The evolution of pathogens is widely believed to be one of the major challenges facing agriculture and medicine [1
]. Experimental studies focused on the evolution of pathogens, including the emergence of virulence and pathogen adaptation to changing agricultural and medical practices, can provide critical information for more effective management of infectious diseases. In medicine, infectious diseases are mitigated mainly through the application of antimicrobial substances such as antibiotics. While pesticides such as fungicides are widely utilized in agricultural ecosystems, host resistance imposes many fewer environmental costs and is a more cost efficient approach to control plant diseases. In both agriculture and medicine, the efficacy of host resistance and antimicrobials usually decays over time as a result of the continuous evolution and adaptation of pathogens.
Plant pathogens are thought to evolve faster in agricultural ecosystems than in natural ecosystems [3
]. Wild relatives are the primary sources of host resistance bred into cultivated crops. The disease resistance genes carried by these wild relatives of modern crop plants have coexisted with their pathogens for many thousands or millions of years in natural ecosystems. However, when these resistance genes are introgressed into modern crops and deployed in agricultural ecosystems, their value in controlling infectious diseases usually does not last for more than 10 years [7
]. Temporal analysis of population dynamics also is consistent with the hypothesis of rapid pathogen evolution in agricultural ecosystems. For example US-1 was the predominant genotype in Phytophthora infestans
populations around the world until the 1990s [8
] but this genotype is rarely recovered since 2000. In the UK, a single P. infestans
genotype called Blue-13-A2 was first detected in southern England in 2003 at a very low frequency. By 2007, this genotype was detected in all populations sampled across the UK and accounted for more than half of 1000 isolates assayed (J. Zhan & D. Cooke, unpublished data). These P. infestans
examples illustrate how pathogen populations can experience rapid turnover as new genotypes with greater fitness emerge, spread, out-compete and replace earlier genotypes.
The evolution of pathogens can be influenced by the type of resistance, the amount of diversity found in host populations and the type of cropping system [9
]. Modern agriculture is dominated by species monocultures grown at a high density. In these agricultural ecosystems, it is common for a single host cultivar or genotype carrying a major resistance gene to be grown over a large area. The limited genetic diversity in the host populations coupled with intensive use of major resistance genes can lead to rapid shifts in associated pathogen populations. A mutant with higher fitness that emerges in a pathogen population as a result of a single mutation event can quickly increase in frequency through strong directional selection and spread across entire fields or regions through natural or human-mediated migration.
Multi-cropping, where the same annual crop is grown in the same field more than once during the same year, is another common practice in modern agriculture, especially in countries experiencing a shortage of arable land. This practice may further accelerate the evolution of plant pathogens because locally adapted pathogen genotypes with a high parasitic fitness can steadily increase in frequency due to the year-around availability of the living host (i.e. a "green bridge" allows the parasitic phase of the pathogen life cycle to occur continuously).
It is hypothesized that the evolution of plant pathogens in agricultural ecosystems can be retarded by increasing genetic diversity of the host populations, by using partial resistance encoded by several genes and by avoiding multi-cropping systems. Increasing genetic diversity in host populations by mixing plants carrying different major resistance genes (e.g. cultivar mixtures or multilines) is thought to be an ecologically and evolutionarily sound approach to control plant diseases, particularly for airborne pathogens of cereals [13
]. Increasing host diversity by using cultivar mixtures will impose disruptive selection on pathogen populations, i.e. pathotypes that are favored on one host will have lower fitness on the other hosts in the mixture [14
], impeding their ability to evolve towards higher virulence, here defined as the damage a pathogen causes to its host [17
]. On the other hand, because many fungal pathogens have large effective population sizes [18
] and exhibit a mixture of sexual and asexual reproduction [19
], they can rapidly obtain new pathogenicity factors through mutation or new combinations of pathogenicity factors through recombination and then maintain the novel combinations of pathogenicity factors through asexual reproduction. Thus extensive use of cultivar mixtures could lead to the development of complex races [21
] that would be able to infect a large number of host genotypes carrying different major resistance genes.
Though less efficient, partial resistance is thought to offer a more durable method to control plant diseases than major-gene resistance because it works against all pathogen strains and selects equally against all pathotypes [23
]. Partial resistance mediated by multiple genes is generally inherited as a quantitative trait [16
], where each gene makes a minor but additive contribution to the overall resistance [27
]. But selection can increase the frequencies of genes encoding higher virulence in pathogen populations infecting partially resistant hosts and reduce the effectiveness of quantitative resistance [9
] though possibly at a slower pace compared to major resistance genes [32
In contrast to multi-cropping, in single cropping systems an annual crop is grown for only 6-9 months of the year or different crops are rotated annually, forcing pathogens to undergo a saprophytic phase in their life cycle in which different strains not only compete with each other but also with other microbial species for nutrients and habitats. Pathogen genotypes that have a high parasitic fitness on living hosts may have a low saprophytic fitness on the dead host biomass. This trade-off could delay the emergence of highly parasitic pathogen strains in agricultural ecosystems characterized by single cropping and regular crop rotations.
Much of our knowledge of pathogen evolution in agricultural ecosystems is drawn from theory or through historical inference from population surveys [33
]. Experimental tests of pathogen evolution are limited and usually are conducted in controlled environments under laboratory or greenhouse conditions. Here we describe a test of pathogen evolution in a replicated experiment using sensitive molecular markers that could differentiate among pathogen isolates released into an unregulated field setting. This experimental evolution approach based on a mark-release-recapture strategy has now been successfully applied to understand the evolution of cereal pathogens including Mycosphaerella graminicola
], Rhynchosporium secalis
] and Phaeosphaeria nodorum
]. In this study, we used this approach to investigate the evolution of the wheat pathogen Phaeosphaeria nodorum
. The experiment was conducted over two years using five replicated host populations differing in levels of resistance and diversity. In the first year, the pathogen populations were introduced into each host population by artificial inoculation of the hosts with nine P. nodorum
strains tagged with molecular genetic markers and mixed in equal proportions. In the second year, the pathogen populations were established using the infected straw and plant debris saved from the first year's experiment. During the experiment, two fungal collections were made in each of the two years. The recovered pathogen populations were assayed for their molecular markers so that frequencies of the inoculated isolates could be compared across hosts and sampling times (see Figure ). With this experimental design, we were able to determine the effects of host diversity and resistance on the evolution of corresponding pathogen populations. The experimental design also allowed us to detect selection operating during both parasitic and saprophytic phases of the pathogen life cycle. The specific objectives of this experiment were to: i) infer the rate of pathogen evolution in an agricultural system; ii) determine the effect of host resistance on competition among genotypes in P. nodorum
populations; iii) determine the effect of cultivar mixtures on clonal competition in P. nodorum
populations; and iv) compare selection during the parasitic and saprophytic phases of the P. nodorum
life cycle. Our previous data analyses indicated that isolates recovered from the experiment included the asexual progeny of the inoculated genotypes, airborne immigrants from outside of the experimental plots and recombinants arising from crosses between the inoculants and/or immigrants [39
]. The results presented here consider only the effects of host selection and clonal competition among the asexual progeny of the inoculated genotypes. Evolutionary changes in the pathogen populations attributed to recombination and immigration were considered in a separate publication [39
Schematic of experimental design illustrating inoculation and sampling procedures used over the course of the 2-year field experiment conducted in Switzerland during the 2003-2004 and 2004-2005 winter wheat growing seasons.
The heterothallic loculoascomycete Phaeosphaeria nodorum
(Berk.) Castellani and Germano (syn. Septoria nodorum
Berk.), the teleomorph form of Stagonospora nodorum
(E. Müller) Hedjaroude (syn. Leptosphaeria nodorum
E. Muller), causes Stagonospora nodorum leaf and glume blotch on wheat (Triticum aestivum
L.). The pathogen can undergo both sexual and asexual reproduction (see Figure for the life cycle) and has the ability to infect all above-ground plant parts during the parasitic phase [40
]. The pathogen overwinters during its saprophytic phase on infected stubble [44
] and can survive for several months [45
] on wheat straw until the parasitic phase of the disease cycle is re-initiated. The primary inoculum includes infected seeds as well as pycnidiospores and ascospores. Asexual pycnidiospores are dispersed over short distances by rain-splash while sexual ascospores are wind-dispersed, therefore having the potential for long distance movement [46
]. Ascospore-producing perithecia of P. nodorum
can be formed during the host-free period in infested stubble on the soil surface [50
] and during the growing season on infected plants [39
Life cycle of Phaeosphaeria nodorum causing Stagonospora nodorum blotch of wheat.