Autonomous mobility of different copies of the Fot1 element was determined for several strains of the fungal plant pathogen Fusarium oxysporum to develop a transposon tagging system. Two Fot1 copies inserted into the third intron of the nitrate reductase structural gene (niaD) were separately introduced into two genetic backgrounds devoid of endogenous Fot1 elements. Mobility of these copies was observed through a phenotypic assay for excision based on the restoration of nitrate reductase activity. Inactivation of the Fot1 transposase open reading frame (frameshift, deletion, or disruption) prevented excision in strains free of Fot1 elements. Molecular analysis of the Nia+ revertant strains showed that the Fot1 element reintegrated frequently into new genomic sites after excision and that it can transpose from the introduced niaD gene into a different chromosome. Sequence analysis of several Fot1 excision sites revealed the socalled footprint left by this transposable element. Three reinserted Fot1 elements were cloned and the DNA sequences flanking the transposon were determined using inverse polymerase chain reaction. In all cases, the transposon was inserted into a TA dinucleotide and created the characteristic TA target site duplication. The availability of autonomous Fot1 copies will now permit the development of an efficient two-component transposon tagging system comprising a trans-activator element supplying transposase and a cis-responsive marked element.
BY their ability to move from one location in the genome to another, transposable elements act as insertional mutagens. The resulting mutated gene can then be isolated by using the inserted transposon as a tag. Transposons have been used successfully to clone genes from many organisms (for review see Berg and Howe 1989), including different plant species such as maize and snapdragon (Balcellset al. 1991; Gierl and Saedler 1992; Walbot 1992), Arabidopsis (Aartset al. 1993; Bancroftet al. 1993; Longet al. 1993), and Petunia (Chucket al. 1993); animals like nematodes (Greenwald 1985; Moermanet al. 1986), insects (Cooleyet al. 1988; Engels 1989), and yeast (Boeke 1989) or bacteria (Berget al. 1989).
Despite the great importance of filamentous fungi in ecology, animal and plant pathology, or in metabolic production, no transposon-based gene tagging system has so far been developed for these microorganisms. This is probably because laboratory strains of the two best-studied ascomycetes, Neurospora crassa and Aspergillus nidulans, appear to be devoid of active transposons (Kinsey and Helber 1989; Oliver 1992). However, our picture of fungal transposons has recently been enlarged as the result of work on other species such as plant pathogens, industrial and field strains. Active retroelements or DNA transposons exist in such fungal genomes (Oliver 1992; Dobinson and Hamer 1993; Daboussi 1996). We have thus initiated studies to develop a gene tagging system with the Fot1 element of the fungal plant pathogen Fusarium oxysporum, the first active DNA transposon reported in filamentous fungi (Daboussiet al. 1992).
Fot1 was discovered as an insertion within the nitrate reductase (nia) gene. It is 1928 bp long and has short (44 bp) inverted terminal repeats. Fot1 expresses one mRNA of 1.7 kb, which extends over most of the element and encodes a putative transposase of 542 amino acids (Deschampset al. 1999). Like Tc1-mariner elements (Robertson 1995), Fot1 duplicates 2 bp (TA) upon integration, but the transposase encoded by this element shares no significant sequence similarity with the Tc1-mariner transposases. However, the Fot1 transposase has recently been related to other transposases including the Tiggers element from humans and the pogo element from Drosophila (Smit and Riggs 1996). These similarities between the putative transposases and other shared features suggest that Fot1 belongs to the Tc1-mariner superfamily of DNA transposons.
The Fot1 element is widely distributed in the F. oxysporum species with a variable copy number, ranging from 0 to >100 (Daboussi and Langin 1994; this study). To determine whether cloned Fot1 elements can transpose autonomously, mobility was assessed in strains devoid of endogenous elements by use of a phenotypic assay for excision. We present evidence that Fot1 transposes in different strains of F. oxysporum. The success of these assays suggests that Fot1 will provide a valuable tool for tagging genes involved in pathogenicity in this economically important fungus. In addition, we report on the mechanism of Fot1 transposition.
MATERIALS AND METHODS
Fungal strains and media: The following wild-type strains of F. oxysporum (all obtained from C. Alabouvette, INRA, Dijon, France) were used: FOM24, pathogenic on melon, whose genome contains >100 copies of the Fot1 element; two nonpathogenic strains, FO5 with one Fot1 copy and FO47 that is free of Fot1. From these strains, nitrate reductase-deficient mutants were selected on the basis of resistance to chlorate as described in Daboussi et al. (1989). Mutants nia321 (FOM24), nia13 (FO5), and nia1 (FO47), with point mutations in the nitrate reductase structural gene (nia), were used as recipient strains in the transformation experiments. Interestingly, the nia13 (FO5) mutant lost the unique Fot1 copy present in the wild-type FO5 strain (T. Langin, J. M. Daviere, D. Fernandez and M. J. Daboussi, unpublished results) and thus is considered in transformation experiments as a strain free of Fot1. niaD37, niaD136, and niaD62 are the mutants containing a Fot1 insertion within the niaD gene of A. nidulans introduced by transformation into the nia321 FOM24 recipient strain (Daboussi and Langin 1994; Deschampset al. 1999). The minimal and complete media and culture conditions used were described previously (Daboussi-Bareyre 1980). Strains were stored as sporulated mycelium-potato dextrose agar (PDA) plugs under mineral oil at 12°.
Vector constructions: Plasmid p11ΔNdeI was first generated from plasmid pMJT2 (Langinet al. 1990), which contains the complete niaD gene of A. nidulans on a 7.6-kb BamHI fragment, by removing the NdeI fragment, leaving only the 2.7-kb EcoRI fragment containing most of the niaD gene in the BamHI/NdeI fragment. pFox15 was then derived from p11-ΔNdeI by the addition of a 5.0-kb NdeI fragment chosen at random from F. oxysporum strain FOM24 to increase the frequency of single site integration events while keeping the niaD sequence intact. The 2.7-kb EcoRI fragment of pFox15 was then replaced by either the 4.6-kb EcoRI fragment present in pIN62 or pIN136, constructed using the strategy previously described for pIN37 in Daboussi et al. (1992). The 4.6-kb EcoRI fragment contains the corresponding niaD sequence interrupted by the insertion of one Fot1 copy. Fot1-136 is at the same nucleotide position and in the same orientation as the Fot1-37 copy previously described in Daboussi et al. (1992) and Fot1-62 is at another site and in the opposite orientation (Figure 5). These plasmids, named pEC136 and pEC62, were used in cotransformation experiments together with plasmid pAN7-1 (Puntet al. 1987), which carries the hygromycin B phosphotransferase (hph) gene of Escherichia coli under the control of the gpdA promoter of A. nidulans as a selectable marker.
Plasmid pEC62ΔBamHI was constructed by deleting the internal 0.9-kb BamHI fragment of Fot1. Plasmid pEC62-hph C contains the 1.4-kb HpaI fragment from pCB1003 (Carollet al. 1994) encoding the selectable marker hph under the control of the A. nidulans trypC promoter, which was inserted into the SnaBI site in the coding region of Fot1. A third plasmid, pEC62-fr, with a frameshift mutation in the Fot1 coding region, was constructed by digesting pEC62 with XmaCI followed by filling-in with Klenow and recircularizing the construct. These three plasmids contain a EcoRI fragment consisting of the niaD gene interrupted by a defective Fot1-62 copy but differing in size according to the alterations in the transposase gene, e.g., 3.7 kb in pEC62ΔBamHI, 6.0 kb in pEC62-hph C, and 4.6 kb in pEC62-fr.
Transformation experiments: Protoplast preparation and polyethylene glycol-mediated transformation of the nitrate reductase-deficient mutants were conducted according to Langin et al. (1990). A total of 5 μg of pAN7-1 plus 5 μg of one of the plasmids pEC136 and pEC62 was used in each cotransformation. Hygromycin-B-resistant colonies were transferred to minimal medium containing 200 μg/ml of hygromycin B (Sigma, St. Louis) and 20 mm glutamine, and single spore isolates were purified on the same medium. A myceliumagar plug of each hygromycin-B-resistant transformant was placed on a Petri dish containing 10 ml PDA covered by a Hybond N nylon membrane (Amersham, Arlington Heights, IL). After 2 days of incubation at 26° membranes were treated for 30 min in NaOH 0.5 m, 5× SSC (Sambrooket al. 1989) and for 30 min in Tris-HCl 0.5 m, pH 7.5, 10× SSC on a rotary shaker at room temperature. DNA cross-linking was performed by exposing the membranes to UV at 254 nm for 1 min. Cotransformed colonies were identified by hybridization to the 32P-labeled 2.7-kb EcoRI fragment of the niaD gene.
DNA preparation and Southern blot analysis: DNA for Southern blot analysis and for inverse polymerase chain reaction (IPCR) was obtained by a miniprep extraction method (Langinet al. 1990). For Southern blot analysis ∼10 μg of DNA was digested in the presence of 100 units of the restriction enzymes EcoRI, XhoI, BglII, or SacII (Boehringer, Indianapolis) and fractionated through a 0.6% agarose gel. Southern transfer was performed by standard methods (Sambrooket al. 1989) onto Hybond N nylon membranes (Amersham) and the DNA was fixed to the membranes by UV cross-linking. DNA probes correspond to PCR products. Primers used for the amplification of the niaD-specific probe were niaD-144 (5′-GTTCATGCCGTGGTCGCTGCG-3′) and niaD-145 (5′-CCC GGCCAAAGCCTCGAATTCG-3′). For the Fot1-specific probe, a unique primer (Ft1) deduced from the inverted terminal repeat sequence was used (5′-AGTCAAGCACCC ATGTAACC GACCCCCCC-3′). An hph-specific probe was obtained using hph1 (5′-CAGCGAGAGCCTGACCTATTGC-3′) and hph2 (5′-GCCATCGGTCCAGACGGCCGCGC-3′). A single copy sequence consisting of the 5.0-kb NdeI fragment obtained from pEC62 was also used as a probe. These probes were 32P-labeled with a random primer extension kit (Pharmacia, Piscataway, NJ). Hybridization and washing were performed using stringent conditions (65° and 0.2× SSC). The molecular analysis of cotransformants and revertants was assessed using successively the four probes described above. The presence of Fot1 copies in FO5- and FO47-derived strains was determined using Fot1 amplification products as probes, under low stringency conditions (55° and 2× SSC).
Phenotypic selection assay: Single spore isolates from independent cotransformants characterized by integration of the intact 4.6-kb EcoRI fragment of constructs pEC136 or pEC62 were grown for 48 hr in liquid complete medium at 26° on a rotary shaker (150 rpm). Conidial suspensions were obtained by filtering through a 100-mm nylon filter washed twice with sterile distilled water. Two protocols were used: (i) the conidia were plated (102-103 spores per milliliter) onto solid minimal medium (MM) containing nitrate as the sole nitrogen source (MM-nitrate) or (ii) the conidia were inoculated (106-107 spores per milliliter) in liquid MM-nitrate. Excision events were identified after 14-21 days of incubation at 26° by the appearance of the wild-type phenotype. Each patch of aerial mycelium developed on solid MM-nitrate was considered to be an independent reversion event. Usually, 1-10 reversion events per plate were observed, yielding a reversion frequency with respect to the number of spores plated of ∼10-2 to 10-3. In liquid medium selection, only one revertant was isolated from each MM-nitrate-containing tube. Then, single spore colonies from revertants were purified on MM-nitrate and further analyzed by Southern analysis.
DNA sequencing: The Fot1-62 copy was sequenced by the dideoxy chain termination method (Sangeret al. 1977) using the T7 Sequencing Kit (Pharmacia). Oligonucleotides derived from the Fot1-37 sequence (Daboussiet al. 1992) were used as primers. Sequences of empty excision sites were determined directly from one-twentieth of the total amplification product by using the Sequence PCR Product Sequencing Kit (USB) and the niaD-10 oligonucleotide (5′-GGCTTCTCATGGGGC TCGGC-3′) as primer.
Sequencing of the DNA regions flanking reinsertion sites of transposed Fot1 copies was performed using the T7 Sequencing Kit (Pharmacia) and the oligonucleotide pairs Ft2 (5′-CCTTCCTAATGGCGCGTGATCCCCG-3′) + Ft3 (5′-GGCGAT CTTGATTGTATTGTGGTG-3′) and Ft4 (5′-CTCTGCATTTT TAGCTATTTATTTGAC-3′) + Ft5 (5′-CGTCCGCAGAGTAT TACCGGCATTGTAG-3′) as primers.
Polymerase chain reaction (PCR): Putative Fot1 sequences in the FO5 and FO47 strains were sought using the unique Ft1 primer or the combination of the Ft2/Ft3 internal primers, and the nia gene was used as a positive control. To analyze the footprints left by excision of Fot1 copies in independent revertants, the 2.7-kb EcoRI fragment of the niaD gene was first amplified by primers niaD-144 and niaD-145. PCR was performed according to Saiki et al. (1988) in a Biomed 60 Thermal cycler, using 100-300 ng of total genomic DNA and 1 unit of Appligene Taq polymerase for 30 cycles (1.30 min at 94°, 1 min at 60°, and 1.30 min at 72°), followed by an additional extension step of 15 min at 72°.
IPCR analysis: To isolate the regions that flank the reinsertion sites of the transposed Fot1 copies, 2 μg of DNA from independent revertants was digested with either one of the following enzymes, EcoRI, XhoI, BglI, or SacII (Boehringer), which do not have restriction sites within the Fot1 element. The digested DNA was separated by agarose gel electrophoresis and a gel slice containing the reinserted Fot1 copy was obtained. DNA (∼100 ng) was eluted by the Jetsorb kit (Bioprobe, Montreuil, France) and directly self-circularized overnight in a total volume of 100 μl either at 16° using a ligation kit (Amersham) or at 9° by incubating with 1 unit of T4 ligase (Boehringer). After self-ligation, DNA was ethanol precipitated, resuspended in 20 μl of H2O, and ∼50 ng was used as template in IPCR by using the expand long template PCR system (Boehringer) and the oligonucleotides Ft2 and Ft3 as inverse primers. PCR conditions were 2 min at 94° followed by 10 cycles at 94° for 10 sec, 65° for 30 sec, 68° for 6 min, 20 cycles at 94° for 10 sec, 65° for 30 sec, and 68° for 10 min and a final elongation step at 68° for 17 min. A second cycle of nested IPCR was generally performed by using the oligonucleotides Ft4 and Ft5 as primers. IPCR products were purified by the Jetsorb kit (Bioprobe), and cloned with the pGEM-T vector system (Promega, Madison, WI).
Chromosomal assignment of transposed Fot1 copies: To determine the transposition pattern of Fot1 copies introduced into F. oxysporum, analysis of the electrophoretic karyotypes of transformant T272 from FO47 and of four derived revertants was performed. Intact chromosomal DNAs of F. oxysporum strains were obtained from protoplasts prepared by growing and digesting mycelia as described by Langin et al. (1990). The washed protoplasts were treated as described by Mäntylä et al. (1992). Pulsed-field gel electrophoresis was performed using the contour-clamped homogeneous electric field (CHEF)-mapper system (Bio-Rad, Richmond, CA). Chromosome-sized DNA bands were separated in 0.9% PFCA (pulsed field certified agarose, Bio-Rad) gels, by using 1× TBE (Sambrooket al. 1989) as running buffer, at 12°, 1.6 V/cm with switching intervals increasing from 30 to 60 min during 180 hr. After electrophoresis, gels were stained in ethidium bromide (1 μg/ml) for 1 hr and destained by soaking overnight at room temperature in distilled H2O. Size estimates were determined using chromosomal DNAs from Schizosaccharomyces pombe and Hansenula wingei (Bio-Rad) as molecular size reference markers. CHEF gels were soaked successively (2 × 15 min) in 0.25 m HCl, 0.5 m NaOH/1.5 m NaCl, and 1.5 m NaCl/0.5 m Tris-HCl (pH 7.5)/1 mm EDTA. The neutralized DNA was transferred onto Hybond-N blotting membranes (Amersham) for 2 hr in 20× SSC (Sambrooket al. 1989) by using a LKB (Piscataway, NJ) vacuum blotting system. DNAs were fixed using UV light. Hybridization was performed as described for Southern blots.
Screening for Fot1-free strains: Because resident genomic Fot1 elements could support the excision of transformed Fot1 copies inserted into the niaD gene by supplying transposase in trans, a genetic context free of active Fot1 was required to test the ability of the transformed elements to transpose autonomously. A collection of 50 pathogenic and nonpathogenic isolates of F. oxysporum was examined by Southern blot analysis using Fot1 as a probe (Daboussi and Langin 1994). Two strains, FO5 and FO47, both nonpathogenic, were shown to lack Fot1 on the basis of the absence of hybridization to Fot1 under moderately stringent conditions (data not shown). Moreover, a PCR was carried out using Fot1 primers corresponding either to the Fot1 inverted terminal sequences (Ft1) or to internal regions (complementary to Ft2, Ft3). No amplified Fot1 sequence was obtained from strains FO5 or FO47, confirming the hybridization studies.
Experimental design: Figure 1 summarizes the strategy adopted in our experimental design. We have constructed (see materials and methods) two plasmids, pEC136 and pEC62, which contain the A. nidulans niaD gene inactivated by the insertion of a Fot1 element. These elements (Fot1-136 and Fot1-62) are inserted into the third intron of the niaD gene, at different nucleotide positions and in opposite orientations (see Figure 5). The two plasmids were introduced by cotransformation with the pAN7-1 plasmid into different nia- recipient strains. Hygromycin-B-resistant transformants (5-20 transformants/microgram of pAN7-1) were screened for the presence of niaD sequences on colony blots with the niaD probe. Cotransformants were detected with a frequency of 50-90% in the different experiments. Southern hybridization analysis revealed that most cotransformants had the 4.6-kb EcoRI fragment from plasmid pEC136 or pEC62, which hybridizes to both 32P-labeled niaD and Fot1 probes (Figure 2), indicating that intact niaD::Fot1 constructs have been integrated. Some of these cotransformants contain only one copy of the construct as exemplified by T272, T280, and T281 in Figure 3, while cotransformants T183 and T229 contain at least two copies.
Selection of excision events: The ability of the plasmid-borne Fot1 to excise was first tested in the FOM24 strain, which contains active Fot1 elements. After 7-14 days of growth at 26° on solid MM-nitrate, FOM24-derived cotransformants with a sparse growth typical of niaD inactivation gave rise to patches of aerial mycelium indicating restoration of the wild-type phenotype at a relatively high frequency, usually in the range of 10-2 to 10-3 (see materials and methods).
In contrast, in FO5- and FO47-derived cotransformants free of endogenous transposons, reversion events could be detected only when spores were inoculated at the highest concentration (106-107/ml) into liquid MM-nitrate. To exclude the possibility that excision was promoted by an endogenous source of transposase supplied by other transposable elements present in the genome, as observed for Drosophila hobo elements introduced in Musca domestica (Atkinsonet al. 1993; Handler and Gomez 1995), we analyzed FO5 and FO47 transformants containing the different constructs in which the transposase has been inactivated by either frameshift, deletion, or disruption with the hph marker (see materials and methods). Reversion to the wild-type phenotype as a consequence of Fot1 excision from these constructs was never observed in the FO5 and FO47 genetic backgrounds, even after 2 mon of growth in liquid MM-nitrate medium. In contrast, it has been observed in the FOM24 strain, which contains a Fot1-transposase source. However, the excision frequency of inactivated Fot1 elements was very low compared to that of pEC62, suggesting that the endogenous source of transposase and/or transactivation by genomic copies is limited.
Fot1 reintegrates frequently after excision: EcoRI-digested DNA from independent revertants consistently showed a wild-type fragment of 2.7 kb hybridizing to the niaD probe (Figure 3, top), which confirmed the excision of at least one Fot1 copy from the niaD gene. In 22 independent FO47-deriving revertants that contain a single copy of the transgene (see Table 1), excision of Fot1-136 or Fot1-62 copies led to the disappearance of the 4.6-kb signal (examples shown in Figure 3, left). In contrast, for the 2 FO47- and the 17 FO5-derived independent revertants that contain two copies of the transgene (see Table 1), either (1) one of the two Fot1 copies integrated into the genome excised from the niaD gene, giving rise to the 2.7-kb band hybridizing to the niaD probe, while the other did not excise (e.g., Rt183, Ra,b,d 229, in Figure 3, right), or (2) both copies excised (e.g., Rb,c,u 183 and Rg229, in Figure 3, right).
The strains bearing somatic excisions were analyzed using a Fot1 probe to check for possible Fot reintegration (Figure 3, bottom). The results presented in Table 1 show that in the two genetic backgrounds, the excised Fot1 element reinserted in 70-75% of the revertants analyzed. Our data thus suggest that transposition of Fot1 likely occurs by a “cut-and-paste” mechanism, which includes excision of an element from its original site and reinsertion of the same element at another site.
Chromosomal assignment of transposed copies: The locations of the transposed copies of Fot1 in some revertants was determined to provide insight into the process of transposition of the Fot1 element.
The single copy of the pEC62 construct in a FO47-derived transformant (strain T272) was mapped on the 3.5-Mb chromosome (Figure 4). In this strain, following excision, the Fot1 element is associated with different chromosomes in the three out of four cases that were analyzed (Figure 4C). In the fourth case, insertion occurred on the same chromosome. These data suggest that Fot1 does not transpose preferentially to proximal sites and that the high frequency of reinsertion events is likely to result from transposition into unreplicated sites.
Sequences of excision footprints and reinsertion site: To prove that Fot1 transposes correctly into the two genetic backgrounds, we cloned and sequenced both the empty niaD sites and the flanking sequences of reinserted Fot1 elements.
DNA fragments, obtained after PCR amplification of genomic DNA of 18 independent revertants derived from FO5, FO47, and FOM24 cotransformants, were sequenced (Figure 5). As expected, the presence of a typical footprint consisting of the same three or four base pairs composed by the TA duplication site, plus one or two additional nucleotides from one or both ends of Fot1 (Daboussiet al. 1992), was consistently detected. Footprints did not display extensive variation (no more than four-nucleotide insertions) probably because larger footprints would be incompatible with gene expression. The type of footprint was distributed similarly in the different genetic contexts.
To determine the molecular basis of Fot1 integration events, amplification of Fot1 flanking sequences was performed by IPCR. Six Fot1 flanks were cloned and sequenced (Figure 6). The sequence in all clones contained the ends of the inverted terminal repeats of Fot1 flanked by a TA dinucleotide. After the TA, which most probably represents the duplication site, the sequences differed from each other and showed no significant similarity with any other sequence in the data bases. No strong consensus of the target sites, other than the TA dinucleotide, could be detected by comparison of the different insertion sequences. Comparison of the sequences of the integration sites obtained in our study with the consensus sequence for the flanking sequence of Tc1 integration sites (van Luenen and Plasterk 1994; Korswagenet al. 1996) did not reveal any clear resemblance.
Sequencing of the Fot1-62 copy: The complete sequence of the autonomous Fot1-62 copy has been determined (data not shown) and compared to that of the Fot1-37 copy (Daboussiet al. 1992). One base pair substitution (A/G at position 1772, outside the open reading frame) was found in Fot1-62. In addition, one error in the sequence of Fot1-37 was corrected (CG instead of GC at position 1732).
An autonomous Fot1 element is identified: We have described a strategy for identifying autonomous copies of the Fot1 element from F. oxysporum, which relies on the use of a phenotypic assay based on the expression of the niaD gene that is restored upon Fot1 excision. The expression of the niaD gene is easily detected by dense growth on medium with nitrate as the sole nitrogen source. Using this assay, we showed that Fot1 excision can occur in two F. oxysporum strains free of endogenous Fot1 elements. Moreover, the use of defective Fot1 elements unable to excise in these backgrounds demonstrates that excision of Fot1 from the niaD gene is not due to a cross-mobilizing system but is promoted by the transposase encoded by the cloned Fot1 copies. These two Fot1 copies (Fot1-62 and Fot1-136) represent the first autonomous DNA transposable element ever described in filamentous fungi.
Fot1 may transpose by a cut-and-paste mechanism: This study has permitted a detailed analysis of Fot1 integration at both the molecular and chromosomal level. The data presented here suggest that transposition of Fot1 appears to occur via a cut-and-paste mechanism in which the Fot1 element excises from the donor site and reinserts in a new location with a high frequency (∼75%).
The chromosomal locations of four transposed Fot1 demonstrate that Fot1 is able to jump from one chromosome to another, indicating that there seems to be no preference to transpose near the donor site. This pattern of distribution is quite different from that observed in the FOM24 strain, in which the numerous copies of Fot1 are concentrated on some chromosomes, suggesting a preference for transposition to genetically linked sites (J. M. Daviere, unpublished results). This difference suggests that patterns of transposition of Fot1 can vary with the host. This has been reported for transposed Ac/Ds elements that transpose preferentially to linked sites (60-70%) in maize (Dooner and Belachew 1989), tobacco (Joneset al. 1990; Dooneret al. 1991), and Arabidopsis (Bancroft and Dean 1993; Kelleret al. 1993), with pattern of distribution varying from locus to locus. In contrast, many transposition events in tomato were to unlinked sites (Osborneet al. 1991; Bhatt and Dean 1992). Another explanation for the difference in the Fot1 distribution pattern is the assumption that amplification of Fot1 in some strains may not result exclusively from conservative transposition but may also result from a replicative process involving extrachromosomal Fot1 copies detected in FOM24 (M. J. Daboussi, unpublished results). The amplification could also result from unequal crossing over between tandemly integrated Fot1 copies as described in cell lines (Hyrienet al. 1988; Toledoet al. 1993) and yeast (Paquinet al. 1992).
The origin of Fot1 footprints: Fot1, like many other transposons in animals and plants (Berg and Howe 1989), generally leaves a footprint of a few nucleotides when it excises. The analysis of these altered sequences addresses questions about their origin. Two possible explanations might account for footprints: (i) precise excision followed by incomplete DNA repair synthesis or (ii) imprecise excision. Detailed analysis of P transposition in Drosophila (Engelset al. 1990; Kaufman and Rio 1992) and Tc 1 in Caenorhabditis elegans (Plasterk 1991; Plasterk and Groenen 1992) showed evidence that the excision of these elements is followed by template-dependent double-stranded break (DSB) repair. Precise excision of the transposon followed by DSB repair that uses the sister chromatid or homologous chromosome as template usually results in the reintroduction of a new copy of the transposon into the old sites. In rare cases, the repair process is interrupted prematurely, resulting in partial reinsertion of sequences encoded by the template DNA (the so-called footprint).
From our data on somatic excision and integration events, we believe that the double-stranded gap repair mechanism may also be applicable to Fot1. Indeed, the structural similarities between Fot1 and Tc1, i.e., short size, TA site duplication, and conservation of the DD35E motif (Smit and Riggs 1996), as well as similarity of products arising from the excision of the Tc1 (Eide and Anderson 1988) and Fot1 elements, suggest that these elements may share common features in their transposition mechanism. However, additional experiments are required to demonstrate the involvement of DSB in F. oxysporum. One way is to investigate whether the presence of a wild-type ectopic template has any influence on the reversion frequency caused by Fot1 insertion. It is also possible to provide a marked transgenic template as described for P and Tc1 (Glooret al. 1991; Plasterk and Groenen 1992) to determine if specific alterations can be shuttled.
Development of a transposon-based gene tagging system: Identification of autonomous Fot1 copies offers the opportunity to develop a gene tagging system based on the use of a transposable element, a strategy not yet available in filamentous fungi. Although the present work was carried out with nonpathogenic strains of F. oxysporum, this approach is likely to be of great value in the molecular dissection of pathogenicity in phytopathogenic isolates of the fungus. The ability of Fot1 to excise and reinsert frequently at many sites distributed throughout the genome should be exploited to mutagenize the whole genome to generate mutants affected in pathogenesis. Mutants impaired in their pathogenic potential can be selected among Fot1-transposed strains, and genes mutated in this way can be isolated by using the transposon as a tag; the mutated allele, in turn, can be used as a probe to clone the wild-type gene. This strategy will clearly be advantageous over other strategies, given the ability of the introduced transposon to keep moving to new sites in the genome, generating many new mutations. However, to use the Fot1 element as a gene tag efficiently, one should be able to select as many independent somatic transposition events as possible in each transformant. Although somatic excision can be monitored easily through the nia excision marker and reinsertion was shown to occur at a frequency of 70%, the level of activity of the introduced Fot1 element may be insufficient in other strains for use directly in transposon tagging experiments. More active autonomous elements are being constructed by linking the open reading frame of Fot1-62 to strong promoters from A. nidulans genes, both constitutive (gpdA) or inducible (alcA, Felenbok 1991). These constructs will be associated with the nonautonomous element carrying the hph-resistant marker presented here, thus providing the basis for a two-component system as described for plants (Bancroftet al. 1992; Honmaet al. 1993).
Can Fot1 elements be used in other species? The extraordinary utility of transposons as tools for genome manipulation in plants, Drosophila, and nematodes leads one to imagine what might be achieved in filamentous fungi in similar ways. The phenotypic excision assay developed to monitor transposition excision in F. oxysporum, which consists of an autonomous Fot1 element inserted into an intron of the niaD gene, rendering it inactive until the element is excised, can be directly applied to different fungal species in which nia- mutants are available. The fact that these nia- mutants can easily be recovered through their resistance to chlorate (Cove 1976; Daboussiet al. 1989) provides the possibility to test their activity in many heterologous species.
Finally, the fact that Fot1-like elements have recently been discovered in different species of Ascomycetes, such as Magnaporthe grisea (Kachrooet al. 1994; Farmanet al. 1996), Aspergillus niger (Nyyssönenet al. 1996), A. nidulans (R. Prade, personal communication), Botrytis cinerea (Leviset al. 1997), N. crassa (Margolin et al. 1995), and Cochliobolus carbonum (Panaccioneet al. 1996) suggests that the host factors required for Fot1 transposition in these species should be present. This might be achieved either by checking for transposition or by transactivation of defective endogenous Fot1-like elements.
We thank Steven Langrell for constructing some plasmids and Dr. Masayuki Nozue for participating in the IPCR experiments. We also thank Charles White and Philippe Silar for helpful comments on the manuscript. Part of this research was supported by funds from the Centre National de la Recherche Scientifique (URA 2225). Q.M. and F.K. acknowledge the receipt of a fellowship under Human Capital and Mobility Project ERBCHRXCT930244.
Communicating editor: M. J. Simmons
- Received July 10, 1998.
- Accepted November 20, 1998.
- Copyright © 1999 by the Genetics Society of America