Free

Resistance Correlations Influence Infection by Foreign Pathogens

Abstract

Reciprocal selection promotes the specificity of host-pathogen associations and resistance polymorphisms in response to disease. However, plants and animals also vary in response to pathogen species not previously encountered in nature, with potential effects on new disease emergence. Using anther smut disease, we show that resistance (measured as infection rates) to foreign pathogens can be correlated with standing variation in resistance to an endemic pathogen. In Silene vulgaris, genetic variation in resistance to its endemic anther smut pathogen correlated positively with resistance variation to an anther smut pathogen from another host, but the relationship was negative between anther smut and a necrotrophic pathogen. We present models describing the genetic basis for assessing resistance relationships between endemic and foreign pathogens and for quantifying infection probabilities on foreign pathogen introduction. We show that even when the foreign pathogen has a lower average infection ability than the endemic pathogen, infection outcomes are determined by the sign and strength of the regression of the host’s genetic variation in infection rates by a foreign pathogen on variation in infection rates by an endemic pathogen as well as by resistance allele frequencies. Given that preinvasion equilibria of resistance are determined by factors including resistance costs, we show that protection against foreign pathogens afforded by positively correlated resistances can be lessened or even result in elevated infection risk at the population level, depending on local dynamics. Therefore, a pathogen’s emergence potential could be influenced not only by its average infection rate but also by resistance variation resulting from prior selection imposed by endemic diseases.

Introduction

The emergence of novel infectious diseases is a rising threat to domesticated and wild species, due in part to increased rates of anthropogenic species redistributions that result in novel host-pathogen contacts (Jones et al. 2008; Lindahl and Grace 2015). Well-known natural history observations show that hosts are affected by disease from only a small fraction of the organisms that are potential pathogens in their local area (Gilbert and Webb 2007). The factors that determine whether a pathogen can exploit a host are hugely consequential, mediating which pathogens cause disease outbreaks (Webby et al. 2004) and which infectious diseases play a role in shaping host abundance, host evolution, and community structure (Woolhouse et al. 2002, 2005). Many studies have highlighted species-level risk factors for the emergence of new diseases resulting from cross-species transmission, emphasizing primarily the roles of phylogenetic distance and the degree of sympatry among potential hosts (Davies and Pedersen 2008; Parker et al. 2015; McDonald and Stukenbrock 2016). Still, our understanding of the local genetic factors that predispose host populations to the emergence of novel pathogen associations remains incomplete. In particular, we address here the question of how local disease resistance structure in the host population affects its initial interactions with a novel foreign pathogen.

The local interactions of hosts with an endemic pathogen is the focus of a large body of theoretical work on population genetics and coevolutionary dynamics of host resistance, resistance costs, and pathogen infectivity (Antonovics and Thrall 1994; Boots et al. 2014; White et al. 2018; Ashby et al. 2019). However, such coevolutionary studies have usually been viewed as independent from the factors that define the risk of disease emergence resulting from invasion of a foreign pathogen. Heath (1981) was the first to effectively articulate the idea that genetically determined resistance to a local endemic pathogen might be different from the defense against a foreign pathogen that is not normally harbored by a host, and she termed this “nonhost resistance.” Indeed, it has since been argued that the ability of both plant and animal hosts to reject a foreign pathogen may involve fixed, generalized mechanisms of detecting molecular profiles of bacterial, fungal, or viral groups (Nürnberger et al. 2004; Cui et al. 2015; De Schutter and Van Damme 2015; Thines 2019). Moreover, because of the expectation that nonhost resistance involves generalized mechanisms, a number of authors (e.g., Bettgenhaeuser et al. 2014; Stam et al. 2014) have pointed out the assumption that such nonhost resistance is uniform across individuals of the potential host species and across strains of the foreign pathogen.

However, there is now abundant evidence for standing variation within plant and animal species for resistance to pathogens that they do not normally harbor in nature (e.g., Best and Kerr 2000; Marmor et al. 2006; Gilbert et al. 2018; for additional examples, see table S1), implicating resistance variation to endemic diseases as a possible contributor to a host’s response to foreign pathogens. This possibility is furthermore supported by substantial evidence for correlated variation in resistance to different endemic pathogens that are phylogenetically quite distant (e.g., Gross et al. 1980; Loiseau et al. 2008; Nsabiyera et al. 2012; see table S2). Similar relationships are seen for genetic variation in plant resistance to different types of insect pests (Leimu and Koricheva 2006). Indeed, it has recently been shown that some of the same molecular pathways for resistance against endemic plant pathogens can be involved in defense against foreign pathogens (Stam et al. 2014; Ayliffe and Sørensen 2019). Thus, selection imposed by endemic pathogens could affect the range and frequencies of host genotypes suitable to infection by newly introduced pathogens.

In this study, we investigate the relationship between resistance polymorphism that is characteristic of endemic host-pathogen interactions and the resistance that may occur to foreign (nonendemic) pathogens. Here we restrict our analysis to situations where correlated resistances are due to genetic causes, that is, genetic correlations (resulting from pleiotropy or linkage disequilibrium), distinct from correlations in infection by two pathogens resulting from shared local environments, cross-immunity to past infection, or pathogen interference during coinfection. We first experimentally tested the relationship between variation in infection rates by endemic and foreign pathogens using material from natural populations of the host plant Silene vulgaris. We included one specialized endemic anther smut fungus (Microbotryum silenes-inflatae), one foreign (nonendemic) anther smut fungus (Microbotryum lychnidis-dioicae, normally found on Silene latifolia), and a generalist foliar leaf-spot pathogen (Stemphylium solani) encountered by S. vulgaris in its introduced range. Second, we developed a theoretical framework for understanding how resistance correlations arise, using a simple one-locus model. In particular, we show that the genetic regression of host variation in infection rates by the foreign pathogen on variation in infection rates by the endemic pathogen provides insight into the likelihood of invasion by a foreign pathogen. We use the term resistance “transitivity” for such a genetic regression, and we show how its slope and intercept predict disease transmission by a foreign pathogen introduced into a host population. In particular, we show that average resistance to a foreign pathogen provides an incomplete picture of the potential for new disease transmission by foreign pathogens.

Empirical Investigations

Study System

Anther smut disease is caused by fungi in the genus Microbotryum, and affected plants are primarily in the Caryophyllaceae (Kido and Hood 2020). Infections are biotrophic (requiring living host cells for colonization), systemic (colonizing all or many parts of the host), and persistent (infected hosts do not recover), and infection results in complete sterilization of the host through the replacement of pollen with fungal spores and the failure of female structures to mature. While the majority of perennial Silene species are likely to harbor anther smut disease (Hood et al. 2010), there is a high degree of host specificity such that most pathogen species can persist on only one host species (Refrégier et al. 2008). Instances of cross-species disease transmissions have been reported in sympatric host populations (Antonovics et al. 2002; Gladieux et al. 2010; Hood et al. 2019), and cophylogenetic studies indicate that new disease emergence via host shifts have been important in shaping the larger-scale distribution of Microbotryum species among host lineages (de Vienne et al. 2013; Hartman et al. 2019).

Relatively few studies have examined the detailed genetic mechanisms of resistance to anther smut disease, but variation and/or heritability of resistance has been demonstrated in the hosts Silene latifolia (Alexander 1989; Alexander et al. 1993; Alexander and Antonovics 1995), S. dioica (Carlsson-Granér 1997), and S. vulgaris (Antonovics et al. 2002; Cafuir et al. 2007) along with its close sister species S. uniflora (Chung et al. 2012). Silene vulgaris has been observed as the recipient of local cross-species disease transmission events from Microbotryum on S. latifolia and S. dioica (Antonovics et al. 2002; Hood et al. 2019). Silene vulgaris is a very broadly distributed herbaceous perennial, occupying habitats from ruderal low-elevation sites to subalpine meadows (Abbate et al. 2018). The species is native to Europe and introduced to North America (Keller et al. 2009).

Collections and Sampling

The collection region for S. vulgaris seeds was bounded approximately by longitudes 44.281–44.395 and latitudes 7.539–7.696 (table S3) in southern Cuneo Province, Italy, where the species is widespread (Bruns et al. 2018). Seeds were sampled from natural populations as 28 maternal half-sib families (seeds from a single maternal plant, but where multiple sires are possible), with two families from each of 14 localities. To minimize confounding resistance variation in the host with different levels of variation in the pathogen species, the endemic and foreign pathogens were sampled as single-genotype isolates. The endemic pathogen, M. silenes-inflatae, was collected as the spore contents of a diseased S. vulgaris flower from one of the populations (coordinates: 44.212, 7.672). The foreign pathogen, M. lychnidis-dioicae, was collected from a diseased S. latifolia plant in Lamole, Italy (coordinates: 43.551, 11.358), approximately 300 km away. This Microbotryum species has been shown to infect S. vulgaris at a low rate compared with the endemic pathogen, and there has been occasional cross-species disease transmission in sympatric populations (Antonovics et al. 2002; Hood et al. 2019). The necrotrophic foliar pathogen, Stemphylium solani (Weber 1930), was isolated from an S. vulgaris plant in the introduced range in Bernardston, Massachusetts (coordinates: 42.682, −72.544), and cultured on malt extract agar. Its identity was confirmed by the DNA sequence of the internal transcribed spacer region of the ribosomal RNA (GenBank accession no. MN856663). Stemphylium solani is a North American native pathogen (Gilbert and Parker 2010) and is common on cultivated tomato but occasionally is reported from leaf lesions on other plant families, including one report in the Caryophyllaceae (Orieux and Felix 1968).

Experimental Inoculations

If resistance assessments to alternate pathogens are measured independently on individuals sampled randomly from within-host groups (e.g., family groups of host siblings), the correlations in resistance can represent genetic correlations in a way analogous to the correlation of trait variation measured among individuals in two environments (Falconer 1952; Via and Lande 1985). In this study, resistance of each S. vulgaris family was based on infection rates following inoculation of 60 seedlings per family with the endemic M. silenes-inflatae and inoculation of 120 seedlings per family with the foreign M. lychnidis-dioicae; more seedlings were used in the latter because we expected lower infection rates with the foreign pathogen. Seeds of S. vulgaris were surface sterilized (2 min in 2% sodium hypochlorite, 20% ethyl alcohol, and 0.2% Triton X-100), rinsed with sterile water, and germinated on 0.8% agar containing Murashige and Skoog basal medium at one-tenth strength. Seeds from different families were placed on the agar medium in random order, with approximately 36 seeds per 25 × 150-mm petri plate. When the cotyledons had separated (after 10 days of incubation at 25°C), a 4-μL suspension of 500 pathogen spores per microliter was placed on the apical meristem (as in Hood and Antonovics 2000), with application of each pathogen type (endemic vs. foreign pathogen) being randomly assigned to each seedling. Following 48-h incubation at 15°C to promote spore germination and infection (see Day 1979), seedlings were transferred to soil in 3.81-cm Cone-tainers (Stuewe and Sons, Corvallis, OR). The positions of individual plants were randomized within the greenhouse, where they were maintained until flowering.

Disease status of plants was scored at flowering by the presence of pathogen spores in the anthers. Any diseased plants were removed from the greenhouse to help prevent secondary infections. Plants that were healthy at first flowering were cut back and scored again upon reflowering to include rare latent infections. Microbotryum infections were confirmed to be M. silenes-inflatae or M. lychnidis-dioicae by morphology in pathogen culture; M. silenes-inflatae produces much smaller colonies compared with the other species (Gold et al. 2009).

To assess resistance to the necrotrophic foliar pathogen, Stemphylium solani, leaves from uninoculated greenhouse-grown plants of the same S. vulgaris families were used to assess lesion development following inoculation. Detached leaves of similar size from separately grown S. vulgaris plants were inoculated by placing 0.6 cm diameter agar culture plugs onto the underside of leaves following wounding with approximately 10 punctures with sterile insect pins (as in Gilbert and Webb 2007). The inoculated leaves and wound-only control leaves were maintained for 4 days in humid chambers at 25°C, and necrotic lesion area was calculated using ImageJ software (Rueden et al. 2017). Ten individual plants per family were used, with three leaves per individual plant, from each of 24 S. vulgaris families.

Data Analysis

We assessed the resistance transitivity by determining the slope and intercept of the regression of family-level variation in infection rates by the foreign pathogen (M. lychnidis-dioicae, or St. solani) on family-level variation in infection rates by the endemic pathogen (M. silenes-inflatae) for nontransformed data. We used total least squares regression rather than the usual least squares regression to determine the transitivity slope, because there is potential error in the independent variable, that is, the family-level resistance estimates for the endemic pathogen inoculations. We also performed Pearson correlation tests on arcsine-square-root-transformed proportional data to assess the significance of the correlations.

A prior study on resistances to two endemic anther smut fungi in S. uniflora (Chung et al. 2012) found strongly correlated infection rates among the plant families (fig. S1) but with a small number of significant outliers that were very susceptible to one pathogen but resistant to the other. For all regressions in the current study, outlier analysis was similarly performed based on the influence of individual family means on the total least squares regression coefficient using dfbeta values and boxplots for the regression residuals (Cousineau and Chartier 2015; table S4; fig. S2A). In this way, three S. vulgaris families, each one from a different collection site, were identified as potential outliers for the regression of host variation in infection rates by the two Microbotryum pathogens. To confirm their status as outliers, we conducted a further independent inoculation experiment that again showed the same relative rates of infection as observed in the main experiment (fig. S3). The regression analyses were therefore carried out with and without outlier families. For the regression of infection by M. silenes-inflatae versus St. solani, no outlier families were found (fig. S2B).

Experimental Inoculation Results

There was wide variation in infection rates among S. vulgaris families to both the endemic and foreign anther smut pathogens. At the family level, infection rates by the endemic pathogen, M. silenes-inflatae, ranged from 0 to 0.88, and the average infection rate was 0.55 ($n=28$). Infection rates by the foreign anther smut pathogen, M. lychnidis-dioicae, ranged from 0 to 0.54, with a lower average infection rate of 0.13 ($n=28$). When all families were included in the analysis, total least squares regression using nontransformed data showed a slope of −0.183 ($SE=0.133$) and intercept of 0.23 (fig. 1A), and the correlation between infection rates by the endemic pathogen and infection rates by the foreign pathogen was not significant (proportional data transformed, Pearson $r=−0.096$, $P=.627$). However, analysis of the 25 families remaining after removal of the three outliers resulted in a total least squares regression with a slope of 0.197 ($SE=0.074$) and intercept of −0.02, with a significant positive correlation between family-level infection rates by the endemic and foreign pathogens (proportional data transformed, Pearson $r=0.471$, $P=.017$; fig. 1A).

Resistance to the necrotrophic generalist pathogen, St. solani, showed the opposite relationship; there was a significant negative correlation of lesion size with the infection rates by endemic M. silenes-inflatae (proportional data transformed, Pearson $r=−0.481$, $P=.017$; fig. 1B). A total least squares regression using nontransformed data provided a slope of −0.242 ($SE=0.084$) and intercept of 0.48.

Theoretical Investigation

Our results described above show that the direction and magnitude of resistance correlations between endemic and foreign pathogens can depend on the pathogens being investigated and may also be confounded by host genotypes that are outliers to the general trends. Naively, we might expect infection rates by foreign pathogens to be lower where cross-species correlations in resistance are positive and higher where these are negative. However, to explore and quantify these expectations, we provide first a simple one-locus model to establish a theoretical framework for resistance correlations with regard to endemic and foreign pathogens, identifying the main variables and the conditions predicting the likelihood of infection by the foreign pathogen. This is followed by consideration of how the numerical dynamics of a host population maintaining an endemic pathogen at equilibrium might determine the probability of infection by a foreign pathogen on contact with a random healthy host individual, as well as the potential for a foreign pathogen to subsequently spread within the host population.

Model of Resistance Transitivity

To investigate how resistance variation to an endemic pathogen may affect infection by a foreign pathogen, we first develop a simple one-locus model to show how single-locus effects translate into the relationship between endemic pathogen infection rates and foreign pathogen infection rates, expressed as a regression. We use the term resistance “transitivity” to describe this regression, and show that both the slope and intercept provide information on the probability of infection by a foreign pathogen for a healthy individual in a host population. On the assumption that gene effects are additive, the model would also reflect resistance as a multilocus quantitative trait. We then apply this theory to the anther smut disease example, where in this system there is strong theory on the conditions allowing for genetic polymorphism for resistance to endemic disease (Antonovics and Thrall 1994).

We assume that host resistance is controlled by a single locus with two alleles, which affects the likelihood of infection on exposure to a pathogen. The relationship of resistance to the endemic and foreign pathogens is therefore the result of pleiotropy of the allele effects, rather than linkage disequilibrium as might be the case if separate genes were involved in resistance to the two pathogen types. Also, for simplicity, we assume haploid allele expression and that both the endemic pathogen and the foreign pathogen are genetically uniform; this is equivalent to considering a diploid case with no dominance variation. While we do not have information on the frequency and effects of the specific genes involved in resistance of S. vulgaris to anther smut, if additive gene effects are assumed, our analysis below shows that the slope and intercept of the regression based on family means, as in our experimental study, is fundamentally informative about the likelihood of a foreign pathogen infecting and spreading in a polymorphic host population.

Infection rates for an endemic pathogen and a foreign pathogen on two resistance genotypes of the host, A1 and A2, are represented in figure 2A (A2 is more resistant, i.e., lower infection rate, than A1 to the endemic pathogen), where the infection rate, bounded by 0 and 1, is symbolized as βij for pathogen type j on host type i. Denoting $S^=(β11+β21)/2$ as the average infection rate by the endemic pathogen and $S^′=(β12+β22)/2$ as the average infection rate for the foreign pathogen, $S^−S^′$ represents the difference between their average infection rates. This quantity represents “nonhost resistance” as usually conceptualized. However, as shown below, this does not take into account the variation in effect or in the frequency of resistance alleles and therefore does not reflect the relative probability of infection by a foreign pathogen in a population setting. Additionally, the proportional difference between the two host genotypes to infection rates by the endemic pathogen relative to their difference in infection rates by the foreign pathogen (the relative slopes of the lines in fig. 2A) can be given by the parameter $Q=β22β11−β12β21$, bounded between 1 and −1. This parameter, which is employed below, has also been shown to be important in determining dynamical outcomes in two-predator two-prey systems (the parameter inverse delta in Abrams and Cortez 2015).

The transitivity slope (T) for resistance, the regression coefficient for the variation in infection rate of the foreign pathogen on the variation in infection rate of the endemic pathogen, is shown in figure 2B. This reflects a one-locus representation of the regression coefficient measured in our experimental study. For the line $y=a+Tx$, the regression coefficient is calculable as the covariance of infection rates by the two pathogens divided by the variance in infection rates by the endemic pathogen, which reduces to

$(1)T=β12−β22β11−β21.$

This is equivalent to the ratio of the difference in infection rates (betas) of the two host genotypes to the foreign pathogen divided by the difference in their infection rates to the endemic pathogen, and this is a genetic regression because the infection rates by the foreign and endemic pathogen are assumed to have been measured independently.

Infection by the foreign pathogen on introduction into a completely healthy host population will depend on the frequency of the two host resistance genotypes. We let p be the frequency of the A1 allele and $1−p$ be the frequency of the A2 allele. Then, if a random healthy individual in the host population is challenged by a foreign pathogen (as in our empirical studies), the probability that the host will become infected is

$(2)pβ12+(1−p)β22.$

However, if populations are diseased and only healthy individuals are available for inoculation, then the probability of foreign pathogen infection is dependent on the frequency of A1 and A2 among only the healthy individuals, which we term ph and $1−ph$, respectively. This was the case in our inoculation experiments described above, because seeds used for estimating transitivity could only be sampled from healthy individuals (individuals diseased by anther smut are sterilized).

It can be shown from equation (2) that even if the foreign pathogen has a lower average infection rate ($S^−S^′>0$), the average host individual may actually still be more susceptible to the foreign pathogen than to the endemic pathogen, depending on the frequency of the resistance allele. A necessary (but not sufficient) condition for this to happen is that $β12>β11$ or $β22>β21$, which results in the lines in figure 2A crossing and the genetic regression line in figure 2B crossing the diagonal one-to-one line (unit slope and zero y-intercept). An additional necessary condition is that the intersection of the regression line with the diagonal one-to-one line occurs within the host’s genotypic range of resistance values to the endemic pathogen (i.e., between values for β21 and β11). This latter condition means that there can be resistance allele frequencies with a greater infection probability by the foreign pathogen than the endemic pathogen on a randomly selected host, even if the foreign pathogen’s average infection rate across the host genotypes is less than for the endemic pathogen ($S^′). While the y-intercept of the regression line is $a=S^′−TS^$, the intersection of the regression line with the diagonal one-to-one line is at coordinates $(x,y)=a/(1−T)$. In our experimental results in figure 1A, the genetic regression line crossing the diagonal one-to-one line only occurs when the outlier genotypes are included in estimates of resistance transitivity; these genotypes have higher infection rates by the foreign pathogen than the endemic pathogen, and the intersection with the diagonal ($[S^′−TS^]/[1−T]=0.195$) is well within the range of resistance variation in the host to its endemic pathogen. While this condition of $β22>β21$ would be most easily met with a negative transitivity slope, it is also possible for greater average infection rates for the foreign pathogen with some combinations of $S^′ and positive transitivity slope, depending on allele frequencies (fig. S4). The important conclusion is that a lower average infection ability by a foreign pathogen does not necessarily indicate average infection outcomes in a host population with genetic variation for resistance.

Resistance Transitivity with Numerical Dynamics

To examine the effect of resistance transitivity on the likelihood of infection by a foreign pathogen under a more realistic situation of arrival into a population maintaining endemic disease, we used the model of Antonovics and Thrall (1994) to establish equilibrium conditions prior to introduction. In this model, the variation in resistance allele frequencies is generated by endemic disease dynamics reflecting the biology of anther smut, including frequency-dependent transmission, no coinfection, and complete sterility of infected individuals. Thus, before the introduction of the foreign pathogen, this system with the endemic pathogen alone and two host genotypes differing in resistance ($β11>β21$) is described by the equations

$(3)dX1dt=X1(b1−μ−γN)−X1β11Y/N,(4)dX2dt=X2(b2−μ−γN)−X2β21Y/N,(5)dYdt=Y(X1β11/N+X2β21/N−μ),$
where X1 and X2 are numbers of healthy individuals with alleles A1 and A2, respectively; Y is the number infected, $N=X1+X2+Y$; b1 and b2 are birth rates of the respective genotypes with a cost of resistance ($C=b1−b2>0$); μ is the host death rate; and γN represents the density-dependent reduction in birth rate.

We varied the equilibrium frequencies of the resistance alleles by varying the cost of resistance that is expected of coevolved host-pathogen systems (for derivation, see Antonovics and Thrall 1994). Parallel to the approach in our experimental work, we then estimated the probability that a random healthy host (with genotype frequency $ph=X1/[X1+X2]$) would become infected when challenged by exposure to the foreign pathogen. We explored a range of transitivity slopes (depicted in fig. 3A) that result from regression of infection rates by a foreign pathogen on infection rates by an endemic pathogen on host genotypes with alleles A1 or A2. For simplicity, throughout we assumed that the average infection rate of the foreign pathogen was always reduced from the average infection rate of the endemic pathogen by $S^−S^′=0.1$.

Using the equations for equilibrium frequencies of A1 and A2, it can be shown (see the supplemental PDF) that the probability of infection by the foreign pathogen on challenge to a random healthy host is

$(6)QB−µTBC−B,$
where, as before, $T=(β12−β22)/(β11−β21)$ is the transitivity slope, μ is the host death rate, $C=b1−b2$ is the reproductive cost of resistance, $B=β11−β21$ is the benefit of resistance to the endemic pathogen, and $Q=β22β11−β12β21$ is the proportional difference of the two host genotypes to infection rates by the endemic pathogen relative to their difference in infection rate by the foreign pathogen.

Varying the costs of resistance (C), and thus mediating resistance allele frequencies among healthy individuals at equilibrium, affected how the transitivity slope affected the probability of infection by the foreign pathogen (fig. 3B). With low resistance costs (small C), the risk of foreign pathogen infection is greater for negative transitivity slopes than for positive transitivity slopes. This outcome is because low resistance costs lead to a high equilibrium frequency of the genotype resistant to the endemic pathogen, which under a negative transitivity slope is susceptible to the foreign pathogen. However, with high resistance costs (large C), where the genotype resistant to the endemic pathogen is rare at equilibrium, the risk of a foreign pathogen infection is greater for positive values of the transitivity slope than for negative values of the transitivity slope.

Additionally, the results (fig. 3C) also show that there are conditions under which the infection rate of healthy individuals by the foreign pathogen is likely to be greater than for the endemic pathogen (the points above the shaded gray area). Indeed, it is possible to show that the conditions for the foreign pathogen to be more likely than the endemic pathogen to infect healthy individuals is

$(7)T>1−Q(1−C/B)µB.$

Note that the formulations of T and Q shown above mean that these terms cannot combine such that both have negative values nor have values of $T≥1$ with positive values of Q (see the supplemental PDF). Thus, values of $T≥1$ occur with negative Q values where there is a less proportional difference between host genotypes in infection rates by the endemic pathogen than their difference in their infection rates by the foreign pathogen (e.g., fig. S4C, S4D). Conversely, negative T values occur with positive Q values where there is a greater proportional difference between host genotypes in infection rate by the endemic pathogen than by the foreign pathogen (e.g., fig. S4E, S4F).

The effects of varying resistance costs (C) in inequality (7) depend on the value and sign of the transitivity slope, T. We first consider where $T=1$ (transitivity slope is parallel to the one-to-one line), which represents cases where resistance decreases the infection rate of the endemic and foreign pathogens by the same amount. With $T=1$, varying the cost of resistance (increasing or decreasing C) cannot determine conditions where the foreign pathogen is more likely than the endemic pathogen to infect healthy individuals (noting the constant vertical distance from black dotted line to the boundary of the gray shaded area in fig. 3C). This outcome when $T=1$ is because negative Q values cannot produce quantities <1 on the right side of inequality (7), and Q does not occur as positive values with $T≥1$. We next consider where $T>1$ (transitivity slope is more positive than the one-to-one line), which represents cases where resistance decreases the infection rate for the foreign pathogen more than for the endemic pathogen. With $T>1$, and thus Q is negative, decreasing resistance costs (lower C) disfavors conditions where the foreign pathogen is more likely than the endemic pathogen to infect healthy individuals; this is consistent with the expectation that lower resistance costs lead to a higher equilibrium frequency of the resistance allele, A2, that also protects against the foreign pathogen (fig. 3B, 3C). We then consider where $0 (transitivity slope is less positive than the one-to-one line), which represents cases where resistance decreases the infection rate of a foreign pathogen less than the endemic pathogen. Values of T between 0 and 1 can have positive or negative Q values, but there are not conditions with a negative Q where the foreign pathogen is more likely to infect than the endemic pathogen for the same reason described above when $T=1$. With $0 and a positive Q value, decreasing resistance costs (lower C) similarly disfavors the foreign pathogen being more likely than the endemic pathogen to infect healthy individuals due to a higher equilibrium frequency of the resistance allele, A2, that also protects against the foreign pathogen. We finally consider where $T<0$ (transitivity slope is negative), which represents cases where resistance to the two pathogens is negatively correlated. When $T<0$, and thus Q is positive, decreasing resistance costs (lower C) has the contrasting effect from when $T>0$, favoring conditions where the foreign pathogen is more likely to infect due to increasing the frequency of the A2 allele that is more susceptible to the foreign pathogen.

The potential for a foreign pathogen to subsequently spread within the host population will also be influenced by the transitivity slope. Denoting the equilibrium frequency of A1 as $p˜$, it can be shown that the condition for the foreign pathogen to increase following introduction (with frequency-dependent transmission) is

$(8)β12p˜h+β22(1−p˜h)−µ>0,$
namely, that the infection rates of the foreign pathogen, weighted by the equilibrium host genotype frequencies, has to be greater than the disease-independent mortality rate (Antonovics and Thrall 1994). Substituting the foreign pathogen transmission rates in inequality (8) as $β12=(S^′+T[β11−β21])/2$ and $β22=(S^′−T[β11−β21])/2$, by using the foreign pathogen’s average infection rate Ŝ′, the transitivity slope T, and the difference in resistance of the host genotypes to its endemic pathogen, the condition for foreign pathogen spread can be rewritten (see the supplemental PDF) as
where $q˜h=1−p˜h$. This outcome is consistent with the results shown in figure 3C. Considering that the equation component ($β21−β11$) will be negative by definition (see fig. 2A), when the resistance allele is rare, such that ($p˜h−q˜h$) is positive, then a positive transitivity slope, T, favors spread of the foreign pathogen. Correspondingly, when the resistance allele is more common, such that ($p˜h−q˜h$) is negative, a negative transitivity slope favors the foreign pathogen. These consequences of positive or negative transitivity slopes for favoring the foreign pathogen spread are similar to when the foreign pathogen is more likely than the endemic pathogen to infect a random healthy individual described above and can be viewed in terms of resistance costs because of the effects of costs on the equilibrium frequency of the resistance allele, as seen in figure 3B.

Our results illustrate that both positive and negative transitivity slopes have potential to determine an infection probability by a foreign pathogen that is larger than predicted by that foreign pathogen’s average infection rate value alone. We also show that the relevant equilibrium resistance allele frequencies for this to occur can be found within the range of values predicted by a model of anther smut disease. Moreover, these results are consistent with conditions under which the foreign pathogen will increase in frequency following arrival in a host population already maintaining an endemic pathogen. More generally, if we assume additive gene effects, the theory confirms that even where specific resistance loci have not been identified, a regression line for resistance transitivity that crossed the diagonal of the one-to-one line within the range of the observed genetic variants would provide qualitative information on the likely infection success of a foreign pathogen.

Discussion

These empirical and theoretical studies show that host responses to newly introduced foreign pathogens can be mediated by the resistance variation maintained through selection imposed by their endemic diseases. Reports of genetic variation for resistance to pathogens that the hosts have not previously experienced in nature are common (table S1). However, even though correlations in resistance to different endemic pathogens are well established (table S2; Leimu and Koricheva 2006), only a few studies on disease emergence have incorporated intraspecific variation for susceptibility in the new host as a factor, and then mostly in the context of a host’s evolutionary response to the emergent disease (Bonneaud et al. 2011; Alves et al. 2019). We also show experimentally that naturally occurring genetic variation to an endemic pathogen can be important to resistance variation against both related and unrelated foreign pathogens.

Prior research in the anther smut system provided evidence of positively correlated resistance variation to two endemic Microbotryum species on Silene uniflora (Chung et al. 2012), where the regression slope for among-host-family infection rates by the two pathogens was close to 1 (fig. S1). Those two endemic pathogens on S. uniflora are sometimes sympatric (Chung et al. 2012; Abbate et al. 2018) and thus could have selected for a generalized resistance against the dual threat of co-occurring diseases. However, such an adaptive explanation for positively correlated resistances would be unlikely to apply to the case of a foreign pathogen. The relatively infrequent spillover of anther smut disease from Silene latifolia onto S. vulgaris (Hood et al. 2019) suggest that the foreign pathogen is unlikely to have imposed significant selection on S. vulgaris that could explain the resistance transitivity that we observed. In addition, a substantial degree of specialization by the foreign pathogen, M. lychnidis-dioicae, has been shown to its host of origin, S. latifolia (Antonovics et al. 2002; Sloan et al. 2008).

Positive resistance transitivity may be expected when resistance mechanisms are effective against related pathogens with similar physiology or ecological modes of parasitism, in the same way that plant defenses can protect against different insect pests (Leimu and Koricheva 2006). However, we also found a strongly negative relationship between resistances to the endemic Microbotryum species and St. solani, a nectrotrophic generalist pathogen, where a trade-off in defensive mechanisms may be indicated. A negative relationship between resistance to biotrophic and necrotrophic plant pathogens has been described before and may be based on allocation to alternative resistance pathways (Glazebrook 2005; Spoel et al. 2007; Faris et al. 2010; Wolpert and Lorang 2016).

For a host population maintaining an endemic disease, our theoretical investigation showed that the slope and intercept of resistance transitivity can be a major determinant of the transmission success of an introduced foreign pathogen. The empirical study indicates that the range of parameter values we investigated theoretically are realistic to the anther smut system. While we found a positive correlation among family-level infection rates for most S. vulgaris families, we interestingly observed a small number of significant outliers from the overall correlation in resistance. These outlier families fell above the one-to-one line in the plot of infection rates by the endemic and foreign pathogens. This reflected that they are more susceptible to the foreign pathogen than to the endemic pathogen. Interpreting such outliers and the other S. vulgaris families based on single-locus effects (or additive effects of multiple loci) is difficult, and these significant deviations of family infection rates from the regression slope strongly suggest the presence of loci with epistatic interactions. These outliers may be similar to those also found in Chung et al. (2012), where the study of two endemic Microbotryum pathogens on S. uniflora showed a small number of significant outliers from a strongly positive regression among most families. A model incorporating more than one locus would clearly be needed to understand the interactions between such general and specific resistances, and it would extend our simple one-locus model to situations involving more complex infection genetics.

Regardless of the specific mechanisms of resistance to endemic and foreign pathogens, our basic theory shows that predicting the likelihood of infection by a foreign pathogen in natural populations depends on the frequency of resistance among the healthy host individuals. This extends the general concept of nonhost resistance by considering how the host’s population genetics contribute to variation in the average infection rates by foreign and endemic pathogens. Furthermore, if selection imposed by endemic pathogens varies among host populations, this could in turn lead to variable risk to foreign pathogen invasion. Even with positive resistance transitivity, there can be facilitation of foreign pathogen infection when equilibrium resistance frequencies are low prior to foreign pathogen introduction, as may be the case when resistance costs are large relative to the potential harm caused by the endemic disease (Antonovics and Thrall 1994).

While prior field surveys for cross-species disease transmission among hosts to anther smut disease (e.g., Hood et al. 2019) could have been influenced by selection imposed by their endemic pathogen species, the described theoretical framework provides some prediction about the risk of new disease transmission. The negative transitivity slope ($T<0$) seen for the host’s resistances to anther smut and the necrotophic fungus, S. solani, suggests that harboring endemic anther smut disease increases the risk of infection by such a foliar pathogen, especially when there is a high equilibrium frequency of resistance to the endemic disease (i.e., with low resistance costs). Among different anther smut fungi, the predictions for how the endemic disease affects the risk of infection by the foreign pathogen is complicated by the occurrence of the outlier families. When the outliers are included, and we observe a negative transitivity ($T<0$), the expectations are similar to the case of S. solani in that endemic disease favors infections by the foreign pathogen when there is a high equilibrium resistance frequency (i.e., with low resistance costs). However with the outliers omitted, the transitivity that we observed, where $0>T>1$ and with a negative y-intercept, suggests that the foreign anther smut pathogen is less likely to infect or spread than the endemic pathogen under any resistance frequency. In this system, more studies that quantify resistance costs and the frequencies of resistance, including the outlier type, may help us to understand the distributions of different anther smut fungi among host species and where cross-species disease transmission is most likely to occur.

The current study has only addressed aspects of transmission following initial exposure to a foreign pathogen and not its rate of increase or eventual equilibrium level. The ensuing dynamics will involve changes in resistance frequencies from simultaneous selection imposed by the new and preexisting diseases. Indeed, some of the best documented cases of the evolution of disease resistance are responses to introduced pathogens, as seen in myxoma virus in rabbits (Best and Kerr 2000; Alves et al. 2019) or conjunctivitis in house finches (Bonneaud et al. 2011, 2018). It is unclear in such cases whether correlations in resistance affected the dynamics of selection imposed by the new disease or then may have influenced the hosts’ resistance interactions with their endemic diseases. There may be additional important numerical dynamics to consider in future investigations. In particular, with the importance that resistance costs have in determining whether positive or negative resistance transitivity favors infection by the foreign pathogen, costs may also affect equilibrium levels of disease prevalence (Bower et al. 1994; Boots et al. 2009) and thus the availability of healthy hosts for colonization. Another interesting possibility is that negative resistance transitivity, while generally favoring foreign pathogen transmission when resistance costs are low, may subsequently contribute to the maintenance of polymorphism in resistance to the endemic disease. This effect on resistance polymorphism could occur because foreign pathogen invasion risk represents an additional type of cost of resistance to the endemic disease, similar to what has been posited for other systems in which there is a within-species trade-off in resistances to multiple endemic pathogen genotypes or species (Lively 1996; Hedrick 2002).

In conclusion, we show that a pathogen’s host range could be constrained not only by fixed mechanisms of baseline disease resistance (i.e., the average nonhost resistance) but also by resistance variation resulting from selection imposed by an endemic pathogen. Where resistance polymorphisms are maintained, it is important to determine their correlated effects on resistance to other foreign pathogens. Understanding how selection by an endemic disease might protect or predispose populations from the invasion by foreign pathogens may have important implications for disease threats to conservation goals and domestic populations. These effects remain to be directly investigated in most host-pathogen systems.

We are grateful to the anonymous reviewers and the editors for their insightful comments and suggestions that improved the manuscript. M.E.H. and E.B. were supported by a National Institutes of Health grant (R01GM140457).

N.L and V.L. designed and implemented the empirical studies with input from M.E.H. N.L. conceived and initiated the theoretical investigation, and J.A. further developed the model with input from E.B. and M.E.H. The manuscript was derived from a draft by N.L., with the writing by M.E.H. and J.A., and important input from E.B. and V.L.

Literature Cited

• Abbate, J. L., P. Gladieux, M. E. Hood, D. M. de Vienne, J. Antonovics, A. Snirc, and T. Giraud. 2018. Co-occurrence among three divergent plant-castrating fungi in the same Silene host species. Molecular Ecology 27:3357–3370.

• Abrams, P. A. and M. H. Cortez. 2015. The many potential indirect interactions between predators that share competing prey. Ecological Monographs 85:625–641.

• Alexander, H. M. 1989. An experimental field study of anther-smut disease of Silene alba caused by Ustilago violacea: genotypic variation and disease incidence. Evolution 43:835–847.

• Alexander, H. M., and J. Antonovics. 1995. Spread of anther-smut disease (Ustilago violacea) and character correlations in a genetically variable experimental population of Silene alba. Journal of Ecology 83:783–794.

• Alexander, H. M., J. Antonovics, and A. W. Kelly. 1993. Genotypic variation in plant disease resistance-physiological resistance in relation to field disease transmission. Journal of Ecology 81:325–333.

• Alves, J. M., M. Carneiro, J. Y. Cheng, A. L. de Matos, M. M. Rahman, L. Loog, P. F. Campos, et al. 2019. Parallel adaptation of rabbit populations to myxoma virus. Science 363:1319–1326.

• Antonovics, J., M. E. Hood, and J. Partain. 2002. The ecology and genetics of a host shift: Microbotryum as a model system. American Naturalist 160(suppl.):S40–S53.

• Antonovics J., and P. H. Thrall. 1994. Cost of resistance and the maintenance of genetic-polymorphism in host-pathogen systems. Proceedings of the Royal Society B 257:105–110.

• Ashby, B., R. Iritani, A. Best, A. White, and M. Boots. 2019. Understanding the role of eco-evolutionary feedbacks in host-parasite coevolution. Journal of Theoretical Biology 464:115–125.

• Ayliffe, M. and C. K. Sørensen. 2019. Plant nonhost resistance: paradigms and new environments. Current Opinion in Plant Biology 50:104–113.

• Best, S. M. and P. J. Kerr. 2000. Coevolution of host and virus: the pathogenesis of virulent and attenuated strains of myxoma virus in resistant and susceptible European rabbits. Virology 267:36–48.

• Bettgenhaeuser, J., B. Gilbert, M. Ayliffe, and M. J. Moscou. 2014. Nonhost resistance to rust pathogens: a continuation of continua. Frontiers in Plant Science 5:664.

• Bonneaud, C., S. Balenger, A. F. Russell, J. Zhang, G. E. Hill, and S. V. Edwards. 2011. Rapid evolution of disease resistance is accompanied by functional changes in gene expression in a wild bird. Proceedings of the National Academy of Sciences of the USA 108:7866–7871.

• Bonneaud, C., L. Tardy, M. Giraudeau, G. Hill, K. McGraw, and A. Wilson. 2018. Evolution of host resistance and damage-limiting mechanisms to an emerging bacterial pathogen. BioRxiv p.486365.

• Boots, M., A. Best, M. R. Miller, and A. White. 2009. The role of ecological feedbacks in the evolution of host defence: what does theory tell us? Philosophical Transactions of the Royal Society B 364:27–36

• Boots M, A. White, A. Best, and R. Bowers. 2014. How specificity and epidemiology drive the coevolution of static trait diversity in hosts and parasites. Evolution 68:1594–1606.

• Bowers, R. G., M. Boots, and M. Begon. 1994. Life-history trade-offs and the evolution of pathogen resistance: competition between host strains. Proceedings of the Royal Society B 257:247–253

• Bruns, E., J. Antonovics, and M. E. Hood. 2018. Is there a disease-free halo at species range limits? the co-distribution of anther-smut disease and its host species. Journal of Ecology 107:1–11.

• Cafuir, L., J. Antonovics, and M. E. Hood. 2007. Tissue culture and quantification of individual-level resistance to anther-smut disease in Silene vulgaris. International Journal of Plant Sciences 168:415–419.

• Carlsson-Granér, U. 1997. Anther-smut disease in Silene dioica: variation in susceptibility among genotypes and populations, and patterns of disease within populations. Evolution 51:1416–1426.

• Chung, E., E. Petit, J. Antonovics, A. B. Pedersen, and M. E. Hood. 2012. Variation in resistance to multiple pathogen species: anther-smuts of Silene uniflora. Ecology and Evolution 2:2304–2314.

• Cousineau, D., and S. Chartier. 2015. Outliers detection and treatment: a review. International Journal of Psychological. Research 3:58–67.

• Cui, H., K. Tsuda, and J. E. Parker. 2015. Effector-triggered immunity: from pathogen perception to robust defense. Annual Review of Plant Biology 66:487–511.

• Davies, T. J., and A. B. Pedersen. 2008. Phylogeny and geography predict pathogen community similarity in wild primates and humans. Proceedings of the Royal Society B 275:1695–1701.

• Day, A. W. 1979. Mating type and morphogenesis in Ustilago violacea. Botanical Gazette 140:94–101.

• De Schutter, K. and E. Van Damme. 2015. Protein-carbohydrate interactions as part of plant defense and animal immunity. Molecules 20:9029–9053.

• de Vienne, D. M., G. Refrégier, M. López-Villavicencio, A. Tellier, M. E. Hood, and T. Giraud. 2013. Cospeciation vs. host-shift speciation: methods for testing, evidence from natural associations and relation to coevolution. New Phytologist 198:347–385.

• Falconer, D. S. 1952. The problem of environment and selection. American Naturalist 86:293–298.

• Faris, J. D., Z. Zhang, H. Lu, S. Lu, L. Reddy, S. Cloutier, J. P. Fellers, et al. 2010. A unique wheat disease resistance-like gene governs effector-triggered susceptibility to necrotrophic pathogens. Proceedings of the National Academy of Sciences of the USA 107:13544–13549.

• Gilbert, G. S., and I. M. Parker. 2010. Rapid evolution in a plant-pathogen interaction and the consequences for introduced host species. Evolutionary Applications 3:144–156.

• Gilbert, G. S., and C. O. Webb. 2007. Phylogenetic signal in plant pathogen–host range. Proceedings of the National Academy of Sciences of the USA 104:4979–4983.

• Gilbert, B., J. Bettgenhaeuser, N. Upadhyaya, M. Soliveres, D. Singh, R. F. Park, M. J. Moscou, and M. Ayliffe. 2018. Components of Brachypodium distachyon resistance to nonadapted wheat stripe rust pathogens are simply inherited. PLoS Genetics 14:p.e1007636.

• Gladieux, P., E. Vercken, M. C. Fontaine, M. E. Hood, O. Jonot, A. Couloux, and T. Giraud. 2010. Maintenance of fungal pathogen species that are specialized to different hosts: allopatric divergence and introgression through secondary contact. Molecular Biology and Evolution 28:459–471.

• Glazebrook, J. 2005. Contrasting mechanisms of defense against biotrophic and necrotrophic pathogens. Annual Review of Phytopathology 43:205–227.

• Gold, A., T. Giraud, and M. E. Hood. 2009. Within-host competitive exclusion among species of the anther-smut pathogen. BMC Ecology 9:11.

• Gross, W. B., P. B. Siegel, R. W. Hall, C. H. Domermuth, and R. T. Duboise. 1980. Production and persistence of antibodies in chickens to sheep erythrocytes: II. resistance to infectious 6 diseases. Poultry Science 59:205–210.

• Hartmann, F. E., R. C. Rodríguez de la Vega, F. Carpentier, P. Gladieux, A. Cornille, M. E. Hood, and T. Giraud. 2019. Understanding adaptation, coevolution, host specialization, and mating system in castrating anther-smut fungi by combining population and comparative genomics. Annual Review of Phytopathology 57:431–457.

• Heath, M. C. 1981. A generalized concept of host-parasite specificity. Phytopathology 71:1121–1123.

• Hedrick, P. W. 2002. Pathogen resistance and genetic variation at MHC loci. Evolution 56:1902–1908.

• Hood, M. E., and J. Antonovics. 2000. Intratetrad mating, heterozygosity, and the maintenance of deleterious alleles in Microbotryum violaceum = Ustilago violacea. Heredity 85:231–241.

• Hood, M. E., J. Antonovics, M. Wolf, Z. L. Stern, T. Giraud, T., and J. L. Abbate. 2019. Sympatry and interference of divergent Microbotryum lineages. Ecology and Evolution 9:5457–5467.

• Hood, M. E., N. Lerner, V. Luizzi, J. Antonovics, and E. Bruns. 2021. Data from: Resistance correlations influence infection by foreign pathogens. American Naturalist, Dryad Digital Repository, https://doi.org/10.5061/dryad.dr7sqv9xs.

• Hood, M. E., J. I. Mena-Alí, A. K. Gibson, B. Oxelman, T. Giraud, R. Yockyeng, M. T. K. Arroyo, et al. 2010. Distribution of the anther-smut pathogen Microbotryum on species of the Caryophyllaceae. New Phytologist 187:217–229.

• Jones, K. E., N. G. Patel, M. A. Levy, A. Storeygard, D. Balk, J. L. Gittleman, and P. Daszak. 2008. Global trends in emerging infectious diseases. Nature 451:990–993.

• Keller, S. R., D. R. Sowell, M. Neiman, L. M. Wolfe, and D. R. Taylor. 2009. Adaptation and colonization history affect the evolution of clines in two introduced species. New Phytologist 183:678–690.

• Kido, A., and M. E. Hood. 2020. Mining new sources of natural history observations for disease interactions. American Journal of Botany 107:3–11.

• Leimu, R., and J. Koricheva. 2006. A Meta-analysis of genetic correlations between plant resistances to multiple enemies. American Naturalist 168:E15–E37.

• Lindahl, J. F., and D. Grace. 2015. The consequences of human actions on risks for infectious diseases: a review. Infection Ecology and Epidemiology 5:30048.

• Lively, C. M. 1996. Host-parasite coevolution and sex. BioScience 46:107–114.

• Loiseau, C., R. Zoorob, S. Garnier, J. Birard, P. Federici, R. Julliard, and G. Sorci. 2008. Antagonistic effects of a Mhc class I allele on malaria-infected house sparrows. Ecology Letters 11:258–265.

• Marmor, M., K. Hertzmark, S. M. Thomas, P. N. Halkitis, and M. Volger. 2006. Resistance to HIV infection. Journal of Urban Health 83:5–17.

• McDonald, B. A., and E. H. Stukenbrock. 2016. Rapid emergence of pathogens in agro-ecosystems: global threats to agricultural sustainability and food security. Philosophical Transactions of the Royal Society B 371:20160026.

• Nsabiyera, V., M. Ochwo-Ssemakula, and P. Sseruwagi. 2012. Hot pepper reaction to field diseases. African Crop Science Journal 20:77–97.

• Nürnberger, T., F. Brunner, B. Kemmerling, and L. Piater. 2004. Innate immunity in plants and animals: striking similarities and obvious differences. Immunological Reviews 198:249–266.

• Orieux, L., and S. Felix. 1968. List of plant diseases in Mauritius. Phytopatholical Papers Commonwealth Mycological Institute 7:1–48.

• Parker, I. M., M. Saunders, M., Bontrager, A. P. Weitz, R. Hendricks, R. Magarey, K. Suiter and G. S. Gilbert. 2015. Phylogenetic structure and host abundance drive disease pressure in communities. Nature 520:542.

• Refrégier, G., M. Le Gac, F. Jabbour, A. Widmer, J. A. Shykoff, R. Yockteng, M. E. Hood, and T. Giraud. 2008. Cophylogeny of the anther smut fungi and their caryophyllaceous hosts: prevalence of host shifts and importance of delimiting parasite species for inferring cospeciation. BMC Evolutionary Biology 8:100.

• Rueden, C. T., J. Schindelin, M. C. Hiner, B. E. DeZonia, A. E. Walter, E. T. Arena, K. W. Eliceiri. 2017. ImageJ2: ImageJ for the next generation of scientific image data. BMC Bioinformatics 18:529.

• Sloan, D. B., T. Giraud, and M. E. Hood. 2008. Maximized virulence in a sterilizing pathogen: the anther-smut fungus and its co-evolved hosts. Journal of Evolutionary Biology 21:1544–1554.

• Spoel, S. H., J. S. Johnson, and X. Dong. 2007. Regulation of trade-offs between plant defenses against pathogens with different lifestyles. Proceedings of the National Academy of Sciences of the USA 104:18842–18847.

• Stam, R., S. Mantelin, H. McLellan, and G. Thilliez. 2014. The role of effectors in nonhost resistance to filamentous plant pathogens. Frontiers in Plant Science 5:582.

• Thines, M. 2019. An evolutionary framework for host shifts–jumping ships for survival. New Phytologist 224:605–617

• Via, S. and R. Lande. 1985. Genotype-environment interaction and the evolution of phenotypic plasticity. Evolution 39:505–522.

• Webby, R., E. Hoffmann, and R. Webster. 2004. Molecular constraints to interspecies transmission of viral pathogens. Nature Medicine 10:S77.

• Weber, G. F. 1930. Gray leaf spot of tomato caused by Stemphylium solani, sp. nov. Phytopathology 20:513–518.

• White, L. A., J. D. Forester, and M. E. Craft. 2018. Dynamic, spatial models of parasite transmission in wildlife: their structure, applications and remaining challenges. Journal of Animal Ecology 87:559–580.

• Wolpert, T. J., and J. M. Lorang. 2016. Victoria Blight, defense turned upside down. Physiological and Molecular Plant Pathology 95:8–13.

• Woolhouse, M. E., D. T. Haydon, and R. Antia. 2005. Emerging pathogens: the epidemiology and evolution of species jumps. Trends in Ecology and Evolution 20:238–244.

• Woolhouse, M. E., J. P. Webster, E. Domingo, B. Charlesworth, and B. R. Levin. 2002. Biological and biomedical implications of the co-evolution of pathogens and their hosts. Nature Genetics 32:569.

References Cited Only in the Online Enhancements

• Fellowes, M. D. E., A. R. Kraaijeveld, and H. C. J. Godfray. 1999. Cross-resistance following artificial selection for increased defense against parasitoids in Drosophila melanogaster. Evolution 53:966–972.

• Fokunang, C. N., T. Ikotun, A. G. Dixon, and C. N. Akem. 2000. Field reaction of cassava genotypes to anthracnose, bacterial blight, cassava mosaic disease and their effects on yield. African Crop Science Journal 8:179–186.

• Hawley, D. M., K. V. Sydenstricker, G. V. Kollias, and A. A. Dhondt. 2005. Genetic diversity predicts pathogen resistance and cell-mediated immunocompetence in house finches. Biology Letters 1:326–29.

• Intemann, C. D., T. Thye, S. Niemann, E. N. Browne, M.A. Chinbuah, A. Enimil, J. Gyapong, et al. 2009. Autophagy gene variant IRGM−261T contributes to protection from tuberculosis caused by Mycobacterium tuberculosis but not by M. africanum strains. PLoS Pathogens 5:e1000577.

• Liu, J. J., A. K. Ekramoddoullah, N. Piggott, and A. Zamani. 2005. Molecular cloning of a pathogen/wound-inducible PR10 promoter from Pinus monticola and characterization in transgenic Arabidopsis plants. Planta 221:159–169.

• McCarthy, A. J., M. A. Shaw, P. D. Jepson, S. M. Brasseur, P. J. Reijnders, and S. J. Goodman. 2011. Variation in European harbour seal immune response genes and susceptibility to phocine distemper virus (PDV). Infection, Genetics, and Evolution 11:1616–1623.

• Mitchell-Olds, T., R. V. James, M. J. Palmer, and P. H. Williams. 1995. Genetics of Brassica rapa (syn. campestris). 2. multiple disease resistance to three fungal pathogens: Peronospora parasitica, Albugo candida, and Leptosphaeria maculans. Heredity 75:362.

• Ostry, M. E., and M. Moore. 2008. Response of butternut selections to inoculation with Sirococcus Clavigignenti-Juglandacearum. Plant Disease 92:1336–1338.

• Ranney, T. G., L. F. Grand, and J. L. Knighten. 1995. Susceptibility of cultivars and hybrids of kousa dogwood to dogwood anthracnose and powdery mildew. Journal of Arboriculture 21:11.

• Savage, A. E., and K. R. Zamudio. 2011. MHC genotypes associate with resistance to a frog-killing fungus. Proceedings of the National Academy of Sciences of the USA 108:16705–16710.

• Wisser, R. J., J. M. Kolkman, M. E. Patzoldt, J. B. Holland, J. Yu, M. Krakowsky, R. J. Nelson, and P. J. Balint-Kurti. 2011. Multivariate analysis of maize disease resistances suggests a pleiotropic genetic basis and implicates a GST gene. Proceedings of the National Academy of Sciences of the USA 108:7339–7344.

Notes

ORCIDs: Hood, https://orcid.org/0000-0003-4379-0711.

Associate Editor: Michael H. Cortez

Editor: Jennifer A. Lau