A small, cysteine-rich protein secreted by Fusarium oxysporum during colonization of xylem vessels is required for I-3-mediated resistance in tomato


  • Martijn Rep,

    Corresponding author
    1. Plant Pathology, Swammerdam Institute for Life Sciences, University of Amsterdam, PO Box 94062, 1090 GB Amsterdam, the Netherlands.
    Search for more papers by this author
  • H. Charlotte Van Der Does,

    1. Plant Pathology, Swammerdam Institute for Life Sciences, University of Amsterdam, PO Box 94062, 1090 GB Amsterdam, the Netherlands.
    Search for more papers by this author
  • Michiel Meijer,

    1. Plant Pathology, Swammerdam Institute for Life Sciences, University of Amsterdam, PO Box 94062, 1090 GB Amsterdam, the Netherlands.
    Search for more papers by this author
  • Ringo Van Wijk,

    1. Plant Pathology, Swammerdam Institute for Life Sciences, University of Amsterdam, PO Box 94062, 1090 GB Amsterdam, the Netherlands.
    Search for more papers by this author
  • Petra M. Houterman,

    1. Plant Pathology, Swammerdam Institute for Life Sciences, University of Amsterdam, PO Box 94062, 1090 GB Amsterdam, the Netherlands.
    Search for more papers by this author
  • Henk L. Dekker,

    1. Mass Spectrometry, Swammerdam Institute for Life Sciences, University of Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, the Netherlands.
    Search for more papers by this author
  • Chris G. De Koster,

    1. Mass Spectrometry, Swammerdam Institute for Life Sciences, University of Amsterdam, Nieuwe Achtergracht 166, 1018 WV Amsterdam, the Netherlands.
    Search for more papers by this author
  • Ben J. C. Cornelissen

    1. Plant Pathology, Swammerdam Institute for Life Sciences, University of Amsterdam, PO Box 94062, 1090 GB Amsterdam, the Netherlands.
    Search for more papers by this author


A 12 kDa cysteine-rich protein is secreted by Fusarium oxysporum f. sp. lycopersici during colonization of tomato xylem vessels. Peptide sequences obtained with mass spectrometry allowed identification of the coding sequence. The gene encodes a 32 kDa protein, designated Six1 for secreted in xylem 1. The central part of Six1 corresponds to the 12 kDa protein found in xylem sap of infected plants. A mutant that had gained virulence on a tomato line with the I-3 resistance gene was found to have lost the SIX1 gene along with neighbouring sequences. Transformation of this mutant with SIX1 restored avirulence on the I-3 line. Conversely, deletion of the SIX1 gene in a wild-type strain results in breaking of I-3-mediated resistance. These results suggest that I-3-mediated resistance is based on recognition of Six1 secreted in xylem vessels.


The innate immune system of plants is responsible for disease resistance through recognition, directly or indirectly, of proteins or other compounds that are secreted by invading microbes. In plants, resistance (R) genes encode proteins that recognize the presence of specific compounds from pathogens. In pathogens, the genes that encode these compounds (directly or indirectly) are called ‘avirulence’ genes. This matching of plant resistance genes with specific pathogen avirulence genes is generally called the ‘gene-for-gene’ model of disease resistance (Van der Biezen and Jones, 1998; Dangl and Jones, 2001).

For the unravelling of molecular mechanisms of disease resistance in plants, it is important to identify the avirulence factors from the pathogen. Identification of such factors may also lead to a better understanding of the molecular basis of pathogenicity, as their secretion in planta could play a positive role in colonization of a susceptible host plant. Indeed, many bacterial avirulence factors turned out to be ‘effector’ proteins that are injected in plant host cells to suppress defence reactions or otherwise promote survival or propagation of bacteria inside host tissues (Van’t Slot and Knogge, 2002). From fungal plant pathogens, few avirulence factors have yet been identified. The majority of these are small cysteine-rich proteins from Cladosporium fulvum, a tomato leaf colonizer that is currently the most developed model for fungus–plant gene-for-gene interactions (Joosten and De Wit, 1999; De Wit et al., 2002; Luderer et al., 2002).

Avirulence factors have also been identified in a few other fungal plant pathogens (Rohe et al., 1995; D’Silva and Heath, 1997; Mandel et al., 1997; Farman and Leong, 1998; Orbach et al., 2000). However, none has yet been found in root-invading fungi. Among these, xylem-colonizing Fusarium oxysporum is a good candidate for producing avirulence factors because it shows apparent gene-for-gene interactions with its hosts. Indications for this are the existence of single dominant resistance genes in various plant species (Huang and Lindhout, 1997; Brotman et al., 2002; Grajal-Martin and Muehlbauer, 2002; Benko-Iseppon et al., 2003) and the existence of cultivar specificity (races) in many formae speciales of F. oxysporum (collections of isolates that share the same host species). However, as F. oxysporum has no known sexual cycle, firm genetic proof of the existence of avirulence genes matching R genes is lacking.

In tomato, the resistance genes I, I-1, I-2 and I-3 have been described (summarized by Huang and Lindhout, 1997). One of these, I-2, has been cloned and belongs to a large group of R genes that encode intracellular proteins with nucleotide binding site (NBS) and leucine-rich repeat (LRR) domains (Simons et al., 1998). The I-2 gene is specifically expressed in tissue surrounding xylem vessels (Mes et al., 2000). This suggests that resistance to xylem-colonizing F. oxysporum is (mainly) mediated by xylem-contact cells, which is in accordance with earlier histological observations (Beckman and Roberts, 1995). These contact cells are likely to respond to avirulence factors that are secreted by the fungus in xylem sap. We therefore decided to analyse the xylem sap proteome of infected tomato plants (Rep et al., 2002; 2003).

Here, we report the identification of the first avirulence factor of F. oxysporum. It is a small, cysteine-rich protein secreted by the fungus into xylem sap, and it triggers disease resistance in tomato plants that contain the I-3 resistance gene.


Identification of Six1

Previously, a protein of 12 kDa was found in xylem sap of tomato infected with F. oxysporum f. sp. lycopersici (Fol) (Rep et al., 2002). Mass spectrometric analysis of this protein (p12) yielded several peptide sequences that did not match any known sequence. In order to obtain the coding sequence for the protein, we designed a set of degenerated primers based on one of the peptides (ACPEGQECTTFNAYNFR). These primers, P12-D1 (corresponding to ACPEGQEC) and a nested primer P12-D2A (corresponding to FNAYNFR) were used together with a vector primer (PACT Reverse) in sequential, nested polymerase chain reactions (PCRs) to amplify the 3′ part of a cDNA from a library of Fol007-infected tomato (see Table 2 for primer sequences). The products of both the first PCR (with P12-D1) and the nested PCR (with P12-D2A) were cloned. Sequence analysis showed that the ≈ 620 bp product of the first PCR encoded the complete peptide used for primer design and that the ≈ 600 bp product from the nested reaction was indeed derived from the first product. Subsequently, with a primer based on the amplified part of the cDNA (P12-R1) and a vector primer (PACT Forward), we amplified the 5′ part of the cDNA from the library. Finally, with primers based on the complete cDNA sequence (P12-F2 and P12-R1), we could amplify the same sequence from the genome of Fol007, thus establishing the fungal origin of the protein as well as the absence of introns. We designated the protein encoded by the cDNA Six1, for secreted in xylem 1. The complete cDNA sequence was submitted to the DDBJ/EMBL/GenBank databases under accession number AJ608702.

Table 2. PCR primers used in this study.
Primer nameSequenceTarget (positions relative to ATG)
P12-mid-FaaaatctagaGCACCACTTTCAATGCGTACSIX1 +503 to +522
TrpC-F3GACTGAGGAATCCGCTCTTG Agrobacterium nidulans TrpC terminator
PACT ForwardTAATACCACTACAATGGATGpACT2 library vector (Clontech)
PACT ReverseGTGCACGATGCACAGTTGpACT2 library vector (Clontech)

The sequence of Six1 and its corresponding peptides from p12 analysed by mass spectrometry is shown in Fig. 1. Three p12 sequence tags and two additional peptide masses match the predicted translation product. As expected for a secreted protein, Six1 contains a potential N-terminal signal peptide for endoplasmic reticulum (ER) translocation. However, after removal of this peptide, the predicted mature product is around 30 kDa, much larger than the 12 kDa protein found in xylem sap (Rep et al., 2002). Moreover, matches to peptides found in mass spectra are limited to the central part of the predicted protein sequence, covering about 10 kDa (Fig. 1). This central part contains six of the eight cysteine residues. The fate of the N- and C-terminal parts (each ≈ 8–10 kDa) is as yet unknown. Three potential N-glycosylation sites are present in the N-terminal part. The protein has no homologues in sequence databases, nor does it have any recognizable sequence motif.

Figure 1.

The translation product of the SIX1 gene and corresponding p12 sequence tags. The predicted translation product of the SIX1 cDNA from Fol007 is shown. Three tryptic peptides of p12 that were sequenced with tandem MS are indicated in bold. All in silico trypsin-generated peptides of the SIX1 translation product that match p12 peptides in MALDI-TOF spectra are underlined. The predicted signal sequence for translocation into the ER is in lower case; the arrow points to the predicted cleavage site (http://www.cbs.dtu.dk/services/SignalP; Nielsen et al., 1997). Potential N-glycosylation sites (*) and cysteine residues (C) are marked below the sequence. Some isolates have a lysine (K) instead of glutamate (E) at position 164.

Occurrence of SIX1 and SIX1-related sequences in formae speciales of F. oxysporum

Southern blots revealed that several isolates of Fol contain two SIX1-related sequences (Fig. 2, lanes 1 and 3–5). Isolation and analysis of BAC clones containing these sequences (see below) showed that one of these is SIX1 itself while the other is a homologous locus (SIX1-H) on a non-overlapping BAC clone that does not encode a functional protein because of insertion of a transposon (M. Rep et al., unpublished). Therefore, the isolates of forma specialis lycopersici contain only one intact copy of SIX1. With the SIX1-specific primers P12-F2 and P12-R1, we could amplify fragments of the expected size from all Fol isolates in our collection, regardless of race. Sequencing of the PCR fragment from several isolates confirmed that SIX1 was amplified and revealed the existence of a single nucleotide polymorphism (G490A, relative to the Fol007 sequence in Fig. 1), which affects the amino acid sequence (E164K) (see Fig. 1). Interestingly, most isolates from other formae speciales tested do not contain any SIX1-related sequence (Fig. 2). These include three isolates of forma specialis radicis-lycopersici, which causes foot and root rot on the same host, tomato. Only one isolate of forma specialis dianthi contains a SIX1-related sequence (Fig. 2, lane 9). As we could not amplify this sequence with PCR from genomic DNA using primers for either SIX1 or SIX1-H, we do not know whether the dianthi sequence encodes a functional protein. No homologue of SIX1 is present in the genome sequence of Fusarium graminearum (http://www-genome.wi.mit.edu/annotation/fungi/fusarium/). Probing of whole chromosomes separated by pulsed-field electrophoresis revealed that SIX1 and SIX1-H are both present on a single chromosome of 2.0–2.2 Mb (Fig. 3).

Figure 2.

SIX1 -like sequences are present only in some isolates of F. oxysporum. Shown is a Southern blot of total DNA of several F. oxysporum isolates, digested with NcoI and XhoI and probed with the SIX1 coding sequence. Lanes 1–5, isolates of forma specialis lycopersici, Fol007 (lane 1), F1-27 (lane 2), Fol004 (lane 3), Fol029 (lane 4), 4287 (lane 5). Lanes 6–8, isolates of forma specialis radicis-lycopersici, C63F (lane 6), C142 (lane 7), C560 (lane 8). Lanes 9 and 10, isolates of forma specialis dianthi, Fod1 (lane 9), Fod2 (lane 10). Lanes 11 and 12, isolates of forma specialis gladioli, Fog G2 (lane 11), Fog G6 (lane 12). Lane 13, non-pathogenic F. oxysporum isolate Fo47. Molecular size markers are indicated on the left (in kb). The position of a fragment containing a SIX1 homologue (SIX1-H) and the 1.8 kb NcoI–XhoI fragment containing SIX1 are indicated.

Figure 3.

SIX1 and SIX1-H are present on the same chromosome. Chromosomes of Fol isolates Fol004 and Fol007, as well as mutant F1-27, were separated with pulsed-field electrophoresis, blotted and probed with the SIX1 coding sequence. This probe recognizes both SIX1 and a homologue, SIX1-H (see Fig. 2). For each fungal isolate, the ethidium bromide-stained gel is shown to the left of the autoradiograph of the SIX1-probed blot. The weaker hybridization of the 2 Mb chromosome of F1-27 is due to the absence of SIX1 (only SIX-H is present). Molecular weight markers are indicated on the left (in Mb).

The genomic region surrounding SIX1 is rich in putative transposable elements

Two overlapping BAC clones containing SIX1 were isolated by PCR screening of a BAC library (see Experimental procedures). From one of the BAC clones, around 15 kb was sequenced including the SIX1 gene. These sequence data have been submitted to the DDBJ/EMBL/GenBank databases under accession number AJ608703. Remarkably, the SIX1 genomic locus shows a high density of potential transposable elements (Fig. 4). Two miniature impalas (mimps; Hua-Van et al., 2000) flanked SIX1 on either side while a third was found 6 kb upstream. Further upstream, an open reading frame (ORF) encoding a Fot1-like transposase was found. Highest similarity of this transposase, here called Fot5, is to Magnaporthe grisea Pot3 (Kang et al., 2001) and two Phaeosphaeria nodorum transposases (accession numbers CAD32687 and CAD32689; all around 50% identity at the protein level). At 86 bp after the stop codon of the transposase coding sequence, the sequence ACGTTA marks the possible end of the transposon (ACGT being a typical end of the terminal inverted repeat of a Fot1-like transposon and TA the target sequence). However, no corresponding inverted repeat sequence was found in the 787 bp upstream of the coding sequence.

Figure 4.

The genomic locus of SIX1. Of a BAC clone, around 15 kb surrounding the SIX1 gene was sequenced. Arrows indicate open reading frames, small black squares indicate mimps (miniature impalas). Small boxes in arrows indicate the position of (putative) introns. ORF1 and ORF2 encode proteins for which no sequence homology was found in databases. SHH1 encodes a protein with homology to salicylate hydroxylases. FOT5 encodes a Fot1-like transposase (see text for details). The known end-point of the deletion in avr3 mutant F1-27 is indicated by the arrow. Restriction sites used for subcloning are indicated.

Three additional ORFs were found in the 15 kb region surrounding SIX1 and were designated SHH1, ORF1 and ORF2. SHH1 (salicylate hydroxylase homologue 1) encodes a protein with homology to fungal and bacterial salicylate hydroxylases (salicylate 1-monooxygenases); it shows 32% identity to NahG from Pseudomonas putida and up to 47% identity to uncharacterized sequences from the genomes of M. grisea, Neurospora crassa, Aspergillus nidulans and Fusarium graminearum (http://www-genome.wi.mit.edu/annotation/fungi/fgi/). ORF1 encodes a protein with a potential N-terminal signal peptide for ER translocation but no sequence homology to other proteins in databases. ORF2 was only partially sequenced. It encodes a protein of at least 698 amino acids, again without similarity to proteins in databases. In the genome sequence of F. graminearum (http://www-genome.wi.mit.edu/annotation/fungi/fusarium/), no orthologues were found of any of the above-mentioned proteins.

Loss of SIX1 breaks I-3-mediated resistance

In a previous study, a mutant of Fol was identified that had lost avirulence on E779, a plant line carrying the I-3 resistance gene (Teunissen et al., 2003a). During further characterization of this mutant, called F1-27, we found that it had lost the SIX1 gene (Fig. 2, lane 2). PCR analysis indicated that the chromosomal deletion in this mutant is at least 12 kb and encompasses additional neighbouring ORFs (Fig. 4). F1-27 still contains the SIX1-H pseudogene on the same chromosome (Fig. 3).

In the progeny of a parasexual backcross of F1-27 to its wild-type parent, Fol007 (Teunissen et al., 2003a), the presence of SIX1 co-segregates with avirulence on E779 (results not shown), suggesting that the loss of SIX1 or another gene on the deleted segment was responsible for the breaking of I-3-mediated resistance by F1-27. To test this, SIX1 was introduced into F1-27 with Agrobacterium-mediated transformation. One transformant was obtained that restored avirulence on the I-3 line (Fig. 5). However, avirulence of this transformant was not complete. The inoculated I-3 plants were slightly smaller than control plants (Fig. 5A, ‘F1-27 + SIX1’), and many displayed vascular browning in the hypocotyl, reflected in an intermediate average disease index (Fig. 5B). Moreover, for unknown reasons, the two other stable SIX1 transformants that we obtained did not restore avirulence. Therefore, to confirm that SIX1 is required for I-3-mediated resistance, Agrobacterium-mediated transformation was used to disrupt the gene through homologous recombination with a deletion cassette containing the hygromycin resistance gene. Three transformants were obtained that had their SIX1 gene replaced by the deletion cassette. All these three six1 deletion strains turned out to be virulent on E779 (Fig. 5, ‘six1-d1–3’). To confirm that this phenotype was indeed caused by the loss of SIX1, the gene was transformed to one of the knock-out mutants (six1-d2 in Fig. 5). Five transformants were obtained and tested in a bioassay (Fig. 6). Although in this experiment disease development was generally less severe than in the experiment shown in Fig. 5, it was clear that avirulence on E779 was fully restored in four transformants, with the remaining one (‘six1-d + SIX1 ♯3’ in Fig. 6) showing partial restoration. Together, these results clearly demonstrate that SIX1 is required for I-3-mediated resistance.

Figure 5.

SIX1 is required for I-3-mediated resistance to Fusarium wilt. A general susceptible tomato line (C32, light grey bars) and a tomato line with only I-3 resistance against Fol (E779, dark grey bars) were inoculated at 10 days with spore suspensions.
A. Plant weight above the cotyledons at 3 weeks after inoculation. The bars indicate the average weight of 20 plants, with error bars indicating the 95% confidence interval. ‘mock’, mock inoculation (water only); ‘WT’, Fol007; ‘F1-27’, avr3 mutant F1-27; ‘F1-27-V’, F1-27 transformed with an empty vector; ‘F1-27 + SIX1’, F1-27 transformed with the SIX1 gene; ‘six1-d1–3’, SIX1-deletion mutants derived from Fol007.
B. Average disease index (black area) of the same strains on C32 and E779.

Figure 6.

SIX1 restores avirulence on I-3 plants in a SIX1 deletion mutant. A general susceptible tomato line (C32, light grey bars) and a tomato line with only I-3 resistance against Fol (E779, dark grey bars) were inoculated at 10 days with spore suspensions.
A. Plant weight above the cotyledons at 3 weeks after inoculation. The bars indicate the average weight of 20 plants, with error bars indicating the 95% confidence interval. ‘mock’, mock inoculation (water only); ‘WT’, Fol007; ‘six1-d + V’, SIX1-deletion mutant transformed with empty vector; ‘six1-d + SIX1♯1–5’, SIX1-deletion mutant transformed with SIX1 (five independent transformants).
B. Average disease index (black area) of the same strains on C32 and E779.


In this study, we have demonstrated through gene knock-out and complementation studies that the SIX1 gene is required for I-3-mediated resistance of tomato towards F. oxysporum f. sp. lycopersici. In mutant F1-27 that had lost SIX1 along with neighbouring sequences, avirulence on I-3 plants could be restored with a construct carrying only SIX1, but complementation was incomplete. We suspect this results from inadequate transgene expression, but have not verified this. Convincing results were obtained through targeted deletion of SIX1 in the wild-type strain and subsequent complementation of a knock-out mutant. SIX1 is the first avirulence gene identified in the Fusarium species complex and, as far as we are aware, in any root-invading pathogen.

The SIX1 gene potentially encodes a protein of 30 kDa after cleavage of the signal peptide. Six of the eight cysteine residues are present in the central part of the protein that corresponds to the 12 kDa protein that was detected in xylem sap of infected tomato plants (Rep et al., 2002). Presumably, proteolytic cleavage is responsible for removal of the N- and C-terminal parts of the primary translation product. Such maturation has been observed before with several in planta secreted proteins of Cladosporium fulvum and was attributed to both fungal and plant proteases (Van den Ackerveken et al., 1993; Joosten et al., 1994).

Although SIX1 can be readily amplified by reverse transcription (RT)-PCR from infected tomato roots, we have as yet been unable to detect SIX1 expression in vitro by RT-PCR under various conditions including nitrogen starvation, which induces the expression of some other genes for in planta secreted proteins (Talbot et al., 1993; Van den Ackerveken et al., 1994; Stephenson et al., 2000). We plan to investigate further at which infection stage SIX1 expression is induced and what the signals for induction are.

The primary biological function of the Six1 protein remains to be established. Proteins that are secreted by a pathogen in host plants are likely to play a role in host colonization at least under some circumstances or in some hosts. Otherwise, such proteins would only constitute a liability through their potential to trigger defence reactions, and natural selection would consequently have eliminated their production (Laugé and De Wit, 1998). Although SIX1 is clearly not required for pathogenicity under the conditions of our bioassay (immersion of damaged roots of 10-day-old seedlings in spore suspensions), we have preliminary evidence that SIX1 is required for full virulence under less favourable conditions, but further investigation is required to confirm this.

The fact that resistance to Fusarium wilt occurs after entry of the fungus into xylem vessels (see Introduction) and that the product of the SIX1 gene can be found in xylem sap together suggests that the protein may be recognized by a receptor-like protein on the surface of xylem parenchyma cells. This resembles the situation with Cladosporium fulvum, a colonizer of the leaf apoplast. Resistance against Cladosporium depends on secretion by the fungus of small, cysteine-rich proteins and (matching) resistance genes of the Cf class that encode proteins with LRR domains exposed on the outside of the cell envelope (Joosten and De Wit, 1999). Interestingly, two resistance genes against Verticillium, another xylem-colonizing fungus, were cloned recently and also belong to the Cf class (Kawchuk et al., 2001). An alternative scenario is that Six1 is transported into xylem parenchyma cells by a hitherto unknown mechanism and can be recognized by an intracellular mechanism.

Some natural isolates of Fol are virulent on the I-3 line E779, although not as much as mutant F1-27 or the SIX1 deletion mutants (our unpublished results). Sequencing of SIX1 from several virulent and avirulent isolates revealed one single nucleotide polymorphism, but that was not correlated with virulence. For instance, both Fol004 and Fol029 contain the same allele, but only Fol004 is virulent on E779 (Teunissen et al., 2003b). This contrasts with the situation with several other avirulence genes, where the absence of the gene or changes in its sequence were found to correlate with breaking of R gene-mediated resistance (Van den Ackerveken et al., 1992; Joosten et al., 1994; Rohe et al., 1995; Mandel et al., 1997; Orbach et al., 2000; Farman et al., 2002; Luderer et al., 2002). Differences in SIX1 expression may be responsible for differences in virulence on E779. However, the 853 bp upstream of the ATG (which were sufficient for complementation) are identical between Fol004 (virulent) and Fol007 (avirulent). We are currently investigating this issue further with expression analysis and gene knock-out studies.

The absence of SIX1 in several isolates of F. oxysporum outside forma specialis lycopersici is reminiscent of studies on Nectria haematococca (anamorph Fusarium solani), showing that some isolates contain a dispensable chromosome that confers high virulence on pea (Funnell and VanEtten, 2002; Temporini and VanEtten, 2002). On one such chromosome, genes involved in virulence were found to be clustered in a so-called pathogenicity island, a chromosomal region that typically contains a high number of repetitive elements interspersed between genes that are involved in virulence (Han et al., 2001). Some characteristics of the genomic region containing SIX1 resemble those of pathogenicity islands and/or dispensable chromosomes. It contains a high number of transposons (Fig. 4) and the chromosome is relatively small. In addition, deletions can be tolerated without impact on vegetative growth. Examples of this are the deletion of more than 12 kb in mutant F1-27, described in this study, an even larger deletion in another isolate (≈ 250 kb including the SIX1-H pseudogene; results not shown) and the absence of both SIX1 and SIX1-H in non-lycopersici isolates of F. oxysporum. An intriguing possibility is that the chromosomal region containing SIX1, or even the whole chromosome, encodes functions that promote tomato xylem colonization.

Experimental procedures

Plant material, fungal isolated, infection assay

The F. oxysporum isolates used for this study are presented in Table 1. The race nomenclature of F. oxysporum f. sp. lycopersici is as follows (see also Mes et al., 1999): race 1 is avirulent on I/I-1 lines, race 2 is virulent on I/I-1 but avirulent on I-2 lines, and race 3 is virulent on I/I-1 and I-2 lines but avirulent on I-3 lines. Fol004 (race 1) is avirulent on I/I-1 lines and (modestly) virulent on I-2 and I-3 lines, Fol007 and 4287 are virulent on I/I-1 lines and avirulent on I-2 and I-3 lines, and Fol029 is (by definition) avirulent only on I-3 lines.

Table 1. F. oxysporum strains used in this study.
IsolateOriginal designationsOrigin/referencea
  • a

    . Not necessarily the reference to the original description.

Fol004 (race 1) IPO1530 The Netherlands (Mes et al., 1999)
Fol007 (race 2)D2France (Mes et al., 1999)
Fol029 (race 3)5397, BE1Florida, USA (Marlatt et al., 1996)
4287 (race 2) Spain (Di Pietro and Roncero, 1996)
F1-27 (Teunissen et al., 2003a)
F1-27-SIX1-1 This study
Fol007-six1-d1 This study
Fol007-six1-d2 This study
Fol007-six1-d3 This study
Fo47 France (Lemanceau and Alabouvette, 1991)
Fod1 WCS829, F101Italy (Baayen et al., 1997)
Fod2 WCS816 The Netherlands (Baayen et al., 1997)
Fog G2 France (Mes et al., 1994)
Fog G6 the Netherlands (Mes et al., 1994)
Forl C63F Israel (Katan et al., 1991)
Forl C142 Israel (Katan et al., 1991)
Forl C560 Israel (Katan et al., 1991)

The plant lines used in this study are C32, which is susceptible to all races of F. oxysporum f. sp. lycopersici (Kroon and Elgersma, 1993), and E779, which is resistant only to race 3 and contains the I-3 resistance gene introgressed from the Lycopersicon pennellii accession LA716 (Scott and Jones, 1989).

To test virulence of Fol isolates or mutants on tomato lines, the root dip method was used (Wellman, 1939). Briefly, spores were collected from 5-day-old cultures in potato dextrose broth (PDB; Difco) and used for root inoculation of 10-day-old tomato plants at a spore density of 107 ml−1. The seedlings were then potted individually. Three weeks after inoculation, plant weight above the cotyledons was measured, and the extent of browning of vessels in the remaining part of the stem was scored. Disease index was scored on a scale of 0–4 [0, no symptoms; 1, slightly swollen and/or bent hypocotyl; 2, one or two brown vascular bundles in hypocotyl; 3, at least two brown vascular bundles and growth distortion (strong bending of the stem and asymmetric development); 4, all vascular bundles are brown, plant either dead or very small and wilted]. All isolates, mutants and transformants of Fol were tested in at least two independent experiments.

Isolation and sequencing of SIX1 cDNA and genomic locus

Collection of xylem sap and protein mass spectrometry were described before (Rep et al., 2002). Based on one sequence tag obtained from a 12 kDa protein found in xylem sap of tomato infected with Fol007 [ACPEGQECTTFNAYNFR], the following degenerated primers were designed: P12-D1, P12-D2A and P12-D2B (Table 2). Primer P12-D1 is based on the amino acid sequence ‘CPEGQEC’. The last two primers are both based on the amino acid sequence ‘FNAYNFR’; the last two nucleotides are based on the two alternative codon sets for arginine (CGN and AGR respectively). With P12-D1 together with vector primer PACT Reverse (Table 2), an ≈ 620 bp PCR fragment was obtained from a cDNA library of Fol-infected tomato roots. In a nested PCR with the 620 bp fragment as input, a ≈ 600 bp PCR fragment was obtained only with P12-D2A – not with pP12-D2B – in combination with PACT Reverse. Both 620 and 600 bp fragments were cloned into pGEM-T-easy (Promega) and sequenced. The sequence immediately following the P12-D1 primer in the 620 bp fragment encoded the remaining part of the tryptic peptide, thus confirming that a partial cDNA encoding p12 was isolated. Based on this sequence, the reverse primer P12-R1 was designed. With this primer, in combination with a primer on the library vector (PACT Forward), the 5′ part of the cDNA was amplified, cloned into pGEM-T-easy and sequenced. Based on the full cDNA sequence, two additional primers were designed (P12-F1 and P12-F2). With P12-F2 in combination with P12-R1, a 1 kb PCR product was obtained from genomic DNA of Fol007. This product was cloned in pGEM-T-easy and sequenced.

To isolate the genomic locus of SIX1, a Fol BAC library (Teunissen et al., 2003b) was screened with P12-F1 and P12-R1. This resulted in the identification of two clones that contained SIX1 and one (non-overlapping) clone containing SIX1-H, a homologous sequence. From one of the SIX1-containing BAC clones, three fragments were subcloned in pBluescript KS+ (Stratagene). The first is an ≈ 6 kb BamHI–ApaI fragment containing sequences upstream of SIX1 and the 5′ part of SIX1 (up to the internal ApaI site). The second is a ≈ 4 kb ApaI–BamHI fragment containing sequences downstream of SIX1 and the 3′ part of SIX1. The BamHI site of this fragment turned out to be derived from the BAC vector; the BAC insert ends with a HindIII site (Fig. 4). The third is a ≈ 11 kb ApaI fragment that encompasses the 6 kb BamHI–ApaI fragment plus further sequences upstream of SIX1. All fragments were sequenced, together making up about 15 kb of sequence surrounding SIX1.

Transformation of Fol and targeted deletion of SIX1

Fol was transformed with Agrobacterium-mediated transformation, with a protocol adapted from Mullins et al. (2001). Briefly, 105 fungal spores were mixed with the same volume of an Agrobacterium tumefaciens suspension (OD660 = 0.45) in induction medium [IM: 10 mM glucose, 10 mM K2HPO4, 10 mM KH2PO4, 2.5 mM NaCl, 4 mM (NH4)2SO4, 0.7 mM CaCl2, 2 mM MgSO4, 9 µM FeSO4, 0.5% (w/v) glycerol, 910 ml of 40 mM MES, pH 5.3] supplemented with 100 µM or 200 µM acetosyringone, depending on the strain of A. tumefaciens. The mixture was transferred to filters (ME25; Schleicher and Schuell) on co-cultivation plates (composition as IM, but with 5 mM glucose and 1.5% agar). Plates were incubated at 25°C for 2 days. Transformants were selected by transfer of the filters to Czapek Dox agar (CDA; Oxoid) with cefotaxime (100 µg ml−1; Duchefa) and either hygromycin B (100 µg ml−1; Duchefa) or zeocin (100 µg ml−1; InvivoGen). Media containing zeocin were buffered with 0.1 M Tris, pH 8.

For transformation of SIX1 into Fol, a DNA fragment was needed that includes sequences upstream and downstream of the ORF. As we had not obtained such a fragment directly from a BAC clone, we reconstructed it in the following way. First, the complete ORF was amplified by PCR from genomic DNA of Fol007 with P12-F2 and P12-R1 and cloned into pGEM-T-easy. Several clones were sequenced to ensure that no base changes were introduced by PCR. Then, a ≈ 900 bp NcoI–BspEI fragment with upstream flanking sequences was isolated from the 6 kb BAC subclone described above and inserted in front of the ORF (the BspEI site lies 24 bp upstream of the ATG, downstream of primer P12-F2). Finally, a ≈ 700 bp XhoI–SpeI fragment with downstream flanking sequences was isolated from the 4 kb BAC subclone described above and inserted behind the ORF (the XhoI site lies 70 bp downstream of the stop codon, upstream of primer P12-R1). From this construct, a ≈ 2.5 kb HindIII–SpeI fragment (with 853 bp upstream of the start codon and 748 bp downstream of the stop codon) was transferred to the binary vector pRD803, between the HindIII and SpeI sites. This plasmid was called pRD803-SIX1. pRD803 was derived from pBHt2 (Mullins et al., 2001) through insertion of a 3 kb EcoRI–XbaI fragment with the phleomycin resistance gene (ble) from pAN8.1 (Punt and van den Hondel, 1992) and the subsequent insertion of extra unique restriction sites (EcoRI, KpnI, HpaI, BamHI, SpeI, ApaI, ClaI) between the XbaI and HindIII sites. For Agrobacterium-mediated transformation of Fol, pRD803 and pRD803-SIX1 were transformed to A. tumefaciens strain COR309 (UIA143 with pMOG101 and pCH32; Hamilton, 1997).

For deletion of SIX1 in Fol, a deletion cassette was constructed as follows. First, in pGEM-T-easy containing the NcoI–SpeI fragment with SIX1 (see above), most of the SIX1 ORF (from +1 to +522) was deleted and replaced by an XbaI restriction site. This was done by PCR amplification of vector plus SIX1 flanking sequences using primers P12-ATG-R and P12-mid-F (Table 2) followed by XbaI digestion of the linkers and religation. In the resulting XbaI site, a 3.15 kb NheI–XbaI fragment containing the hygromycin B resistance gene (hph) from pAN7.1 (Punt and van den Hondel,, 1992) was cloned with the coding sequence in the opposite orientation with respect to the SIX1 gene. The deletion cassette was then transferred as a HindIII–SpeI fragment to pPZP200-X, resulting in pSIX1d::HYG. pPZP200-X was derived from pPZP200 (Hajdukiewicz et al., 1994) through insertion of extra restriction sites in the polylinker (ApaI, NcoI, AatII, between SalI and XbaI). pSIX1d::HYG was transformed to A. tumefaciens strain EHA105 (Hood et al., 1993) for Agrobacterium-mediated transformation of Fol. Transformants of Fol were tested for deletion of SIX1 by PCR analysis, using primer P12-F19, which anneals upstream of the promoter sequence present in the construct used for transformation, and primer TrpC-F3, which anneals to the TrpC terminator behind the hygromycin resistance gene (Table 2). Conversely, primers P12-F2 and P12-R1 (Table 2) were used to confirm the loss of the wild-type gene. This was only the case in the three transformants (out of 94) that scored positive in the first PCR.

CHEF gel and Southern analysis

Genomic DNA of F. oxysporum was isolated according to the method of Raeder and Broda (1985). The CHEF blots with chromosomes of Fol used in this study were described by Teunissen et al. (2003a). Southern blotting was done according to Sambrook et al. (1989), using 10 µg of genomic DNA for each lane. PCR with primers P12-F2 and P12-R1 was used to generate a 1 kb fragment containing the entire SIX1 ORF. This fragment was radioactively labelled with [α-32P]-dATP using the DecaLabelTM DNA labelling kit from MBI Fermentas. Hybridization was done overnight at 65°C in 0.5 M phosphate buffer, pH 7.2, containing 7% SDS and 1 mM EDTA. Blots were washed at 65°C with 0.5× SSC, 0.1% SDS. The position of sequences hybridizing to the SIX1 sequence was visualized by phosphorimaging (Molecular Dynamics).


We thank Claude Alabouvette and Antonio di Pietro for generously providing strains of F. oxysporum, Hedwich Teunissen for use of her CHEF blots, and Michel Haring for critical reading of the manuscript. The research of Dr M. Rep has been made possible by a fellowship from the Royal Netherlands Academy of Arts and Sciences.