Portal de Periódicos da CAPES

Fungal epidemics – does spatial structure matter?

2004; Wiley; Volume: 163; Issue: 1 Linguagem: Inglês

10.1111/j.1469-8137.2004.01116.x

1469-8137

A. Drenth,

Mycotoxins in Agriculture and Food

Humans have been plagued by a plethora of pathogens and pests throughout history. Although our crops and surrounding environment are under continuous attack from pests and diseases, only in a limited number of cases do we experience an epidemic, defined as a rapid and devastating spread of a pathogen or pest. Scientists have studied epidemics for a long time to try to understand the factors involved which give rise to the sudden spread of a disease. A number of factors have been understood in this process for a long time, such as the presence of a virulent pathogen, presence of a susceptible host, and suitable environmental conditions for spread, infection and colonization by the pathogen. However, countless times conditions are suitable for rapid spread of a plant disease but no epidemic occurs. Could spatial distribution of susceptible hosts be responsible for this? The paper by Otten et al. in this issue (pp. 125–132) is part of series which aims to provide a conceptual framework that explains the observed disease spread of soil-borne fungal plant pathogens. ‘How long can we continue to ignore the erosion of genetic and spatial structure in our food production?’ Van der Plank (1963) was among the first to develop a solid theoretical framework for the epidemiology of plant diseases in the 1960s. Models were developed and tested which described the development of an epidemic. Van der Plank put forward a model that defined three key parameters (infection rate, latent period and infectious period), which allowed one to determine if the threshold for an epidemic to take place would be reached. Over the years the models put forward by van der Plank and others have been refined and adapted to different host pathogen systems in an effort to use them to support decision making in disease management (Jeger & van den Bosch, 1994; Jeger, 2000). The development of a theoretical framework by van der Plank, underpinned by experimental approaches in many different host pathogen systems, moved the field of plant disease epidemiology – especially of fungal foliar pathogens – forward significantly. The theoretical framework developed allowed us to model and determine thresholds for epidemics to take place in a homogeneous environment such as a single field. Because of the additional complex the soil medium, the epidemiology of soil-borne plant pathogens has not advanced as much as for their foliar counterparts. Many soil-borne diseases have a rather different mode of dispersal compared to foliar pathogens and one which is also much slower. A typical soil-borne pathogen may spread from its initial point of infection slowly over a number of years to invade a single field. Their spread is often helped along by agricultural practices such as cultivation, irrigation and harvesting. Scientists working on the epidemiology of soil-borne plant pathogens have long suspected that spatial distribution of susceptible host plants in a single field or over large areas was important. However, practical tools to determine, rapidly and accurately, the distribution of pathogens in the soil were hard to come by and a theoretical framework to test the observed data against was not available. Significant advances in our understanding of the nature of host–pathogen interactions at the molecular level and the increasing availability of DNA-based tools for detecting, identifying and quantifying pathogen populations in the soil medium have occurred in the past decade. Thus we may expect some advances concerning the epidemiology of soil-borne organisms to take place in the years ahead. In order for these tools, such as DNA-based diagnostics, which allow accurate tracking of soil-borne pathogens, to be put to effective use there is a need for the development of a theoretical framework to model the spread of these pathogens. An important issue in epidemiology is the determination of the relative importance of the biological parameters that control the epidemic. Especially in the case of soil-borne diseases, which rely to some extent on root contact between susceptible host plants, the spatial distribution of the host plants may be one of the most important parameters that determines whether or not invasion by these pathogens takes place. A long-standing and much-debated question in plant pathology is whether there is less disease in diverse ecological systems as a result of increased distance between susceptible host plants, which dilute the inoculum and thus slow down the epidemic. However, we rarely try to address this question as we tend to determine disease development and effectiveness of disease control measures in replicated, small-scale experimental plots. The effectiveness of the introduction of spatial heterogeneity involving host plants with different levels of susceptibility for a particular disease is rarely tested in large-scale field experiments. The scale of this diversity can be from mixed lines to alternating rows, alternating strips to a patchwork of fields with different cultivars or host plants. An excellent example where the effectiveness of crop heterogeneity on the severity of the rice blast disease, was tested on a large scale, was conducted in China and published in Nature by Zhu et al. (2000). The question as to whether or not invasion and establishment of a pathogen will take place is asked typically after a new disease occurs in a region. In order to predict what will happen we need to understand what sort of dynamics underlie invasion by a plant pathogen. The relevance of spatial threshold levels lies in the fact that this is vital information to manage invasive pathogens. Most eradication campaigns of invasive pathogens are based on removal of susceptible host material within a certain radius of the initial infection point. Needless to say, it would be beneficial to have a way to determine the threshold probability of spread between susceptible hosts as a guide to determine such a radius. Although the earlier work by van der Plank advanced the insight into the epidemiology of foliar plant pathogens and led to the creation of numerous models using a basic set of underlying principles, a theoretical basis to predict invasion in a heterogeneous environment, such as cultivar mixtures, alternating rows, intercropping, to a patchwork of small fields in an area, was lacking. Therefore in order to understand the processes that influence invasion of a disease in a soil medium or heterogeneous environment where the spatial structure is an important parameter a new theoretical framework is needed. The theoretical basis for plant disease epidemics has not been established to the same extent as in human and animal epidemiology (Michael, 1993; Jeger, 2000). Gilligan (2002) correctly points out that the spatial structure of epidemics is widely acknowledged for its importance, and then widely ignored. Spatial structure can have a significant influence on ecological and evolutionary dynamics in host pathogen systems (Thompson, 1994). In the discipline of plant pathology there is a need to develop the theory for soil-borne plant pathogens further, instead of stumbling on empirical data, to underpin development of effective disease management practices. Invasion may be defined as the introduction and subsequent increase of infection in a host population. Invasion starts with the arrival of the initial inoculum, expansion to a patch, then several patches in a single field to spread of a disease over an entire region. Another important factor is the persistence threshold which indicates whether there are sufficient susceptible host plants to maintain both the pathogen and host plant population over a long period of time in some form of coexistence. In the case of soil-borne pathogens it makes sense to assume that not only the number but also the spatial distribution of these susceptible host plants is important for both invasion and persistence of a pathogen population over time. Spatial heterogeneity has received a lot of attention in human and animal populations (Swinton & Gilligan, 1998) and theories for invasion and persistence have been developed (Anderson & May, 1991; Grenfell & Dobson, 1995; Mollison & Levin, 1995), but spatial heterogeneity has received relatively little attention in plant pathogen systems. Some of the theories which are relevant to plant diseases are the metapopulation concept and the percolation theory. The metapopulation concept (Hanski & Gilpin, 1997) assumes the existence of a number of smaller subpopulations with disease spread occurring within and between these stratified subpopulations. These subpopulations can occur at different spatial scales from single plants, for example trees in an orchard, to fields within regions. Although this concept has been widely used in epidemiological studies on mammals it has received relatively little attention in plant pathology (Thrall & Burdon, 1997). A pathogen may be present and persist in some fields while it is absent in others. Thus the basic concept is that we do not have a single epidemic or invasion but many of these events take place at different points in time and space. The existence of many subpopulations instead of a single large population affects the probability of invasion and persistence (Levin & Durrett, 1996). The percolation model allows one to determine the critical threshold probability for the transmission of infection between neighboring plants or fields (Grassberger, 1983). Above the threshold the disease spreads while below the threshold the disease fails to advance. Thus this model can be used to estimate the probability for the transmission of infection between different small subpopulations of the host plant. Relevant questions which this model can address include: will a pathogenic strain invade?; how long will it take to invade?; can the pathogen invade and persist?; how is this affected by the spatial structure of the host plants? The metapopulation and percolation theory work well in concert whereby we look at the epidemiology of the small subpopulations in the large metapopulation and then use the percolation model to determine the probability of spread of disease between these small subpopulations. The disease dynamics within each subpopulation typically do not explain the overall disease spread in a region but if we add the percolation model to determine the probability of spread between different subpopulations we can simulate epidemics over a much larger scale. The concept of spatial distribution of the different subpopulations is now introduced into the epidemiological model. The theory developed by Otten et al. in this issue is especially relevant for fungi which have hyphal strands that have the ability to grow from plant to plant through root contact in soil. The authors apply the percolation theory to develop a model that can determine the spatial threshold needed for fungal invasion. An increased understanding of spatial threshold levels for pathogenic microorganisms is important for scientists working on the epidemiology of important plant diseases (Shea et al., 2000). It also can provide useful information in eradication campaigns of invasive plant diseases where decisions have to be made about the distance from an infection that susceptible host plants have to be removed. Invasion theories can also be applied to estimate the effect of the introduction of partially resistant host cultivars and determine the spread of fungicide resistant strains of a pathogen. Another area where spatial structure is relevant is the application of soil-borne biological control agents to control weeds. Because of the soil medium it is to be expected that epidemics of soil-borne pathogens take place at a smaller spatial scale than airborne pathogens (Thrall & Burdon, 1997). We also have to keep the limitations of developed models in mind as spatial heterogeneity is mostly nonrandom, multiple modes of reproduction of fungal pathogens exist and of course humankind itself, by any measure a significant contributor to spread of plant diseases as a result of agricultural pursuits. The spatial dynamics which existed at various levels of scale in mixed landraces, and patchworks of small fields with different crops and different varieties are quickly disappearing in our modern agriculture. The mechanization of agriculture, the green revolution, chemical-based crop protection, increase in the scale of farming operations and at present the introduction of genetically modified crops are all moving us rapidly in the direction of a more homogeneous agricultural environment. To a large degree we are focusing on yield quantity and quality. Levels of disease resistance are established in small experimental field plots and we often do not hesitate to extrapolate this data from a relatively small area to a continental scale. The danger of this is that we ignore the bigger picture as we take on ever greater risks by increasing the genetic and spatial uniformity of our food supply. A classic textbook example of this is the large-scale use of T-cytoplasm to produce male-sterile plants in hybrid corn production in the US. This cytoplasm also carried a gene that made these hybrids susceptible to Helminthosporium maydis race T which resulted in the destruction of over a billion dollar worth of corn in a single year (Zadoks & Schein, 1979). A recent example is orange rust in sugarcane caused by Puccinia kuehnii, which decimated variety Q124 which made up 87% of the area under cultivation in the Mackay region in Queensland, Australia. For how long can we continue to ignore the erosion of genetic and spatial structure in our food production? It is in this light that we should think about spatial issues and discuss and encourage the development of a theoretical framework, underpinned by solid experimental science, to provide guidance on how we can exploit spatial structure most effectively to secure a reliable food supply in the future.