New Phytologist Research The quick and the deadly: growth vs virulence in a seed bank pathogen Susan E. Meyer1, Thomas E. Stewart2 and Suzette Clement1 1US Forest Service, Rocky Mountain Research Station, Shrub Sciences Laboratory, 735 North 500 East, Provo, UT 84606, USA; 2Department of Plant and Wildlife Science, Brigham Young University, Provo, UT 84601, USA Summary Author for correspondence: • We studied the relationship between virulence (ability to kill nondormant Susan E. Meyer Bromus tectorum seeds) and mycelial growth index in the necrotrophic seed Tel: +1 801 356 5125 pathogen Pyrenophora semeniperda. Seed pathosystems involving necrotrophs Email: [email protected] differ from those commonly treated in traditional evolution-of-virulence models in Received: 13 January 2010 that host death increases pathogen fitness by preventing germination, thereby Accepted: 23 February 2010 increasing available resources. Because fast-germinating, nondormant B. tectorum seeds commonly escape mortality, we expected virulence to be positively New Phytologist (2010) 187: 209–216 correlated with mycelial growth index. doi: 10.1111/j.1469-8137.2010.03255.x • We performed seed inoculations using conidia from 78 pathogen isolates and scored subsequent mortality. For a subset of 40 of these isolates, representing a range of virulence phenotypes, we measured mycelial growth index. Key words: Bromus tectorum (cheatgrass), Drechslera campanulata, necrotrophic • Virulence varied over a wide range (3–43% seed mortality) and was significantly 2 pathogenesis, Pyrenophora semeniperda negatively correlated with mycelial growth index (R = 0.632). More virulent (black fingers of death), seed bank pathogen, isolates grew more slowly than less virulent isolates. virulence evolution. • We concluded that there is an apparent tradeoff between virulence and growth in this pathogen, probably because the production of toxins necessary for necrotrophic pathogenesis competes with metabolic processes associated with growth. Variation in both virulence and growth rate in this pathosystem may be maintained in part by seasonal variation in the relative abundance of rapidly germinating vs dormant host seeds available to the pathogen. Introduction production, this clearly represents excessive virulence with a fitness cost. The evolution of virulence has been the subject of intensive Many effects of life history and population dynamics on research, starting with the seminal paper of Anderson & the shape of the ESS curve have been examined using May (1982), who proposed a model which assumed that modeling approaches. These include the idea that both virulence is an unavoidable negative consequence of parasite coinfection by genetically different strains with equal prob- multiplication within the host. In this model, the tradeoff abilities of transmission and superinfection, or the sup- between increased reproductive rate and longevity of infec- planting of one strain by another that is competitively tion results in the prediction of an evolutionarily stable superior, will lead to the evolution of increased virulence strategy (ESS) of optimal virulence that maximizes total relative to that predicted for a single-strain infection pathogen reproductive output (reviewed by Alizo´n et al., (Nowak & May, 1994; May & Nowak, 1995). Another 2009). This early tradeoff model narrowly defined virulence idea is that vertical transmission will lead to decreased levels as host death rate caused by the pathogen. Death of an of virulence, because the parasite requires the host to repro- animal host ends the infection for a biotroph and, if it duce in order to be transmitted, but this simple result is takes place prematurely, before the pathogen has completed complicated if horizontal transmission and multiple strains its life cycle or has had multiple cycles of propagule are present (Lipsitch et al., 1996). Another model deals No claim to original US government works New Phytologist (2010) 187: 209–216 209 Journal compilation Ó New Phytologist Trust (2010) www.newphytologist.com New 210 Research Phytologist with host population structure, and leads to the prediction digested enzymatically to provide nutrition to the fungus. that virulence will decline when pathogen dispersal, and This mode of pathogenesis has been extensively character- thus contact with new hosts, is reduced (Lipsitch et al., ized for other necrotrophs, such as Cochliobolus 1995). (Panaccione, 1993) and Botrytis (van Kan, 2006). The concept of tradeoffs in virulence evolution has been Pyrenophora semeniperda also produces toxins, including broadened to include factors other than longevity as the cytochalasin B and other more unusual cytochalasins tradeoff and host death rate as the virulence measure (Evidente et al., 2002; Capio et al., 2004). Toxin produc- (Combes, 1997; Poulin & Combes, 1999). For example, tion in this organism is known to vary among isolates, and O’Keefe & Antonovics (2002) modeled virulence evolution levels of culture filtrate toxicity in wheat seedling bioassays in castrating pathogens that have no impact on mortality, have been positively correlated with host leaf disease severity but where host fecundity and pathogen transmission are after needle conidial inoculation (Campbell et al., 2003a). negatively correlated. They found that virulence could Culture filtrates in this study had no effect on intact seeds. evolve to high levels in a spatially unstructured model, Pyrenophora semeniperda has been considered a weak pushing pathogen and sometimes host populations towards pathogen because the cereal crop seeds that often manifest extinction. The introduction of population structure the disease in laboratory viability tests usually escape mor- allowed for stabilization at intermediate levels of virulence. tality and develop into normal seedlings (Campbell & The evolution of virulence has been researched exten- Medd, 2003). We have shown, however, that the fate of a sively for plant pathogens (reviewed by Sacrista´n& seed infected by this pathogen is largely determined by its Garcı´a-Arena´l, 2008). Virulence has been defined as both germination rate (Beckstead et al., 2007). Fast-germinating pathogenicity (infectivity), which is the ability to infect seeds, such as cereal crop seeds and nondormant seeds of specific plant genotypes, and aggressiveness, which is mea- weedy annual grasses, usually escape mortality and germi- sured as the degree of damage to the host plant (Sacrista´n nate normally, although the fungus is often able to sporulate & Garcı´a-Arena´l, 2008). Here, we use the latter definition, on germinated seeds (Medd & Campbell, 2005). By con- which includes the concept that virulence can be measured trast, slow-germinating or dormant seeds of susceptible in terms of negative fitness consequences for the host. species are usually killed (Kreitlow & Bleak, 1964; These consequences often do not include outright mortal- Beckstead et al., 2010). ity and can be difficult to quantify. Plant pathogens pos- We have documented the major impact of this pathogen sess a very broad spectrum of mechanisms of pathogenesis, in seed banks of the invasive winter annual grass Bromus and it has been suggested that these mechanisms need to tectorum in semiarid North America. In the extensive be considered explicitly in order to develop more refined monocultures of this annual grass that have replaced native models of virulence evolution (Frank & Schmid-Hempel, vegetation over millions of hectares in the region, densities ) 2008). of pathogen-killed seeds as high as 20 000 m 2 are not Seed bank pathogens are a poorly studied group of organ- uncommon (Meyer et al., 2007). By contrast, the seed isms, in spite of their undeniable importance in plant popu- banks of native grasses generally contain low to very low lation biology and the structuring of natural plant densities of pathogen-killed seeds (Beckstead et al., 2010), communities (Gilbert, 2002). These pathogens present an indicating that evolution on the host B. tectorum is probably interesting variation on the theme of virulence vs pathogen the primary force shaping virulence patterns in the system reproductive output. Unlike an actively growing plant or studied. The impact of the disease on B. tectorum is most animal, a host seed has finite resources, and, if the seed ger- severe in spring, when remaining ungerminated seeds have minates, the resultant seedling will garner most of these entered a state of secondary dormancy and can germinate resources. This should mean that a more virulent seed slowly if at all; nondormant seeds in autumn can germinate pathogen will have higher fitness, because rapid seed death very rapidly and usually escape mortality (Meyer et al., leads to increased resources available for pathogen growth, 2007). and presumably for reproduction as well. In this study, we defined pathogen virulence as the ability In this study, we investigated virulence in the generalist to kill nondormant B. tectorum seeds. We chose this mea- ascomycete seed bank pathogen Pyrenophora semeniperda sure because dormant seeds generally suffer mortality (Medd et al., 2003; anamorph Drechslera campanulata). regardless of strain identity, whereas nondormant seeds pro- This organism produces macroscopic, fingerlike, black stro- vide a more precise method of detecting virulence variation mata, so that diseased seeds in soil seed bank samples and in the pathogen. Because speed is apparently of the essence inoculation trials can readily be discerned. Like its relatives,