Portal de Periódicos da CAPES

The effect of fungal pathogens on the water and carbon economy of trees: implications for drought‐induced mortality

2014; Wiley; Volume: 203; Issue: 4 Linguagem: Inglês

10.1111/nph.12857

1469-8137

Jonàs Oliva, Jan Stenlid, Jordi Martínez‐Vilalta,

Tree-ring climate responses

Drought-induced forest mortality is emerging as a widespread phenomenon with potentially large implications for forest function and dynamics (Allen et al., 2010; Anderegg et al., 2012; Martínez-Vilalta et al., 2012). Although the physiological mechanisms underlying tree mortality are still not completely understood, there is agreement that they involve the storage and transport systems of water and carbohydrates (McDowell et al., 2008; Sala et al., 2010; McDowell, 2011). The xylem of plants is susceptible to drought-induced embolism and severe water deficits may result in the complete loss of xylem hydraulic conductivity and cause tree mortality (hydraulic failure; cf. Tyree & Sperry, 1988; McDowell et al., 2008; Choat et al., 2012). Drought also has detrimental effects on the carbon (C) economy of plants, and it has been hypothesized that reduced assimilation due to stomatal closure may lead to a depletion of stored C reserves and, eventually, to tree death due to C starvation (Waring, 1987; Martínez-Vilalta et al., 2002; Bréda et al., 2006; McDowell et al., 2008). However, only in recent studies has a direct link between reduced C reserves and tree mortality been established (Adams et al., 2009, 2013; Galiano et al., 2011; Hartmann et al., 2013; Mitchell et al., 2013; Quirk et al., 2013; Sevanto et al., 2014). Finally, phloem transport could also become impaired due to the inability of plants to maintain phloem turgor under extremely low xylem water potentials, limiting the local availability of carbohydrates for metabolic functions (Sala et al., 2010; Sevanto et al., 2014). We postulate that tree mortality research has suffered from a false dichotomy of drought vs biotic attack (McDowell et al., 2013). Pests and pathogens cause tree mortality and it is well known that drought may predispose forests to attacks by insects (Mattson & Haack, 1987; Gaylord et al., 2013) and fungal pathogens (Desprez-Loustau et al., 2006; La Porta et al., 2008). The interaction between drought stress and the damage caused by forest pests and pathogens has been addressed in a recent meta-analysis (Jactel et al., 2012), and the connection between the physiological status of the tree and disease development has motivated a number of reviews in the past (Schoeneweiss, 1975; Boyer, 1995). Biotic agents have also been included in theoretical models for drought induced mortality (Martínez-Vilalta et al., 2002; McDowell et al., 2008, 2011). However, previous reports have not fully acknowledged the diversity of trophic interactions that microorganisms establish with the host trees and how this diversity has direct consequences in terms of the physiological mechanisms leading to mortality. Tree mortality can result directly from a toxic effect from metabolites produced by pathogens, but pathogens can also disrupt the xylem and phloem of the infected hosts and affect their C economy through the consumption of C reserves and the induction of C-expensive defences. Here, we develop a new framework that brings together the effects of pathogens and drought on the water and C economy of trees, and explore the implications for the process of drought-induced mortality. We argue that predictions of drought-induced mortality under pathogen attack can be improved by taking into account the type of trophic interaction that the pathogen establishes with the host. Three main types of trophic interactions can be distinguished amongst tree pathogens: biotrophs, necrotrophs and vascular wilts (Deacon, 1997) (Fig. 1). In general terms, biotrophs drain C and nutrients from living cells – the host response is based on recognition followed by programmed cell death (Glazebrook, 2005). Necrotrophs instead interact with the host through the defence response and get C and nutrients from dead cells – the host response is based on C-based constitutive and induced responses from living cells surrounding the infection (Glazebrook, 2005). A third category includes vascular fungi that colonize the vascular system systemically, often aided by toxins (Yadeta & Thomma, 2013) – the host responses are based on blocking vertical and lateral spread in the xylem. In this letter, we describe how each of these pathogen types interact with the water and C transport systems of trees, and by which mechanisms they may contribute to drought-induced mortality (Fig. 2). As a basis for our rationale, we use the mechanistic model of McDowell et al. (2011) to represent the mortality process of trees subjected to drought stress. We show that during drought pathogens may disrupt the C balance of trees through three non-exclusive processes: (1) by directly depleting non-structural carbohydrate (NSC) reserves, (2) by forcing consumption of NSC reserves by the host or (3) by increasing repair costs (Fig. 2a). Our model makes explicit predictions on the changes in photosynthesis, growth and respiration; as well as on the impacts on the NSC budget, phloem, and xylem transport during a drought episode leading to tree death. The amount of C allocated to defence and the impact on a biotic agent's biomass are also included. Tree death is represented as the point in which no C for sustaining the basic metabolism is available (i.e. zero C available point), regardless of the process leading to this point. Death occurs when C available falls below (intersects) the amount of C needed for osmotic adjustment and maintenance of phloem and xylem transport. By considering different trophic interactions, two novel perspectives for current mortality models are put forward: (1) we show the fundamental differences among the mechanisms leading to tree mortality between biotrophs, necrotrophs and vascular wilt pathogens, and (2) we predict how different type of pathogens affect the timing of the zero C available point, and consequently whether they contribute or not to drought-induced mortality. We have considered the timing of the interaction between drought and pathogens in two ways. Either the pathogen acts simultaneously with drought, as an opportunistic agent taking advantage of the effects of reduced water availability on the host (inciting or contributing factor following Manion's (1981) theory of decline); or else acts before the drought episode, causing a long-term effect weakening the tree (predisposing factor). Our framework focuses on drought as the stress condition of the host, and we do not discuss drought as the weather phenomenon that could facilitate/impair the pathogen spore dispersal or germination and competition with other microorganisms. Biotrophic pathogens have evolved mechanisms to derive C directly from living cells with specialized structures named haustoria, which tap into host cells and create a local C sink (Fig. 1). Some well-studied biotrophic pathogens are Erysiphe alphitoides, Phaeocryptopus gaeumannii, and rust fungi of the genus Melampsora. Trees have evolved defence mechanisms that shut the flow of C towards the pathogen. The defence is based on a fast recognition of the threatening agent that triggers a programmed cell death (PCD) that kills the infected cells and withdraws the C and nutrients before they are assimilated by the pathogen (Fig. 1) (Glazebrook, 2005). Together with PCD, trees also trigger salicylic-acid mediated defence responses (Fig. 2d). With effector molecules, biotrophs manipulate the defence machinery of the host in order to delay defence responses in order to gain enough time to multiply and spread into neighbouring cells (Fig. 2e). Biotrophs mainly affect the C cycle by reducing assimilation and, compared with hemi-biotrophs and necrotrophs, they produce little disruption of the water and C transport systems of the host (Fig. 2c) (Bassanezi et al., 2002). Known mechanisms of reduction of photosynthetic capacity involve the reduction of stomatal conductance by physically occluding of stomata with mycelia or fruiting bodies, as well as other not-yet-understood mechanisms of fungal interference with RuBisCO activity (Manter et al., 2000; Hajji et al., 2009). During fungal establishment and especially when fruiting bodies are produced, C is drained from the leaves, which become C sinks (Hewitt & Ayres, 1976), hence early leaf-shedding is a common tree reaction to reduce C losses (Manter et al., 2003). In those cases in which stomatal functions are heavily impaired (Manter et al., 2000), damages can be very severe, leading to significant growth reductions of infected trees (Kimberley et al., 2010). Damages can also accumulate over several years by, for instance, eliciting recurrent early leaf-shedding processes, reducing NSC reserves and increasing the chances of death in the long run (Marcais & Bréda, 2006). During acute drought C assimilation decreases and leaf cells may decrease their non-structural C reserves (e.g. Adams et al., 2013). Increasing demands of sucrose by leaves decrease the influx of C into the biotrophs, which cannot compete for sugars with living cells under drought conditions (Wyness & Ayres, 1987). Low C accessibility during drought slows down fungal multiplication, and deters further damages (Fig. 2e). Sporulation and mycelial growth, for example, has been shown to be negatively affected by previous water stress (Ayres, 1977; Woolacott & Ayres, 1984), and the link between low disease levels of biotrophs and low NSC availability has been established in model plant systems (Engelsdorf et al., 2013). The strong connection between the nutritional status of the host and the pathogen makes us hypothesize that drought will negatively affect biotrophs during pathogen attack and therefore no worsening effects on tree death are anticipated. As shown in our framework, no significant changes on the timing of the zero C available point are predicted (Fig. 2b). Our hypothesis is supported by the fact that biotrophs tend to be more prevalent in well watered and fertilized sites (Toome et al., 2010), and are expected to decrease in current climate scenarios including increased drought conditions (Desprez-Loustau et al., 2007; La Porta et al., 2008; Sturrock et al., 2011; Marçais & Desprez-Loustau, 2012). Still, we predict worsening effects of drought on host survival in those cases in which biotrophs attain significant population levels on the tree before the drought onset (Fig. 2b). Depleted carbohydrate reserves may impair the subsequent capacity of trees to cope with water stress. Furthermore, if early leaf-shedding has followed the biotroph attack, a C-expensive crown restoration may also accelerate tree death. Necrotrophic pathogens obtain nutrients from dead cells and from structural C sources such as cellulose and hemicellulose. Necrotrophs can attack leaves, twigs, branches, the stem or the root system where they can destroy cambium and the vascular tissue and hence affect both C and water transport systems. Tree defence is activated upon pathogen contact with living cells and is mainly directed at compartmentalizing the pathogen within C-expensive barriers (Fig. 1). Compartmentalization also implies the sacrificial conversion of vascular tissues in the sapwood (Oliva et al., 2012), and, in the case of pathogens causing cankers, in the cambial zone and the phloem. Necrotrophs neutralize tree defences and kill living cells by secreting enzymes and toxins (Fig. 1). Some well-known necrotrophic pathogens include many root rots such as Heterobasidion annosum or Armillaria sp. and canker pathogens such as Cryphonectria parasitica or Cytospora chrysosperma. The accessibility to C by both the tree and the pathogen determines the outcome of the interaction by simultaneously affecting the capacity of the pathogen to build up further inoculum and counteract tree defences, and the capacity of the tree to build up a sufficiently strong response (Fig. 2h). Some necrotrophic root pathogens gain access to the C sources within the host by degrading constitutive and induced defence barriers, like bark or lignin. In these cases pathogens use C from external sources like neighbouring infected or dead trees (Stenlid, 1987; Cleary et al., 2012). In the case of necrotrophs affecting branches or the main stem, the pathogen must gain access to C rich tissues of the phloem passively, either via airborne infection of wounded tissues or by entering the tree as endophytes (Manion & French, 1967). In any case, the outcome of the interaction depends on the host's C availability in order to react fast and compartmentalize the pathogen (Guyon et al., 1996). The magnitude of C needed for defence is large and it has been shown to have a negative impact on tree radial growth (Bendz-Hellgren & Stenlid, 1995; Krokene et al., 2008; Cruickshank et al., 2011; Oliva et al., 2012). By forcing the tree to invest C in defence, necrotrophs affect water transport and storage indirectly by inducing low growth, which results in lowering the overall conductivity of diseased tissues (Joseph et al., 1998) and reducing sapwood storage (Oliva et al., 2012). Necrotrophs can also destroy functional tissues in leaves, stem and roots, which may require repair, and thus they can increase further the C needs from the host. Under favourable conditions for the host, necrotrophic interactions may persist for decades until trees ultimately die (Cherubini et al., 2002). Indeed, large cankers are often seen in trees and are the result of many years of seasonal variations in the capacity of the tree to prevent the pathogen advance (Manion, 1981; Solla et al., 2006). The outcome of necrotrophic interactions is influenced by external stress factors such as drought affecting C availability in the host. Severe and prolonged drought periods usually reduce C reserves (Galiano et al., 2011, 2012; McDowell, 2011), limiting the availability of C to support defences and preventing the establishment or the expansion of previously established necrotrophs (Kane & Kolb, 2010; Anderegg & Anderegg, 2013; Gaylord et al., 2013). Decreased tree defences facilitate the access of necrotrophic pathogens to C sources, from which they build up further inoculum and produce further damages (Fig. 2) (Manion & French, 1967; Lygis et al., 2005; Marcais & Bréda, 2006). Defoliation frequently occurs during drought periods, and degradation of starch into readily usable/transportable sugar compounds to restore the crown can also facilitate C access to necrotrophic root pathogens (Wargo, 1972). As lesions enlarge, the size of the front, where host and pathogen interact, increases and with it the C costs to contain the pathogen's progression (Fig. 2h). As with biotrophs, C used before the drought for repairing infected tissues or for building up defences can also contribute to accelerating tree mortality (Fig. 2f). Overall, necrotrophs accelerate drought-induced mortality either by depleting resources and creating repair needs in advance or by making trees run out of C at a faster rate (Fig. 2f) . Consistent with our framework, increased damages have often been observed/expected under drought conditions by necrotrophic canker (Luque et al., 2000; Desprez-Loustau et al., 2006; Waldboth & Oberhuber, 2009) and root rot pathogens (La Porta et al., 2008; Sturrock et al., 2011). Vascular wilt pathogens thrive inside xylem conduits, releasing toxic compounds and disturbing water transport (Fig. 1). Some examples of vascular wilt pathogens include some Ophiostoma species, remarkably O. novo-ulmi and also several Ceratocystis and Leptographium species. These type of pathogens feed on xylem sap sugars, C leakages, defence compounds and sugars from cell-wall degradation processes (Hammerbacher et al., 2013; Yadeta & Thomma, 2013). Trees block vertical spread by clogging the conduits with tyloses, while lateral spread is prevented by in situ synthesis of C compounds and barrier structures to compartmentalize the infection (Shigo & Tippett, 1981; Bonsen et al., 1985; Yadeta & Thomma, 2013). Defence can be C expensive (Guérard et al., 2007) and result in a reduction of sugars in the vicinity of the lesion (Viiri et al., 2001). Investment in defence can be at the expense of radial growth (Krokene et al., 2008) and also imply a sacrificial loss of conductive tissue (Joseph et al., 1998). In contrast to necrotrophs, vascular wilt pathogens have significant direct effects on water transport and storage in trees (Fig. 2). Xylem disruption has immediate effects and may cause sudden mortality on adult trees (Tyree & Zimmermann, 2002). Conduit clogging results in foliage wilting that impacts current and future C reserves by cutting downstream C supply and by reducing autumn re-assimilation of nutrients from leaves. Under these conditions, xylem, phloem and foliage damage become very costly to repair (Fig. 2j). Wilt diseases are often associated with bark beetles that feed on the phloem, increasing even further the costs of repair and reducing the capacity to allocate C to the crown and restore foliage. Nevertheless, insect phloem damage has been shown to be of lesser importance compared with xylem dysfunction induced by insect-vectored wilt pathogens (Hubbard et al., 2013), although in some cases disruption of the water balance of the tree is not a pre-requisite for the success of the bark beetles (Wullschleger et al., 2004). In contrast to C starvation-driven mortality in the case of necrotrophs, mortality in trees infected by vascular wilt pathogens seems to be triggered by hydraulic failure (Fig. 2k). Disruption of the vascular system is fast and permanent, hence rapid mortality of the corresponding areas of the crown or the whole tree can be observed. Increased damages by insect bark beetles and their associated vascular wilt pathogens are associated with dryer climatic conditions (Williams et al., 2010), but, contrarily to necrotrophs, drought during the infection/attack may be more important than previous drought events (Croisé et al., 2001). The availability of C for defence at the moment of attack is also of a lesser importance in comparison with necrotrophs (Christiansen & Ericsson, 1986). We thus postulate that vascular wilt pathogens accelerate drought-induced mortality under drought mostly by damaging the xylem vascular system and subsequently causing phloem impairment and foliage wilting. Of special importance is the rapid escalation of repair costs as the attack builds up (Fig. 2j). While C reserves can be reasonably high at the onset of a drought event, they may still not be enough for rebuilding a sufficient amount of foliage, phloem and xylem for tree survival. By increasing repair costs, wilt pathogens can also accelerate drought induced mortality processes (Fig. 2j). The presented framework sets the ground for predicting the role of pathogens on tree mortality under drought based on the type of trophic interaction established with the host. Although most pathogens fall within the three categories described in the previous sections, some might establish more than one type of trophic interaction. This is the case of the so-called hemibiotrophs, a category that includes many Phytophthora species that share characteristics with both biotrophs and necrotrophs. In these cases, we suggest that the type of trophic interaction that contributes more to the pathogen's inoculum build-up should be considered. Would pre-inoculation water stress (Marçais et al., 1993) or C starvation (Engelsdorf et al., 2013) favour disease development, these pathogens should be considered for their necrotrophic phase and thus be expected to accelerate drought-induced-mortality. Other pathogens can display a behaviour in between a wilt pathogen and a necrotroph. These pathogens are typically secondary pathogens affecting woody tissues, like shoots and twigs (Jactel et al., 2012), and while they can cause disease under negative water potentials, tree resistance is typically restored when water stress is remediated (Crist & Schoeneweiss, 1975; Schoeneweiss, 1975; Johnson et al., 1997). The fact that the pathogenicity of these fungi is strongly dependent on xylem colonization (Luchi et al., 2005), and that the necrotrophic phase precedes the wilting of the infected tissue, makes them similar to the 'vascular wilt pathogens' in our framework. The same reasoning can be applied to similar pathogens for which pre-inoculation water stress and C limitation would contribute little to host susceptibility (Madar et al., 1989). Future climate scenarios predict an impact on water and C balance of trees (Wang et al., 2012). At the same time, forest pathogens are pervasive in forest ecosystems all over the globe and are known to cause tree mortality and have a major role in forest dynamics (Worrall et al., 2005). Carbon and water systems are inevitably connected and both are affected by drought and by pathogens. Pathogens can accelerate drought-induced mortality by directly depleting NSC, accelerating NSC consumption by the host or by increasing repairing costs (Fig. 2a). These three processes are tightly connected with the type of trophic interactions established between the host and the pathogen. We describe how these types of pathogens would interact with the host, and by which mechanisms would cause the death of the tree. This theoretical framework allows us to predict that some pathogens such as necrotrophs or vascular wilts can benefit from drought events, and thus contribute to drought induced mortality; and that some, like biotrophs are very unlikely to cause significant damages under drought. Considering their different effects on the host and the contrasted interaction with drought, determining under what environmental conditions the previous trophic interactions will be favoured (or disfavoured) is pivotal to predictions of how forests will respond to warmer and drier conditions in the future. Future research needs to quantify the contribution of pathogens to direct drought effects in the context of drought-induced tree mortality. Manipulative experiments controlling both drought and pathogen inoculum can be used to assess the extent to which pathogens accelerate mortality by comparing the time needed to kill trees under drought with and without specific pathogens (Fig. 2a). This study was supported by the Spanish government through grant CGL2010-16373, and by the Swedish Research Council FORMAS through the grant 215-2012-1255.