Trichothecene Nonproducer Gibberella Species Have Both Functional and Nonfunctional 3-O-Acetyltransferase Genes

Corresponding author: Makoto Kimura, Plant Science Center, RIKEN, 2-1 Hirosawa, Saitama 351-0198, Japan., mkimura{at}postman.riken.go.jp (E-mail)

Communicating editor: M. S. SACHS

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The trichothecene 3-O-acetyltransferase gene (FgTri101) required for trichothecene production by Fusarium graminearum is located between the phosphate permease gene (pho5) and the UTP-ammonia ligase gene (ura7). We have cloned and sequenced the pho5-to-ura7 regions from three trichothecene nonproducing Fusarium (i.e., F. oxysporum, F. moniliforme, and Fusarium species IFO 7772) that belong to the teleomorph genus Gibberella. BLASTX analysis of these sequences revealed portions of predicted polypeptides with high similarities to the TRI101 polypeptide. While FspTri101 (Fusarium species Tri101) coded for a functional 3-O-acetyltransferase, FoTri101 (F. oxysporum Tri101) and FmTri101 (F. moniliforme Tri101) were pseudogenes. Nevertheless, F. oxysporum and F. moniliforme were able to acetylate C-3 of trichothecenes, indicating that these nonproducers possess another as yet unidentified 3-O-acetyltransferase gene. By means of cDNA expression cloning using fission yeast, we isolated the responsible FoTri201 gene from F. oxysporum; on the basis of this sequence, FmTri201 has been cloned from F. moniliforme by PCR techniques. Both Tri201 showed only a limited level of nucleotide sequence similarity to FgTri101 and FspTri101. The existence of Tri101 in a trichothecene nonproducer suggests that this gene existed in the fungal genome before the divergence of producers from nonproducers in the evolution of Fusarium species.

TRICHOTHECENES are a large group of toxic sesquiterpene epoxides that inhibit protein biosynthesis in eukaryotic systems (DESJARDINS et al. 1993 ). Their structural diversity arises from the position, number, and type of the functional groups attached to the 12,13-epoxy-trichothec-9-ene skeleton. The mycotoxin is produced by certain Fusarium species such as Fusarium graminearum (type B trichothecene producer) and F. sporotrichioides (type A trichothecene producer), but not by most other species (e.g., F. oxysporum, F. moniliforme, F. decemcellulare, and F. solani). Recently, about half of the biosynthesis genes have been identified in a gene cluster (i.e., hereafter termed the Tri5-cluster since it contains Tri5, which is responsible for the first key step in the biosynthesis pathway) of F. graminearum and F. sporotrichioides (BROWN et al. 2001 ; LEE et al. 2001 ). However, the rest of the pathway genes were missing from the Tri5-cluster (KIMURA et al. 2001 ); so far, with the exception of only Tri101 (KIMURA et al. 1998A ), genes outside of the Tri5-cluster have not yet been isolated from producers.

Tri101 encodes trichothecene 3-O-acetyltransferase that is involved in the protection of F. graminearum against its own toxin. It appears to be the only functional 3-O-acetyltransferase gene of the producer (KIMURA et al. 1998A , KIMURA et al. 1998C ) and is under the transcriptional control of Tri6, a regulatory gene within the Tri5-cluster (MCCORMICK et al. 1999 ). The resistance gene product acts mainly on the first step after the second cyclization (i.e., conversion of isotrichodermol to isotrichodermin) in F. sporotrichioides (MCCORMICK et al. 1999 ). Although the extent to which Tri101 contributes to self-protection appears to differ between the type A and type B trichothecene producers, the resistance gene is similarly located between the phosphate permease gene (pho5) and the UTP-ammonia ligase gene (ura7) in their respective genomes. These housekeeping genes are also close together in the genome of the nonproducer F. oxysporum, suggesting that pho5 and ura7 may be linked in other Fusarium species (KIMURA et al. 1998B ). Since F. oxysporum appeared not to have apparent homologs of Tri101 as assessed by standard hybridization conditions in genomic Southern analysis, the resistance gene has previously been assumed to have been acquired horizontally as a defense option in the evolution of Fusarium species (KIMURA et al. 1998B ).

Horizontal transfer of fungal nuclear genes has so far been suggested from (1) the occurrence of the gene in a particular strain or species and not in a large number of their close relatives and (2) the inconsistency of the phylogeny of the gene in question from the traditionally accepted phylogeny of the organism (ROSEWICH and KISTLER 2000 ). However, instances of horizontal gene transfer are mostly considered "anecdotal" in fungi (WALTON 2000 ). Clear evidence for horizontal transfer is expected as more genomic sequences become available for analysis.

To find clear evidence of horizontal gene transfer without whole genome sequences, we sought to verify previous Southern analysis for the lack of Tri101 in several Fusarium species by cloning and sequencing their pho5-to-ura7 regions. Sequence and functional analysis of these regions led to an unexpected finding with interesting implications regarding the evolutionary history of Tri101 in that the nonproducers contained related pseudogenes in this region. Further analysis revealed that trichothecene nonproducer Fusarium strains possessed an additional functional 3-O-acetyltransferase gene different from Tri101 (designated as Tri201).

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Fungal strains:
F. graminearum F15 is a 3-acetyldeoxynivalenol (3-ADON) producer strain used in our previous studies (KIMURA et al. 1998A , KIMURA et al. 1998C ). Other strains were purchased from the Institute for Fermentation (IFO; Osaka, Japan): F. oxysporum IFO 31983, F. moniliforme (Gibberella fujikuroi) IFO 31251, F. decemcellulare IFO 30918, F. solani IFO 31094, Fusarium species (sp.) IFO 7772 (strain IFO 7772 was incorrectly deposited as "Gibberella zeae"), F. sporotrichioides IFO 9955, and Trichothecium roseum IFO 5772. In the following paragraphs, a gene name is preceded by the abbreviated name of the fungus (e.g., Fopho5 for F. oxysporum pho5).

Cloning of a portion of pho5 and ura7 from the trichothecene nonproducing Fusarium species:
To obtain the nucleotide sequences of pho5 and ura7 from the nonproducers, we first amplified a portion of these genes by PCR. Primers that successfully worked for the amplifications were (1) stPho51 (5'-ACTACTTTCATCGTCCCTGGTGAA-3') and stPhod1 (5'-GATACCAAGAAGCATGAAAAGAGC-3') for Fopho5, (2) stUra72 (5'-GGCGAGTGCTTTGTTCTCGATGAC-3') and stUrad2 (5'-CTGAGTGGGCTTTGTCTTCTCCTC-3') for Foura7, (3) stPho53 (5'-CTTCTGTGTCTCCATGTTGACCATCATGCTCGG-3') and stPhod5 (5'-GGGCCAAAGTTGAAAAAGAATTGGGCAAGAACG-3') for Fmpho5, (4) stUra74 (5'-TGCTGGTCTTCTTCTCAAGACTCTTGGTCTTCG-3') and stUrad2 for Fmura7, (5) stPho53 and stPhod3 (5'-CTCCGAAGCCAACAAGTGTTCGCCACATCTTGT-3') for Fsppho5, and (6) stUra73 (5'-CCGATCTCGATCTTGGAAACTACG-3') and stUrad2 for Fspura7. The PCR products were cloned in pGEM-TEasy (Promega, Madison, WI) and sequenced.

Amplification of the syntenic regions by long PCR:
On the basis of the partial sequence of phos5 and ura7, the syntenic region of each nonproducer was amplified by long PCR from the genomic DNA using the LA-PCR kit (Takara Biomedicals, Kusatsu, Japan). Primers that worked well for the amplifications were (1) UP-P3 (5'-CGCTGCCTCAGGAAAGATTGGTTCCATCATTGG-3') and rUP-U2 (5'-GCATCCGCATAAAACAACAATCTCAAACAAAAC-3') for F. oxysporum, (2) GfUP-P3 (5'-TTCCTGGCTACTGGGTCTCGGTTGCTACTATTG-3') and rGfUP-U3 (5'-GCCAAGCAATCGGTAATCGGTAATCGGTAGGAG-3') for F. moniliforme, and (3) GzUP-P1 (5'-AGCTCCCACACGCTCTGACAATGCTATCAAGC-3') and GzUP-U2 (5'-CTGAGCATCAGCGTTTCCTACGTCCCTATTATC-3') for Fusarium sp. IFO 7772. The amplified fragments (>8 kb) were cloned in pGEM-TEasy. Sequence templates were generated using the GPS1 genome priming system (New England Biolabs, Beverly, MA). Primers were synthesized on the basis of the above provisional sequences and used for both the PCR amplification and direct DNA sequencing.

Sequence comparison of the syntenic regions:
Sequence identities were calculated by using the CLUSTALW program in the LASERGENE software (DNAstar, Madison, WI). The syntenic regions that contain all four (pseudo)genes oriented in the same direction (i.e., F. graminearum vs. Fusarium sp. IFO 7772 and F. oxysporum vs. F. moniliforme) were analyzed by Harr plot analyses (HARR et al. 1982 ). The plots were made using the Harrplot 2.0 program (GENETYX-MAC 10.1, Software Development, Tokyo) with a window of 100 and a stringency of 75.

Construction of a phylogenetic tree:
The phylogenetic relationships of the Fusarium species used in this study were confirmed by constructing a neighbor-joining tree (CLUSTALW at DDBJ) based on the 28S rDNA sequences. Fifteen of the reference 28S rDNA sequences were retrieved from GenBank (accession numbers indicated), 7 sequences were generated from strains obtained from the IFO, and the final sequence was generated from F. graminearum F15. Primers F65 and R635 (see MULE et al. 1997 ) were used for the PCR amplification. The sequences are deposited with GenBank under accession nos. AB084297, AB084298, AB084299, AB084300, AB084301, AB084302, AB084303, AB084304. Bootstrap tests were performed with 1000 replications for the consensus tree. The values are shown only at the nodes of major clades for simplicity.

Cloning of FoTri201 from F. oxysporum:
The cDNA library of F. oxysporum was constructed using the Schizosaccharomyces pombe expression vector pcDSP21 (KIMURA et al. 1994 ); mRNA was isolated from spores germinated in the presence of 100 µg/ml T-2 toxin, and the SalI-NotI directional library was constructed using a Superscript cDNA synthesis kit (Invitrogen, Carlsbad, CA). S. pombe cells carrying the cDNA expression library were selected with T-2 toxin as described previously (KIMURA et al. 1998A ). The cDNA in the shuttle vector was recovered from a resistant colony and its nucleotide sequence was determined. Using the LA PCR in vitro cloning kit (Takara Biomedicals), the genomic regions upstream and downstream of FoTri201 were amplified by vectorette PCR from MunI and XhoI cassette libraries, respectively. The following primers were used for the amplifications: primer Fo201-longD1 (5'-CATTTTTGATGGTGGCGGTGATGGCAGGATGTG-3') and nested primer Fo201-longD2 (5'-TCGGTGATTGGGAAGCAGAAGCTGACTTGGGTGTAGGATC-3') for cloning of the upstream region and primer Fo201-longU1 (5'-CCGATGGAGAAATTACTGCTGCGTTATCGCTG-3') and nested primer Fo201-longU2 (5'-TATCGCTGAGGGATGAGGATACGGAGACATTG-3') for cloning of the downstream region.

Cloning of FmTri201 from F. moniliforme:
On the basis of the sequence of FoTri201, we designed primers Fo201-cloneU1 (5'-GACCTAGACATAGAGCTGGACATCATCG-3') and Fo201-cloneD2 (5'-GCCAATGAATTGCCCATACTTTTTCCAC-3') and used them to amplify a portion of FmTri201 from F. moniliforme. The PCR product was cloned in pGEM-TEasy and sequenced. The regions upstream and downstream of this fragment were amplified by vectorette PCR from EcoT22I and FbaI cassette libraries, respectively. Primer Gf201-longD1 (5'-TGAATACGCCAGTGCCATCATCTTTAACCTGG-3') and nested primer Gf201-longD2 (5'-ACCCAAGGGAAGCTCTGAGATAGACGCTGTAG-3') were used for amplification of the upstream region, and primer Gf201-longU1 (5'-CTTGGATTGGGGAATCCTGAGAGTGTGAGGAG-3') and nested primer Gf201-longU2 (5'-TCATGCCCAAGAGACCAGACGGAGAAATAAC-C-3') were used for amplification of the downstream region.

In vitro acetyltransferase assays:
The coding regions of Tri101 and Tri201 were amplified by PCR using the following primers (mismatched bases are underlined): Nco-Fm101 (5'-TACACCATGGTCGATCTAGACGTCGAAATCAACC-3') and Bam-Fm101 (5'-GAATGGATCCTGTGATCAATGCTGATCGACGTCTG-3') for FmTri101; Nco-Fsp101 (5'-TTCTACACCATGGCCAACATAGATACTCCTCATC-3') and Bam-Fsp101 (5'-ATGGGATCCAAGCTCATCAATGCGCAGGCCCAAG-3') for FspTri101; and Nde-Fo201 (5'-GTTACACATATGGTCGACCTAGACATAG-3') and Bam-Fo201 (5'-CATTGGATCCCCGCGAGCAACCGGAAAC-3') for FoTri201. The NcoI-BamHI or NdeI-BamHI fragments of the PCR products (without nonsynonymous substitutions in the coding region) were cloned in pET-19b or pET-12a (Novagen), respectively, and propagated in Escherichia coli DE3 (BL21). Recombinant enzyme was prepared from the bacteria and assayed for 3-O-acetyltransferase activity as described (KIMURA et al. 1998B ); the sample was developed on a TLC plate (Merck silica TLC) using ethylacetate/toluene (3:1) as the solvent.

Northern blot analysis and reverse transcriptase-PCR:
Total RNA was isolated from mycelia as described previously (KIMURA et al. 1998B ). Ten micrograms of RNA samples were denatured, separated on agarose gels containing formaldehyde, and transferred to nylon membranes (Schleicher & Schuell, Dassel, Germany). Probes were prepared by labeling the DNA fragments (entire length of FspTri101 or FoTri201 cloned in pGEM-TEasy) with digoxigenin using the RNA labeling kit (Roche Applied Science). Hybridization, washing, and detection were accomplished according to the manufacturer's protocol. For reverse transcriptase (RT)-PCR, the cDNA was synthesized from total RNA (treated with the RNase-free DNase I) using the Superscript first-strand synthesis system (Invitrogen, San Diego). The following primers were used for the amplification of the Tri201 cDNA: FoTri201-U (5'-CACTAAACGTTACAAACATGGTCGAC-3') and FoTri201-D (5'-GCCAATGAATTGCCCATACTTTTTCCAC-3').

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Cloning of the pho5-to-ura7 regions from trichothecene nonproducing Fusarium species:
A portion of pho5 and ura7 could be amplified from all nonproducers examined by PCR based on the corresponding gene sequences of F. graminearum (KIMURA et al. 1998B ). Primers were designed on the basis of the sequence of each fungal strain and used for long PCR. By this strategy, we were able to amplify the pho5-to-ura7 regions from F. oxysporum, F. moniliforme, and Fusarium sp. IFO 7772, but no PCR products were obtained from F. decemcellulare and F. solani (see also supplemental figures at http://www.genetics.org/supplemental/).

In sequence analyses of the regions between pho5 and ura7, an unexpected result was obtained. The BLASTX search (similarity at the translated level) generated significantly high alignment scores of these regions against the FgTRI101 polypeptide (accession no. BAA24430; i.e., FoTri101, 143 bits, E = 4 x e^-33; FmTri101, 125 bits, E = 1 x e^-27; and FspTri101, 612 bits, E = 1 x e^-174). In addition, pseudo succinyl-CoA synthase genes (scs) were identified between pho5 and ura7 in F. oxysporum (293 bits, E = 3 x e^-78, to CAB11045) and F. moniliforme (242 bits, E = 7 x e^-71, to CAB11045); this gene was not present between pho5 and ura7 in F. graminearum, but was located nearby. We were able to amplify the region upstream of ura7 from F. graminearum (1188 bp; GenBank accession no. AB076252) and Fusarium sp. IFO 7772 using primer rstSCS2 (5'-GACACTGGTTTGGGGCGATGATACCAGGACAAT-3'), which is based on the sequence of Foscs. In this manner, the regions from all four Fusarium species proved to be syntenic; they all contained pho5, Tri101, ura7, and scs (Fig 1A). All exon and intron regions were determined by RT-PCR of individual (pseudo)genes. The final sequences deposited with GenBank contain the regions flanked by the following primer annealing sites: between stPho53 and stUrad2 for the sequences of F. oxysporum (8856 bp; AB076253) and F. moniliforme (8435 bp; AB076254) and between stPho53 and rstSCS2 for the sequence of Fusarium sp. IFO 7772 (9362 bp; AB076255).

View larger version (32K): In this window In a new window Download PPT slide   Figure 1. Comparative analyses of trichothecene producer and nonproducer Fusarium species. (A) The pho5-to-ura7 syntenic regions from four Fusarium species are illustrated. Exons are shown in solid boxes. Arrows (over the solid boxes) represent directions of the genes and pseudogenes. X indicates that the gene function is inactivated by mutations (e.g., the presence of stop codons or substitution of active site residues of the encoded enzyme). A set of imperfect inverted repeats (IR) was found adjacent to FmTri101 (total 63 bp with 9 bp of mismatches; AB076254), but not adjacent to FgTri101, FoTri101, and FspTri101. Crossed lines denote inversions of the indicated regions. Inverted repeat motifs (including imperfect ones) of significant length were found only upstream and downstream of FmTri101 from F. moniliforme, but not around other (pseudo) genes examined in this study. (B) Harr plot analysis showing the nucleotide sequence similarity of the syntenic region; F. graminearum vs. Fusarium sp. (left) and F. oxysporum vs. F. moniliforme (right). The comparison was performed with a base window of 100 and stringency of 75. A local region of significant nucleotide sequence similarity (identities >80%) was found in the intergenic region of the four strains (marked by region S; see also Fig 1A in which the directions of alignments are indicated by arrows).

Characterization of the syntenic regions:
Four genes (including pseudogenes) were in the syntenic region of these Fusarium strains; the directions and orders of these genes, however, were different as shown in Fig 1A. Such rearrangements involve inversion processes (marked by crossed lines between gene lines of Fusarium sp. IFO 7772 and F. oxysporum), which are often associated with inversely oriented repeat sequences situated close to the inversion end points. However, the inverted repeat motifs (marked by IR in Fig 1A) were evident only upstream and downstream of FmTri101 (bp 1959–2021 and bp 3918–3856 of AB076254, with 9 bp of mismatches). RT-PCR of candidate housekeeping genes (i.e., pho5, ura7, and scs) revealed that all of them are transcribed; subsequent sequence analyses indicated that the corresponding coding regions of Fgscs, Foscs, and Fmscs were interrupted by stop codons as determined from cDNA sequencing (Fig 1A).

The syntenic regions of Fusarium strains were compared by Harr plot analyses with pairs of F. graminearum vs. Fusarium sp. IFO 7772 (Fig 1B, left) and F. oxysporum vs. F. moniliforme (Fig 1B, right). The coding regions of pho5 and ura7 were significantly conserved among Fusarium strains (see also supplementary tables at http://www.genetics.org/supplemental/). No significant similarities were found between the scs sequences of F. graminearum and Fusarium sp. IFO 7772; small regions of similarities, however, were noticed downstream (labeled region S; Fig 1A and Fig B) and at the C terminus of Tri101. In alignment of the syntenic regions of F. oxysporum and F. moniliforme, gaps of significant sizes appeared upstream and downstream of Tri101. Except for these gaps, these two strains showed relatively high similarities over the entire length of these regions (including the scs-ura7 intergenic region).

Phylogenetic position of Fusarium sp. IFO 7772:
Since the phylogenetic position of Fusarium sp. IFO 7772 was not certain, we examined the evolutionary relationships of the fungal strains used in the study using 28S rDNA sequence data. As shown in Fig 2A, Fusarium sp. IFO 7772 was separated from F. decemcellulare and F. solani and clustered together with other strains that belong to the teleomorph genus Gibberella. Strain IFO 7772 did not fall into a group that produces trichothecenes, which is in accordance with our failure to detect Tri5 by Southern blot analysis (data not shown).

View larger version (42K): In this window In a new window Download PPT slide   Figure 2. The nonproducer strains that belong to the teleomorph genus Gibberella have either a functional or a nonfunctional Tri101. (A) Phylogenetic relationships of the Fusarium strains used in the study determined from 28S rDNA sequence comparisons. The neighbor-joining tree was constructed by using the reference sequences in GenBank (accession numbers are shown). The type of trichothecenes produced by the fungal strains is indicated as type A or type B; the teleomorph genus (e.g., Gibberella) is shown on the right part of the tree. The scale bar indicates one base change per 100 nucleotide positions. Bootstrap values from 1000 replications are shown at the nodes of major clades. Fusarium sp. IFO 7772 was positioned as a nonproducer strain that belongs to the teleomorph genus Gibberella. (B) In vitro acetylation assay of trichothecene by recombinant TRI101. Lane 1, T-2 toxin standard; lanes 2 and 3, ethylacetate extract of T-2 toxin incubated with acetyl-CoA and the crude extract of bacteria expressing FmTri101 (lane 2) or FspTri101 (lane 3); lane 4,3-acetylT-2 toxin (3-AT-2) standard. (C) Northern blot analysis. Spores of Fusarium sp. IFO 7772 were germinated with or without 100 µg/ml of T-2 toxin on YG medium. Total RNA samples were run on a 0.8% formaldehyde-agarose gel (stained with ethidium bromide; bottom) and then hybridized with the FspTri101 riboprobe after transfer to a nylon membrane (top).

Characterization of Tri101 from the trichothecene nonproducing strains:
The putative coding region of FoTri101 was interrupted by an ochre codon and an opal codon. This observation indicates that the gene occurs as a pseudogene in F. oxysporum. In contrast, FmTri101 and FspTri101 did not contain any stop codons; each revealed an open reading frame (ORF) that encodes a predicted protein of 458 and 453 amino acids, respectively. Among these sequences, the FspTri101 ORF, but not the FmTri101 ORF, maintained a perfect conservation of the "HXXXDG" motif, which is known to be the catalytic domain of an acetyltransferase family (ST-PIERRE et al. 1998 ). No TRI6-binding consensus sequences (HOHN et al. 1999 ) were upstream of the initiation codons.

In accordance with the structural features of the encoded proteins, only recombinant FspTRI101 expressed in E. coli showed trichothecene 3-O-acetyltransferase activity (Fig 2B). To determine if the resistance (pseudo)genes were transcribed in the three nonproducers, we isolated total RNA from the spores germinated in the presence or absence of 100 µg/ml T-2 toxin. As shown in Fig 2C, induction of FspTri101 expression by T-2 toxin was demonstrated by Northern blot analysis. However, the transcripts of FoTri101 and FmTri101 were not detected even by the sensitive RT-PCR assay under either culture condition (data not shown).

Cloning and characterization of FoTri201 from F. oxysporum:
Although inactivating mutations are present in the coding regions of FoTri101 and FmTri101, germinating spores of F. oxysporum and F. moniliforme showed resistance to trichothecenes and also possessed activities to transfer an acetyl group to the C-3 position of trichothecenes in vivo (data not shown). This implied that they possess another unidentified 3-O-acetyltransferase gene. We thus attempted to clone the responsible gene from F. oxysporum as described in MATERIALS AND METHODS.

S. pombe transformants carrying the cDNA expression library were transferred to YEA plates containing 50 µg/ml T-2 toxin. Among 5000 Ura⁺ transformants, one T-2 resistant clone was obtained. Analysis of the nucleotide sequence of the cDNA insert revealed a putative ORF (designated as FoTri201) with significant similarity to FgTRI101 protein: this cDNA clone was subsequently determined to specify a truncated but active form (7 bp deleted from the first A of the initiation codon, which resulted in deletion of 8 amino acid residues from the N terminus; see below). The active site motif HXXXDG was conserved in FoTRI201 (Fig 3). On the basis of the cDNA sequence, the coding region of FoTri201 was amplified by PCR from F. oxysporum genomic DNA. Sequence analysis of the product showed that the coding region was not interrupted by introns. The promoter and terminator regions of FoTri201 were cloned by vectorette PCR (GenBank accession no. AB083516). The upstream region contained an original initiation codon 24 bp upstream from the first ATG of the cloned cDNA; the transcription of this upstream ATG codon was confirmed by RT-PCR analysis (data not shown). The complete FoTri201 ORF encoded a predicted protein of 454 amino acids with 63.0% similarity to the FgTRI101 polypeptide (BLASTP score; 582 bits, E = 1 x e^-165). As for the Tri101-related genes of the nonproducers, the TRI6-binding consensus sequence (HOHN et al. 1999 ) was not found in the promoter region of FoTri201.

View larger version (34K): In this window In a new window Download PPT slide   Figure 3. Amino acid sequence of TRI201 from the nonproducer F. oxysporum. The deduced amino acid sequence of F. oxysporum Tri201 (FoTRI201; bottom) was compared to TRI101 from F. graminearum (FgTRI101; top) and Fusarium sp. IFO 7772 (FspTRI101; middle). The first ATG of the truncated FoTri201 clone recovered in S. pombe corresponds to the ninth Met residues (see text). Asterisks and dots indicate identical and similar amino acids, respectively. Residues that comprise the catalytic triad (i.e., the HXXXDM motif) are boxed.

In the Northern blot analysis, only a small amount of the FoTri201 transcript was detected from spores germinated in the presence of T-2 toxin (Fig 4A); induction of the resistance gene by addition of trichothecene was not as evident as we have observed in F. graminearum (KIMURA et al. 1998C ) and Fusarium sp. IFO 7772 (this study). A sensitive and semiquantitative RT-PCR assay also showed slightly elevated expression of FoTri201 in the trichothecene-treated culture (Fig 4B). As expected from the conservation of the active site sequence motif, recombinant FoTRI201 produced in E. coli could specifically transfer an acetyl group to C-3 of T-2 toxin, nivalenol (NIV), and deoxynivalenol (DON) (Fig 4C).

View larger version (31K): In this window In a new window Download PPT slide   Figure 4. Expression analyses of Tri201. (A) Northern blot analysis. The RNA probe of FoTri201 was hybridized to the total RNA of the fungus in either the presence (+) or the absence (-) of T-2 toxin. (B) RT-PCR. Amplification of FoTri201 was done using the cDNA samples (RT; +) from both the trichothecene-treated (T-2 toxin; +) and untreated (T-2 toxin; -) cultures. Control reactions (RT; -) did not yield the PCR products. (C) TLC of in vitro acetylation assay. Lanes 1–9, standard samples of T-2 toxin (lane 1), 3-AT-2 (lane 3), NIV (lane 4), 3-acetylnivalenol (3-ANIV; lane 6), DON (lane 7), and 3-ADON (lane 9); ethylacetate extract of T-2 toxin (lane 2), NIV (lane 5), or DON (lane 8) after 3 hr of incubation with acetyl-CoA and recombinant FoTRI201.

Cloning of FmTri201 from F. moniliforme:
The coding region of Tri201 was highly conserved between F. oxysporum and F. moniliforme: FmTri201 from F. moniliforme could easily be amplified by PCR with primers based on the sequence of FoTri201. The EcoT22I-FbaI fragment (2216 bp; AB089164) containing the upstream and downstream regions of FmTri201 was cloned by vectorette PCR as described in MATERIALS AND METHODS. The complete ORF of FmTri201 coded for a polypeptide of 454 amino acid residues with high similarity to FoTri201 (85.8 and 89.2% identities at the nucleotide and peptide sequence level, respectively); the active site motif was also conserved.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

This study indicated that Tri101 is located in the pho5-to-ura7 syntenic region of Gibberella strains that do not produce trichothecenes. Our previous false conclusion regarding the horizontal acquisition of Tri101 (KIMURA et al. 1998B ), which was based on the suggestion that nonproducers lacked Tri101, could be attributed to the lack of significant nucleotide sequence similarities between the corresponding sequences of producers and nonproducers. Since long PCR (using several pairs of perfect match long primers) has not been successful for cloning the corresponding region from F. decemcellulare and F. solani, we have not isolated Tri101 from these fungal strains. It remains to be elucidated whether Tri101 was acquired after the major diversification of Gibberella from Calonectria or earlier in the evolutionary history of Fusarium species (Fig 3C).

Of the three Tri101 regions cloned from trichothecene nonproducing Fusarium, only FspTri101 proved functional with expression responsive to the addition of T-2 toxin to the culture. This feature is similar to that of F. graminearum (KIMURA et al. 1998C ), in which the resistance gene is under the transcriptional control of Tri6 (HOHN et al. 1999 ). However, no TRI6-binding consensus sequences were in the promoter region of FspTri101. One possible explanation is that the extant producer species (e.g., F. graminearum) evolved with fortuitous mutations in the Tri101-promoter regions to allow a noncluster trichothecene gene to be efficiently coregulated with other pathway genes. Alternatively, it is possible that the ancestor of Fusarium sp. IFO 7772 was a trichothecene producer and fortuitous mutations wiped out existing TRI6-binding sites or that it elaborated an alternative system for transcriptional activation of FspTri101.

Surprisingly, F. oxysporum and F. moniliforme possessed trichothecene 3-O-acetyltransferase activities although Tri101 was a nonfunctional pseudogene in these nonproducers. To search for the gene responsible for the acetylation, we isolated FoTri201 from F. oxysporum by cDNA expression cloning. On the basis of this sequence, FmTri201 was then cloned from F. moniliforme by PCR techniques. At the nucleotide sequence level, the similarity of Tri201 to FgTri101 was so low that it failed to be detected by standard Southern blot analysis. Perhaps the presence of functional (i.e., Tri201) and nonfunctional (i.e., Tri101) 3-O-acetyltransferase genes represents a classical example of gene duplication (KISTLER et al. 1995 ) and the absence of significant nucleotide sequence similarity reflects a long evolutionary history subsequent to the duplication event. Analyses of the neighbor genes to Tri201 may provide additional clues concerning function (other than conferring resistance to added trichothecenes) of this gene in F. oxysporum and F. moniliforme.

The aflatoxin biosynthesis genes exist in the aflatoxin nonproducer Aspergillus oryzae, which has traditionally been used as a koji mold for food and beverage fermentation in Asian countries (KLICH et al. 1995 ; KUSUMOTO et al. 1996 ). In this context we should not work on an implicit assumption that the trichothecene biosynthesis gene cluster has been absent from the genome in the evolutionary history of F. oxysporum and F. moniliforme. It is also possible that the ancestor of the extant nonproducer had once acquired a cluster of functional biosynthesis genes and subsequently lost functions and similarities by the accumulation of random mutations (due to the lack of selective advantage to having this gene cluster for F. oxysporum and F. moniliforme). In contrast, there is a selective pressure to maintain the functions of the cluster genes in F. graminearum [i.e., trichothecenes have a role as a pathogenicity factor for F. graminearum during its parasitism to the sensitive host plant (PROCTOR et al. 1995 )].

Complete genome databases will contribute significantly in understanding the phytopathogenicity of fungi and the molecular biology of trichothecene biosynthesis (PENNISI 2001 ). In addition, as we have learned from the unexpected discovery of trichothecene biosynthesis genes (i.e., pseudo-Tri101 and Tri201), the genomics of related nonproducer species may also lead to understanding the evolutionary history of the trichothecene biosynthesis genes. In fact, recent progress of the Magnaporthe grisea genome sequence project has revealed a region that is syntenic to a portion of the Neurospora crassa genome (HAMER et al. 2001 ). In view of the conservation of gene organization across those distant genera, the existence of syntenic regions within the genome of a single genus is not surprising. Hence, comparative studies of the trichothecene producers and closely related nonproducers will be important to fully understand the evolution of the genes necessary for trichothecene biosynthesis.

	FOOTNOTES

Sequence data from this article have been deposited with the EMBL/GenBank Data Libraries under accession nos. AB076252, AB076253, AB076254, AB076255, AB083516, AB084297, AB084298, AB084299, AB084300, AB084301, AB084302, AB084303, AB084304, and AB089164.

	ACKNOWLEDGMENTS

We thank Dr. Kerry O'Donnell for confirmation of Fusarium sp. IFO 7772 as a trichothecene nonproducer strain.

Manuscript received July 23, 2002; Accepted for publication October 30, 2002.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

BROWN, D. W., S. P. MCCORMICK, N. J. ALEXANDER, S. P. PROCTOR, and A. E. DESJARDINS, 2001  A genetic and biochemical approach to study trichothecene diversity in Fusarium sporotrichioides and Fusarium graminearum. Fungal Genet. Biol. 32:121-133.[Medline]

DESJARDINS, A. E., T. M. HOHN, and S. P. MCCORMICK, 1993  Trichothecene biosynthesis in Fusarium species: chemistry, genetics, and significance. Microbiol. Rev. 57:595-604.[Abstract/Free Full Text]

HAMER, L., H. PAN, K. ADACHI, M. J. ORBACH, and A. PAGE et al., 2001  Regions of microsynteny in Magnaporthe grisea and Neurospora crassa.. Fungal Genet. Biol. 33:137-143.[Medline]

HARR, R., P. HAGBLOM, and P. GUSTAFSSON, 1982  Two-dimensional graphic analysis of DNA sequence homologies. Nucleic Acids Res. 10:365-374.[Abstract/Free Full Text]

HOHN, T. M., R. KRISHNA, and R. H. PROCTOR, 1999  Characterization of a transcriptional activator controlling trichothecene toxin biosynthesis. Fungal Genet. Biol. 26:224-235.[Medline]

KIMURA, M., T. KAMAKURA, Q. Z. TAO, I. KANEKO, and I. YAMAGUCHI, 1994  Cloning of the blasticidin S deaminase gene (BSD) from Aspergillus terreus and its use as a selectable marker for Schizosaccharomyces pombe and Pyricularia oryzae.. Mol. Gen. Genet. 242:121-129.[Medline]

KIMURA, M., I. KANEKO, M. KOMIYAMA, A. TAKATSUKI, and H. KOSHINO et al., 1998a  Trichothecene 3-O-acetyltransferase protects both the producing organism and transformed yeast from related mycotoxins. Cloning and characterization of Tri101. J. Biol. Chem. 273:1654-1661.[Abstract/Free Full Text]

KIMURA, M., G. MATSUMOTO, Y. SHINGU, K. YONEYAMA, and I. YAMAGUCHI, 1998b  The mystery of the trichothecene 3-O-acetyltransferase gene. Analysis of the region around Tri101 and characterization of its homologue from Fusarium sporotrichioides. FEBS Lett. 435:163-168.[Medline]

KIMURA, M., Y. SHINGU, K. YONEYAMA, and I. YAMAGUCHI, 1998c  Features of Tri101, the trichothecene 3-O-acetyltransferase gene, related to the self-defense mechanism, in Fusarium graminearum. Biosci. Biotechnol. Biochem. 62:1033-1036.[Medline]

KIMURA, M., H. ANZAI, and I. YAMAGUCHI, 2001  Microbial toxins in plant-pathogen interactions: biosynthesis, resistance mechanisms, and significance. J. Gen. Appl. Microbiol. 47:149-160.

KISTLER, H. C., U. BENNY, E. W. BOEHM, and T. KATAN, 1995  Genetic duplication in Fusarium oxysporum.. Curr. Genet. 28:173-176.[Medline]

KLICH, M. A., P.-K. CHANG, E. J. MULLANEY, D. BHATNAGAR, and T. E. CLEVELAND, 1995  Hybridization of genes involved in aflatoxin biosynthesis to DNA of aflatoxigenic and non-aflatoxigenic aspergilli. Appl. Microbiol. Biotechnol. 44:439-443.[Medline]

KUSUMOTO, K., K. MORI, Y. NOGATA, H. OHTA, and M. MANABE, 1996  Homologs of an aflatoxin biosynthetic gene ver-1 in the strains of Aspergillus oryzae and its related species. J. Ferment. Bioeng. 82:161-164.

LEE, T., D.-W. OH, H.-S. KIM, J. LEE, and Y.-H. KIM et al., 2001  Identification of deoxynivalenol- and nivalenol-producing chemotypes of Gibberella zeae using PCR. Appl. Environ. Microbiol. 67:2966-2972.[Abstract/Free Full Text]

MCCORMICK, S. P., N. J. ALEXANDER, S. E. TRAPP, and T. M. HOHN, 1999  Disruption of TRI101, the gene encoding trichothecene 3-O-acetyltransferase, from Fusarium sporotrichioides.. Appl. Environ. Microbiol. 65:5252-5256.[Abstract/Free Full Text]

MULE, G., A. LOGRIECO, G. STEA, and B. ANTONIO, 1997  Clustering of trichothecene-producing Fusarium strains determined from 28S ribosomal DNA sequences. Appl. Environ. Microbiol. 63:1843-1846.[Abstract]

PENNISI, E., 2001  The push to pit genomics against fungal pathogens. Science 292:2273-2274.[Free Full Text]

PROCTOR, R. H., T. M. HOHN, and S. P. MCCORMICK, 1995  Reduced virulence of Gibberella zeae caused by disruption of a trichothecene toxin biosynthetic gene. Mol. Plant-Microbe Interact. 8:593-601.[Medline]

ROSEWICH, U. L. and H. C. KISTLER, 2000  Role of horizontal gene transfer in the evolution of fungi. Annu. Rev. Phytopathol. 38:325-363.[Medline]

ST-PIERRE, B., P. LAFLAMME, A. M. ALARCO, and V. DE LUCA, 1998  The terminal O-acetyltransferase involved in vindoline biosynthesis defines a new class of proteins responsible for coenzyme A-dependent acyl transfer. Plant J. 14:703-713.[Medline]

WALTON, J. D., 2000  Horizontal gene transfer and the evolution of secondary metabolite gene clusters in fungi: an hypothesis. Fungal Genet. Biol. 30:167-171.[Medline]

	NOTE ADDED IN PROOF

Region S (Fig 1) proved to correspond to a recently released EST (GenBank accession no. BU065253) of F. graminearum (nitrogen-starved mycelia).

This article has been cited by other articles:

				T. Igawa, N. Takahashi-Ando, N. Ochiai, S. Ohsato, T. Shimizu, T. Kudo, I. Yamaguchi, and M. Kimura Reduced Contamination by the Fusarium Mycotoxin Zearalenone in Maize Kernels through Genetic Modification with a Detoxification Gene Appl. Envir. Microbiol., March 1, 2007; 73(5): 1622 - 1629. [Abstract] [Full Text] [PDF]

				T. Tokai, M. Fujimura, H. Inoue, T. Aoki, K. Ohta, T. Shibata, I. Yamaguchi, and M. Kimura Concordant evolution of trichothecene 3-O-acetyltransferase and an rDNA species phylogeny of trichothecene-producing and non-producing fusaria and other ascomycetous fungi Microbiology, February 1, 2005; 151(2): 509 - 519. [Abstract] [Full Text] [PDF]