Review

Blackwell Publishing Ltd.

Research review
Wanted: pathogenesis-related marker
molecules for Fusarium oxysporum

Author for correspondence:

Ghislaine Recorbet
Tel: +33 (0)380693348
Fax: +33 (0)380693753
Email: recorbet@epoisses.inra.fr
Received: 27 January 2003
Accepted: 1 April 2003

Ghislaine Recorbet, Christian Steinberg, Chantal Olivain, Vronique Edel,

Sophie Trouvelot, Eliane Dumas-Gaudot, Silvio Gianinazzi and Claude
Alabouvette
Unit Mixte de Recherche 1088, INRA / Universit de Bourgogne: Biochimie, Biologie Cellulaire et
Ecologie des Interactions Plantes/Micro-Organismes, INRA-CMSE, BP 86510, 21065 Dijon Cedex,
France

doi:10.1046/j.1469-8137.2003.00795.x

Summary
Key words: phytopathogenic fungi,
pathogenicity factors, random insertional
mutagenesis, targeted disruption,
biocontrol, impala, Fot1.

Although Fusarium oxysporum pathogens cause severe wilts in about 80 botanical

species, the mechanisms of pathogenicity and symptom induction are poorly understood. Knowledge about the genetic and biochemical pathways involved in the
pathogenesis of F. oxysporum would be invaluable in getting targets for both fungicide development and search for biocontrol agents. In this respect, we described
the main approaches that have been developed to identify some mechanisms underlying the pathogenesis of F. oxysporum. During the last decades, the potential functions triggering of F. oysporum pathogenicity have mainly been investigated by
comparing soilborne pathogenic strains with nonpathogenic ones with regards to
the analysis of the pre- and infection stages and of the resulting plantfungus interactions. The relatively recent progress in the molecular biology of this fungus has
allowed complementary approaches to be developed in order to identify key factors
involved in F. oxysporum pathogenicity. Screening mutants of F. oxysporum for loss
of virulence led to the successful identification of some pathogenesis-related factors,
such as hydrophobicity or attachment of germlings. Taken together, the strategies
described above support the idea that changes in fungal metabolism is also of importance in triggering of F. oxysporum pathogenesis.
New Phytologist (2003) 159: 7392

Introduction
Fusarium oxysporum (Sacc.) Snyder & Hans. is a widespread
soilborne plant pathogen which causes wilt disease in a broad
range of agricultural and ornamental plant species. Pathogenic
strains are each regarded as specialized parasite of a limited
host range and are grouped together into formae speciales
according to the plant species they infect (Armstrong &
Armstrong, 1981). The F. oxysporum species also includes
nonpathogenic strains which have been extensively studied
for their antagonistic activity against various formae speciales

New Phytologist (2003) 159: 73 92 www.newphytologist.com

of F. oxysporum (Mandeel & Baker, 1991; Alabouvette &

Couteaudier, 1992; Alabouvette et al., 2001). Although nonpathogenic isolates of F. oxysporum are efficient colonizers
of plant rhizosphere and roots, they do not induce any
symptoms (Elias et al., 1991; Olivain & Alabouvette, 1999).
Despite an increasing amount of literature during the last
2 yr, little is known about the traits unique to the pathogenic
F. oxysporum which allow infection (Di Pietro et al., 2001a;
Namiki et al., 2001; Inoue et al., 2002; Jain et al., 2002;
Madrid et al., 2003). At the present time, the only valid
method to identify formae speciales or races is through a

73

74 Review

Research review

bioassay in which putative host-plants are confronted to the

fungus. As stressed by Alabouvette et al. (2001), the so-called
nonpathogenic F. oxysporum are strains that failed to provoke
the disease on a limited number of plant species to which
they have been inoculated. This negative definition does
not help plant pathologists willing to use them for biological
control. Numerous studies conducted at the population
level by comparing strains isolated either from roots or soils,
have demonstrated the large diversity existing among strains
of F. oxysporum. Unfortunately, they do not provide any
information on the significance of this diversity in relation
to the physiological, ecological or pathogenic behaviour of
these strains (ODonnell et al., 1998; Baayen et al., 2000;
Alabouvette et al., 2001). The identification of pathogenesisrelated marker molecules for F. oxysporum is thus expected
to be relevant in a broad range of phytopathology-related
features. With the scheduled phasing out of methyl bromide
currently used to control Fusarium wilt diseases, there is an
urgent need for alternative methods (Ma et al., 2001). Both
the search for fungicide molecules and the use of nonpathogenic F. oxysporum strains for the biocontrol of Fusarium wilt
require to understand their interactions with the pathogen,
host plant, and surrounding soil microbial communities.
Knowledge about the genetic and biochemical bases involved
in the pathogenesis of F. oxysporum would be invaluable in
getting clues helping to discriminate between pathogenic and
nonpathogenic strains and may lead to the development of
more effective, long-term, and ecologically safe fungal control
strategies.
Many fungi are economically important plant pathogens.
The need to understand the mechanisms leading to pathogenesis has driven a considerable amount of research on plant
disease for many years. In order to identify fungal pathogenicity
determinants, several molecular approaches have recently
been developed, as reviewed in Oliver & Osbourn (1995),
Hensel & Holden (1996), Hamer & Holden (1996), Brown
& Holden (1998), Dufresne et al. (1998), Gold et al. (2001)
and Lorenz (2002). The targeted strategy concerns the disruption of a gene whose function is postulated to play a role in
pathogenesis on the basis of either biochemical or cytological
data. To investigate its importance, a mutant strain lacking
the function is constructed and its pathogenicity is then compared with that of the corresponding wild-type-strain. Black
box approaches aim to identify factors involved in fungal
pathogenesis without any a priori knowledge of their function. They are based on either the generation of mutants
and the characterization of their defects or the analysis of
genes which are differentially expressed during pathogenesis.
A third strategy, the candidate approach, consists of supposing
that a gene is involved in the infection process by analogy with
other organisms or patho-systems. Taken together, those strategies have led during the past decade to a tremendous literature concerning the successful identification of determinants
involved in the pathogenesis of numerous fungal plant pathogens

as reviewed in Idnurm & Howlett (2001). Pathogenicity

determinants for nonobligate pathogens have been defined
by Oliver & Osbourn (1995) as being necessary for disease
development but not for normal growth on artificial media.
Keeping this point in mind, controversy may thus emerge as
to whether or not in vitro nutrional requirements and morphological defects should be regarded as pathogenicity factors
or as housekeeping factors.
Different attributes may be important for different fungi to
cause disease in view of the great diversity existing among
pathogenic fungi in their lifestyles and the symptoms which
they produce (Oliver & Osbourn, 1995; Deacon, 1996;
Dufresne & Osbourn, 2001). Phytopathogenic fungi can be
divided between those infecting leaves and stems of plants
and those proliferating in root tissues (Agrios, 1997). Infection of these tissues and survival within them require distinct
strategies, but attention has mainly focused on the factors
required by aerial fungal pathogens for successful infection.
Among the 80 pathogenicity functions tabulated by
Idnurm & Howlett (2001), few have been identified as
being implicated in root infection. In this respect, reviewing
the approaches which recently allowed us to identify key processes in fungal pathogenesis in order to give new prospects
in studying the pathogenicity of the soilborne pathogen
F. oxysporum should be considered with regard to the biology
and the time course of infection of this fungus. For this
purpose, it was our aim to first summarize the models of the
hostparasite interactions in Fusarium wilt disease which
provided clues about the potential mechanisms underlying
F. oxysporum pathogenesis and to further integrate the methods
which recently identified functions involved in fungal pathogenicity. With regard to the view that mutation may convert
a fungal plant pathogen to a nonpathogenic mutualist
(Freeman & Rodriguez, 1993b; Redman et al., 1999;
Freeman et al., 2002), the potential for using the F. oxysporum
mutants described in this review in biocontrol strategies is
finally discussed.

Biological models for potential targets

Besides the identification of F. oxysporum pathogenicity
determinants, considerable attention has been paid to the
nature and genetic basis for resistance and tolerance of plants
to fungal wilt diseases. In this respect, the model of the host
parasite interactions in wilt disease has been updated by
Beckman & Roberts (1995) who depicted the time course
of the major events during pathogenesis by F. oxysporum.
The life-cycle of F. oxysporum includes: the saprophytic
phase when the fungus survives in soil by producing resting
structures; the germination and growth by nutrients that are
released from the extending roots of a variety of plants; the
penetration into the epidermal and cortical cells of host or
nonhost plants; and the parasitic growth into host plant after
entry into the xylem elements.

www.newphytologist.com New Phytologist (2003) 159: 73 92

Research review

The preinfection stages

The potential existence of host-specific recognition events
during the preinfection stage for F. oxysporum has mainly
been addressed by comparing soilborne pathogenic and
nonpathogenic strains in many ways. The preinfection stages,
as defined by Deacon (1996), include all the events leading to
attempted penetration such as triggering of germination of
spores, orientation of pathogen growth to a suitable infection site, and adhesion to the host surface. Edel et al. (1997)
showed that some plant species cultivated in the same soil
influenced the structure of F. oxysporum populations. Different components of the population became dominant
depending on the plant species, suggesting a particular
relationship between the plant and F. oxysporum strains and
rising the question as to whether a host plant could modify
the behaviour of either pathogenic or nonpathogenic strains.
By studying the mycelial development (length and branching)
of five strains of F. oxysporum in the vicinity of the roots of a
host and nonhost plant, Steinberg et al. (1999a) did not
detect a chemotropic response towards or away from the
root but showed that fungal development of all five strains
was stimulated in the presence of roots, irrespective of plant
species. One can expect therefore that it is possible to discriminate between pathogenic and nonpathogenic strains
of F. oxysporum on more specific host or nonhost root components. This point was investigated by testing the effect of
soluble root exudates and insoluble root materials on mycelial
development. All fractions were found to be able to stimulate
branching and hyphal extension of all the strains, and none of
the strains was consistently more stimulated than the others
(Steinberg et al., 1999a). Whether fractions of root could
account for differences in pathogenicity remains to be further
investigated with regard to the search for specific host-plant
components, if any. The nonpathogenic strain Fo47 was
further used to test the effect of glucose and nitrogen (casamino
acids) supplies on mycelial development. Whereas increasing
the concentration of glucose from 0.1 to 10 g l1 had no
effect, the hyphal length and number of branches increased
with concentrations of organic nitrogen equal or higher than
0.125 g l1 (Steinberg et al., 1999a). Although this test has
been performed in vitro, the organic nitrogen present in root
exudates is likely to play a role in the triggering of the mycelial
development of F. oxysporum. In this respect, studying differences among strains in their ability to metabolize these
compounds could be relevant to their ability to colonize roots.
In addition, qualitative or quantitative changes in the organic
nitrogen content of root exudates are likely to occur between
plant species and may account for differences in the ability of
a strain to colonize different hosts. As part of an investigation
into the factors influencing the colonization of the roots of the
host plant by pathogenic and nonpathogenic strains of F.
oxysporum, the effect of carbon sources on the growth habits
of five strains has also been studied. Because utilisation of

New Phytologist (2003) 159: 73 92 www.newphytologist.com

Review

polysaccharides as carbon sources can reduce the growth rate

of fungi as depolymerisation can be energetically expensive
(Anderson & Salomons, 1984), it has been proposed that
pathogenic and nonpathogenic strains of F. oxysporum could
be discriminated between when examining polysaccharides
representative of root cell wall components. The radial extention
rates on solid media of nonpathogenic and pathogenic strains
on tomato were thus measured, using solubles saccharides
(glucose, xylose and D galacturonic acid), polysaccharides
(carboxymethyl cellulose, xylan and pectin) and soluble and
insoluble extracts of tomato roots, as carbon sources. The
growth parameters were shown to be unique features of each
F. oxysporum strain, whatever the carbon source (Steinberg
et al., 1999b). In vitro growth habits on different carbon
sources do not allow discrimination between pathogenic
and nonpathogenic strains of F. oxysporum, but no conclusion
can be drawn from this study as to whether or not carbon
utilization by F. oxysporum is a trait reflecting pathogenicity.
To compare the overall patterns of external root colonization between a pathogenic and a nonpathogenic strain of
F. oxysporum, microscopic analyses were carried out in situ on
tomato roots (Olivain & Alabouvette, 1997, 1999). Both a
pathogenic strain of F. oxysporum f.sp. lycopersici (Fol32) and
a nonpathogenic soilborne isolate of F. oxysporum (Fo5a4)
were transformed with the -D-glucuronidase (GUS ) gene to
enable their histochemical localization on tomato roots in the
presence of 5-bromo-4-chloro-3-indolyl -D-glucuronide
(Couteaudier et al., 1993). In addition, because the GUS gene
was under the control of the promoter of glyceraldehyde-3phosphate dehydrogenase, an enzyme catalyzing the second
step of glycolysis, the intensity of the blue staining resulting
from the GUS activity gives an indication of the metabolic
activity (Eparvier & Alabouvette, 1994). Both pathogenic
and nonpathogenic strains were found to be able to colonize
the root surface in the same way. Observations of the whole
root showed that colonization occurs very quickly, as
abundant hyphae were observed at the root surface 24 h after
inoculation. No preferential zone for root colonization was
observed, although after 24 h, colonization was mainly
observed in the zone with root hairs; after 48 h, hyphae were
observed everywhere on the root surface, even at the tip. Most
of these hyphae did not adhere closely to the root surface. This
pattern of colonization was suggested to be related to the
experimental design: the young tomato radicule has been
immersed in a conidial suspension before being allowed to
grow in a nutrient solution without the fungus. Therefore, all
hyphae are likely to result from the germination of the
microconidia that were able to bind to the root surface within
this hour. Indeed, the occurrence of a site-specific binding
process between F. oxysporum conidia and the surface of
tomato roots has been demonstrated and no difference was
observed in terms of binding affinity among several of
F. oxysporum differing either in pathogenicity or in host range
(Recorbet & Alabouvette, 1997). The GUS activity was

75

76 Review

Research review

found to decrease quickly and 4 d after inoculation, bluestained hyphae outside the roots were only observed in
discontinuous spots, especially at the apex and at the emergence of secondary root, where exudation and slough are
important, supporting the activity of the fungus. Most of the
hyphae did not show any GUS activity at the root surface,
suggesting that the fungus may enter a resting stage resulting
from nutrient starvation (Olivain & Alabouvette, 1997,
1999).
The penetration step
Studying the infection process of F. oxysporum in cotton root
tips, Rodriguez-Glvez & Mendgen (1995) showed that once
having fully colonized the root surface, surface hyphae
produced branches which immediately penetrated the root
inter- or intracellularly. These branches of penetrating hyphae
were characterized by a reduced diameter compared to the
mycelium forming hyphae. This corresponds to the observations reported by Bishop & Cooper (1983) who showed that
the penetration process of F. oxysporum started with the
development of specialized penetration hyphae. Based on
cytological data, the penetration process was found to be
similar for both pathogenic and nonpathogenic strains
(Olivain & Alabouvette, 1997, 1999). Hyphae penetrating
into the root were observed as early as 24 h. Preferential sites
of infection were not detected and pathogenic and nonpathogenic strains were able to penetrate into both the well
differentiated and apical zones. The number of penetrating
hyphae appeared to be very limited in comparison with the
intensity of the colonization of the root surface. In view of the
previous observations, it can be suggested that once having
fully colonized the root surface, the fungi have to cope with
starvation and that hyphal branching within the root provides
a mean of exploring new regions for sources of nutrients
(Cooke & Rayner, 1984; Prosser, 1995; Steinberg et al.,
1999a). Cell wall degradation may be important to fungi
not only for penetration and ramification inside the plant
tissue but also for releasing, from the wall polysaccharides,
nutrients necessary for growth. Many studies, as reviewed
in de Lorenzo et al. (1997) and ten Have et al. (2002), have
adressed the role of plant cell wall-degrading enzymes
(CDWEs) in fungal pathogenesis. The action of pectic
enzymes and, in particular, of endopolygalacturonases on
cell walls appears to be a preriquisite for wall degradation
by other CWDEs such as hemicellulases and cellulases (Karr
& Albersheim, 1970). It has been proposed that polygalacturonases (PGs) play a key role in fungal pathogenicity
by depolymerising homogalacturonan, a major component
of the plant cell wall. Shieh et al. (1997) showed that an
endopolygalacturonase from Aspergillus flavus is involved in
the infection of cotton balls. ten Have et al. (1998) reported
that the deletion of one out of six endopolygalacturonase
genes of the necrotroph Botrytis cinerea leads to a reduced

virulence of the fungus in tomato and apple. Recently,

polygalacturonase was found to be a pathogenicity factor
in the Claviceps purpurearye interaction (Oeser et al., 2002).
Di Pietro et al. (2001b) studied the secretion of the endopolygalacturonase PG1 in different formae speciales of F.
oxysporum. Two PG1-deficient isolates obtained in vitro were
identified in F. oxysporum f. sp. melonis. The results indicated
that at least in this forma specialis, PG1 was not essential for
pathogenicity because a strong PG1-producer was as virulent
as the deficient one. Garcia-Maceira et al. (2001) showed
that the tomato vascular wilt pathogen F. oxysporum f. sp.
lycoperci produced an array of pectinolytic enzymes susceptible to contribute to penetration and colonization of
the host plant. They described PG5, a novel extracellular
endoPG. As shown by reverse transcription polymerase chain
reaction (RT-PCR), PG5 is expressed in tomato roots during
the initial stages of infection, but the targeted inactivation of
pg5 has no detectable effect on virulence towards tomato
plants. Because F. oxysporum produces PGs other than PG1
and PG5 in vitro, the possibility that PG plays a role in
pathogenicity can not be ruled out. In addition, in planta
expression of CWDE may differ greatly from in vitro
expression (Fraissinet-Tachet et al., 1995). Thus, whether
polygalacturonase activities are important for the colonization
of plant tissues by F. oxysporum remains an open question,
because of the possibility that other enzymes (some of which
may be produced by the fungus only in planta) may
compensate for the loss of the products of the disrupted genes
(de Lorenzo et al., 1997). Available evidence indicates that
fungal endopolygalacturonases are also responsible for the
induction of plant defense responses by releasing oligogalacturonides from the plant cell wall (Anderson, 1989).
Oligogalacturonides of chain length varying between 10
and 15, which are transiently produced by the action of
the endopolygalacturonases on homopolygalacturonan, are
elicitors of plant defense (Hahn et al., 1989; Darvill et al.,
1992). Concerning F. oxysporum, it has been suggested that
oligogalacturonides are released in planta at strategic sites
during the infection by the pathogen (Benhamou et al.,
1990). Inhibitors of fungal enzymes that degrade plant cell
walls have been proposed to be part of the plant defenses that
limit the development of disease symptoms. Polygalacturonase
inhibitor proteins (PGIPs) are glycoproteins located in plant
cell walls that specificically inhibit fungal PGs (Cook et al.,
1999). It has been shown that PGIP prepared from tomato
cell walls in vitro protected these walls from degradation by a
complex CWDE mixture produced in culture by F. oxysporum
f. sp. lycopersici (Jones et al., 1972). In addition to endoPGs,
exo--1,4-polygalacturonases (exoPGs) may have an important function in pathogenplant interactions as they degrade
elicitor-active oligogalacturonides released by endoPGs
(Favaron et al., 1988) and are generally not inhibited by plant
polygalacturonase-inhibiting proteins (Cervone et al., 1990).
A gene encoding an exoPG of F. oxysporum f. sp. lycopersici,

www.newphytologist.com New Phytologist (2003) 159: 73 92

Research review

pgx4, has been isolated by Garcia-Maceira et al. (2000).

Despite being present in a single copy, the targeted
inactivation of pgx4 has no detectable effect on virulence
towards tomato plants. As stated above for PG1, the most
likely explanation for the unaltered pathogenicity of the
F. oxysporum mutants is that the lack of PG1 activity is
complemented by other pectinolytic enzymes. As part of an
investigation into the CWDEs secreted by plant pathogenic
fungi upon contact with the host tissue, the gene pl1 encoding
an endo-pectate lyase was shown to be expressed during plant
infection (Huertas-Gonzlez et al., 1999), and the enzyme
activity band corresponding to the gene product was detected
in extracts of tomato roots infected with F. oxysporum
(Di Pietro & Roncero, 1996). However, pl1 was proven not
to be essential because its targeted disruption does not affect
invasive growth of F. oxysporum (Di Pietro et al., unpublished).
Gmez-Gmez et al. (2001, 2002) further investigated the
role of fungal xylanases in the pathogenesis of F. oxysporum
f. sp. lycopersici. The genes xyl3, xyl4 and xyl5 were characterized
and found to be expressed in tomato roots during some stages
of infection but their targeted inactivation had no effect on
virulence.
Although made up mainly of polysaccharides, plant cell
walls contain different classes of structural proteins and
enzymes. Degradation of these proteins by extracellular
fungal proteases may be critical for pathogens to successfully
colonize plants host. Skovgaard & Rosendahl (1998) have
compared isolates of F. oxysporum from different habitats with
respect to intra- and extracellular isozyme banding patterns. A
protease was found to be present in F. oxysporum isolated from
plants with root rots only, suggesting it may be a pathogenicityrelated enzyme. Di Pietro et al. (2001c) isolated from the
tomato vascular wilt fungus F. oxysporum f. sp. lycopersici
the gene prt1 for which the predicted amino acid sequence
showed significant homology with subtilisin-like fungal proteinases. Prt1 was found to be a single-copy gene with a structure which was highly conserved among different formae
speciales of F. oxysporum. The targeted inactivation of ptr1 in
F. oxysporum f.sp. lycopersici had no detectable effect on pathogenicity on tomato plants. It was suggested that Prt1 did not
contribute significantly to total extracellular protease activity
in F. oxysporum, probably as a result of low level of gene expression. As it is the case in carbohydrate-degrading enzymes
(Di Pietro & Roncero, 1998; Garcia-Maceira et al., 2000),
functional redundancy of proteases in F. oxysporum poses a
major challenge in understanding the role of these enzymes
in pathogenesis.
An alternative approach to the isolation and disruption of
individual genes encoding cell wall-degrading enzymes aims
at identifying the genetic regulatory elements for which mutation results in the simultaneous loss or downregulation of
multiple enzymes. The production of wall depolymerases
by plant pathogenic fungi is under catabolite repression.
In Saccharomyces cerevisiae, the SNF1 gene is required for

New Phytologist (2003) 159: 73 92 www.newphytologist.com

Review

expression of catabolite-repressed genes when glucose is limiting.

An ortholog of SNF1, ccSNF1, was isolated from the maize
pathogen Cochliobolus carbonum (Tonukari et al., 2000). Targeted
disruption of ccSNF1 resulted in the decrease in the transcription of 10 cell wall-degrading enzymes and the corresponding
mutants were reduced in pathogenicity. It was suggested
that the biochemical processes controlled by ccSNF1 probably
included the ability to degrade polymers of the plant cell wall
and to take up and metabolize the sugars produced. A second
regulatory gene involved in glucose repression has been
implicated as a pathogenicity factor in Magnaporthe grisea
(Sweigard et al., 1998). Whether such regulatory genes are
involved in F. oxysporum pathogenicity remains unknown.
Overcoming the plant defense responses
Plants possess a broad spectrum of basic defense mechanisms,
pre-established or induced, which render them resistant to
most potential colonizers. The ability of F. oxysporum strains
to tolerate or delay host defenses has been proposed as a trait
that may reflect pathogenicity on a specific host. Olivain
& Alabouvette (1997, 1999) demonstrated, on the basis of
cytological observations, that whereas the colonization of the
nonpathogenic GUS-marked strain was restricted to the
superficial layers of cells by the defense reactions of the plant,
the GUS-marked pathogen grew rapidly, centripetally
towards the stele, which appeared intensely blue stained 7 d
after inoculation. The ability of the pathogenic F. oxysporum
to infect its host is thus likely to depend on its capacity to
obtain adequate nutrients while avoiding or negating host
defenses. It has been suggested that the hypodermis might be
an important barrier to colonization of the inner cortex. In
the case of the nonpathogenic strain (Fo5a4), the defense
reactions risen in the hypodermis were efficient in limiting the
colonization of the root. On the contrary, in the case of the
pathogen, F. oxysporum f. sp. lycopersici strain Fol32, this
barrier seemed ineffective, because it was discontinuous and
formed too late. From these observations, it was concluded
that the decisive phenomenon for the success or failure of a
vascular infection is the kinetics at which defense reactions
create barriers that are efficient or not in preventing the fungus
from reaching the stele (Olivain & Alabouvette, 1997, 1999).
Apart from cytological observations, few studies have
adressed the expression of plant defenses in response to both
pathogenic and nonpathogenic F. oxysporum. Among them,
special attention has been paid to the secondary metabolites
produced by plants, many of which have antifungal activity.
Saponins, which are plant glycosides, have been implicated as
preformed chemical barriers against pathogen attacks (Osbourn,
1996). The saponin avenacin A-1 from oat roots is degraded
by avenacinase from the take-all fungus, Gaeumannomyces
graminis var. avenae. Targeted disruption of the corresponding
gene prevented mutants from attacking oats while maintaining
full pathogenicity on wheat, a host that does not produce

77

78 Review

Research review

avenacin A-1 (Bowyer et al., 1995). -tomatine, a phytoanticipin from tomato, is likely to have been the most studied
saponin with regards to F. oxysporum. The fungitoxic effect
of -tomatine has been attribuated to its interaction with
3-hydroxyl sterols, which results in an increase in fungal cell
permeability (Safe et al., 1997). Because the amount of
-tomatine found in roots and in stems is sufficient to inhibit
the growth of F. oxysporum (Sandrock & vanEtten, 1998) and
to induce its tomatinase gene which catalyses the hydrolysis of -tomatine into less fungitoxic molecules (Lairini &
Ruiz-Rubio, 1997), -tomatine resistance may therefore be
a prerequisite for successful infection. A potential relationship
between tomatine tolerance and pathogenicity on tomato has
been investigated among 17 pathogenic or nonpathogenic
isolates of F. oxysporum by Suleman et al. (1996). Correlation
of virulence with tomatine tolerance was only found under
limited experimental conditions. In addition, whereas tomatinase activities have been thought to be exclusively associated
with fungi which are pathogenic on tomato, the occurrence of
-tomatine detoxifying enzymes has been demonstrated in
other formae speciales nonpathogenic on tomato such as
tuberosi, melonis, niveum and gladioli (Lairini et al., 1997).
However, it is not known from this study if those formae speciales are able to colonize tomato roots without inducing any
symptoms or if their tomatinase are expressed in situ. Because
whether a plant is susceptible or immune to systemic invasion
was found to be unrelated to the ability of the fungus to
colonize the root cortex (Beckman & Roberts, 1995; Olivain
& Alabouvette, 1997, 1999), a potential role for tomatinase
in pathogenicity can not be ruled out. In this respect, the cloning and the characterization of the cDNA and genomic DNA
encoding tomatinase from the vascular pathogen of tomato
F. oxysporum f. sp. lycopersici have been reported by RoldnArjona et al. (1999). The tomatinase from F. oxysporum is
encoded by a single gene (FoTom1) with an expression which
is induced by -tomatine and fully repressed in the presence
of glucose. RT-PCR experiments in planta indicated that the
FoTom1 gene is expressed throughout the entire disease cycle
of F. oxysporum f. sp. lycopersici suggesting that this enzyme
may be important for pathogenicity. In addition, the high
sequence homology between tomatinase and glycosyl hydrolases, which are enzymes thought to be required for nutritional purposes, and the fact that tomatinase gene expression
is likely to be sujected to carbon catabolite repression
(Roldn-Arjona et al., 1999) suggest a nutritional function for
tomatinase. The in situ expression of FoTom1 throughout the
entire disease cycle raises the possibility that F. oxysporum f. sp.
lycopersici may face with carbon limitation in planta. To
understand the role of tomatinase in pathogenicity, Ito et al.
(2002) attempted the transformation-mediated disruption of
the tomatinase gene in F. oxysporum f. sp. lycopersici. Mutants
obtained showed increase in sensitivity to -tomatine and
decrease in tomatinase activity in vitro. Inoculation assay
showed reduced pathogenicity of the mutants on tomato

plants, suggesting an important role of tomatinase in pathogenicity of the fungus. However, the resident tomatinase gene
of the mutants was not disrupted but the gene replacement
vector was ectopically integrated into the genome of the
mutants. Northern blot analysis revealed low levels of the
tomatinase gene expression and also the presence of antisense
RNA in the ectopic mutant, suggesting that post-transcriptional
gene silencing might be involved in the suppression of
tomatinase. Besides the detoxification of constitutive phytoanticipins, one trait which has also been investigated as to
potentially reflect pathogenicity on a specific host is the ability
of fungi to tolerate phytoalexins, the so-called antimicrobial
compounds synthesized de novo by plants in response to
microbial infection (VanEtten et al., 1994). Detoxification of
pisatin and maackiain, the phytoalexins from pea and chickpea, respectively, has been implicated in the pathogenicity of
Nectria haematococca on the corresponding host (Wasmann &
VanEtten, 1995; Enkerli et al., 1998). Concerning F. oxysporum, much attention has been paid to rishitin, a phytoalexin
from tomato, whose level produced in tomato roots and
stems was found to approach or exceed inhibitory level for
F. oxysporum f. sp. lycopersici (Fond et al., 1977). A potential
relationship between rishitin tolerance and pathogenicity on
tomato has been investigated in vitro among 17 pathogenic or
nonpathogenic isolates of F. oxysporum by Suleman et al.
(1996). When pathogenic isolates were compared to nonpathogenic ones on tomato, the nonpathogens did not
display the highest sensitivity to rishitin. It was concluded
from this study that although the results did not prove a role
for rishitin tolerance in pathogenicity, the sufficient natural
variation found to exist in this trait may contribute to an isolates disease potential on tomato. Among the antimicrobial
metabolites produced by plants in order to prevent fungal
penetration, the plant-encoded enzymes chitinases and -1,
3-glucanases have been investigated as potential pathogenesisrelated molecules for F. oxysporum. One important defenserelated feature of those hydrolases is their ability to degrade
chitin and -1,3-glucan, the major components of most fungal cell walls (Wessels & Sietsma, 1981). The degree to which
these enzymes are lethal or directly injurious to pathogens is
still unclear, but they could be effective in cleaving from fungal wall oligosaccharides which could in turn elicit defense
responses (Krebs & Grumet, 1993). The expression of plant
hydrolases in response to both pathogenic and nonpathogenic
strains of F. oxysporum has been addressed on celery roots.
Inoculation of celery roots with incompatible or compatible
races of the pathogen F. oxysporum f. sp. apii resulted in similar
increases in chitinase and -1,3-glucanase activities, but the
nonpathogen F. oxysporum f. sp. cepae did not stimulate either
enzyme (Krebs & Grumet, 1993). It could not be concluded
from this study whether this may reflect a difference in hostpathogen recognition at the formae specialis level or a failure of
the nonpathogen to adequately germinate on the time-course
study using inoculated soil. In cases where induction of

www.newphytologist.com New Phytologist (2003) 159: 73 92

Research review

hydrolytic enzyme activity by nonpathogenic strains of F.

oxysporum has been observed, pruned tissues were used or root
colonization by the nonpathogen was not quantified (Matta
et al., 1988; Tamietti et al., 1993; Fuchs et al., 1997). The
time-course inoculation patterns of -1,3-glucanase and chitinase isoforms in susceptible tomato roots inoculated with
either pathogenic or nonpathogenic strains of F. oxysporum
were investigated with regard to the fungal dynamics of root
colonization. As detected by anodic polyacrylamide gel
electrophoresis for acidic proteins 11 d after inoculation, the
strains of F. oxysporum f.sp. lycopersici were found to induce at
least one additional -1,3-glucanase (Recorbet et al., 1998).
This isoform may serve as a biochemical marker of pathogenicity in fusarium wilt of tomato but its occurrence was
detected too late during the infection stage for discrimination
purposes. Taken together, those results have prompted the
search for the detection of earlier plant defense reactions with
potential to discriminate between pathogenic and nonpathogenic strains of F. oxysporum.
Active oxygen (AO) production is considered as a preliminary step in the activation of coordinated enzyme systems that
build defense barriers. De Donato et al. (1997) have demonstrated that AO was induced in melon cells by F. oxysporum
f. sp. melonis elicitors only when a Fusarium resistant melon
genotype reacted with a hyphal wall preparation from an
avirulent race of the pathogen. Using cell suspensions of a
susceptible flax variety, Olivain et al. (2001a) have measured the
H2O2 production in the presence of germinated microconidia
of either the pathogenic strain F. oxysporum f. sp. lini (Foln3)
or the nonpathogenic strain Fo47. The nonpathogenic strain
induced a biphasic production of H2O2 whereas interaction
with the pathogenic strain only induced the first oxydative
burst. No difference in AO production could be recorded
when heat-treated or conidia without germ tubes were used,
suggesting the existence of an eliciting compound which is
produced by hyphae or germlings. The nature of the eliciting
compound and the elucidation of the resulting signalling
pathway remain to be investigated.
Putting all together
Although the approaches described above, including the
analysis of the pre- and infection steps by F. oxyporum and
of the resulting plantfungus interactions, do not provide
direct evidence of the involvement of a pathogenicity-related
function or molecule, they do allow, as listed in Table 1,
speculations to be made about the processes which are likely
to be required for a successful infection.

Black box approaches and other patho-systems

to explore F. oxysporum pathogenesis
Among the complementary approaches described for the
identification of functions with potential to discriminate

New Phytologist (2003) 159: 73 92 www.newphytologist.com

Review

between pathogenic and nonpathogenic F. oxysporum, much

attention has been paid to the generation of pathogenicity
mutants by either random or targeted insertional mutagenesis
with subsequent characterization of the induced mutations.
Random mutagenesis
Generation of such mutants has first been attempted by using
chemical or UV treatments. Sanchez et al. (1975) reported
the generation by chemical mutagenesis of single and multiple
auxotrophic mutants of the F. oxysporum f. sp. lycopersici
strains which cause Fusarium crown-rot and Fusarium wilt of
tomato. Changes in pathogenicity were observed when the
auxotrophs were compared to the wild-type strain but no
correlation was observed between changes in pathogenicity
and the particular nutritional requirement. More recently,
Akiyama et al. (2000) isolated nonpathogenic mutants of
F. oxysporum f. sp. cucumerinum following mutagenesis with
5-azacytidine. Analysis of the protein expression pattern
revealed a protein that is present in yeast-form cells of the
mutants but absent in those of the wild-type strain. This
protein was found to be identical to a region of a polypeptide
encoded by a cDNA clone, sti35, a heat shock gene in F.
oxysporum. The targeted disruption of sti35 in F. oxysporum
f. sp. cucumerinum was reported by Thanonkeo et al.
(2000). Disruption of sti35 had no effect on normal growth
and development in nutrient medium. However, all disruptants retained pathogenicity to cucumber plants, suggesting
that sti35 is dispensable for fungal pathogenicity to host
plants, although it is induced by the phytoalexin produced by
the infected plants. Using UV mutagenesis, Freeman et al.
(2002) isolated a nonpathogenic mutant of F. oxysporum f. sp.
melonis by a continuous dip-inoculation method. In crossprotection experiments with muskmelon, significant disease
reduction was achieved when comparing mutant-colonized
wild-type challenged seedlings with those of wild-typechallenged alone. Preliminary work initiated regarding the
mechanisms involved in the nonpathogenic traits of the
mutant showed that there is a reduction in cellulolytic activity
of the mutant compared with that of the wild-type strain.
However, regarding chemical and UV mutagenesis, it has
been shown that strains selected in such a way often present
low stability and that their analysis can be complicated by the
presence of multiple site-mutation events (Bouhot, 1981).
The relatively recent development in the molecular biology
of F. oxysporum offers new perspectives for cloning genes
involved in different steps of the pathogenic process with
random insertional mutagenesis approaches. These methods
consist in the generation of mutants after insertion of a DNA
fragment into a gene of interest. The mutated copy of the gene
can be isolated by the inserted DNA as a tag. The advantage
of this technique is the possibility to clone a gene without
prior knowlegde of the gene product. The so-called black
box approach proved valuable in understanding bacterial

79

80 Review

Event

Observation

Potential mechanism

Reference(s)

Germination of conidia

Hyphal length and number of branches

increase with addition of casamino acids.
Mycelium fully colonizes the root surface.
Hyphae do not show GUS activity at the
root surface.
Surface hyphae produce specialized
penetration hyphae.
The targeted disruptions of:
endopolygalacturonases
exo-polygalacturonase
endo-pectate lyase
xylanases
protease
have no effect on pathogenicity.
A nonpathogenic F. oxysporum induces a
second oxydative burst as compared to a
pathogenic one.
Degradation of phytoanticipin(s):
correlation of virulence with tomatine
tolerance under some experimental conditions
mutants with low level of tomatinase gene
expression show reduced pathogenicity on
tomato plants
the FoTom1 gene, encoding tomatinase,
induced by tomatine and repressed by glucose,
is expressed in planta.
Degradation of phytoalexin(s):
existence of a sufficient natural variation
in rishitin tolerance between soilborne isolates.

Potential role of organic nitrogen in root

exudates in triggering of mycelial development.
Potential resting stage at the root surface.

Steinberg et al. (1999a)

Hyphal branching within the root may provide

a mean of exploring new sources of nutrients.
Studying the role of cell wall degrading enzymes
in pathogenicity is likely to be impaired by their
functionnal redudancy either in F. oxysporum
or in planta.

Rodriguez & Mendgen (1995)

Potential production by germlings of nonpathogenic

F. oxysporum of eliciting compound(s) leading to the
activation of the building of defense barriers.
Potential role of phytoanticipin-detoxifying
enzymes in pathogenicity.

Olivain et al. (2001a)

Root colonization

Penetration of the root

Degradation of cell
wall components

www.newphytologist.com New Phytologist (2003) 159: 73 92

Overcoming the plant

defense responses

Olivain & Alabouvette (1997, 999)

Garcia-Maceira et al. (2001); Di Pietro et al. (2001b)

Garcia-Maceira et al. (2000)
Huertas-Gonzlez et al. (1999)
Gmez-Gmez et al. (2001, 2002)
Di Pietro et al. (2001c)

Suleman et al. (1996)

Ito et al. (2002)

Potential carbon limitation in planta.

Roldn-Arjona et al. (1999)

Potential role of phytoalexin-detoxifyng

enzymes in pathogenicity.

Suleman et al. (1996)

Research review

Table 1 Speculations about the processes which are likely to be required for a successful infection by Fusarium oxysporum on the basis of cytological and biochemical data as listed in part II

Research review

pathogenesis (Hensel & Holden, 1996). In filamentous fungi,

transformation with DNA that does not exhibit homology
with the fungal genome usually results in heterologous integration of transforming DNA into the genome, which makes
it possible to use the transforming DNA to disrupt genes potentially involved in pathogenicity. Using such an approach,
Madrid et al. (2003) managed to isolate from a virulent strain
of F. oxysporum f. sp. lycopersici the transformant Tr5 that was
unable to cause disease symptoms on tomato plants. The
molecular analysis of the insertional mutant identified the
chitin synthase-encoding gene, chsV, as an essential pathogenicity determinant. F. oxysporum strains lacking this chitin
synthase are viable but dramatically reduced in virulence. The
chsV insertional mutant and a gene replacement mutant of F.
oxysporum display morphological abnomalies such as hyphal
swellings that can be partially restored by osmotic stabilizer.
The mutants are unable to infect and colonize tomato plants
or to grow invasively on tomato fruit tissue. They are also
hypersensitive to plant antimicrobial defense compounds such
as the tomato phytoanticipin -tomatine or H2O2. Reintroduction of a functional chsV copy into the mutant restores the
growth phenotype of the wild-type strain. It was proposed
that a natural role of ChsV in Fusariumtomato interaction
is to ensure survival of the pathogen within the host by protecting it from host toxic compounds.
To overcome complex plasmid integration events, an insertional mutagenesis technique, termed restriction enzymemediated integration (REMI) has been developed and was
successfully used for the transformation of diverse filamentous
fungi and for the isolation of genes important for fungal plant
pathogenicity (Kahmann & Basse, 1999). In REMI, the
fungus is transformed with plasmid DNA in the presence of
a restriction endonuclease that generates compatible ends,
resulting in plasmid integration at the corresponding restriction site in the genome. Other advantages of REMI often
include increased efficiency of transformation, single copy
insertion and random integration events. Namiki et al. (2001)
and Inoue et al. (2002) used REMI mutagenesis to tag genes
required for pathogenicity in F. oxysporum f. sp. melonis. Of
the 2929 REMI transformants tested, 43 showed reduced
pathogenicity on susceptible melon cultivars. One of the
mutants (FMMP95-1) was found to be an arginine auxotroph
(Namiki et al., 2001). Structural analysis confirmed that
arginine mutation in FMMP95-1 resulted from the integration
of the BamHI-linearized pSH75 plasmid in an open reading
frame (ORF), named ARG1, encoding an arginosuccinate
lyase gene involved in the last step for arginine biosynthesis.
Genetic and chemical complementation of FMMP95-1
demonstrated that ARG1 gene mutation was the direct cause
of reduced pathogenicity in FMMP95-1. The ARG1 gene
expression was found to be suppressed in a nutrition-rich
condition in vitro. Because it has been shown that nutrient
limitation may induce the expression of genes involved in
pathogenicity (Talbot et al., 1997; Segers et al., 2001), it can

New Phytologist (2003) 159: 73 92 www.newphytologist.com

Review

be suggested that F. oxysporum may be faced with starvation

in situ. However, evidence is missing from this study as to
whether or not the auxotroph mutant is able to grow in planta
without inducing wilt symptoms. The characterization of
B60, another pathogenicity REMI mutant of F. oxysporum
f. sp. melonis, was reported by Inoue et al. (2002). Molecular
analysis of B60 identified the affected gene, designated FOW1,
which encodes a protein with strong similarity to mitochondrial carrier proteins (MCPs) of yeast. MCPs are small transport
proteins of the mitochondrial inner membrane that catalyze
the transport of metabolites across the inner membrane with
a high degree of substrate specificity (Belenkiy et al., 2000).
The FOW1 gene was found to be conserved in other formae
speciales of F. oxysporum, causing wilt of different plants and
FOW1 targeting in F. oxysporum f. sp. lycopersici also resulting
in the reduction of virulence to tomato plants. Despite being
expressed during growth in culture, FOW1 was dispensable
for the saprophytic growth and conidiation of F. oxysporum.
The infection behavior of wild-type and fow1 mutant strains
was observed using transformants that constitutively express
green fluorescent protein (GFP). Fow1 was not found essential for hyphal growth on the root surface and penetration of
the epidermal cells. Fluorescence microscopy observation
performed 10 d after plant inoculation allowed detection
hyphae of the wild-type strain in the vascular tissue of crowns
and roots and RT-PCR experiments showed that FOW1 is
expressed by the pathogen in the plant. By contrast, no mycelial structures of the mutant were observed within the plant
tissue and no F. oxysporum colonies could be isolated from
plants inoculated with fow1 mutants. It was concluded from
these results that FOW1 had a function that is required for the
colonization of plant tissue by F. oxysporum, probably as a
response to the host environment (Inoue et al., 2002). No
conclusion can be drawn from the reported data about
both the efficiency of the REMI method to generate single
plasmid integration event in the F. oxysporum genome and the
number of tagged genes in the pathogenicity reduced mutants
recovered.
Although an efficient tool for tagging and cloning pathogenicity genes from fungal pathogens, a significant limitation
of REMI is that a large portion of generated mutants appeared
to be untagged by the transforming DNA as a result from
spontaneous mutations either generated during protoplast
formation or caused by REMI-induced DNA repair. As an
alternative tool to REMI, Agrobacterium tumefaciens-mediated
transformation (ATMT) offers a potential for insertional
mutagenesis. Mullins et al. (2001) have reported the optimization of an AMTT protocol which permitted an efficient
genetic manipulation of F. oxysporum to be achieved. Fungal
conidia were cocultivated with A. tumefaciens cells and the
transformation efficiency was found to correlate strongly with
the duration of cocultivation of fungal spores with A. tumefaciens cells and significantly with the number of bacteria
present during the cocultivation period. Growing bacterial

81

82 Review

Research review

cells in the presence of acetosyringone (AS) before cocultivation shortened the time required for the formation of
transformants but decreased to 53% the percentage of transformants containing a single T-DNA insert per genome.
However, this increased to over 80% when A. tumefaciens cells
grown in the absence of AS were used. The host sequences
flanking the inserted T-DNA were isolated by a thermal asymetric interlaced PCR (TAIL-PCR) technique and the use of
only one arbitrary primer resulted in the successful amplification of desired products in 90% of those transformants
analyzed. The insertion event was found to be a random
process, suggesting the suitability of ATMT as an insertional
mutagenesis tool for F. oxysporum. Whether ATMT generates
a significant percentage of untagged mutants remains to be
determined. However, the huge advantage offered by
A. tumefaciens-mediated transformation over conventional
protoplast-based transformation techniques is the versatility
it provides in choosing which starting material to transform:
A. tumefaciens has been shown to be able to transform protoplasts, spores, hyphae and blocks of mycelial tissue (Romaine,
unpublished).
Another tagging system is represented by the use of transposons, which are segments of DNA capable of moving from
one location to another in the genome. The potential use of
the naturally occuring F. oxysporum transposons impala and
Fot1 to generate random mutations in filamentous fungi has
been stressed in Daboussi (1996) and Migheli et al. (1999).
The transposon impala was recently found to be able to tag a
pathogenicity gene in the rice blast fungus Magnaporthe grisea
(Villalba et al., 2001). From a collection of 350 revertants,
mutants either altered for their mycelial growth (rev2) or
nonpathogenic (rev77) were obtained. Complementation of
rev77 with a 3-kb genomic fragment from a wild-type locus
was successful, demonstrating the tagging of a pathogenicity
gene by impala. This gene, called ORP1, was found to be
essential for penetration of host leaves by M. grisea and has
no sequence homology to known gene. Although only one
pathogenicity mutant was recovered, the mutant frequency
(0.3%) was found similar to the mutant frequency (0.5%)
obtained with REMI in M. grisea. However, it can not be concluded from this study as to whether or not impala generates
a significant percentage of tagged mutants. The ability of the
transposon impala to inactivate genes involved in the pathogenicity of F. oxysporum f. sp. melonis has been investigated by
Hua-Van (1998) and Migheli et al. (2000). Transposon mutagenesis has been expected to generate less untagged mutants
compared to the other insertional methods because a transformation step in not required. Somatic excision of an impala
copy inserted in the nitrate reductase reductase-encoding
niaD gene was positively selected through a phenotypic assay
based on the restauration of nitrate reductase activity. Independent excision events were analyzed molecularly and shown
to carry reinserted impala in more of 70% of the cases and
impala was demonstrated to transpose randomly. By screening

747 revertants on melon plants, a high proportion (3.5%) of

mutants impaired in their pathogenic potential could be
recovered. According to the kinetics of wilt symptom development, the strains that were impaired in pathogenicity were
clustered in three classes: class 1 grouped two strains (rev127
and rev157) that never induced Fusarium wilt symptoms on
the host plant; class 2 and class 3 grouped 15 and 9 revertants
which caused symptoms more than 50 d and 30 d after inoculation, respectively (Migheli et al., 2000). When pathogenic
deficient mutants were compared to the wild type strain for
their radial growth development, no significant difference
could be recorded, suggesting that they were not impaired
in hyphal extention (Recorbet, unpublished). Revertants
grouped in class 2 showed a very long lag phase before causing
symptoms, but once this phase elapsed, they showed a degree
of virulence similar to that presented by the pathogenic wildtype strain Fom24. Revertants grouped in class 3 showed a
shorter lag phase but symptoms appeared more gradually. It
was thus concluded that the pathogenicity deficiency of revertants from class 2 and class 3 resulted from different mechanisms which remain to be elucidated, such as impairment in
saprophytic competence, root colonization, or root penetration abilities. Among the two nonpathogenic mutants recovered, rev157 was found to be able to colonize and penetrate
root tissue but was unable to reach the vessels, explaining why
it did not cause symptoms of Fusarium wilt. The nonpathogenic revertant 127 was seldom reisolated from root tissue,
suggesting it could be impaired in root colonization ability.
The molecular analysis of the pathogenicity deficient revertants showed that in the majority of cases (26 of 30), the
excised impala copy was reintegrated into a new genomic
position (Migheli et al., 2000). To assess whether or not the
reinserted impala copy had tagged a gene involved in F. oxysporum pathogenicity, flanking regions to the impala reinsertion
site were rescued by inverse-PCR (IPCR) and used to probe
the genomic DNA. In most cases, reinsertion sites were found
to occur in monocopy except for revertants 69 and 127 in
which impala reinsertion happened in a genomic sequence
detected in four copies (Hua-Van, 1998). The IPCR products
generated from revertants 156, 427, 463, 473, 134 and 157
have been sequenced and database searches were performed.
No significant ORF suggesting the existence of a gene disrupted by impala could be detected (Hua-Van, 1998). Similar
results were obtained for revertants 389, 435 and 468. The
fact that impala reinsertion did not occur in 10% of the pathogenicity deficient mutants recovered indicates that other
mechanisms may be responsible for the mutant phenotype.
Among them, the mobilization of other transposable elements,
the genetic instability of Fom24 or the activation of DNArepair systems have been suggested by Hua-Van (1998). To
assess whether instability of the transposed impala copy may
occur during the interaction with the host plant, inoculated
revertants were reisolated from root and shoot and were
compared along with the initial revertant for both their pattern

www.newphytologist.com New Phytologist (2003) 159: 73 92

Research review

of impala reinsertion and their pathogenicity on melon

plants. Three striking cases could be detected in this respect
(Steinberg, Olivain, Recorbet & Daboussi, unpublished).
Whereas the initial revertant 468 was found to be pathogenic
on melon plants, isolates from root and shoot tissues happened to be impaired in pathogenicity and both could be
grouped within class 2. Whereas both the initial revertant
468 and its corresponding root isolate exhibited similar
impala reinsertion patterns, impala was found to have transposed into a new genomic position in the shoot isolate. When
this analysis was performed on rev505 belonging to class 2,
the root isolate did not induce Fusarium wilt symptoms
any more on melon plants (class1), whereas the shoot isolate
was as virulent as the wild type strain Fom24. Impala was
found to have transposed into a new genomic position as
compared to the initial revertant 505, and the reinsertion loci
had the same size for both the root and the shoot isolates.
Concerning the revertant 493 (class 2), its root isolate exhibited similar kinetics of wilt symptom development, but the
shoot isolate was more virulent than Fom24. Whereas both
the initial revertant 493 and its corresponding root isolate
exhibited similar impala transposition patterns, the reintegrated copy was lost in the shoot isolate. Although the possibility that impala may transpose during the subculture step
(necessary to recover isolates from root or shoot) cannot be
ruled out, those results strongly argue for the occurrence of
transposition events of impala in planta. In order to overcome
the instability of the mutagenic impala copy after transposition, the development of a two-component-based system, as
described for plants, has been proposed by Migheli et al.
(2000).
As stressed by Hua-Van (1998), the suitability of transposons to act as gene tags happens to be dependent upon the
number of their endogenous copies because the reinserted
element has to be detected for the recovery of the flanking
sequences. With regards to this point, Edel et al. (2001) have
addressed the genomic distribution of two naturally occuring
F. oxysporum transposable elements, impala and Fot1, within
nonpathogenic soilborne populations. The impala element
was shown to be present in all of the 47 isolates tested, irrespective of their soil of origin. Concerning the transposon
Fot1, Edel et al. (2001) demonstrated that this element was
undetected in 38% of the nonpathogenic isolates tested, some
of which, originating from the suppressive soil from Chteaurenard, (France) proved to be efficient antagonistic strains
against Fusarium wilt diseases.
Fot1 has thus emerged as a potential tool to inactivate genes
involved in the antagonist activity of some nonpathogenic
strains of F. oxysporum. In this respect, much attention has
been paid to the strain Fo47, free of endogenous Fot1 element, whose genetic context combined to a cotransformation
procedure was used to demonstrate the ability of Fot1 to transpose autonomously and to reinsert frequently at different sites
distributed throughout the fungal genome (Migheli et al.,

New Phytologist (2003) 159: 73 92 www.newphytologist.com

Review

1999). Trouvelot et al. (2002) further tested Fot1 as a mutagenesis tool to investigate the biocontrol mechanisms by
which strain Fo47 is active against Fusarium wilt. Ninety
strains in which Fot1 has transposed were obtained and
screened for their antagonistic activity against F. oxysporum
f. sp. lini in glasshouse experiments. Sixteen strains were found
to be affected in their biocontrol activity, either positively (10
revertants more antagonist than Fo47 were clustered within
class 1) or negatively (six revertants less antagonist were clustered within class 2). The phenotypic characterization of the
two more affected revertants, rev83 and rev94 belonging to
class 1 and 2, respectively, showed that they were impaired
neither in their in vitro growth habits nor in their competitiveness in soil as compared to the wild-type strain Fo47. It
was thus suggested that the altered biocontrol phenotype of
the mutants was unlikely to be expressed during the saprophytic phase but rather during the interaction with the
plant. Despite being mainly known to control Fusarium wilt
through saprophytic competition for nutrients (Alabouvette
& Couteaudier, 1992), the isolate Fo47 has also been shown
to induce resistance in tomato and watermelon plants (Fuchs
et al., 1997; Larkin & Fravel, 1999). However, the extent of
the role of ISR (induced systemic resistance) in the biological
control by Fo47 was reported to be marginal (Larkin &
Fravel, 1999). In view of the results reported in Trouvelot
et al. (2002), it can be suggested that the increase in the
biocontrol activity of revertant 83, which happened to be
efficient even at a population ratio of 10 : 1 (nonpathogen :
pathogen), may reflect an increase in a ISR-related mechanism. It can also be speculated that the decrease in the biocontrol activity of revertant 94, may be linked to a reduced ability
to induce resistance in flax plants. Split-root experiments
should be performed to test for this hypothesis. Fo47 is
known to control Fusarium wilt through at least three different mechanisms (Mandeel & Baker, 1991). On the contrary,
some soilborne nonpathogenic strains of F. oxysporum, such as
Fo50, have been shown to reduce wilt disease without colonizing the plant root, implicating that they reduce the pathogen density before the parasitic stage (Eparvier, 1992). In view
of the low copy number of both Fot1 and impala within this
strain (Edel et al., 2001), it can be suggested that transposonbased approaches should be able to generate mutants of Fo50
affected in their biocontrol activity during the saprophytic
phase of the interaction. No conclusion can be drawn from
the data reported by Trouvelot et al. (2002) as to whether or
not the altered biocontrol phenotypes resulted from Fot1
transposition.
Thus, with the exception of their trapping within niaD
(Langin et al., 1995; Migheli et al., 1999), there is not yet any
demonstration that Fot1 or impala can insert into genes of
F. oxysporum. Anyway, transposon-based strategies proved to
be efficient in generating pathogenesis or antagonist affected
mutants of F. oxysporum (Migheli et al., 2000; Trouvelot
et al., 2002). The availability of such mutants allows for the

83

84 Review

Research review

first time a comparison isogenic strains, overcoming the

potential bias linked to the large diversity existing among
soilborne pathogenic and nonpathogenic strains of F. oxysporum (Edel et al., 2001). One of the most promising black
box approaches may rely on in planta differential expressionbased strategies (Pieterse et al., 1993; Talbot et al., 1993;
Benito et al., 1996). Indeed, both the pathogenicity deficient
(revertants 157 and 127) and antagonist affected (revertants
83 and 94) phenotypes of the F. oxysporum mutants happened
to be detected during the parasitic stage of the interactions
(Trouvelot et al., 2002; Migheli et al., 2000). Thus, the
analysis of genes dif-ferentially expressed in planta between
both interactions (mutant vs wild-type strain) may lead to
the identification of pathogenesis or biocontrol-related
determinants. Despite being studied mainly at the gene
level, differential expression can also be assessed at the protein
level (Marmeisse et al., 1994; Pain et al., 1994; Perfect et al.,
1998). The techniques of proteome analyses and differential
protein display have the power and advantages over genomic
approaches to monitor global changes that occur in a tissue or
an organism (Bestel-Corre et al., 2002). With regards to such
methods, Recorbet et al. (2001) compared the protein
content of melon roots 2 d after inoculation by either the
pathogenic wild-type strain F. oxysporum f. sp. melonis
Fom24TR7 or the nonpathogenic mutant rev157. Similar
protein patterns were observed between control roots and
roots inoculated with the nonpathogenic mutant rev157.
Additionnal proteins in the range of 35 kDa were detected
in melon roots inoculated with the pathogenic wild-type.
Whether these proteins originate from the plant or the
pathogen remains to be further investigated. A similar
approach was undertaken by Rep et al. (2002) who investigated the protein content of tomato xylem sap after colonization by F. oxysporum. Peptide mass fingerprinting and mass
spectrometry sequencing were used to identify the most
abundant proteins appearing during compatible or incompatible interactions. A new member of the PR-5 family was
identified that accumulated early in both types of interactions.
Other pathogenesis-related proteins such as -1,3-glucanases
appeared in compatible interactions only, concomitantly with
disease development. In addition to plant defense-related
proteins, a pathogen-derived protein was also identified that
may be a good candidate for virulence/or avirulence factor
(Rep, unpublished).
Targeted inactivation of a candidate gene
As described in the first part, the targeted disruption of a gene
whose function is postulated to play a role in F. oxysporum
pathogenesis, on the basis of either biochemical or cytological
data, has not led until now to the conclusive identification of
functions involved in pathogenicity. As an alternative strategy,
the candidate approach, consists of supposing that a gene is
involved in the infection process by analogy with other

organisms or patho-systems. Results in this emerging field

essentially concern signalling genes. In order to recognize
suitable hosts, penetrate and invade plant tissues, fungal plant
pathogens are likely to perceive signals from the environment
and respond with the appropriate metabolic and morphogenetic
changes required for pathogenicity.
In many foliar plant pathogens, it has been shown that such
responses depended on the mitogen activated protein (MAP)
kinase signal transduction pathway. Among the MAPKs,
the yeast and fungal extracellular signal-regulated kinase
(YERK1) subfamily plays a key role in infection-related
morphogenesis and pathogenicity. MAPKs belonging to this
subfamily are involved in infection structure formation and in
invasive growth of phytopathogenic fungi attacking the aerial
parts of the plant, including the cereal leaf pathogens Magnaporthe grisea and Cochliobolus heterotrophus (Xu & Hamer,
1996; Lev et al., 1999), the cucumber leaf pathogen Colletotrichum lagenarium (Takano et al., 2000), the maize pathogen
Ustilago maydis (Mller et al., 1999) and the broad-host-range
necrotroph Botrytis cinerea (Zheng et al., 2000). Di Pietro
et al. (2001a) investigated MAPKs as candidate pathogenesisrelated functions in F. oxysporum. They managed to identify,
from F. oxysporum f. sp. lycopersici, the gene fmk1 (Fusarium
MAP kinase 1) encoding a mitogen-activated protein kinase
that belongs to the YERK1 family. The targeted inactivation
of fmk1 was found to abolish pathogenicity of F. oxysporum on
tomato plants. The fmk1 gene was reported to be dispensable
for vegetative growth and conidiation under laboratory conditions, suggesting that fmk1 acted specifically in signalling
during the interaction of F. oxysporum with the host plant and
was not required under conditions encountered by the fungus
grown in culture. The wild-type and the mutant strains
were thus compared for their ability to accomplish different
pathogenicity-related functions in the initial stages of infection. Both the wild-type strain and the fmk1 mutant happened
to germinate 1 d after immersion of tomato roots in the respective conidial suspensions. Whereas most of the wild-type
germlings had attached to the root and initiated hyphal
growth on the root surface, the propagules of the mutants
remained in solution. In planta experiments revealed that
whereas wild-type conidia differentiated a thin germ tube that
elongated and attached to the root surface, most conidia from
the fmk1 mutant failed to show germ tube elongation and
became swollen. It was concluded that FMK1 was essential
for complete differentiation of infection hyphae, attachement
to the root surface and penetration. In addition, FMK1 was
found to be required for maintaining full hydrophobicity of
aerial hyphae (Di Pietro et al., 2001a). It was speculated that
the easily wettable phenotype of the fmk1 mutant and the
inability of the hyphae to escape the water tension of water
droplets may be related to a failure to express hydrophobin
genes, leading to an impairment in the attachment process to
the root surface. Additional factors, such as the occurrence of
fungal mucin-like proteins, have been suggested to mediate

www.newphytologist.com New Phytologist (2003) 159: 73 92

Research review

adhesion to the host plant. The data from Di Pietro et al.

(2001a) also demonstrated that the MAPK Pmk1 from the
rice blast fungus M. grisea, which differs from FMK1 in eight
amino acid residues, can partially restore pathogenesis-related
functions in the fmk1 mutants. Only strains carrying multiple
copies of Pmk1 regained surface hydrophobicity to the same
extent as the wild-type, and those with single or double Pmk1
copies showed partial complementation of the mutant phenotype. Pmk1 also partially restored root attachment and pathogenicity in the fmk1 mutant. Because M. grisea is regarded as
a foliar pathogen, it was thus concluded that, in spite of the
diversity of infection mechanisms developed by plant pathogenic fungi, the key steps leading to plant infection and their
underlying signalling pathways were essentially the same. As
stated by Hamer & Holden (1996), candidate gene knock-out
experiments have revealed regulatory networks that appear to
be specific to pathogenesis. The challenge for pathologists will
be to link these signalling pathways to the host environment
and to the expression of determinants that advance the disease
state. In this respect, Dufresne & Osbourn (2001) demonstrated that the leaf blast pathogen M. grisea could also infect
the roots of cereals and caused lesion development. To establish whether M. grisea could be used to investigate the genetic
requirements for infection of different plant tissues, a collection of mutants was tested for the ability to cause disease on
barley and wheat roots. Mutants with leaf-specific, root-specific,
and general pathogenicity defects could be identified. The
MAP kinase PMK1 was found to be required for lesion development on both leaves and roots, suggesting it may belong
to a core set of basic pathogenicity factors required for plant
colonization. Whereas the M. grisea mutant nut1 is pathogenic on leaves, it happened to fail to give symptoms on
roots, indicating it was defective specifically in root infection
(Dufresne & Osbourn, 2001). The NUT1 gene, a homolog of
the Aspergillus nidulans AREA gene, is a global regulator of
nitrogen utilization (Froeliger & Carpenter, 1996) and is also
involved in the induction of the nitrate-reductase structural
gene under nitrogen-starvation conditions. Whether such a
gene may be involved in the pathogenesis of the root pathogen
F. oxysporum remains an open question. The targeted disruption of a putative NUT1 homolog in this fungus should help
to address this point. Likewise, the data reported in Dufresne
& Osbourn (2001) raise the possibility of a link between
nitrogen starvation and fungal pathogenicity. Regulation of
nitrogen assimilation is important for disease development
by M. grisea on leaves and roots. Under nitrogen-starvation
conditions, the NPR1 gene of M. grisea is required for the
expression of the hydrophobin-encoding gene MPG1, which
is essential for pathogenicity to leaves (Talbot et al., 1993).
There is further evidence that many of basic compounds
required for fungal growth are lacking during infection and
must therefore be synthesized. The importance to pathogenic
fungi of in vivo methionine synthesis has been demonstrated
in M. grisea (Balhadre et al., 1999) and Cladosporium fulvum

New Phytologist (2003) 159: 73 92 www.newphytologist.com

Review

(Solomon et al., 2000). It has also been shown in M. grisea

that a histidine auxotroph is significantly less pathogenic than
the wild-type. In C. fulvum, a study of fungal genes expressed
during carbon starvation in vitro identified several genes that
were up regulated during growth in planta (Coleman et al.,
1997; Oliver et al., 2000). These included genes predicted to
encode acetaldehyde dehydrogenase (Aldh1) and alcohol
oxidase (Aox1). Alcohol oxidase was recently found to be a
pathogenicity factor for C. fulvum, but aldehyde dehydrogenase was shown to be dispensable (Segers et al., 2001). The
fact that many genes important for pathogenicity happened
to be induced during in vitro starvation conditions raises the
possibility of such a screening to be attempted in F. oxysporum.
In this respect, unusual requirements for vitamins by strains
of F. oxysporum f. sp. melonis and f. sp. vasinfectum have been
pointed out by Meyer & Maraite (1967) and El-Abyad &
Ramadan (1979), respectively. In addition, an understanding
of how fungal metabolism changes during host colonization
is also of importance in establishing the role of nutritional
clues in triggering of pathogenesis or infection structure
formation and may lead to the development of new disease
control strategies. The use of nutrient-regulated promoterGFP fusions to sense the fungal microenvironment in planta
should help to address this question (Bowyer et al., 2000;
Lorang et al., 2001; Rohel et al., 2001). Based on a probable
in planta nutrient deficiency, the possibility exists that
these essential biosynthetic pathways may be targets for
fungicide development (Daniel & Lucas, 1995; Solomon
et al., 2000).
Among the components of the signal transduction pathways, G proteins are heterotrimeric GTP-binding proteins
involved in tranducing signals from activated membrane
receptors to a variety of intracellular targets. G protein
subunits have been shown to be implicated in the pathogenicity of Magnaporthe grisea (Liu & Dean, 1997), Cryphonectria
parasitica (Choi et al., 1995), Colletotrichum trifolii (Truesdell
et al., 2000) and Botrytis cinerea (Gronover et al., 2001). Jain
et al. (2002) investigated the subunit of G proteins as a
potential pathogenesis-related function in F. oxysporum.
They managed to isolate from F. oxysporum f. sp. cucumerinum the gene fga1 encoding a G subunit that showed high
identity to those of G protein alpha family proteins from
other filamentous fungi. The targeted inactivation of fga1
resulted in altered colony morphology, reduced conidiation,
increased heat resistance and reduced pathogenity after rootdip inoculation. All these observations suggested an important
role played by G protein signalling during growth, development and infection processes in F. oxysporum.
Updating the biological models
Screening mutants of F. oxysporum for loss of virulence led
to the successful identification of some pathogenesis-related
functions, as listed in Table 2, allowing us to update the data

85

Method

www.newphytologist.com New Phytologist (2003) 159: 73 92

Generation of auxotrophic mutants of

F. oxysporum f. sp. lycopersici with changes
in pathogenicity.
Generation of nonpathogenic mutants of
F. oxysporum f. sp. cucumerinum.
UV mutagenesis
Generation of a nonpathogenic mutant of
F. oxysporum f. sp. melonis.
Plasmid-mediated
Recovery of a nonpathogenic chsV insertional
transformation
mutant of F. oxysporum f. sp. lycopersici.
ChsV, a class V chitin synthase, is required for:
cell wall structure integrity
colonization of tomato plants
growth on tomato fruit tissue
overcoming sensitivity to -tomatine and H2O2.
REMI
Recovery of an arginine auxotroph mutant
of F. oxysporum f. sp. melonis deficient in pathogenicity.
Mutation in ARG1 encoding an argino-succinate lyase.
The ARGI gene expression is suppressed in
nutrient-rich conditions.
Recovery of a nonpathogenic fow1 insertional
mutant of F. oxysporum f. sp. melonis.
FOW1, a mitochondrial carrier protein, is dispensable
for saprophytic growth and root penetration but is
required for the colonization of plant tissue.
FOW1 is conserved among formae speciales
of F. oxysporum and is required for the virulence
of F. oxysporum f. sp. lycopersici.
Transposon-based
Recovery of F. oxysporum f. sp. melonis mutants
mutagenesis
impaired in pathogenicity clustered within 3
classes according to the kinetics of wilt symptom
development.
Targeted inactivation of a candidate gene
Inactivation of fmk1
The fmk1 mutant is nonpathogenic. FMK1
encoding a MAP kinase
is required for:
in F. oxysporum
complete differenciation of infection hyphae
f. sp. lycopersici
full hydrophobicity of aerial hyphae
attachement of germlings to the root surface.
Inactivation of fga1
The fga1 mutant shows a reduced pathogenicity.
encoding a subunit of G
FGA1 is dispensable for vegetative
protein in F. oxysporum
growth but not for conidiation.
f. sp. cucumerinum

Potential mechanism

Reference(s)

Nutrient limitation may induce the expression

of genes involved in pathogenicity. Potential
nutrient limitation in planta.
Different expression patterns of sti35
in the mutants and wild-type strain.
Potential reduction in cellulolytic activity of the
mutant compared with that of the wild-type strain.
ChsV may contribute to the structural defence
function of the cell wall by preventing the access
of antifungal plant compounds to their cellular targets.

Sanchez et al. (1975)

Akiyama et al. (2000)

Freeman et al. (2002)
Madrid et al. (2003)

Nutrient limitation may induce the expression

of genes involved in pathogenicity. Potential
nutrient limitation in planta.

Namiki et al. (2001)

FOW1 plays a role in fungal pathogenesis probably

as a response to host environment.

Inoue et al. (2002)

Different mechanisms are involved in F. oxysporum

pathogenicity.
Potential transposition in planta.

Migheli et al. (2000)

Potential existence of hydrophobin(s) or

mucin-like proteins in F. oxysporum.

Di Pietro et al. (2001a)

fga1 may play a role in the infection process.

Jain et al. (2002)

Research review

Random mutagenesis
Chemical mutagenesis

Observation

86 Review

Table 2 Updating speculations about the processes which are likely to be required for a successful infection by Fusarium oxysporum by using black box approaches or other patho-systems as
listed in part III

Research review

generated on the basis of cytological and biochemical

observations (Table 1). Although untagged mutations are
likely to occur whatever the mutagenic procedure used, the
availability of such mutants allows a comparison of isogenic
strains, overcoming the potential bias linked to the large
diversity existing among soilborne pathogenic and nonpathogenic strains of F. oxysporum. Taken together, the strategies
described above support the idea that changes in fungal
metabolism is of importance in triggering of pathogenesis.
Furthermore, the targeted disruption of fmk1 in F. oxysporum
demonstrated the occurrence of previously unknown pathogenecity factors, such as hydrophobicity or attachment of
germlings, with promising biochemical determinants which
remain to be further dissected.

Pathogenesis-related marker molecules and

biocontrol activity
As previously stated, the method to identify pathogenic formae
speciales or races of F. oxysporum is through a bioassay in
which putative host-plants are confronted to the fungus. As
compared to other patho-systems, assessing the pathogenicity
or the nonpathogenicity of F. oxysporum strains or mutants is
impaired by a time-consuming and cumbersome screening
procedure. A rapid inoculation technique has been proposed
by Freeman & Rodriguez (1993a). The method, adapted
from a procedure used for determining pathogenicity of
Colletotrichum magna on cucurbits, involves constant exposure
of seedlings to conidial suspensions. This continuous-dip
inoculation technique in vials was found to be reliable and
consistent with the standard pot-inoculation procedure
for determining pathogenicity of various Fusarium isolates,
requiring less time, labor and space. Initially, the continuousdip inoculation bioassay was devised for large-scale screening
and isolation of nonpathogenic mutants of C. magna
(Freeman & Rodriguez, 1992). Following UV mutagenesis,
this procedure allowed to recover the nonpathogenic mutant
(path-1) mutated at a single locus (Freeman & Rodriguez,
1993b). The mutant maintained wild-type levels of in
vitro sporulation, spore adhesion, appressorial formation and
infection. It was also shown that a phytopathogenic fungus
could be modified by mutation to a nonpathogenic endophyte.
Freeman & Rodriguez (1993b) and Redman et al. (1999)
demonstrated that the nonpathogenic (path-1) mutant of
C. magna was capable of colonizing the root and stems but not
the cotyledons of cucurbits. This mutant was able to protect
watermelon seedlings from disease caused by the wild-type
C. magna, the wilt pathogen F. oxysporum f. sp. niveum, and the
agressive foliar pathogen C. orbiculare without inducing ISRrelated necrosis. Biochemical analyses indicated that path-1colonized plants were able to mount a rapid defense response
that resulted in protection from pathogen ingress. A model
designated endophyte-associated resistance, based on the fact
that path-1 either activates plant defenses to low levels or

New Phytologist (2003) 159: 73 92 www.newphytologist.com

Review

primes the defense system without activation, was thus

proposed by Redman et al. (1999). The fact that mutation
at a single locus can convert a fungal plant pathogen to a
nonpathogenic, endophytic mutualist supports the idea that
a nonpathogenic mutant is expected to be antagonist against
the wild-type strain. In this respect, Freeman et al. (2002) by
using UV mutagenesis and a continuous dip-inoculation
procedure isolated a nonpathogenic mutant of F. oxysporum
f. sp. melonis that was able to reduce muskmelon seedlings
mortality in cross-protection experiments. However, biocontrol
experiments performed with the nonpathogenic endophyte
revertant 157 from F. oxysporum f. sp. melonis Fom24 (Migheli
et al., 2000) showed that rev157 did not protect melon and
flax plants, inoculated with their respective pathogens Fom24
and Foln3 (Olivain et al., 2001b). Taken together, these data
address several questions about the compatibility of the
procedures used to screen for nonpathogenic mutants and
biocontrol activity. A standard pot-inoculation procedure
has been used to recover the nonpathogenic mutant rev157
(Olivain et al., 2001b). On the contrary, both Freeman &
Rodriguez (1993b) and Freeman et al. (2002) used the
continuous-dip inoculation technique in vials to recover the
path-1 mutant of C. magna and the nonpathogenic mutant of
F. oxysporum f. sp. melonis. Soilless conditions added to the
inoculation method may have maximized the probability to
recover a mutant unaffected in its competitive potential and
able to trigger localized protection. The nonpathogenic
revertant 157 should be tested for its competitiveness in soil
as described by Couteaudier & Alabouvette (1990) in order
to asses whether it is less competitive than the wild-type
strain Fom24. In this respect, Larkin & Fravel (1999) have
investigated the mechanisms of action and doseresponse
relationships governing biological control of Fusarium wilt by
nonpathogenic Fusarium spp. Differences in doseresponse
relationships among the biocontrol isolates were attributed
to differences in their mechanisms of action, with CS-20
and CS-1 acting primarily by inducing resistance and Fo47
functioning primarily by competition for nutrients. With
regard to inundative biocontrol situations, they underlined
that one of the primary benefits of ISR as a mechanism of
action for these biocontrol isolates was that much lower
inoculum levels were required than when competition was the
mechanism of action. In view of these data, it is likely that
knowledge about the F. oxysporum determinants triggering
ISR should help in the selection of efficient antagonist strains
by using the split-root screening procedure. On the contrary,
understanding the mechanisms by which Fo47 is efficient
against Fusarium wilt involves a screening for antagonist
affected mutants by using the standard pot-inoculation
procedure in order to take into account all the components
that may trigger biocontrol (Trouvelot et al., 2002). Such a
method allowed Migheli et al. (2000) to distinguish at least
three mechanisms involved in the pathogenesis of F.
oxysporum. Likewise, in view of the data supporting the idea

87

88 Review

Research review

that changes in fungal metabolism may be of importance in

both biocontrol activity and pathogenesis, the inoculum
composition and production should be considered when
generating and screening for F. oxysporum mutants. In this
respect, the high yielding mass production technique for F.
oxysporum chlamydospores is likely to be a useful tool in
exploring the events triggering of their germination (Bedi &
Sauerborn, 1999).
Much attention has been paid to the screening of insertional fungal mutants in order to tag genes involved in pathogenicity or biocontrol activity. However, untagged mutants
are also likely to be useful in getting a better understanding of
plant /microorganisms interactions as described for C. magna
path-1 and mutants from F. oxysporum f. sp. melonis and Fo47
(Redman et al., 1999; Migheli et al., 2000; Freeman et al.,
2002; Trouvelot et al., 2002). New trends are emerging in the
study of these mutants such as search for proteins differentially expressed either in vitro or in planta that may serve as
diagnostic tools to discriminate between a pathogenic and a
nonpathogenic plant/fungal interaction.

Note added in proof

During the editing process of this review, two additional
Fusarium oxysporum genes involved in the regulation of its
pathogenicity have been identified. Fgbl encodes a G protein
beta subunit ( Jain et al., 2003. Current Genetics 43: 7986) and
pacC encodes a PH signalling transcription factor (Caracuel
et al., 2003. Molecular Microbiology 48: 765779).

References
Agrios GN. 1997. Plant pathology. San Diego, CA, USA: Academic Press.
Akiyama K, Thanonkeo P, Ogawa H, Ohguchi T, Takata R. 2000.
Detection and cloning of the gene encoding a protein produced by
nonpathogenic mutant of Fusarium oxysporum. Journal of Bioscience and
Bioengineering 90: 302307.
Alabouvette C, Couteaudier Y. 1992. Biological control of fusarium wilts
with nonpathogenic Fusaria. In: Tjamos EC, Cook RJ, Papavizas GC, eds.
Biological control of plant diseases. New York, USA: Plenum Press,
415426.
Alabouvette C, Edel E, Lemanceau P, Olivain C, Recorbet G, Steinberg C.
2001. Diversity and interactions among strains of Fusarium oxysporum :
Application to biological control. In: Jeger MJ, Spence NJ, eds. Biotic
interactions in plantpathogen associations. Wallingford, UK: CAB
International, 131158.
Anderson AJ. 1989. The biology of glycoproteins as elicitors. In: Kosuge T,
Nester E, eds. Plantmicrobe interactions molecular and genetic perspectives.
New York, USA: McGraw-Hill, 87130.
Anderson C, Salomons GL. 1984. Primary metabolism and biomass
production from Fusarium. In: Moss MO, Smith JE, eds. The applied
mycology of Fusarium. Cambridge, UK: Cambridge University Press,
231250.
Armstrong MJ, Armstrong JK. 1981. Formae speciales and races of
Fusarium oxysporum causing wilt diseases. In: Nelson PE, Toussoun TA,
Cook RJ, eds. Fusarium: diseases, biology, and taxonomy. London, UK:
Pennsylvania State University Press, 391399.
Baayen RP, ODonnell K, Bonants PJM, Cigelnik E, Kroon LPNM,

Roebroeck EJA, Waalwijk C. 2000. Gene genealogies and AFLP analyses

in the Fusarium oxysporum complex identify monophyletic and
nonmonophyletic formae speciales causing wilt and rot disease.
Phytopathology 90: 891900.
Balhadre PV, Foster AJ, Talbot NJ. 1999. Identification of
pathogenicity mutants of the rice blast fungus Magnaporthe grisea
by insertional mutagenesis. Molecular PlantMicrobe Interactions 12:
129142.
Beckman CH, Roberts EM. 1995. On the nature and genetic basis for
resistance and tolerance to fungal wilt diseases of plants. Advances in
Botanical Research 21: 3677.
Bedi JS, Sauerborn J. 1999. A new technique for clamydospore production
by Fusarium oxysporum f. sp. orthoceras, a mycoherbicidal agent for
Orobanche cumana. Plant Disease Research 14: 207209.
Belenkiy R, Haefele A, Eisen MB, Wohlrab H. 2000. The yeast
mitochondrial transport proteins: New sequences and consensus residues,
lack of direct relation between consensus residues and transmembrane
helices, expression patterns of the transport protein genes, and protein
protein interactions with other proteins. Biochemica et Biophysica Acta
1467: 207218.
Benhamou N, Chamberland H, Pauz FJ. 1990. Implication of pectic
components in cell surface interactions between tomato root cells and
Fusarium oxysporum f. sp. lycopersici. A cytochemical study by means of a
lectin with polygalacturonic acid-binding specificity. Plant Physiology 92:
9951003.
Benito EP, Prins T, van Kan JAL. 1996. Application of differential display
RT-PCR to the analysis of gene expression in a plant fungus interaction.
Plant Molecular Biology 32: 947957.
Bestel-Corre G, Dumas-Gaudot E, Poinsot V, Dieu M, van Tuinen D,
Remacle V, Gianinazzi-Pearson V, Gianinazzi S. 2002. Identification
of symbiosis-related proteins from Medicago truncatula Gaert. by
two-dimentional electrophoresis and matrix-assisted laser desorption
ionization time-of-flight mass spectrometry. Electrophoresis 23:
122137.
Bishop CD, Cooper RM. 1983. An ultrastructural study of vascular
colonization in three vascular wilt diseases. I. Colonization of susceptible
cultivars. Physiological and Molecular Plant Pathology 23: 323343.
Bouhot D. 1981. Some aspects of the pathogenic potential in formae
speciales and races of Fusarium oxysporum on Cucurbitae. In: Nelson PE,
ToussounTA, Cook RJ, eds. Fusarium: diseases, biology, and taxonomy.
London, UK: The Pennsylvania State University Press, 391399.
Bowyer P, Clarke BR, Lunness P, Daniels MJ, Osbourn AE. 1995. Host
range of a plant pathogenic fungus determined by a saponin detoxifying
enzyme. Science 267: 371374.
Bowyer P, Mueller E, Lucas J. 2000. Use of an isocitrate lyase promoterGFP fusion to monitor carbon metabolism of the plant pathogen Tapesia
yallundae during infection of wheat. Molecular Plant Pathology 1:
253262.
Brown JS, Holden DW. 1998. Insertional mutagenesis of pathogenic fungi.
Current Opinion in Microbiology 1: 390394.
Cervone F, De Lorenzo G, Pressey R, Darvill AG, Albersheim P. 1990. Can
Phaseolus PGIP inhibit pectic enzymes from microbes and plants?
Phytochemistry 29: 447449.
Choi GH, Chen B, Nuss DL. 1995. Virus-mediated or transgenic
suppression of a G-protein subunit and attenuation of fungal
virulence. Proceedings of the National Academy of Sciences, USA 92:
305309.
Coleman M, Henricot B, Arnau J, Oliver RP. 1997. Starvation-induced
genes of the tomato pathogen Cladosporium fulvum are also induced in
planta. Molecular PlantMicrobe Interactions 10: 11061109.
Cook BJ, Clay RP, Bergman CW, Albersheim P, Darvill AG. 1999. Fungal
endoplolygalacturonases exhibit different substrate degradation patterns
and differ in their susceptibilities to polygalacturonase-inhibiting proteins.
Molecular PlantMicrobe Interactions 12: 703711.
Cooke RC, Rayner ADM. 1984. Growth forms and responses. In: Cooke

www.newphytologist.com New Phytologist (2003) 159: 73 92

Research review
RC, Rayner ADM, eds. Ecology of saprophytic fungi. London, UK:
Longman, 1451.
Couteaudier Y, Alabouvette C. 1990. Quantitative comparison of Fusarium
oxysporum competitiveness in relation to carbon utilization. FEMS
Microbiology Ecology 74: 261 268.
Couteaudier Y, Daboussi MJ, Eparvier A, Langin T, Orcival J. 1993. The
GUS gene fusion system, a useful tool in studies on root colonization by
Fusarium oxysporum. Applied and Environmental Microbiology 59:
17671773.
Daboussi MJ. 1996. Fungal transposable elements: Generators of diversity
and genetic tools. Journal of Genetics 75: 325 339.
Daniels A, Lucas JA. 1995. Mode of action of the anilino-pyrimidine
fungicide pyrimethanil.1. In vivo activity against Botrytis fabae on broad
bean (Vicia faba) leaves. Pesticide Science 45: 33 41.
Darvill A, Augur C, Bergmann C, Carlson RW, Cheong JJ, Ebehard S,
Hahn MG, Lo VM, Marfa V, Meyer B, Mohnen D, ONeill MA,
Spiro MD, van Halbeek H, York WS, Albersheim P. 1992.
Oligosaccharins-oligosaccharides that regulate growth, development and
defence responses in plants. Glycobiology 2: 181198.
De Donato M, Mozzetti C, Chiavazza P, Rubino R, Matta A. 1997. Active
oxygen production in cell-free melon homogenates elicited with hyphal
walls components of Fusarium oxysporum f. sp. melonis. I. A response
correlated with race-specific resistance. Journal of Plant Pathology 78:
3543.
de Lorenzo G, Castoria R, Bellicampi D, Cervone F. 1997. Fungal
invasion enzymes and their inhibition. In: Carroll G, Tudzynski P,
eds. The Mycota V Part a: Plant relationships. Berlin, Germany:
Springer-Verlag, 61 83.
Deacon JW. 1996. Ecological implications of recognition events in the
pre-infection stages of root pathogens. New Phytologist 133: 135145.
Di Pietro A, Garcia-Maceira FI, Mglecz E, Roncero MIG. 2001a. A MAP
kinase of the vascular wilt fungus Fusarium oxysporum is essential for
pathogenicity. Molecular Microbiology 39: 1140 1152.
Di Pietro A, Garcia-Maceira FI, Huertas-Gonzlez MD, Ruiz-Roldn MC,
Caracuel Z, Barbieri AS, Roncero MIG. 2001b. Endoplolygalacturonase
PG1 in different formae speciales of Fusarium oxysporum. Applied and
Environmental Microbiology 64: 19671971.
Di Pietro A, Huertas-Gonzlez MD, Gutierrez-Corona JF, MartinezCadena G, Mglecz E, Roncero MIG. 2001c. Molecular characterization
of a subtilase from the vascular wilt fungus Fusarium oxysporum. Molecular
PlantMicrobe Interactions 14: 653 662.
Di Pietro A, Roncero MIG. 1996. Endopolygalacturonase from Fusarium
oxysporum f. sp. lycopersici: purification, characterization, and production
during infection on tomato plants. Phytopathology 86: 1324 1330.
Di Pietro A, Roncero MIG. 1998. Cloning, expression, and role in
pathogenicity of pg1 encoding the major extracellular
endopolygalacturonase of the vascular wilt pathogen Fusarium oxysporum.
Molecular PlantMicrobe Interactions 11: 9198.
Dufresne M, Bailey JA, Dron M, Langin T. 1998. Clk1, a serine/threonine
protein kinase encoding gene, is involved in pathogenicity of
Colletotrichum lindemuthianum on common bean. Molecular
PlantMicrobe Interactions 11: 98 108.
Dufresne M, Osbourn AE. 2001. Definition of tissue-specific requirements
for plant infection in a phytopathogenic fungus. Molecular PlantMicrobe
Interactions 14: 300307.
Edel V, Steinberg C, Gautheron N, Alabouvette C. 1997. Populations of
nonpathogenic Fusarium oxysporum associated with roots of four plant
species compared to soilborne populations. Phytopathology 87:
693697.
Edel V, Steinberg C, Gautheron N, Recorbet G, Alabouvette C. 2001.
Genetic diversity of Fusarium oxysporum populations isolated form
different soils in France. FEMS Microbiology and Ecology 36: 6171.
El-Abyad MS, Ramadan ZM. 1979. Vitamin requirements of Fusarium
oxysporum f. sp. Vasinfectum Zentralblatt Fur Bakteriologie 134:
412418.

New Phytologist (2003) 159: 73 92 www.newphytologist.com

Review

Elias KS, Schneider RW, Lear MM. 1991. Analysis of vegetative

compatibility groups in nonpathogenic populations of Fusarium
oxysporum isolated from symptomless tomato roots. Canadian Journal
of Botany 69: 20892094.
Enkerli J, Bhatt G, Covert SF. 1998. Maackiain detoxification contributes
to the virulence of Nectria haematococca MP VI on chickpea. Molecular
PlantMicrobe Interactions 11: 317326.
Eparvier A. 1992. Comptition entre Fusarium oxysporum pathognes et
F. oxysporum non pathognes pour la colonisation des tissus racinaires: mise
au point et utilisation de techniques srologique et gntique de marquage.
PhD thesis, Universit C. Bernard Lyon I, France.
Eparvier A, Alabouvette C. 1994. Use of ELISA and GUS-transformed
strains to study competition between pathogenic and non-pathogenic
Fusarium oxysporum for root colonization. Biocontrological Science and
Technology 4: 3547.
Favaron F, Alghishi P, Marciano P, Magro P. 1988. Polygalacturonase
isozymes and oxalic acid produced by Sclerotinia sclerotinium in soybean
hypocotyls as elicitors of glyceollin. Physiological and Molecular Plant
Pathology 33: 385395.
Fond JE, McCance DJ, Drysddle RS. 1977. The detoxification of d.
tomatine by Fusarium oxysporum f. sp. lycopersici. Phytochemistry 16:
545546.
Fraissinet-Tachet L, Reymond-Cotton P, Fvre M. 1995. Characterization
of a multigene family encoding an endopolygalacturonase in Sclerotinia
sclerotinium. Current Genetics 29: 9699.
Freeman S, Rodriguez RJ. 1992. A rapid reliable bioassay for pathogenicity
of Colletotrichum magna on cucurbits and its use in screening for
nonpathogenic mutants. Plant Disease 76: 901905.
Freeman S, Rodriguez RJ. 1993a. A rapid inoculation technique for assessing
pathogenicity of Fusarium oxysporum f. sp. niveum and F. o. melonis on
cucurbits. Plant Disease 77: 11981201.
Freeman S, Rodriguez RJ. 1993b. Genetic conversion of a fungal pathogen
to a non-pathogenic, endophytic mutualist. Science 260: 7578.
Freman S, Zveibil A, Vintal A, Maymon M. 2002. Isolation of nonpathogenic
mutants of Fusarium oxysporum f. sp. melonis for biological control of
Fusarium wilt in cucurbits. Phytopathology 92: 164168.
Froeliger EH, Carpenter BE. 1996. NUT1, a major nitrogen regulatory gene
in Magnaporthe grisea, is dispensable for pathogenicity. Molecular and
General Genetics 251: 647656.
Fuchs JG, Monne-Locoz Y, Dfago G. 1997. Non-pathogenic Fusarium
oxysporum strain Fo47 induces resistance to Fusarium wilt in tomato. Plant
Disease 81: 492496.
Garcia-Maceira FI, Di Pietro A, Huertas-Gonzles MD, Ruiz-Roldn MC,
Roncero MI. 2001. Molecular characterization of an
endopolygalacturonase from Fusarium oxysporum expressed during
early stages of infection. Applied and Environmental Microbiology 67:
21912196.
Garcia-Maceira FI, Di Pietro A, Roncero MIG. 2000. Cloning and
disruption of pgx4 encoding an in planta expressed exopolygalacturonase
from Fusarium oxysporum. Molecular PlantMicrobe Interactions 13:
359365.
Gold SE, Garcia-Pedrajas MD, Martinez-Espinoza AD. 2001. New (and
used) approaches to the study of fungal pathogenicity. Annual Review of
Phytopathology 39: 337365.
Gmez-Gmez E, Roncero MIG, Di Pietro A, Hera C. 2001. Molecular
characterization of a novel endo--1,4-xylanase gene from the vascular wilt
fungus Fusarium oxysporum. Current Genetics 40: 268275.
Gmez-Gmez E, Ruiz-Roldn MC, Di Pietro A, Roncero MI, Hera C.
2002. Role in pathogenesis of two endo--1,4-xylanase genes from the
vascular wilt fungus Fusarium oxysporum. Fungal Genetics and Biology 35:
213222.
Gronover CS, Kasulke D, Tudzynski P, Tudzynski B. 2001. The role of G
protein subunits in the infection process of the gray mold fungus Botrytis
cinerea. Molecular PlantMicrobe Interactions 14: 12931302.
Hahn MG, Bucheli P, Cervone F, Doares SH, ONeill RA, Darvill A,

89

90 Review

Research review

Albersheim P. 1989. Roles of cell wall constituents in plantpathogen

interactions. In: Kosuge T, Nester EW, eds. Plantmicrobe interactions
molecular genetic perspectives. New York, USA: McGraw-Hill,
131181.
Hamer JE, Holden DW. 1996. Linking approaches in the study of fungal
pathogenesis: a commentary. Fungal Genetics and Biology 21:
1116.
ten Have A, Mulder W, Visser J, van Kan JAL. 1998. The
endopolygalacturonase gene Bcpg1 is required for full virulence of
Botrycis cinerea. Molecular PlantMicrobe Interactions 11: 1009 1016.
ten Have A, Tenberge KB, Benen JAE, Tudzynski P, Visser J, van Kan JAL.
2002. The contribution of cell wall degrading enzymes to pathogenesis of
fungal plant pathogens. In: Kempken F, ed. Application in agriculture. The
Myctota XI. Berlin, Germany: Springer, 341 358.
Hensel M, Holden DW. 1996. Molecular genetic approaches for the study
os virulence in both pathogenic bacteria and fungi. Microbiology 142:
10491058.
Hua-Van A. 1998. Caractrisation de la famille dlments transposables impala
et dveloppement dun outil de mutagense insertionnelle chez le champignon
phytopathogne Fusarium oxysporum. PhD thesis, Universit Paris VI,
France.
Huertas-Gonzlez MD, Ruiz-Roldn MC, Garcia-Maceira FI,
Roncero MIG, Di Pietro A. 1999. Cloning and characterization of pl1
encoding an in planta-secreted pectate lyase of Fusarium oxysporum.
Current Genetics 35: 36 40.
Idnurm A, Howlett BJ. 2001. Pathogenicity genes of phytopathogenic
fungi. Molecular Plant Pathology 2: 241255.
Inoue I, Namiki F, Tsuge T. 2002. Plant colonization of the vascular wilt
fungus Fusarium oxysporum requires FOW1, a gene encoding a
mitochondrial protein. Plant Cell 14: 1869 1883.
Ito S, Takahara H, Kawaguchi T, Tanaka S, Kameya-Iwaki M. 2002.
Post-transcriptional silencing of the tomatinase gene in Fusarium
oxysporum f. sp. lycopersici. Journal of Phytopathology 150: 474 480.
Jain S, Akiyama K, Mae K, Ohguchi T, Takata R. 2002. Targeted disruption
of a G protein subunit gene results in reduced pathogenicity in Fusarium
oxysporum. Current Genetics 41: 407 413.
Jones TM, Anderson AJ, Albersheim P. 1972. Hostpathogen interactions.
IV. Studies on the polysaccharide-degrading enzymes secreted by Fusrium
oxysporum f. sp. lycopersici. Physology and Plant Pathology 2: 153 166.
Kahmann R, Basse C. 1999. REMI (Restriction Enzyme Mediated
Integration) and its impact on the isolation of pathogenicity genes in fungi
attacking plants. European Journal of Plant Pathology 105: 221229.
Karr AL, Albersheim P. 1970. Polysaccharide-degrading enzymes are unable
to attack plant cell walls without prior action by a wall-modifyingenzyme. Plant Physiology 46: 69 80.
Krebs SL, Grumet R. 1993. Characterization of celery hydrolytic enzymes
induced in response to infection by Fusarium oxysporum. Physiological and
Molecular Plant Pathology 43: 193 208.
Lairini K, Prez-Espinoza A, Ruiz-Rubio M. 1997. Tomatinase induction in
formae speciales of Fusarium oxysporum non-pathogenic of tomato plants.
Physiological and Molecular Plant Pathology 50: 3752.
Lairini K, Ruiz-Rubio M. 1997. Detection of tomatinase from Fusarium
f. sp. lycopersici in infected tomato plants. Phytochemistry 45: 13711376.
Langin T, Capy P, Daboussi MJ. 1995. The transposable element impala,
a fungal member of the Tc1-mariner superfamily. Molecular and General
Genetics 246: 1928.
Larkin RP, Fravel DR. 1999. Mechanisms of action and doseresponse
relationships governing biological control of Fusarium wilt of tomato by
nonpathogenic Fusarium spp. Phytopathology 89: 1152 1161.
Lev S, Sharon A, Hadar RMH, Horwitz BA. 1999. A mitogen-activated
protein kinase of the corn leaf pathogen Cochliobolus heterotrophus is
involved in conidiation, appressorium formation, and pathogenicity,
diverse roles for mitogen-activated protein kinase homologs in foliar
pathogens. Proceedings of the National Academy Sciences, USA 96:
1354213547.

Liu S, Dean RA. 1997. G protein subunit genes control growth,

development, and pathogenicity of Magnaporthe grisea. Molecular
PlantMicrobe Interactions 10: 10751086.
Lorang JM, Tuori RP, Martinez JP, Sawyer TL, Redman RS, Rollins JA,
Wolpert TJ, Johnson KB, Rodriguez RJ, Dickman MB, Ciuffetti LM.
2001. Green fluoreescent protein is lighting up fungal biology. Applied and
Environmental Microbiology 67: 19871994.
Lorenz MC. 2002. Genomic approaches to fungal pathogenesis. Current
Opinion in Microbiology 5: 372378.
Ma Q, Gan J, Becker JO, Papiernik SK, Yates SR. 2001. Evaluation of
propargyl bromide for control of barnyardgrass and Fusarium oxysporum in
three soils. Pesticide Management Science 57: 781786.
Madrid MP, Di Pietro A, Roncero MIG. 2003. Class V chitin synthase
determines pathogenesis in the vascular wilt fungus Fusarium oxysporum
and mediates resistance to plant defence compounds. Molecular
Microbiology 47: 257266.
Mandeel Q, Baker R. 1991. Mechanisms involved in biological control
of fusarium wilt of cucuber with strains of nonpathogenic Fusarium
oxysporum. Phytopathology 81: 462469.
Marmeisse R, van den Ackerveken G, Goosen T, de Wit P, van den
Broek H. 1994. The in-planta induced ecp2 gene of the tomato pathogen
Cladosporium fulvum is not essential for pathogenicity. Current Genetics
26: 245250.
Matta A, Abbatista-Gentile I, Ferraris L. 1988. Stimulation of -13glucanase and chitinase by stresses that induce resistance to Fusarium wilt
in tomato. Phytopathology Mediterranean 27: 4550.
Meyer JA, Maraite H. 1967. Unusual requirement for a vitamin by strains
of Fusarium oxysporum f. sp. melonis. Nature 216: 7071.
Migheli Q, Laug R, Davire JM, Gerlinger C, Kaper F, Langin T,
Daboussi M. 1999. Transposition of the autonomous Fot1 element in
the filamentous fungi Fusarium oxysporum. Genetics 151: 10051013.
Migheli Q, Steinberg C, Davire JM, Olivain C, Gerlinger C,
Gautheron N, Alabouvette C, Daboussi MJ. 2000. Recovery of mutants
impaired in pathogenicity after transposition of impala in Fusarium
oxysporum. Phytopathology 90: 12791284.
Mller P, Aichinger C, Feldbrgge M, Kahmann R. 1999. The MAP kinase
Kpp2 regulates mating and pathogenic development in Ustilago maydis.
Molecular Microbiology 34: 10071017.
Mullins ED, Chen X, Romaine P, Raina R, Geiser DM, Kang S. 2001.
Agrobacterium-mediated transformation of Fusarium oxysporum: An
efficient tool for insertional mutagenesis and gene transfer. Phytopathology
91: 173180.
Namiki F, Matsunaga M, Okuda M, Inoue I, Nishi K, Fujita Y, Tsuge T.
2001. Mutation of an arginine biosynthesis gene causes reduced
pathogenicity in Fusarium oxysporum f. sp. melonis. Molecular
PlantMicrobe Interactions 14: 580584.
ODonnell K, Kistler HC, Cigelnik E, Ploetz RC. 1998. Multiple
evolutionary origins of the fungus causing Panama disease of banana:
concordant evidence from nuclear and mitochondrial gene genealogies.
Proceedings of the National Academy of Sciences, USA 95: 20442049.
Oeser B, Heidrich PM, Mller U, Tudzynski P, Tenberge KB. 2002.
Polygalacturonase is a pathogenicity factor in the Claviceps purpurea/rye
interaction. Fungal Genetics and Biology 36: 176186.
Olivain C, Alabouvette C. 1997. Colonization of tomato root by a
non-pathogenic strain of Fusarium oxysporum. New Phytologist 137:
481494.
Olivain C, Alabouvette C. 1999. Process of tomato root colonization by a
pathogenic strain of Fusarium oxysporum f. sp. lycopersici in comparison
with a non-pathogenic strain. New Phytologist 141: 497510.
Olivain C, Trouvelot S, Pugin A, Alabouvette C. 2001a. Rponse
diffrentielle des cellules de lin en interaction avec des souches pathognes
ou non pathognes d Fusarium oxysporum. 5e Congrs de la Socit
Franaise de Phytopathologie. Angers, France: City Press, 109.
Olivain C, Trouvelot S, Stawiecki K, Pugin A, Alabouvette C. 2001b.
Perte conjointe du pouvoir pathogne et du pouvoir

www.newphytologist.com New Phytologist (2003) 159: 73 92

Research review
antagoniste par mutation transpositionnelle chez. Fusarium
oxysporum f. sp. melonis. 5e Congrs de la Socit Franaise de
Phytopathologie. Angers, France: City Press, 143.
Oliver RP, Henricot B, Segers G. 2000. Cladosporium fulvum, cause of leaf
mould of tomato. In: Kronstad J, ed. Fungal pathology. Dortrecht, The
Netherlands: Kluwer Academic Publishers, 65 91.
Oliver R, Osbourn A. 1995. Molecular dissection of fungal
phytopathogenicity. Microbiology 141: 1 9.
Osbourn AE. 1996. Preformed antimicrobial compounds and plant defense
against fungal attack. Plant Cell 8: 18211831.
Pain NA, OConnell RJ, Mendgen K, Green JR. 1994. Identification of
glycoproteins specific to biotrophic intracellular hyphae formed in the
Colletotrichum lindemuthianumbean interaction. New Phytologist 127:
233242.
Perfect SE, OConnel RJ, Green EF, Doering-Saad C, Green JR. 1998.
Expression cloning of a fungal prolin-rich glycoprotein specific to the
biotrophic interface formed in the Colletotrichumbean interaction. Plant
Journal 15: 273279.
Pieterse CMJ, Riach MBR, Bleker T, Van den Berg-Velthuis GCM,
Govers F. 1993. Isolation of putative pathogenicity genes of the potato late
blight fungus Phytophtora infestans by differential screening of a genomic
library. Physiological and Molecular Plant Pathology 43: 69 79.
Prosser JI. 1995. Kinetics of filamentous growth and branching. In: Gow
NAR, Gadd GM, eds. The growing fungus. 14. London, UK: Chapman &
Hall, 301318.
Recorbet G, Alabouvette C. 1997. Adhesion of Fusarium oxysporum conidia
to tomato roots. Letters in Applied Microbiology 25: 375 379.
Recorbet G, Bestel-Corre G, Dumas-Gaudot E, Gianinazzi S,
Alabouvette C. 1998. Differential accumulation of -1,3-glucanase and
chitinase isoforms in tomato roots in response to either pathogenic or
non-pathogenic strains of Fusarium oxysporum. Microbiological Research
153: 257263.
Recorbet G, Gane A, Dumas-Gaudot E, Alabouvette C, Gianinazzi S.
2001. Comparaison des interactions plante Fusarium oxysporum
pathognes et plante-F. oxysporum non pathognes par analyse protique
diffrentielle. 5e Congrs de la Socit Franaise de Phytopathologie. Angers,
France: City Press, 90.
Redman RS, Ranson JC, Rodriguez RJ. 1999. Conversion of a pathogenic
fungus Colletotrichum magna to a nonpathogenic endophytic mutalist by
gene disruption. Molecular PlantMicrobe Interactions 12: 969 975.
Rep M, Dekker HL, Vossen JH, de Boer AD, Houterman PM, Speijer D,
Back JW, de Koster CG, Cornelissen BJC. 2002. Mass spectrometric
identification of isoforms of PR proteins in xylem sap of fungus-infected
tomato. Plant Physiology 130: 904 917.
Rodriguez-Glvez E, Mendgen K. 1995. The infection process of Fusarium
oxysporum in cotton root tips. Protoplasma 189: 6172.
Rohel EA, Payne AC, Fraaije BA, Hollomon DW. 2001. Exploring
infection of wheat and carbohydrate metabolism in Mycosphaerella
graminicola transformants with differentially regulated green fluorescent
protein expression. Molecular PlantMicrobe Interactions 14:
156163.
Roldn-Arjona T, Prez-Espinoza A, Ruiz-Rubio M. 1999. Tomatinase
from Fusarium oxysporum f. sp. lycopersici defines a new class of
saponinases. Molecular PlantMicrobe Interactions 12: 852 861.
Safe LM, Safe SH, Subden RE, Morris DC. 1997. Sterol content and
polyene antibiotic resistance in isolates of Candica krusei, Candida
parakrusei, and Candida tropicalis. Canadian Journal of Microbiology 23:
398401.
Sanchez LE, Leary JV, Endo RM. 1975. Chemical mutagenesis of
Fusarium oxysporum f. sp. lycopersici: non-selected changes in pathogenicity
of auxotrophic mutants. Journal of General Microbiology 87:
326332.
Sandrock RW, VanEtten HD. 1998. Fungal sensitivity to and enzymatic
degradation of the phytoanticipin d-tomatine. Phytopathology 88:
137143.

New Phytologist (2003) 159: 73 92 www.newphytologist.com

Review

Segers G, Bradshaw N, Archer D, Blisset K, Oliver RP. 2001. Alcohol

oxidase is a novel pathogenicity factor for Cladosporium fulvum, but
aldehyde dehydrogenase is dispensable. Molecular PlantMicrobe
Interactions 14: 367377.
Shieh M-T, Brown RL, Whitehead MP, Cary JW, Cotty PJ, Cleveland TE,
Dean RA. 1997. Molecular genetic evidence for the involvement of a
specific polygalacturonase, Pc2, in the invasion and spread of Aspergillus
flavus in cotton bolls. Applied and Environmental Microbiology 63:
35483552.
Skovgaard K, Rosendahl S. 1998. Comparison of intra- and extracellular
isozyme banding patterns of Fusarium oxysporum. Mycological Research
102: 10771084.
Solomon PS, Nielsen PS, Clark AJ, Oliver RP. 2000. Methionine synthase,
a gene required for methionine synthesis, is expressed in planta by
Cladosporium fulvum. Molecular Plant Pathology 1: 315323.
Steinberg C, Whipps JM, Wood D, Fenlon J, Alabouvette C. 1999a.
Mycelial development of Fusarium oxysporum in the vicinity of tomato
roots. Mycological Research 103: 769778.
Steinberg C, Whipps JM, Wood DA, Fenlon J, Alabouvette C. 1999b.
Effect of nutritional sources on growth of one non-pathogenic strain and
four strains of Fusarium oxysporum pathogenic on tomato. Mycological
Research 103: 12101216.
Suleman P, Tohamy AM, Saleh AA, Madkour MA, Straney DC. 1996.
Variation in sensitivity to tomatine and rishitin among isolates of Fusarium
oxysporum f. sp. lycopersici, and strains not pathogenic on tomato.
Physiological and Molecular Plant Pathology 48: 131144.
Sweigard JA, Carrol AM, Farral L, Chumley F, Valent B. 1998.
Magnaporthe grisea pathogenicity genes obtained through
insertional mutagenesis. Molecular PlantMicrobe Interactions 11: 404
412.
Takano Y, Kikucki T, Kubo Y, Hamer JE, Mise K, Furusawa I. 2000. The
Colletotrichum lagenarium MAP kinase gene CMK1 regulates diverse
aspects of fungal pathogenesis. Molecular PlantMicrobe Interactions 13:
374383.
Talbot NJ, Ebbole DJ, Hamer JE. 1993. Identification and characterization
of MPG1, a gene involved in pathogenicity from the rice blast fungus
Magnaporthe grisea. Plant Cell 5: 15751590.
Talbot NJ, McCafferty HRKMM, Moore K, Hamer JE. 1997. Nitrogen
starvation in the rice blast fungus Magnaporthe grisea may act as an
environmental cue for disease symptom expression. Physiological and
Molecular Plant Pathology 50: 179195.
Tamietti G, Ferraris L, Matta A, Abbatista-Gentile I. 1993. Physiological
responses of tomato plants grown in Fusarium suppressive soil. Journal of
Phytopathology 138: 6676.
Thanonkeo P, Akiyama K, Jain S, Takata R. 2000. Targeted disruption of
sti35, a stress-responsive gene in phytopathogenic fungus Fusarium
oxysporum. Current Microbiology 41: 284289.
Tonukari NJ, Scott-Craig JS, Walton JD. 2000. The Cochliobolus carbonum
SNF1 gene is required for cell wall-degrading enzyme expression and
virulence in maize. Plant Cell 12: 237247.
Trouvelot S, Olivain C, Recorbet G, Migheli Q, Alabouvette C. 2002.
Recovery of Fusarium oxysporum Fo47 mutants affected in their biocontrol
activity after transposition of the Fot1 element. Phytopathology 92:
936945.
Truesdell GM, Yang Z, Dickman MB. 2000. A G alpha subunit gene from
phytopathogenic fungus Colletotrichum trifolii is required for conidial
germination. Physiological and Molecular Plant Pathology 56:
131140.
VanEtten HD, Mansfield JW, Bailey JA, Farmer EE. 1994. Two classes
of antibiotics: Phytoalexins versus phytoanticipins. Plant Cell 6: 1191
1192.
Villalba F, Lebrun M-H, Hua-Van A, Daboussi MJ, GrosjeanCournoyer M-C. 2001. Transposon impala, a novel tool for gene tagging
in the rice-blast fungus Magnaporte grisea. Molecular PlantMolecular
Interactions 14: 308315.

91

92 Review

Research review

Wasmann CC, VanEtten HD. 1995. Transformation-mediated

chromosome loss and disruption of a gene for pisatin demethylase decrease
the virulence of Nectria haematococca on pea. Molecular PlantMicrobe
Interactions 9: 793803.
Wessels JGH, Sietsma JH. 1981. Fungal cell walls: a survey. In: Tanner W,
Loewus A, eds. Encyclopedia of plant physiology, news series, plant
carbohydratess II. 13B. Berlin, Germany: Springer-Verlag, 352 394.

Xu J-R, Hamer JE. 1996. MAP kinase and cAMP signaling regulate
infection structure formation and pathogenic growth in the rice
blast fungus Magnaporthe grisea. Genes and Development 10: 2696
2706.
Zheng L, Campbell M, Murphy J, Lam S, Xu J-R. 2000. The BMP1 gene
is essential for pathogenicity in the gray mold fungus Botrytis cinerea.
Molecular PlantMicrobe Interactions 13: 724732.

www.newphytologist.com New Phytologist (2003) 159: 73 92