1991A , MIAO et al. 1991B ; COVERT et al. 1996 ; KISTLER et al. 1996 ; WASMANN and VANETTEN 1996 ; VANETTEN et al. 1998 ; HAN et al. 2001 ). In this pathogen, genes for phytoalexin detoxification as well as other virulence determinants are on the 1.6-Mb supernumerary chromosomes. Inherent instability of the supernumerary chromosomes does not affect growth, but does affect the disease-causing capacity on host plants (MIAO et al. 1991A , MIAO et al. 1991B ; KISTLER et al. 1996 ; WASMANN and VANETTEN 1996 ; VANETTEN et al. 1998 ; HAN et al. 2001 ). The apparent loss of a supernumerary chromosome following transformation was also reported in this pathogen (WASMANN and VANETTEN 1996 ). The chromosome-deletion transformants showed markedly reduced virulence on pea. Fungal supernumerary chromosomes, which are not required for growth but confer an advantage for colonizing certain ecological niches, have been termed conditionally dispensable (CD) chromosomes (COVERT 1998 ). The 1.05-Mb chromosome controlling AF-toxin biosynthesis and pathogenicity was found dispensable for saprophytic growth of strain NAF8. Thus, we propose that the 1.05-Mb AFT chromosome is a CD chromosome.

AKAMATSU et al. 1999 observed that all strains of A. alternata pathotypes have small chromosomes of <1.7 Mb, but that nonpathogenic isolates do not have such small chromosomes. In the apple pathotype, the AMT gene encoding a cyclic peptide synthetase was identified to be required for biosynthesis of cyclic peptide AM-toxin (JOHNSON et al. 2000A ). This gene was found to be located on a small chromosome of 1.1–1.7 Mb, depending on the strains (AKAMATSU et al. 1999 ; JOHNSON et al. 2000A ). JOHNSON et al. 2001 found an AM-toxin-minus, nonpathogenic isolate from laboratory stocks of the apple pathotype isolates and showed that it had no small chromosomes. This suggests that AMT is also located on CD chromosomes in the apple pathotype (JOHNSON et al. 2001 ). We observed previously that the AKT genes are located on the 4.1-Mb chromosome of strain 15A of the Japanese pear pathotype (TANAKA and TSUGE 2000 ). However, other strains of the Japanese pear pathotype tested were found to have the genes on small chromosomes of <1.8 Mb (A. TANAKA and T. TSUGE, unpublished results). Thus, most of the strains from the strawberry, apple, and Japanese pear pathotypes carry the genes for host-specific toxin biosynthesis on small chromosomes. It is possible that the 4.1-Mb AKT chromosome of 15A might result from translocation of an original smaller chromosome.

The patterns of repeated DNA sequences on certain supernumerary chromosomes of fungi suggest that they have a different evolutionary history from the essential chromosomes in the same genome and that they may have been introduced into the genome by horizontal transfer from another species (ENKERLI et al. 1997 ; COVERT 1998 ; ROSEWICH and KISTLER 2000 ). Supernumerary chromosomes have been shown to have the capacity for transfer between otherwise genetically isolated strains of Colletotrichum gloeosporioides (MASEL et al. 1996 ; HE et al. 1998 ). HE et al. 1998 provided experimental evidence that the 2.0-Mb supernumerary chromosome can be selectively transferred from the A-biotype strain to a vegetatively incompatible B-biotype strain in this fungus. Two biotypes both cause anthracnose diseases on the tropical legumes of Stylosanthes spp., but are different in host range and symptoms. Transfer of the 2.0-Mb chromosome did not affect pathogenicity of the recipient B-biotype strain, indicating that the chromosome might not carry genes determining pathogenicity (MASEL et al. 1996 ; HE et al. 1998 ). Although the pathological role of the 2.0-Mb chromosome is unknown, this observation indicates that the horizontal transfer of supernumerary chromosomes across incompatibility barriers can occur in fungi.

We previously measured genetic relatedness among A. alternata pathotypes on the basis of three DNA markers: restriction fragment length polymorphisms (RFLPs) of nuclear rDNA, nucleotide sequence variation in rDNA internal transcribed spacer regions, and RFLPs of mitochondrial DNA (KUSABA and TSUGE 1994 , KUSABA and TSUGE 1995 , KUSABA and TSUGE 1997 ). These analyses showed that a single pathotype population does not form a monophyletic group. The AKT homologs are unique in the Japanese pear, strawberry, and tangerine pathotypes (TANAKA et al. 1999 ; MASUNAKA et al. 2000 ; TANAKA and TSUGE 2000 ); the AMT gene is only in the apple pathotype (JOHNSON et al. 2000B ). These data indicate that the evolution of toxin production in A. alternata may involve horizontal transfer of DNA. All strains of the strawberry pathotype tested share the 1.05-Mb chromosome encoding the AFT genes. Thus, the ability to produce AF-toxin in the strawberry pathotype could be potentially imparted by transfer of the 1.05-Mb chromosome, rather than the DNA segment, containing the AFT genes to a strain of A. alternata. Horizontal transfer has also been proposed for the acquisition of genes for host-specific toxin biosynthesis in Cochliobolus carbonum (NIKOLSKAYA et al. 1995 ; AHN and WALTON 1996 ; WALTON et al. 1998 ; WALTON 2000 ) and C. heterostrophus (YANG et al. 1996 ; YODER 1998 ; KODAMA et al. 1999 ).

AFT1-1, AFTR-1, and AFT3-1 are 94, 96, and 94% identical in nucleotide sequence to AKT1, AKTR-1, and AKT3-1, respectively. The organization of the genes, however, is different between the AFT and AKT clusters. Furthermore, strains of both pathotypes have multiple sets of different clusters in the gene organization. Thus, it is unlikely that variability between the AFT and AKT clusters was originated by simple mutation. We found a new ORF, designated ORFS1, in the AFT cluster. DNA gel blot analysis revealed that ORFS1 was present in the strawberry pathotype strains, but not in strains of the Japanese pear and tangerine pathotypes (K. ITO, R. HATTA, A. TANAKA, M. YAMAMOTO, K. AKIMITSU and T. TSUGE, unpublished results). The AKT cluster in pcAKT-1 of the Japanese pear pathotype was found to contain two additional genes, AKT4 and AKTS1, upstream of AKT1 (TSUGE et al. 2001 ). AKTS1 is present only in the Japanese pear pathotype (TSUGE et al. 2001 ). Thus, the AFT and AKT clusters contain the pathotype-specific genes as well as the genes common in the strawberry, Japanese pear, and tangerine pathotypes. Comparison of the structure, function, and chromosomal distribution of the gene clusters from these pathotypes is of great importance in the study of the evolution of toxin biosynthesis and the origin of genes for toxin biosynthesis.

	ACKNOWLEDGMENTS

We are grateful to Sally A. Leong for providing pMLF2. We thank Motoichiro Kodama, Hirofumi Yoshioka, Kazuhito Kawakita, and Noriyuki Doke for valuable suggestions and the Radioisotope Research Center, Nagoya University, for technical assistance. This work was supported by Special Coordination Funds for Promoting Sciences from the Ministry of Education, Science, Sports, Culture and Technology of Japan, Grant-in-Aid for Scientific Research (B) from Japanese Society for Promotion of Sciences, and Daiko Foundation.

Manuscript received October 11, 2001; Accepted for publication February 4, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ADACHI, Y., H. WATANABE, and T. TSUGE, 1996  Relationships between genetic polymorphisms and fungicide resistance within Alternaria alternata.. Phytopathology 86:1248-1254.

AHN, J.-H. and J. D. WALTON, 1996  Chromosomal organization of TOX2, the complex locus controlling host-selective toxin biosynthesis in Cochliobolus carbonum.. Plant Cell 8:887-897[Abstract].

AKAMATSU, H., M. TAGA, M. KODAMA, R. D. JOHNSON, and H. OTANI et al., 1999  Molecular karyotypes for Alternaria plant pathogens known to produce host-specific toxins. Curr. Genet. 35:647-656[Medline].

ALTSCHUL, S. F., T. L. MADDEN, A. A. SCHÄFER, J. ZHANG, and Z. ZHANG et al., 1997  Gapped BLAST and PSI-BLAST: a new generation of protein database search programs. Nucleic Acids Res. 25:3389-3402[Abstract/Free Full Text].

AN, Z., M. L. FARMAN, A. BUDDE, S. TAURA, and S. A. LEONG, 1996  New cosmid vectors for library construction, chromosome walking and restriction mapping in filamentous fungi. Gene 176:93-96[Medline].

BRUCHEZ, J. J. P., J. EBERLE, and V. E. RUSSO, 1993  Regulatory sequences in the transcription of Neurospora crassa gene: CAAT box, TATA box, intron, poly(A) tail formation sequences. Fungal Genet. Newsl. 40:89-96.

CHU, G., D. VOLLRATH, and R. W. DAVIS, 1986  Separation of large DNA molecules by contour clamped homogeneous electric fields. Science 234:1582-1585[Abstract/Free Full Text].

COVERT, S. F., 1998  Supernumerary chromosomes in filamentous fungi. Curr. Genet. 33:311-319[Medline].

COVERT, S. F., J. ENKERLI, V. P. W. MIAO, and H. D. VANETTEN, 1996  A gene for macckiain detoxification from a dispensable chromosome of Nectria haematococca.. Mol. Gen. Genet. 251:397-406[Medline].

ELGERSMA, Y., A. VOS, M. VAN DEN BERG, C. W. T. VAN ROEMUND, and P. VAN DER SLUIJS et al., 1996  Analysis of the carboxyl-terminal peroxisomal targeting signal 1 in a homologous context in Saccharomyces cerevisiae.. J. Biol. Chem. 271:26375-26382[Abstract/Free Full Text].

ENKERLI, J., G. BHATT, and S. F. COVERT, 1997  Nht1, a transposable element cloned from a dispensable chromosome in Nectria haematococca. Mol. Plant-Microbe Interact. 10:742-749[Medline].

FENG, B.-N., S. NAKATSUKA, T. GOTO, T. TSUGE, and S. NISHIMURA, 1990  Biosynthesis of host-selective toxins produced by Alternaria alternata pathogens. I. (8R,9S)-9,10-epoxy-8-hydroxy-9-methyl-deca-(2E,4Z,6E)-trienoic acid as a biological precursor of AK-toxins. Agric. Biol. Chem. 54:845-848.

HAN, Y., X. LIU, U. BENNY, H. C. KISTLER, and H. D. VANETTEN, 2001  Genes determining pathogenicity to pea are clustered on a supernumerary chromosome in the fungal plant pathogen Nectria haematococca.. Plant J. 25:305-314[Medline].

HAYASHI, N., K. TANABE, T. TSUGE, S. NISHIMURA, and K. KOHMOTO et al., 1990  Determination of host-selective toxin production during spore germination of Alternaria alternata by high-performance liquid chromatography. Phytopathology 80:1088-1091.

HE, C. Z., A. G. RUSA, A. M. POLAWSKI, J. A. G. IRWIN, and J. M. MANNERS, 1998  Transfer of a supernumerary chromosome between vegetatively incompatible biotypes of the fungus Colletotrichum gloeosporioides.. Genetics 150:1459-1466[Abstract/Free Full Text].

JOHNSON, R. D., L. JOHNSON, Y. ITOH, M. KODAMA, and H. OTANI et al., 2000a  Cloning and characterization of a cyclic peptide synthetase gene from Alternaria alternata apple pathotype whose product is involved in AM-toxin synthesis and pathogenicity. Mol. Plant-Microbe Interact. 13:742-753[Medline].

JOHNSON, R. D., L. JOHNSON, K. KOHMOTO, H. OTANI, and C. R. LANE et al., 2000b  A polymerase chain reaction-based method to specifically detect Alternaria alternata apple pathotype (A. mali), the causal agent of Alternaria blotch of apple. Phytopathology 90:973-976.

JOHNSON, L., R. D. JOHNSON, H. AKAMATSU, A. SALAMIAH, and H. OTANI et al., 2001  Spontaneous loss of a conditionally dispensable chromosome from Alternaria alternata apple pathotype leads to loss of toxin production and pathogenicity. Curr. Genet. 40:65-72[Medline].

JONES, R. N., 1991  B-chromosome drive. Am. Nat. 137:430-442.

KAWAMURA, C., T. TSUJIMOTO, and T. TSUGE, 1999  Targeted disruption of a melanin biosynthesis gene affects conidial development and UV tolerance in the Japanese pear pathotype of Alternaria alternata. Mol. Plant-Microbe Interact.