A Genetic Mechanism for Emergence of Races in Fusarium oxysporum f. sp. lycopersici: Inactivation of Avirulence Gene AVR1 by Transposon Insertion

Compatible/incompatible interactions between the tomato wilt fungus Fusarium oxysporum f. sp. lycopersici (FOL) and tomato Solanum lycopersicum are controlled by three avirulence genes (AVR1–3) in FOL and the corresponding resistance genes (I–I3) in tomato. The three known races (1, 2 and 3) of FOL carry AVR genes in different combinations. The current model to explain the proposed order of mutations in AVR genes is: i) FOL race 2 emerged from race 1 by losing the AVR1 and thus avoiding host resistance mediated by I (the resistance gene corresponding to AVR1), and ii) race 3 emerged when race 2 sustained a point mutation in AVR2, allowing it to evade I2-mediated resistance of the host. Here, an alternative mechanism of mutation of AVR genes was determined by analyses of a race 3 isolate, KoChi-1, that we recovered from a Japanese tomato field in 2008. Although KoChi-1 is race 3, it has an AVR1 gene that is truncated by the transposon Hormin, which belongs to the hAT family. This provides evidence that mobile genetic elements may be one of the driving forces underlying race evolution. KoChi-1 transformants carrying a wild type AVR1 gene from race 1 lost pathogenicity to cultivars carrying I, showing that the truncated KoChi-1 avr1 is not functional. These results imply that KoChi-1 is a new race 3 biotype and propose an additional path for emergence of FOL races: Race 2 emerged from race 1 by transposon-insertion into AVR1, not by deletion of the AVR1 locus; then a point mutation in race 2 AVR2 resulted in emergence of race 3.


Introduction
In the arms race between plants and pathogens, the pathogens can win by circumventing the immune system of host plants, e.g., by avoiding or suppressing defense mechanisms. In general, plants have two types of resistance: polygenic (horizontal), controlled by multiple genes, each with a small phenotypic effect, and monogenic (vertical), controlled by a single resistance (R) gene, which often confers a high level of resistance [1]. Monogenic resistance generates immune responses (e.g. hypersensitive reaction, HR) to particular pathogen(s) [1], and has been effective and practical to use in modern plant breeding. This resistance is described by the 'gene-for-gene theory' [2], which explains the relationship between pathogen races and host plant cultivars by the interaction between an avirulence (AVR) gene in the race and an R gene in the cultivar. When a race possessing an AVR gene attacks a cultivar carrying the corresponding R gene, resistance is induced in the plant and the disease does not occur. A loss of function in an AVR gene allows the pathogen to avoid induction of resistance in the cultivar, the pathogen gains pathogenicity to that cultivar, and a new pathogenic race has emerged.
The ascomycete Fusarium oxysporum Schlecht. emend. Snyd. et Hans. causes vascular diseases of many plant species, yet each strain of this fungus has strictly defined host specificity [3]. Strains that cause wilt disease only on tomato (Solanum lycopersicum L.) are classified as f. sp. lycopersici Snyd. et Hans. (FOL). Three races of FOL have been reported; their relationship with tomato cultivars is explained by the 'gene-for-gene theory' [4]. Original descriptions of FOL races 1, 2 and 3 appeared before 1895 in England, in 1939 in the USA and in 1978 in Australia, respectively [5]. In Japan, races 1, 2 and 3 were reported in Fukuoka in 1905, in 1966 and in 1997, respectively [6].
To date, the R genes I, I2 and I3 are known in tomato cultivars [7]; these R genes correspond to the avirulence genes AVR1, AVR2 and AVR3 in FOL, respectively (Table 1). Historically, race 1resistant cultivars (I i2 i3), races 1 and 2-resistant cultivars (I I2 i3), and races 1, 2 and 3-resistant cultivars (I I2 I3) have been bred sequentially, each genotype corresponds to the emergence of a new race.
AVR3 ( = SIX1), which exists in all races [12], is known to have two silent mutations (lysine or glutamine at amino acid 164) that do not influence avirulence to I3 cultivars [13]. FOL races can be determined by AVR gene combinations [11,14].
Based on the knowledge of AVR genes, it was suggested that FOL races emerged as follows [9]: race 1 (AVR1 AVR2 AVR3) lost the AVR1 locus and became race 2 (-AVR2 AVR3), which escapes recognition by the I gene; a nucleotide substitution in race 2 AVR2 resulted in race 3 (-avr2 AVR3), which evades recognition by both I and I2. Those mutations of AVR genes are consistent in many FOL isolates [11].
Mating type (MAT), vegetative compatibility group (VCG), and phylogeny have been used to characterize genetic relationships among FOL isolates [15,16,17]. MAT and VCG correlate with the phylogenetic relationship [16]. All FOL isolates belong to one of three clades (A1-A3) in the F. oxysporum phylogeny based on the intergenic region of ribosomal DNA (rDNA-IGS), suggesting a polyphyletic relationship with at least three FOL origins [16,17]. In Japanese isolates, race correlates with the phylogenetic relationship; races 1, 2 and 3 belong to clades A2, A1 and A3, respectively [16].
Masunaga el al. first reported emergence of race 3 in Japan in 1997 [18]. It is now the number one wilt disease problem in Japan, since most commercial tomato cultivars are resistant to races 1 and 2 but susceptible to race 3. Japanese race 3 isolates all group in clade A3 and are MAT1-2 and VCG 0033 [16].
In 2008, a new outbreak of Fusarium wilt caused devastating damage to tomato production in greenhouses in Hidaka, Kochi Prefecture, Japan (Fig. S1A, B). The genotype of the affected cultivar was I I2 i3, which suggested the presence of race 3. However, certain characteristics of the pathogenic isolate did not match those reported for previously described Japanese race 3 isolates, suggesting a different biotype, and tomato wilt caused by the novel biotype of FOL race 3 has been occurring in Kochi to date. In this study, the novel biotype was analyzed by phenotypic, genetic and phylogenetic criteria; results suggest a new path for emergence of races.

Results and Discussion
A race 3 isolate, KoChi-1, belongs to a different lineage from the known race 3 isolates in Japan A fungal isolate from the vascular tissues of diseased tomato in a greenhouse in Kochi Prefecture, Japan was identified as F. oxysporum based on morphology [19] and nucleotide sequence of the rDNA-internal transcribed spacer (ITS) region (DDBJ/ EMBL/GenBank accession No. AB675383). Characteristics of the isolate, designated KoChi-1, are summarized in Table 2. In planta assays showed that KoChi-1 caused wilt disease on cvs. Ponderosa (i i2 i3), Momotaro (I i2 i3) and Walter (I I2 i3), but not on cv. Block (I I2 I3), indicating that KoChi-1 was race 3 (Table 2; Fig. 1A, B). This result was consistent with the fact that the commercial cultivar grown in the greenhouse was Momotaro-Fight (I I2 i3, Takii Seed, Kyoto, Japan).
Previous studies found that all race 3 isolates obtained in Japan (representative isolate Chz1-A is presented in Table 2) grouped in the A3 clade [16] (Table 2; Fig. S2), and were MAT1-2 and VCG 0033. However, we found that KoChi-1 belongs to the A2 clade ( Table 2; Fig. S2), and is MAT1-1 and VCG 0030+0032. The A2 clade has been reported to include only race 1 isolates in Japan [16]. Taken together, these characteristics suggest KoChi-1 is a novel biotype of race 3, distinct from the race 3 isolates previously reported in Japan.
KoChi-1 is the first reported race 3 isolate carrying the AVR1 locus, which itself is truncated by a transposon Although previously reported race 3 isolates (e.g., Chz1-A) have no AVR1 locus [8,11], Southern blot analysis using an AVR1 fragment from race 1 isolate MAFF 305121 (733 bp, nt 673-1406 bp, AB674509) as a probe presented that KoChi-1 possessed a single copy of AVR1 in its genome ( Fig. 2A).
Then, we tried to amplify AVR1 from KoChi-1 using a primer set SIX4f-F2/SIX4f-R2 designed by Rep & Houterman to amplify AVR1 from race 1 ( Table 3). The amplicon from KoChi-1 (2685 bp) was longer than that of MAFF 305121 (1924 bp) (Fig. 2B). The sequence of KoChi-1 AVR1 was deposited in DDBJ/EMBL/GenBank databases with the accession No. AB674508. In this paper, nucleotide positions are assigned according to AB674508 unless otherwise stated.
BLASTN searches in the NCBI database suggested that the 759-bp insertion was a transposon with 15-bp terminal inverted repeats (TIRs; 59-CAGGGTTCAAATCCA-39; nt. 1043-1057, 1787-1801; Fig. 2C), and that both TIRs were flanked by 8-bp target site duplication (TSD; 59-CACACCGG-39; nt 1035-1042, 1802-1809; Fig. 2C). The sequence of the TIRs and the 59 region of the transposon were highly homologous to the autonomous transposon Hornet1 from F. oxysporum (AF076626) [20]. These characteristics are consistent with those of the hAT family of class II DNA transposons [21]. Hence, we have designated this transposon Hormin (Hornet1 in miniature). Hormin does not encode transposases (and is therefore not autonomous) and may have emerged from Hornet1 through a series of mutations. A transposon identical to Hormin was previously reported in the alcohol dehydrogenase gene Adh1 in FOL NRRL 34936 [22]. This is the first report of an F. oxysporum AVR gene truncated by a transposon.

KoChi-1 avr1 encodes a defective protein
The deduced amino acid sequence of KoChi-1 AVR1 with Hormin revealed a chimeric protein of 175 amino acids (avr1; Fig. S4) that may be nonfunctional. Here, we designate the AVR1 gene truncated with Hormin as avr1. To investigate the transcription of avr1, total RNA was extracted from tomato roots inoculated with KoChi-1 or MAFF 305121 (race 1, as a control). RT-PCR using primer set SIX4F/SIX4R (designed to amplify AVR1 including its intron) amplified a 734-bp fragment from MAFF 305121 RNA but not from KoChi-1 (Fig. 3). On the other hand, RT-PCR using primer SIX4F with primer hornet-like2 (designed on Hormin, see Table 3, Fig. 2C) generated a 440-bp fragment from KoChi-1 inoculated tomato only (Fig. 3), indicating that KoChi-1 avr1 is expressed in planta. Neither avr1 in KoChi-1 nor AVR1 in MAFF 305121 was expressed in mycelia grown on PDB or MM medium (data not shown). This expression pattern was consistent with that of AVR3 in FOL race 2 Fol007 [23].

Other KoChi-1 AVR genes
KoChi-1 avr2 contains the previously known point mutation G121A; it is one of three mutations known to cause loss of AVR2 function in race 3 isolates [9]. KoChi-1 AVR3 has a glutamine (E) type mutation (Table 2). To date, there have been no reports of E type AVR3 mutations in race 3 [13]. Both avr2 and AVR3 of KoChi-1 were expressed during infection of tomato roots (Fig. 3).
Complementation of KoChi-1 avr1 with AVR1 results in loss of pathogenicity to cultivars carrying the I gene KoChi-1 (avr1 avr2 AVR3) was transformed with the Fol004 (race 1) AVR1 gene. Each of three transformants (K-B-b, K-2-11 and K-2-12) had one copy of AVR1 integrated ectopically into chromosomal DNA to yield strains with the genotype (avr1 AVR1 avr2 AVR3) (Fig. S5); the AVR1 transgene was expressed (Fig. 3). Each of the three transformants lost pathogenicity to tomato cultivars carrying the I gene, e.g., Momotaro (I i2 i3) and Walter (I I2 i3) (Fig. 1A, B). This confirms that avr1 is not functional, and indicates that the mutation can be complemented by AVR1. It also indicates that the integrated AVR1 functioned in spite of coexisting with avr1.
How and where did KoChi-1 emerge?
According to the Broad Institute Fusarium genome database, FOL race 2 isolate NRRL 34936 bears AVR2, AVR3 and genes encoding small proteins secreted into tomato xylem on a small (ca. 2.2 Mb) chromosome. Since the chromosomal location of AVR1 is unknown, we investigated the location of KoChi-1 avr1 by CHEF Southern hybridization (Fig. 4A, B). avr1 was found on a small (ca. 2.5 Mb) chromosome together with avr2 and AVR3 (Fig. 4A, B; lane 8), which was also the case for AVR1 in race 1 isolates MAFF 305121 (1.6 Mb; Fig. 4A, B; lane 1). The small chromosome of each isolate had different size. However, although MAFF 103036 (a Japanese race 1 isolate) was found to carry AVR1 on a ca. 2.5 Mb chromosome, its AVR2 and AVR3 genes were found on a ca. 1.0 Mb chromosome (Fig. 4A, B; lane 2). Perhaps in MAFF 103036, chromosomal fragmentation resulted in relocation of AVR2 and AVR3 to an independent small chromosome. All race 2 and race 3 isolates carried AVR2 or avr2 and AVR3 on chromosomal DNA, but none of them had the AVR1 or avr1.
Mobile elements, together with point mutation in the gene [9,24,25], are involved in the loss-of-function of AVR in fungal plant pathogens such as Magnaporthe oryzae and Cladosporium fulvum [26,27,28,29,30]. Generally, mobile elements play a role in duplication and translocation of the genes/genomic regions in yellowing and wilt and upper leaves are yellowing; 4, all leaves are wilt and yellowing or dead. The symptoms were evaluated after three weeks of inoculation. Four plants were used in each isolate, with three replicates. doi:10.1371/journal.pone.0044101.g001  Figure 3 and previous study [16]. doi:10.1371/journal.pone.0044101.t002 Figure 2. AVR1 in KoChi-1 genome was truncated by a transposon Hormin. (A) Southern blot analysis to investigate the copy number of AVR1 gene. AVR1 probe was prepared using a primer set SIX4F/SIX4R (Table 3), and each gDNA was digested with restriction enzyme, EcoRV or NdeI (Fig. 2C). (B) Detection of AVR1 locus from KoChi-1 using a primer set SIX4f-F2/SIX4f-R2 (Table 3 the genome [20,31], sometimes they cause genetic mutations. AVR genes often locate on mobile element-rich regions in fungal plant pathogens, such as M. oryzae [32], Leptosphaeria maculans [24], Blumeria graminis [33], and F. oxysporum [34]. In Phytophthora infestans, more than five hundreds of potential avirulence genes carrying RxLR motif located in mobile element-rich genomic regions [35]. Moreover, in FOL NRRL 24936 (race 2), a large amount of mobile elements are located on the lineage specific (LS) chromosomes such as Chr03, Chr06, Chr14 (2.2 Mb; the small chromosome carrying AVR2 and AVR3) and Chr15. Of the 72 Hormin elements, 37 are located on LS chromosomes of NRRL 34936 (Fig. S3). Unlike other fungal isolates, it is easy to speculate how races emerged sequentially in FOL due to its simple combinations of AVR genes and the small number of races. Based on the arms race model [36], FOL and its races are considered to have emerged as follows [9] (Fig. S6): First, a nonpathogenic F. oxysporum isolate acquired a small chromosome carrying AVR1, AVR2 and AVR3, and became FOL race 1. The deletion of the AVR1 locus in race 1 resulted in the emergence of race 2 (-AVR2 AVR3), and the point mutation in AVR2 (shown as avr2) in race 2 resulted in the emergence of race 3 (-avr2 AVR3). Refer to Table 2 for relationships among AVR genes, where phylogenetic groups, MAT and VCG of each isolate are also indicated. This study presented an alternative model: AVR1 in a race 1 isolate (AVR1 AVR2 AVR3) lost its function by a transposon insertion, resulting in the emergence of race 2 (avr1 AVR2 AVR3), and race 3 (avr1 avr2 AVR3) emerged from the race 2 as a result of the point mutation (G121A) in AVR2 (Fig. S6). If this scenario describes how KoChi-1 emerged, then where might it have happened? Soilborne pathogens are often carried with seed [1]. KoChi-1 may have been imported on tomato seeds from a production field because we have not found race 2 isolates carrying AVR1 truncated by Hormin, so far, in Japan. There still is the possibility that KoChi-1 evolved via race 2 from a race 1 isolate belonging to the A2 clade in a particular field in Kochi Prefecture. Analysis of more isolates from Kochi, and seed production fields, will be necessary to test these hypotheses.

Pathogenicity assay
Race differential tomato cultivars were used to evaluate FOL pathogenicity. Each isolate was cultured on potato sucrose broth (PSB) for 5 days at 25uC and 120 rpm, and conidial suspensions (1.0610 7 conidia/ml) were prepared. Two seeds of each cultivar were sown to soil (Kureha Soil, Kureha, Iwaki, Japan) in a plastic pot (7 cm-diam.) and were maintained in a growth chamber (16 hours light at 28uC/8 hours dark at 25uC). Roots of 15-dayold tomato were injured, dipped in a conidial suspension for 5 min, and replanted to well-moistened soil. Two weeks later, external symptoms of each plant were evaluated as follows: 0, no wilt or yellowing; 1, lower leaves yellowing; 2, lower and upper leaves yellowing; 3, lower leaves yellowing and wilting and upper leaves yellowing; 4, all leaves wilted and yellowing or dead.

DNA extraction and standard PCR
Fungal genomic DNA (gDNA) was extracted using the protocol described earlier [37,38] with modifications.

Phylogenetic analysis
Nucleotide sequences of the rDNA-IGS fragment from KoChi-1 were aligned with those from other FOL isolates using CLUSTAL X 2.0 [40]. We constructed the phylogeny by the neighbor joining (NJ) method [41] based on Kimura's twoparameter model [42], using MEGA v. 4 [43]. The statistical reliability of each node was assessed using 1000 bootstrap iterations. F. sacchari (synonym, Gibberella sacchari; mating population B of the G. fujikuroi-species complex) FGSC 7610 was used as an outgroup. All sequence data except for KoChi-1 were cited from the NCBI database.

Mating type (MAT) and vegetative compatibility group (VCG) determination
Mating type, MAT1-1 or MAT1-2, was determined by PCR using Gfmat1a/Gfmat1b or GfHMG11/GfHMG12, respectively ( Table 3). The reaction mixture was prepared as described in the section of Standard PCR, reaction conditions were set as follows: One incubation at 94uC for 2 min; 30 cycles of: denaturation at 94uC for 30 s, annealing at 58uC for 30 s, and elongation at 72uC for 45 s; and a final extension at 72uC for 6 min.

Gene expression analysis
Tomato cv. Ponderosa was inoculated with F. oxysporum as described in the section entitled ''Pathogenicity assay''. Eight days after inoculation, we vigorously washed the tomato roots with sterilize water. After drying with paper towels, roots were crushed in liquid nitrogen and total RNA was extracted with the SV Total RNA Isolation System (Promega) following the manufacturer's manual. From the extracted total RNA, cDNA was synthesized using TaKaRa RNA PCR Kit (AMV) Ver. 3.0 (TaKaRa Bio). Expression of target genes was examined with 5 ng of cDNA. To investigate expression of AVR1, avr1, AVR2, AVR3, FEM1 and the tomato actin gene, primer sets SIX4F/SIX4R, SIX4F/hornet-like2, FP962/FP963, and FP157/FP158, and Actin-f/Actin-r (Table 3) were used for PCR, respectively. FEM1 [45,46] and Actin [47] were used as controls for fungal and plant genes, respectively. Negative controls substituted sterile water for conidial suspension. Reaction mixtures were prepared as described above. Thermal conditions were: One incubation at 94uC for 2 min; 35 cycles of: denaturation at 94uC for 30 s, annealing at 57uC for 30 s, and elongation at 72uC for 30 s; and a final extension at 72uC for 7 min.
Complementation with AVR1 using Agrobacterium tumefaciens-mediated transformation (ATMT) The AVR1 gene of FOL race 1 Fol004 was integrated into the KoChi-1 genome ectopically by the ATMT method. Transformation using the binary vector pPHSIX4c (carrying about 2.0 kb of AVR1 locus and phleomycin resistance gene) [8] was carried out following the procedure described earlier [10] with minor modifications. To suppress the growth of Agrobacterium after transformation, we used 25 mg/ml Melopen (Dainippon Sumitomo Phama, Osaka, Japan) and 50 mg/ml Zeocin (Invitrogen, San Diego, USA), respectively.
CHEF gel electrophoresis was performed in 1.0% Sea Kem gold agarose gel (FMC BioProducts, Rockland, USA) with CHEF MapperH XA Pulsed Field Electrophoresis System (BioRad, Hercules, USA). The condition to separate chromosomes was as described earlier [34] with slight modification; 260 hours run at 8uC, 1200-4800 s switch time at 1.5 V/cm. The running buffer 0.5xTBE was refreshed every 2 days. Chromosomes of Schizosaccharomyces pombe (BioRad) and Saccharomyces cereviciae (BioRad) were used as DNA size markers. The gel was stained with ethidium bromide to visualize chromosomes after running electrophoresis.