- Top of page
- Materials and Methods
- Supporting Information
The hyphomycete Botrytis cinerea is widely known as a fungus causing destructive and economically important plant diseases. It is concentrated in the temperate areas of the world where it can infect an extremely wide range of host plants (Jarvis, 1977; Elad et al., 2004). B. cinerea is also noteworthy as a spoilage organism causing considerable losses of plants during storage and transit (Hammer et al., 1990; De Kock & Holz, 1992; Berrie, 1994; Elad, 1998; Droby & Lichter, 2004). Disease control is extremely difficult because the fungus is capable of attacking crops at almost any stage in their growth or storage, and affects all plant parts, including cotyledons, leaves, stems, flowers, fruits and roots. Chemical control remains the main way to reduce B. cinerea disease (Leroux, 2004).
B. cinerea enters a plant via direct penetration or through natural openings or wounds (Verhoeff, 1980). As a necrotroph, it then derives nutrients from dead or dying cells (Agrios, 1997), and the colonization of the plant therefore depends upon the ability of the fungus to kill the host cells. Many factors potentially involved in the pathogenicity of the fungus have been studied (for a review, see Staples & Mayer, 1995). Several genes that code for these factors have been cloned and the corresponding mutants have been analysed (for reviews, see Prins et al., 2000; Kars & van Kan, 2004). The tetraspanin BcPls1 (Gourgues et al., 2004) and the Mitogen-Activated Protein (MAP) kinase BMP1 (Zheng et al., 2000) are two factors involved in penetration, and mutants of either gene are nonpathogenic. Recently, components of some signalling pathways have also been associated with virulence (Schulze Gronover et al., 2001; Klimpel et al., 2002; Viaud et al., 2003) and a mutation in one of them (bcg1) leads to a nonpathogenic strain impaired in plant colonization (Schulze Gronover et al., 2001). Several virulence factors involved in symptom development have also been described, but none of them is required for pathogenicity, probably because of gene redundancy. Among these factors are cell wall degrading enzymes (Ten Have et al., 1998; Valette-Collet et al., 2003), transporter proteins (Schoonbeek et al., 2001; Hayashi et al., 2002) and enzymes protecting the fungus from oxidative stress (Schouten et al., 2002; Rolke et al., 2004).
Production of phytotoxins by B. cinerea has been proposed to promote cell death in the host (Rebordinos et al., 1996; Colmenares et al., 2002). Further, Govrin & Levin (2000) suggested that induced cell death is related to the plant hypersensitive response (HR), and that its triggering by the fungus constitutes an important component of virulence. HR, however, is often associated with plant resistance to pathogens (for a review, see Lam et al., 2001) and a delicate balance between the attack of the pathogen and the defence of the host could control the outcome of the interaction.
Here we report on a nonpathogenic mutant of B. cinerea that fails to overcome the defence system of the plant and whose interaction with its host results in HR-like necrosis. The phenotype of the mutant is different from any other described before. The mutant is blocked after penetration and does not colonize the plant tissue. It is deficient in oxalic acid production and does not produce the nonaspartyl acid protease 1 (ACP1). Moreover, it produces the Botrytis cinerea endopolygalacturonase 1 (BcPG1) elicitor in large amounts relative to the wild-type strain.
- Top of page
- Materials and Methods
- Supporting Information
Total loss of pathogenicity in B. cinerea has been reported only rarely (Zheng et al., 2000; Schulze Gronover et al., 2001; Gourgues et al., 2004). Gene redundancy has frustrated attempts to produce nonpathogenic strains via disruption of a single gene or of small numbers of genes, and little is known today about the master regulations that govern the virulence process of this fungus.
In this paper, we describe the mutant A336 whose ability to infect plant tissues as diverse as apple fruit and grapevine, bean and A. thaliana leaves has been lost. When inoculated as mycelium onto onion epidermis, the mutant is able to penetrate the host but it forms penetration structures less frequently than the wild-type strain. Recent studies by Viaud et al. (2003) and Gourgues et al. (2004) showed that germinated conidia of B. cinerea form appressoria on onion epidermis, without previous ramification of germ tubes. In our study, using mycelium for inoculation, we have shown that the hyphae densely ramify in a claw-like structure, swell at their tips (appressoria) and then penetrate. This claw-like structure does not form when conidia of the same wild-type strain (Bd90) are used (Souliéet al., 2006). To our knowledge, observation of this claw-like structure has not previously been reported, and suggests the existence of different methods of penetration for B. cinerea that have until now not been recognized.
The mutant has completely lost the ability to invade the host after penetration. Penetration and colonization of the plant constitute two distinct stages of the B. cinerea infection process (Tenberge, 2004), and mutant A336 hence appears only to be affected in the second stage of infection. So far, only one nonpathogenic mutant capable of penetration but impaired in tissue colonization has been described (Schulze Gronover et al., 2001); this mutant lacks one subunit of a GTP-binding protein. Interestingly, mutant A336 differs from this mutant in its inability to produce conidia under normal in vitro growth conditions, and different components required for plant colonization are hence likely to be affected in the two strains.
HR-like lesions are visible on both bean and grapevine leaves at the penetration sites of the mutant, but are not visible in infection by the wild-type strain. Strong fluorescence can be observed in the plant cells neighbouring the penetrating mutant hyphae and significant production of hydrogen peroxide can be detected in culture of grapevine cells treated with dialysed fungal culture filtrates of the mutant. The accumulation of fluorescent phenolic compounds has often been observed in relation to efficient plant resistance to pathogens (Koga et al., 1988), and the early production of AOS has been proposed to play a role in the onset of HR (Dixon et al., 1994; Levine et al., 1994; Jabs, 1999). On the one hand, Govrin & Levine (2000) proposed that B. cinerea triggers the HR to facilitate its spread within dead plant tissues. Dickman et al. (2001) also showed that inhibition of HR-related cell death in tobacco leads to enhanced resistance to B. cinerea. On the other hand, Derckel et al. (1999) showed that the aggressiveness of different B. cinerea wild-type strains correlated inversely with the strength of induction of plant defences. Similarly, Keller et al. (1999) showed that transgenic tobacco plants Nicotiana Tabacum L. responding to pathogen attack with localized HR-like necrosis were more resistant to B. cinerea. Finally, Unger et al. (2005) showed that the aggressiveness of B. cinerea is linked to an oxidative burst-suppressing agent that suppresses HR. The precise role of HR in the interaction between B. cinerea and its hosts thus remains to be elucidated. Our results are consistent with a difference in the mutant–plant interaction that leads to the development of a HR-like response that stops the progression of the mutant. This could arise from a stronger and faster reaction of the plant to the mutant than to the wild type, from a weakness of the mutant in overriding the plant defence system, or from both. Indeed, B. cinerea infection may require cell death in the plant to be able to proceed, but the timing and/or the relative strengths of the virulence and defence reactions may also affect the balance between the fungus and its host.
Mutant A336 is greatly affected in the production of oxalic acid, making it the first oxalate-deficient B. cinerea strain to be described. This deficiency explains the poor growth of the mutant on PGA medium as it prevents the acidification of the medium and, consequently, the action of the polygalacturonases needed to degrade the substrate (Favaron et al., 2004; Kars et al., 2005). Although the role of oxalic acid in the pathogenicity of B. cinerea is not yet well defined, it has been reported that oxalic acid chelates Ca2+ ions from pectin in the plant middle lamellae and in doing so facilitates the degradation of the plant cell wall by fungal enzymes (Bateman & Beer, 1965; Marciano et al., 1983). In S. sclerotiorum, a necrotrophic fungus closely related to B. cinerea, a key role has been postulated for oxalic acid in pathogenesis and plant maceration (Godoy et al., 1990, Guimaraes & Stotz, 2004) and its suppressor effect on the oxidative burst of the plant has also been shown (Cessna et al., 2000). Finally, its role in plant pathogenicity has been shown in other fungi from various taxonomic classes (Dutton & Evans, 1996). The oxalate deficiency of mutant A336 therefore could contribute to its lack of virulence.
Mutant A336 is also affected in the production of extracellular enzymes; some of these enzymes, such as phosphatases and the endopolygalacturonase BcPG1, are produced in larger amounts than in the wild-type strain, while others, such as the acid protease ACP1, are not produced at all. This is also likely to contribute to the absence of virulence of the mutant strain. On the one hand, fungal secreted proteases participate in the degradation of the plant tissues during infection and could weaken the plant defence system by targeting the plant defence proteins. The total absence of ACP1 in the mutant could hence diminish its proteolytic potential. On the other hand, BcPG1 has been shown to activate defence reactions against B. cinerea in grapevine (Poinssot et al., 2003) and therefore appears to act both as a lytic enzyme for the fungus (Ten Have et al., 1998; Kars et al., 2005) and as an elicitor for the plant (Poinssot et al., 2003). In this study, BcPG1 was shown to be abundant in the mutant A336 culture filtrates, and these filtrates triggered a strong oxidative burst in grapevine cells that could be largely reduced by the specific BcPG1 inhibitor PGIP2. The overproduction of BcPG1 in the mutant therefore is likely to boost the plant defence reaction and to act against the fungus during the interaction with the plant.
As ACP1 is essentially produced under acidic conditions (Poussereau et al., 2001), and as oxalate production is impaired in the mutant, we contemplated the possibility that a link existed between the modification of the production of this enzyme and the absence of acid secretion during pathogenesis of the mutant. We also investigated whether a perturbation of the pH sensing system resulted in reduced oxalate and altered enzyme production. However, pH regulation of phosphatase production is not disturbed in mutant A336 and addition of oxalic acid alongside the mutant fungus during pathogenicity tests did not restore virulence (Fig. 7). The absence of oxalate production alone therefore does not seem to be responsible for the nonpathogenic phenotype of the mutant, and pH sensing seems to be functional in this strain.
Figure 7. Maceration symptoms on grapevine (Vitis vinifera cv. Chardonnay) leaf discs in the presence of oxalic acid. Mycelium plugs of Botrytis cinerea strain Bd90 and mutant A336 were applied to the upper side of unwounded leaf discs placed on oxalic acid solutions: (a) 0 mm, (b) 1 mm and (c) 10 mm. The oxalic acid solutions had been adjusted beforehand to pH 5. Maceration symptoms are visible as dark tissue around the mycelium plug for wild-type strain Bd90.
Download figure to PowerPoint
One feature of mutant A336 is an inability to produce sclerotia and to show a modified pattern of conidiation. Conidia do not form when the fungus grows on rich medium but do form when the mycelium is placed under conditions in which it cannot grow (on plant tissues or on PGA medium at pH 7). As deficiency in oxalate production is associated with the loss of sclerotia production in S. sclerotiorum (Godoy et al., 1990), the absence of sclerotia in the mutant is also likely to result from its lack of acid production. This strongly supports the existence of a link between oxalate and sclerotia development whose molecular basis is completely unknown. Conidia formation has mainly been studied in Aspeigillus nidulans (for a review, see Adams et al., 1998) and Neurospace crassa (for a review, see Springer, 1993). In the former, it requires the acquisition of developmental competence (Champe et al., 1981), developmental induction (Morton, 1961; Mooney & Yager, 1990) and/or the signal for starvation-related stress (Skromne et al., 1995). In N. crassa, it can be triggered by a drop in nutrients in the growth medium (Toledo & Hansberg, 1995). Two different pathways therefore seem to control this process: one involved in the setting up of a developmental programme, and one involved in the reaction to a stress. The involvement of G-protein signalling and regulators in the first pathway in A. nidulans has been recently demonstrated (for a review, see Yu & Keller, 2005) while components of the stress counterpart await identification. Our results on mutant A336 suggest that this strain is only impaired in the first conidiation-related signalling pathway. The mutant is thus affected in the unfolding of two developmental programmes that lead to either sclerotia or conidia formation. The possibility that oxalate could also play a role in conidiation is ruled out by the failure to restore the production of conidia in the mutant by adding oxalate to the culture medium.
Overall, our data describe a mutant deregulated in several aggressiveness factors, leading to total loss of pathogenicity accompanied by HR-like lesions. We cannot, at present, exclude the possibility that other aggressiveness factors not described in this work are also disturbed in mutant A336. The genetic analysis demonstrated that all the characteristics of the mutant cosegregated and only one locus (BcPTH1) is therefore responsible for the observed phenotype. As this phenotype is pleiotropic, we propose that a gene regulator might be mutated in this strain. We established that the mutation is not tagged by the plasmid used for the mutagenesis, a situation that has previously been reported by Tudzynski & Siewers (2004). Further, the lack of an established system for functional complementation studies in this fungus prevented us from identifying the mutated gene(s). However, the recent advances in B. cinerea complete genome sequencing (http://www.ncbi.nlm.nih.gov/genomes/leuks.cgi?p3=11:Fungi&taxgroup=11:Fungi|12: and http://www.genoscope.cns.fr/externe/English/Projets/Projet_LN/LN.html) will make the identification of the mutated gene in locus BcPTH1 possible. Indeed, microsatellite markers are currently being developed to build a genetic map of the B. cinerea genome. This will allow a positional cloning strategy to identify the genomic region cosegregating with the pathogenicity phenotype. Furthermore, because these markers will be anchored on the B. cinerea complete genome sequence, the search for potential candidate genes for the mutated gene in mutant A336 will be facilitated. It is also possible that the overall effect of the mutation on metabolic and signalling pathways might be elucidated by transcriptome and metabolome analysis. Thus the fungal component(s) revealed by this study, which clearly plays an essential role in virulence, awaits further characterization.