Microbially induced diseases of Agaricus bisporus: biochemical mechanisms and impact on commercial mushroom production
- First Online:
- Cite this article as:
- Largeteau, M.L. & Savoie, J. Appl Microbiol Biotechnol (2010) 86: 63. doi:10.1007/s00253-010-2445-2
- 565 Views
The button mushroom, Agaricus bisporus (Lange) Imbach, the most common cultivated mushroom, is susceptible to a wide range of virus, bacterial, and fungal diseases. However, only some diseases were studied for the mechanisms involved in the host–microorganism interaction. This review deals with biochemical mechanisms related to cavity disease (Burkholderia gladioli) and to the interaction between A. bisporus and the causal agents responsible for the most severe diseases, namely the bacteria Pseudomonas tolaasii and Pseudomonas reactans and the fungi Trichoderma aggressivum and Lecanicillium fungicola.
KeywordsBacterial blotchCavity diseaseDry bubbleGreen mould
The button mushroom, Agaricus bisporus (Lange) Imbach, is the most common cultivated mushroom and is consumed throughout the world. Commercial strains produce substantial yield and exhibit attractive morphology and texture, but they are susceptible to a variety of viral, bacterial and fungal diseases (Fletcher et al. 1989). The first step of A. bisporus cultivation, named spawn run, consists in compost colonisation by the mushroom mycelium. In compost, the vegetative mycelium competes with various microorganisms. Olive-green mould (Chaetomium olivaceum), lipstick mould (Sporendonema purpurescens), sepedonium yellow mould (Sepedonium chrysospermum) and plaster moulds (Scopulariopsis fimicola and Papulaspora byssina) can develop during spawn run, affect mycelial growth and when extensive in compost, reduce yield. Such diseases establish in poor quality substrates. Standardisation of compost composition and composting processes has considerably reduced the occurrence of these moulds in mushroom crops. Especially, computer-assisted management of temperature kinetic and moisture during substrate pasteurisation has considerably improved destruction of competitor and pathogen spores. Selection of productive spawn, fungicide application to the spawn and temperature control during spawn run was found important in disease prevention (Seaby 1996; Rinker and Alm 2000; Boiko et al. 2009). Nevertheless, green mould disease caused by Trichoderma aggressivum is still a major disease in mushroom growing regions worldwide. In recent years, mushroom virus X complex has been reported in the UK mushroom industry (Grogan et al. 2003) and a number of mushroom growing countries. It affects the vegetative mycelium and sporophores growing around bare cropping areas. After spawn run, mushroom production is obtained by covering the colonised compost with a casing soil containing peat moss. Many diseases can affect the sporophore during its development. Cobweb disease (Cladobotryum dendroides), characterised by the growth of coarse mycelium covering affected mushrooms and causing sporophore decay, is not uncommon, but rarely epidemic. Two fungal diseases characterised by anamorphous masses growing in place of sporophores, wet bubble disease (Mycogone perniciosa) and dry bubble disease (Lecanicillium fungicola), are worldwide in distribution, and the latter is responsible for severe outbreaks. Bacterial diseases affect mushroom sporophores as well. Two bacteria, Burkholderia gladioli pv. agaricicola and Janthinobacterium agaricidamnosum sp. nov., are responsible for soft rot, a disease sporadically observed in European farms (Lincoln et al. 1991, 1999; Gill and Tsuneda 1997) but with devastating effects within very short period of time. B. gladioli is also the causal agent of cavity disease. Several A. bisporus diseases are caused by pseudomonads. Pseudomonas agarici is responsible for drippy gills symptoms but also for brown discolouration as recently observed in Italy and the Netherlands (Lo Cantore and Iacobellis 2004; Geels et al. 2008). Pseudomonas gingeri is the causal agent of ginger blotch, Pseudomonas aeruginosa and Pseudomonas fluorescens bv V that of mummy disease reported as sporadic or occasionally endemic. Pseudomonas tolaasii and pathogenic Pseudomonas reactans cause brown blotch disease and are responsible for severe outbreaks throughout the world. In western countries, the average annual damage due to P. tolaasii, L. fungicola, Trichoderma sp. and viruses count for approximately 25% of the total production value, the bacterium and fungi being the major pathogens (Soković and van Griensven 2006). Besides symptoms observed on mushrooms beds, pathogenic bacteria or fungal spores can be present on sporophores visually healthy at harvest and develop during shelf life causing mushroom quality depreciation. Effect of compost improvement and pesticides to reduce the different diseases was intensively studied. Many competitors and pathogens were analysed for diversity and aggressiveness. However, the mechanisms of host–microorganism interaction were only studied for some diseases. This review deals with biochemical mechanisms involved in cavity disease and in the interaction between A. bisporus and the causal agents responsible for the most severe diseases, namely the bacteria P. tolaasii and P. reactans and the fungi T. aggressivum and L. fungicola.
Cavity disease caused by B. gladioli pv. agaricicola (formerly Pseudomonas cepacia then Pseudomonas gladioli pv. agaricicola) was first reported on Agaricus bitorquis (Gill and Cole 1992). Symptoms exhibited by affected sporophores range from mild blotching to deep pitting, where large pervasive cavities extending from the cap surface to the stipe may form. Gill and Tsuneda (1997) hypothesised that the expression of the disease is a combined effect of mushroom tissue-degrading enzyme (chitinase and β-glucanase) and toxin(s). To characterise avirulent mutants of B. gladioli pv. agaricicola, Roy Chowdhury and Heineman (2006) investigated the type II secretion (T2S) pathway, known to be involved in bacteria virulence through the excretion of extracellular enzymes and toxins. The T2S system is apparently responsible for secreting the protease and chitinase that are necessary for disease symptoms. However, whether the two enzyme activities are components of the virulence factors remains undetermined. The avirulent mutants still inhibited mushroom mycelia proving that the mechanism of sporophore degradation is different from the inhibition mechanism of mushroom mycelial growth suggested by Gill and Tsuneda (1997). Among the indigenous flora colonising the mushroom, an Ewingella americana isolate significantly inhibited cavity disease, possibly by resource competition or by expression of a toxin toward B. gladioli. Interestingly, this isolate did not cause any necrotic symptoms (Roy Chowdhury and Heineman 2006) whilst internal stipe necrosis of A. bisporus caused by E. americana was reported by Inglis et al. (1996). These observations are related with the role of pathogenic and not pathogenic P. reactans in brown blotch disease discussed hereafter. Biological control of cavity disease using competitive bacteria is a promising way to mushroom protection, but the development of commercial products needs further works.
Bacterial blotch diseases
Brown blotch of A. bisporus, characterised by discoloured sunken lesions on the mushroom cap and stipe, is a complex disease. Besides P. tolaasii, the major causal agent, other Pseudomonas spp. account for symptom development. P. reactans (Iacobellis and Lo Cantore 2003), Pseudomonas sp. strain NZ17 apparently closely related to Pseudomonas syringae (Godfrey et al. 2001) and Pseudomonas costantinii (Munsch et al. 2002) have been reported to cause blotch symptoms.
Interaction between A. bisporus and P. tolaasii
The bacterium P. tolaasii is endemic in compost and casing soil and is part of the flora associated with the mycelium. The pathogen has flagella, chemotactic and surface active properties which allowed it to attach very rapidly to the host surface. The bacterium can switch between pathogenic and non-pathogenic form, and under certain conditions the pathogenic form becomes dominant (Soler-Rivas et al. 1999a, for a review). Some observations suggested that P. tolaasii switch might be under the influence of A. bisporus metabolites. This was demonstrated with another mushroom, Pleurotus ostreatus, whose components significantly reduced the occurrence of avirulent variants (Murata et al. 1998). The mechanism of A. bisporus infection by P. tolaasii is well documented. Pathogenic P. tolaasii produces extracellular toxins, called tolaasin, responsible for symptom development, which is not secreted by the non-pathogenic bacterium. Tolaasin appeared to be synthesised by a peptide synthetase complex (Rainey et al. 1993) and its production and efficiency is mediated by an extragenomic factor (Mamoun et al. 1997). Nutkins et al. (1991) first identified two components of the toxin, designed Tol I and Tol II. Tol I is a lipodepsipeptide (LDP) composed of 18 aminoacids, 11 of them are in the D-form. The primary structure bears two deshydroaminobutyric acid residues (Δ but), a β-hydroxyoctanoic acid moiety linked to the N-terminus. A lactone ring is formed between the d-Thr14 and the C-terminus. Tol II differs from Tol I in the replacement of Hse16 with Gly16. Later, Bassarello et al. (2004) isolated five other lipodepsipeptides, called tolaasin A–E. These LDP show difference in the peptide moiety and maintain the β-hydroxyoctanoic ø chain at the N-terminus, except tolaasin A which is characterised by a γ-carboxybutanoyl ø moiety. Tolaasin is able to disrupt the A. bisporus plasma membrane and vacuole membranes (Rainey et al. 1991). The ion channel forming activity and the surface active properties of tolaasin were both required to lyse mushroom membranes. Biosurfactant activity was related to the α-helical segment and the presence of the β-hydroxyoctanoic acid (Hutchinson and Johnsone 1993). Two types of ion channels, one more frequent and both inhibited by Zn2+, were identified by the incorporation of tolaasin into lipid bilayer (Cho and Kim 2003), and Ni2+ inhibited the pore activity of tolaasin through an unknown mechanism (Choi et al. 2009). Using model membranes of different lipid composition, Coraiola et al. (2006) proposed a barrel-stave mechanism of action of Tol I based on a valuable increment in the helical content of the LDP which was inserted in the membrane core and oriented parallel to the lipid acyl chain. Using in silico docking calculations, NMR binding assays and in vitro hemolysis activity assays, Lee et al. (2009) identified a detergent, the sorbitololeic acid that can act as Tol I inhibitor, probably by interrupting the multimerisation of tolaasin molecules which is a prerequisite for toxicity.
Membrane disruption by the toxin allows contact between tyrosinases and their substrates, leading to the activation of the AbPPO2 tyrosinase gene (Soler-Rivas et al. 1999a, 2001) and the synthesis of melanin forming a chemical barrier which appeared to prevent infection of other cells (Cole and Skellerup 1986). P. tolaasii lipase (Baral and Fox 1997) and proteinase (Baral et al. 1995) could provoke mushroom membrane breakdown and a switch of latent tyrosinases to their active form as commonly observed in affected mushrooms.
Tyrosinase was not included in the model because its activity did not correlate with the susceptibility of the strains. In addition, no relationship was found between the quantitative trait locus (QTL) of resistance to P. tolaasii and the tyrosinase gene involved in the browning process (Moquet et al. 1999), but activation of latent tyrosinases is a major step after infection. Breeding for A. bisporus resistance using the knowledge presented above is currently a means that mushroom breeders can use to control bacterial blotch disease.
Interaction between A. bisporus and P. reactans
The pathogenic strains of P. reactans, responsible for brown blotch, produce an extracellular substance called the white line-inducing principle (WLIP). Like tolaasin, WLIP is a LDP composed of N-terminal β-hydroxydecanoic acid and a peptide moiety of nine aminoacids, six of which are in the D-form. The molecule contains a lactone ring between D-allo-threonine and N-terminal L-isoleucine (Mortishire-Smith et al. 1991). WLIP caused discoloured and sunken lesions on mushroom sporophores but it was less active than Tol I. Nevertheless, at least in vitro, P. reactans NCPPB1311 produced far more WLIP than the quantity of Tol I produced by the P. tolaasii-type strain NCPPB2192, suggesting that the lower antifungal activity of WLIP was compensated by the higher quantity produced in culture (Lo Cantore et al. 2006). A comparative evaluation of Tol I and WLIP on lipid membranes showed that both LDPs were able to damage biological membranes through the formation of transmembrane pores, but some interesting differences appeared. The conformation of WLIP changed slightly when it passed from the buffer solution to the lipid environment. The LDP had an insufficient length to pass through the entire membrane and exhibited a permeabilizing activity in the same range of that of detergents, suggesting a detergent-like activity for WLIP (Coraiola et al. 2006). The observation that avirulent variants of P. reactans failed to produce WLIP supported the possibility that the LDP may play an important role in the infection of A. bisporus (Lo Cantore et al. 2006). Infection of mushroom caps with pathogenic P. reactans had no effect on the tyrosinase activity of A. bisporus, and discolouration caused by P. tolaasii and P. reactans can be distinguished by chromametric measurements (Soler-Rivas et al. 2000) suggesting that both bacteria induced different mechanisms of response in the host.
Interaction between P. reactans and blotch-causing pseudomonads
Pre-treatment with pure WLIP protected A. bisporus from brown blotch symptoms induced by P. tolaasii. An in vivo mechanism for the action of WLIP may be to sequester tolaasin, removing the ability of the toxin to disrupt the mushroom membrane (Soler-Rivas et al. 1999b). This hypothesis is based on the fact that a specific reaction occurred between WLIP and tolaasin, resulting in a white precipitate (Wong and Preece 1979). This reaction was revealed in vitro through the ‘white line test’ (WLT) and used to identify the pathogenic form of P. tolaasii. The white line production also suggested that WLIP might scavenge tolaasin.
Interestingly, P. costantinii, another agent causing brown blotch symptoms, produced a typical white line when streaked toward P. reactans, and non-pathogenic variants were identified (Munsch et al. 2002). This suggested that the bacterium produced a LDP like-toxin potentially involved in the infectious process and switched from the pathogenic to the non-pathogenic form, as observed for P. tolaasii and P. reactans. A protective effect of WLIP against P. costantinii might be hypothesised.
Pseudomonas ‘NZ,’ negative in the WLT with either P. reactans or P. tolaasii, caused browning not due to tolaasin production but produced factor(s) that might be a protease(s) (Godfrey et al. 2001).
Biological control of bacterial blotch diseases
The use of P. reactans or purified WLIP as control agent of the bacterial blotch disease has been studied, but to our knowledge, no commercial product has been developed to date. Other possibilities of biological control were studied. Munsch and Olivier (1995) proposed the bacteriophage TO1 isolated from the mushroom surface for partial control of the disease. Several authors investigated non-fluorescent Pseudomonas and P. fluorescens protective effect (Soler-Rivas et al. 1999b). Tsukamoto et al. (2002) found, attached to the surface of wild Agaricales, non-pseudomonads bacteria that efficiently detoxified tolaasin and had potential to suppress the occurrence of brown blotch disease. With one exception, the bacteria did not produce extracellular proteases. Importance should be paid to elucidate the detoxifying mechanism, as the bacteria might contribute to the mushroom defence in commercial crops.
More recent work proposed the possibility of phytochemical control of the disease. Extracts of the medicinal plant Salvia miltiorrhiza decreased the incidence of P. tolaasii infection when incorporated in the mushroom compost (Dawoud and Eweis 2006).
Green mould disease
Trichoderma species such as Trichoderma koningii and Trichoderma viride were long associated with mushroom cultivation, but their impact on production was relatively minor. In the mid-1980s and 1990s, T. aggressivum emerged as an aggressive compost mould that decimated mushroom production if it got into freshly spawned compost (Grogan 2008). Two populations of T. aggressivum genetically distinct but closely related, formerly named T. harzianum Th2 and Th4 and reclassified as T. aggressivum f. aggressivum (Ta4) and Trichoderma aggressivum f. europaeum (Ta2), cause green mould disease in North America and Europe, respectively (Ospina-Giraldo et al. 1998; Samuels et al. 2002). Infected crops show a dark green colour indicative of the T. aggressivum sporulation on the compost surface. Infected areas produce no mushrooms and are surrounded by mushrooms infested by red-pepper mites (Pygmephorus spp.) (Krupke et al. 2003). Other Trichoderma species cause less and/or occasional damage in mushroom crops as they are less adapted than T. aggressivum to grow in compost colonised by A. bisporus. Four Ta2 and four poorly aggressive Trichoderma spp. did not differ for the activity of 17 secreted enzymes, but Ta2 growth was affected by a smaller number of bacteria isolated from the compost (Savoie et al. 2001a). The authors concluded that the better adaptation of T. aggressivum to mushroom compost was not related to a specific ability to degrade the compost components but to a higher tolerance towards the inhibitory effect of bacteria and may be of fungi (including A. bisporus) present in compost. From other work, Trichoderma atroviride that generally grows poorly in spawned compost and does not affect mushroom yield showed depolymerase profiles that differed from those of Ta2, Ta4 and T. harzianum Th1. Nevertheless, glycanase and protease total activities detected in compost did not distinguish Ta2 and Ta4 from non-aggressive Trichoderma (Williams et al. 2003), which was in agreement with reports from Savoie et al. (2001a). When grown in vitro in presence of A. bisporus cell walls, T. aggressivum differed from Th1 and T. atroviride by a more rapid production of chymoelastase- and trypsin-like proteases (Williams et al. 2003) suggesting a better ability to attack mushrooms growing at the periphery of the infected compost.
In addition, there is a chemical interaction between A. bisporus and T. aggressivum. Dual cultures showed that the mushroom mycelium is required for intensive sporulation of Ta2 in compost. A simultaneous growth of both fungi was observed before Trichoderma sporulation was stimulated, but as soon as sporulation occurred the mycelial growth of A. bisporus was dramatically reduced and the typical green mould rapidly developed (Mamoun et al. 2000a). A. bisporus mycelium produced metabolites in liquid culture that stimulated the growth of Ta2 but inhibited the growth of the poorly aggressive T. atroviride and T. harzianum Th1 (Mumpuni et al. 1998). The mushroom mycelium exhibited no defence mechanism involving formation of stationary assemblage of brown aerial hyphae and higher laccase activity in the substrate, as observed at the contact between Lentinula edodes or other leaf litter degrading basidiomycetes and non-Ta2 Trichoderma spp. (Savoie and Mata 1999; Savoie et al. 2001b). Mushrooms growing around the infected compost (green area) showed light brown spots. Infection induced degradation of total tyrosinase in mushrooms, but had no effect on the percentage of active tyrosinase (Soler-Rivas et al. 2000). However, enzymes associated with cell wall degradation might be involved in A. bisporus defence (Anderson et al. 2001). The white hybrid U1 and three off-white strains (assumed to be derived from U1) were far more susceptible to green mould disease than three brown strains. A 96 kDa N-acethylglucosaminidase was produced by two brown strains (Amycel 2400 and Sylvan SB65) earlier in growth and at higher specific activity than by the white (U1) or off-white (Sylvan 130) strains grown in vitro in dual cultures with Ta4, suggesting that the enzyme could play a role in the resistance of commercial brown strains to green mould disease due to its action on T. aggressivum cell walls. Ta4 produced also three N-acethylglucosaminidases, of which the 122 kDa chitinase may be an important indicator of antifungal activity. Nevertheless, the induction of Ta4 chitinases by white and brown A. bisporus strains was identical, showing that brown strain resistance cannot be due to the reduced induction of chitinase production by T. aggressivum (Guthrie and Castle 2006). Mamoun et al. (2000b) compared 25 wild strains ranging from cream (L = 84.9) to dark brown (L = 57.7) and six commercial strains including U1 and Amycel 2400. The two last strains exhibited susceptibility in agreement with the findings of Guthrie and Castle (2006), whilst the wild strains showed no correlation between cap colour and susceptibility. That is not surprising, as the crucial step in green mould development is the competition between T. aggressivum and A. bisporus mycelium (white whatever the strain) for compost colonisation. Further investigations would clarify the putative involvement of chitinases in the mushroom resistance, bearing in mind the possibility of a different response of wild and cultivated strains as observed for the resistance to P. tolaasii. Another mechanism could explain the better resistance of brown strains observed by Anderson et al. (2001). Lipoxygenases (LOX) are involved in the formation of 1-octen-3-ol partly responsible for fresh mushroom flavour. Some wild A. bisporus strains are particularly flavoured and considering that LOX are important for plant resistance to pathogens, investigating LOX regulation in the A. bisporus–Trichoderma interaction might give some interesting information.
Conversely, Ta4 cultivated in vitro produced an inhibitor of A. bisporus growth. This antifungal product, identified as 3,4-dihydro-8-hydroxy-3-methylisocoumarin, was neither produced by non-aggressive T. harzianum isolates under the same culture conditions nor detected in green mould infected compost. Another major antifungal compound was produced during the establishment of green mould in compost (Krupke et al. 2003). It might result from a partial degradation or cleavage of the antifungal product identified in vitro. The isocoumarin included a lactone moiety in its structure, and numerous bacteria in the compost are known to produce lactonases, such as Bacillus spp. Krupke et al. (2003) suggested that if the compound found in infected compost is also an isocoumarin, lactonase-producing bacteria should be examined for disease prevention. Despite T. aggressivum resisted numerous bacteria isolated from the compost, it was affected by some isolates (Savoie et al. 2001a), belonging predominantly to Bacillus species. Based on these observations, Serenade® biofungicide (active ingredients: Bacillus subtilis and its lipopeptides), commonly used to prevent Botrytis attack in vineyards, was tested and approved against T. aggressivum in French mushroom farms (Védie and Rousseau 2008).
Plant essential oils that are generally considered less harmful than synthetic chemicals were tested for antimicrobial activity against Trichoderma. Origanum vulgare oil and its major component, carvacrol, as well as Thymus vulgaris oil and its major component, thymol, had very strong activity against Ta2, Th1 and T. atroviride (Soković and van Griensven 2006). However, the effect of these essential oils was determined in vitro and might probably be different in compost, leading to the inhibition of the bacterial flora affecting Trichoderma. Indeed, Bacillus cereus was strongly inhibited by carvacrol and thymol (Gallucci et al. 2009), and Thymus pubescens essential oil, rich in carvacrol, strongly inhibited B. subtilis (Rasooli and Mirmostafa 2002). Carnavacol and thymol were not detected in Mentha piperata which produced a major essential oil, menthol, showing strong inhibitory activity against Trichoderma (Soković and van Griensven 2006). It would be interesting to test menthol in mushroom compost for a putative use as biofungicide. Other investigation showed that addition of Melaleuca alternifolia (tea tree) essential oil to Pleurotus substrate resulted in strong to total inhibition of T. harzianum (Angelini et al. 2008), suggesting to test this product towards various Trichoderma species in A. bisporus compost. These works open another way for the biocontrol of Trichoderma, along with Serenade®.
In 2007, Hatvani et al. reported that T. aggressivum responsible for recent green mould epidemics in Hungary shared the same mtDNA RFLP pattern as the Irish and English isolates collected in the late 1980s. The authors concluded that T. aggressivum f. europaeum had an essential clonal structure and that the epidemic which has started in Ireland has spread eastward to reach Central Europe. Soon after, PCR applied to T. aggressivum f. europaeum from Polish mushroom farms revealed large variations between isolates, all highly aggressive against A. bisporus (Szczech et al. 2008). This variability might have considerable importance due to possible rapid spread of the variants to other countries as previously observed for the variety and also to possible modification of the pathogen susceptibility to the fungicides currently used. The literature provides examples of point or deletion mutation conferring fungicide resistance in fungal pathogens such as Botrytis cinerea and Mycosphaerelle graminicola (Jiang JinHua et al. 2009; Selim 2009).
Besides, the commercial production of P. ostreatus is currently threatened by massive attacks of green mould disease worldwide, caused by Trichoderma pleurotum and/or Trichoderma pleuroticola (Komon-Zelazowska et al. 2007; Gea 2009). The latter showed very similar microarray profile and ability to assimilate the same carbon sources for growth than T. aggressivum. In vitro confrontations with A. bisporus revealed that the two new species exhibited antagonistic potential similar to that of T. aggressivum. The emergence of these new Trichoderma species in A. bisporus mushroom farms might occur. So, it seems of importance to test their susceptibility to the approved fungicides and to putative efficient essential oils.
Dry bubble disease
Dry bubble disease is of common occurrence in all the major mushroom producing countries. The causal agent, commonly known as Verticillium fungicola, was recently renamed as Lecanicillium fungicola (Zare and Gams 2008). L. fungicola var. fungicola is the usual variety reported in Europe, whilst a second subspecies, L. fungicola var. aleophilum, is responsible for the severe outbreaks in North America. The vegetative mycelium is not affected by the pathogen. The disease presents three types of symptoms designed spotty cap, stipe blowout and bubble. The most severe symptom is the bubble, a spherical mass with little or no tissue differentiation, consisting on mycelia of L. fungicola and A. bisporus growing together.
Attachment and recognition
Apparent attachment of the mycoparasite to the surface of A. bisporus vegetative mycelium seemed to occur, possibly by the presence of hydrophobins in the cell walls of both fungi, but there was no recognition and/or union between the two fungi (Calonje et al. 2000, 2002). Specific interaction between the fruiting body and the mycoparasite were observed. The L. fungicola cell wall glucogalactomannan was able to recognise and bind the A. bisporus lectin, a tetrameric glycoprotein. Simultaneously, certain adhesion-like components seemed to be secreted to consolidate the attachment. This lectin could be isolated only from the sporophore and was not present in the vegetative mycelium; that clarified why the latter was not infected (García Mendoza et al. 2003; Bernardo et al. 2004). L. fungicola was detected inside the pin, the first developmental stage of A. bisporus, but not in mycelial aggregates representing the pre-fructification stage (Largeteau et al. 2007, 2009). This observation confirmed that the lectin is only present in the fruiting body tissue and it is involved in the host–pathogen recognition and interaction. Galactomannan isolated from L. fungicola cell walls pretreated with the fungicide prochloraz-Mn was less effective to bind A. bisporus lectin. This effect can be explained by the increase of the terminal galactose residues linked at (1–4) to the (1–6) mannose bone of the galactomannan molecule caused by the fungicide (Bernardo et al. 2004). This action and the partial inhibition of L. fungicola cell wall hydrophobin suggested that the fungicide is affecting the pre-infection mechanisms in A. bisporus (García Mendoza et al. 2003).
Molecular interaction and defences of A. bisporus
The mycoparasite penetrated the host cell walls by the combined effect of wall-lytic enzymes and mechanical pressure (Dragt et al. 1996). As the A. bisporus cell wall consists mainly of chitin and glucans, investigations were performed to identify L. fungicola hydrolytic enzymes able to play an important role in breaking down the host cell walls, allowing infection to develop. In the presence of A. bisporus fruiting body cell walls, L. fungicola produced glucanases and chitinases (Calonje et al. 1997, 2000; Mills et al. 2000). Targeted gene disruption experiments by homologous recombination in L. fungicola showed that L. fungicola β-1,6-glucanase contributed to the pathogenic process used by the mycoparasite to infect A. bisporus mushrooms (Amey et al. 2003). The disruptants in the VfGlu1 β-1,6-glucanase gene produced significantly reduced spot sizes on mushroom cap compared to the wild-type and was less able to utilise chitin as a growth substrate. As complete reduction in virulence did not occur, successful cell wall degradation is likely to be achieved by the activity of more than one enzyme. Higher activity of some extracellular lytic enzymes, such as N-acetyl-β-D-glucosaminidase, β-D-xylosidase, β-D-mannosidase and β-D-cellobiosidase were detected in three L. fungicola var. aleophilum isolates compared with four L. fungicola var. fungicola isolates grown in liquid culture supplemented with A. bisporus cell walls (Juarez del Carmen et al. 2002). Such observations could partly explain the higher aggressivity of L. fungicola var. aleophilum isolates detected by comparing both varieties in pathogenicity tests performed in situ (Largeteau et al. 2005). In addition, L. fungicola var. fungicola were more sensitive to hydrogen peroxide in their environment than L. fungicola var. aleophilum (Juarez del Carmen et al. 2002) whilst the ability of A. bisporus to release H2O2 could contribute to its defence against the pathogen (Savoie and Largeteau 2004). Despites the oxidative burst plays a major role in pathogen–host interaction, Temme and Tudzynski (2009) reported that obviously B. cinerea does not suffer H2O2-induced oxidative stress in planta. The question arises whether L. fungicola, especially the L. fungicola var. aleophilum, can adapt and resist H2O2 when infecting A. bisporus.
Several works investigated the role of phenol oxidases in the defensive strategy employed by A. bisporus during the interaction with the mycopathogen. Infection by L. fungicola increased the peroxidase activity in mushroom in all developmental stages (pin to fully open sporophore) of infected (spotty cap) compared with healthy mushrooms (Thapa and Jandaik 1989). Conversely, the manganese peroxidase gene mnp1 was similarly expressed in bubble white tissue (pin stage to fully developed bubble) and healthy sporophores (unpublished results) suggesting that probably defence is not related to the gene products. But the presence of another Mn peroxidase (MnP) in A. bisporus was envisaged by Bonnen et al. (1994), and more investigation could be performed if new MnPs are found during annotation of the mushroom genome. The presence of discoloured tissues in infected mushrooms raised the question of the effect of the mycopathogen on the polyphenol oxidases (PPO) activities of the host. Indeed, neither tyrosinase nor laccase activity was detected in L. fungicola (Savoie et al. 2004). Mushrooms infection with a L. fungicola suspension inducing brown dots (Soler-Rivas et al. 2000) and in situ inoculation producing bubble (unpublished results) showed that L. fungicola provoked activation of latent tyrosinase independently of the type of symptoms it induced and of the infection procedure. Analyses of white and discoloured bubble tissues showed that browning is not an efficient mechanism of defence but indicated high level of infection leading to tissue alteration (Largeteau et al. 2007). Tyrosinase genes are known to be involved in plant defence mechanisms against microorganismes. The tyrosinase gene AbPPO1 was constitutively expressed in healthy and diseased mushrooms and the down regulation of the tyrosinase gene AbPPO2 in infected compared with healthy pins (very young primordia) showed that the gene is not involved in A. bisporus initial defence. Laccase activities decreased from mycelial strands to mature sporophore, reaching undetectable values in the latter. Native PAGE confirmed the absence of laccase in A. bisporus sporophores, and showed that laccase isozymes in bubbles were significantly different than those characteristic from vegetative mycelium and healthy primordia. L. fungicola modified the production of laccase in the host possibly by the alteration of the primordium laccase through the action of proteases or glycosidases as observed in mycelial interaction between Trichoderma sp. and L. edodes, or alternatively through de novo induction of different laccases (Savoie et al. 2004) as observed after P. tolaasii infection (Soler-Rivas et al. 2001).
The three laccase genes analysed showed comparable expression in healthy and infected pins (Largeteau et al. 2009) but the expression of the other laccase genes present in A. bisporus genome must be estimated before to conclude.
Despite the low involvement of phenol oxidases, pro-oxidants contributed to the defence of A. bisporus against L. fungicola. The overall level of effective H2O2 plus putative generators of active species in bubbles correlated negatively with the susceptibility of the A. bisporus strains, suggesting that oxidative processes were involved in A. bisporus resistance to L. fungicola (Savoie and Largeteau 2004). In the absence of the pathogen, tolerant and susceptible strains of A. bisporus differed in the expression of the heat-shock gene hspA only when the mushroom was at a stage receptive to infection suggesting that the hspA products might be labels of higher stress-resistance abilities and of self-protection against the higher pro-oxidant potential observed in tolerant strains (Largeteau et al. 2009). hspA belongs to the HSP70 family reported to form intermolecular and mixed disulfides with other cytoplasmic proteins, but how these redox modifications influence protein function remains unknown (Cumming et al. 2004).
After infection, the ability of tolerant and susceptible A. bisporus strains to contain the pathogen was similar, as all strains showed a similar range of tissue infection level in bubbles. This observation suggested that susceptibility is related to pathogen recognition and/or penetration at the very early stage of fruiting body development. Inbar and Chet (1995) observed that purified surface lectin from Corticium rolfsii induced T. harzianum chitinase activity suggesting that the induction of chitinolytic enzymes in Trichoderma is an early event elicited by the lectin–carbohydrate interaction, which is the C. rolfsii–T. harzianum recognition signal. The authors postulated that recognition is the first step in a cascade of antagonistic events which triggers the parasitic response in Trichoderma. The above observation concerning L. fungicola susceptibility and the lectin–glucogalactomannan recognition suggest that similar process might occur in A. bisporus–L. fungicola interaction.
Wild A. bisporus strains highly tolerant to L. fungicola were identified in the INRA-CGAB collection (unpublished results) and in the Dutch PPO MRU collection but total resistance was not found (Sonnenberg et al. 2005). The inheritance of the resistance/low sensitivity trait in A. bisporus had the characteristic distribution of a polygenic trait (Kerrigan 2000; Largeteau et al. 2004). Different breading programmes revealed several QTLs and showed that resistance was carried by the wild strains (Sonnenberg et al. 2005; Foulongne, personal communication). The QTLs cover large regions of the wild genome making breeding for resistance rather complex because wild strains generally bear some undesirable phenotypic traits.
Biocontrol of dry bubble diseases
In many countries, the fungicide prochloraz is the only effective chemical to control L. fungicola which tend to become more tolerant. This tolerance has not been associated with any major loss of control probably because of the complex control mechanism of resistance and the fact that the sexual state of Lecanicillium has not been encountered so far, reducing the risk of increased resistance due to sexual recombination (Grogan 2008). Another fungicide, imazalil, applied to spawn or at casing, was proposed by Shamshad et al. (2009) as an alternative to carbendazim, currently under review for use in Australian mushroom farms. Nevertheless, crop protection is moving to totally organic chemicals, and investigations are performed to find natural antagonists. Two isolates of fluorescent pseudomonads isolated from the casing mixtures exhibited strong antagonism against L. fungicola and M. perniciosa during in vitro dual-culture tests (Singh et al. 2000). Despite culture tests are necessary to confirm the efficiency of the bacteria in mushroom crops, this work open the way for biocontrol of both dry and wet bubble disease at the time the major disease in Serbian mushroom farms is caused by M. perniciosa (Glamočlija et al. 2008). Various essential oils showing in vitro antagonistic activity against L. fungicola (Soković and van Griensven 2006) could be tested in situ for mushroom crop protection.
Concluding remarks and perspectives
Comparing several A. bisporus-microorganisms interactions at the biochemical level revealed that some apparent similarities can reflect different mechanisms. Infected mushrooms exhibited brown tissues whatever the microorganism responsible for the disease. However, browning was a defence mechanism in the A. bisporus–P. tolaasii interaction and reflected high tissue colonisation by L. fungicola. All the microorganisms analysed produced lytic enzymes, but the level at which these compounds were involved in the attack of A. bisporus was highly variable. Toxins (LDP) prevailed over lytic enzymes in bacterial diseases.
The hypothesis that oxidative processes are involved in A. bisporus resistance to L. fungicola prompts us to propose further investigation based on related strategies reported for other organisms. NADP oxidases (NOX) are responsible for the ROS production associated with defence response of some fungi. A defence system involving a superoxide occurs in Coprinopsis cinerea, and Podospora anserina requires a NOX and a MAP kinase cascade during hyphal interference (Silar 2005). Increase in NADPH activity has been observed during A. bisporus post-harvest stress (Hammond 1978). Specific NADPH oxidases are probably active in A. bisporus for synthesis of ROS and can serve defence signalling role (Savoie 2008). Measurement of superoxide dismutase and catalase activities and estimation of gene expressions will be ways to analyse A. bisporus-pathogen interaction, as the strong induction of these enzymes in response to enhance levels of ROS might play an important role in the defence of the mushroom. The oxidative stress is known to profoundly impact glutathione and thioredoxin reducing systems of the eukaryotic cell, so it would be interesting to analyse these redox systems.
It was recently shown that RNA silencing plays a role in Arabidopsis defence against bacterial pathogens and vascular fungi of the Verticillium genus (Ellendorf et al. 2009). The role of components of the RNA silencing pathway in A. bisporus defence has to be investigated.
Transposable elements known to contribute to genome plasticity and evolution were identified in A. bisporus (Sonnenberg et al. 1999; Kerrigan et al. 2004), and their role in the mushroom defence against pathogens has to be investigated. B. cinerea manifests genotype and phenotype variation partly caused by transposable elements (Zhao et al. 2009). Loss in P. syringae virulence was found related to a transposable element and numerous homologous sequences were detected in different strains (Comai and Kosuge 1983). Analysing Trichoderma genome for transposons might perhaps contribute to explain the emergence of T. pleurotum and T. pleuroticola and that of T. aggressivum for which no explanation is available to date.
How A. bisporus reacts to face its pathogens still remains largely unresolved. Hairpin-mediated RNA interference (RNAi) was developed in A. bisporus (Eastwood et al. 2008) and many other molecular techniques (gene expression and regulation, homologous integration, point mutation, RNAi, knockout lines) have been developed in A. bisporus and/or other organisms. They can be applied to confirm or not the mechanisms several works suggested to occur in the microbial-induced diseases of the button mushroom and to investigate mechanisms identified in other host–pathogen models. The complete sequence of the mushroom genome, available soon, should greatly facilitate further investigations.