Population dynamics of a non-cultivated biennial plant Tragopogon pratensis infected by the autoecious demicyclic rust fungus Puccinia hysterium
Highlights
► SIR model with characteristics of Tragopogon pratensis – Puccinia hysterium. ► Produced cyclical, stable and extinction dynamics. ► Pathogen-induced mortality had little additional role in population dynamics. ► Limited evidence of density-dependence in seedling emergence. ► Pathogen characteristics contribute to host regulation.
Introduction
Pathogens can play a key ecological role in natural plant communities impacting on species performance, affecting viability and fecundity of individual plants, reducing population size, generating selective forces for genetic change, and altering community structure (Burdon, 1982, Burdon et al., 1989, Dinoor and Eshed, 1990, Thrall and Burdon, 1997, Frantzen and Müller-Schärer, 2002, Burdon et al., 2006). However, there have been few attempts to model epidemics in natural plant populations in non-cultivated systems, with many studies mostly limited to empirical descriptions that ignore theoretical implications or applications (Dobson & Grenfell 1995). Models have been developed describing the population dynamics of annual plant hosts and soil-borne pathogens (Thrall et al. 1997); dispersal characteristics and disease dynamics in populations (Thrall and Burdon, 1999, Frantzen and van den Bosch, 2000); and especially in recent years, combining spatial statistics with spatially explicit models to estimate rates of spread and effects of regional heterogeneity on co-evolutionary processes (Ovaskainen and Laine, 2006, Brooks et al., 2008, Soubeyrande et al., 2009, Smith et al., 2011).
The Park Grass Experiment (PGE) at Rothamsted Research, Harpenden, UK, is the oldest continuous ecological experiment in the UK, with work conducted on long-term plant population dynamics over many years (Silvertown et al. 2002). Plant species within the PGE have shown a range of dynamics including: increases (e.g. Trifolium pratense), decreases (e.g. Veronica chamaedrys), and fluctuations (e.g. Conopodium majus) in population size. Other species (e.g. Tragopogon pratensis) followed an outbreak dynamic in which the population increased and then declined (Dodd et al. 1995). Many of these patterns have been interpreted as consequences of the long-term fertiliser regimes altering the soil nutrient content and pH. However, it has been speculated that the outbreak pattern of dynamics of the biennial species T. pratensis has been due to an autoecious, demicyclic rust, Puccinia hysterium (Silvertown et al. 2006). It is also possible that edaphic factors may affect the plant–pathogen interaction directly, as found in other systems studied (Springer 2009). Unfortunately the long-term PGE data set does not include the recorded incidence of rust infection.
Importantly, the rust is systemic during the wintering of the biennial host. This systemic nature is relatively unusual as systemic rusts are more prevalent in arctic and montane ecosystems (Wennström 1999). The example most cited of a systemic rust is Puccinia punctiformis on Cirsium arvense and other thistle species (Frantzen, 1994, Cripps et al., 2010), with other examples noted including Puccinia pulsatillae, Puccinia pratensis, Puccinia monoica and Puccinia thlaspeos (Jarosz & Davelos 1995). Similarly to P. hysterium, P. thlaspeos is a systemic rust but does not vertically transmit through seeds (Kropp et al. 2002), therefore transmission is solely through spore dispersal. P. hysterium suppressed reproduction of T. pratensis (Salama et al. 2010), whilst in a similar non-cultivated host–pathogen system, the rust Puccinia lagenophorae significantly altered plant fitness by reducing seed production of Senecio vulgaris (Paul and Ayres, 1986a, Paul and Ayres, 1986b). Both P. lagenophorae and P. hysterium are demicyclic and lack an asexual repeating urediospore phase (Wilson & Henderson 1966).
The scientific literature on P. hysterium is sparse. The life history involves teliospores from diseased second year plants infecting first year plants from Jun. to the end of the growing season (Parmelee & Malloch 1972). Symptoms are not normally seen on first year plants. The dispersal period is often interrupted in traditional hay meadows (such as the PGE) by a “hay cut” at the end of Jun. which limits the infective period. Basidia develop from germinated teliospores on first year plants and produce basidiospores which can also be transported to new hosts. The basidiospores germinate to produce systemic mycelia, which enable the pathogen to survive the die-back period during the host’s below-ground overwintering phase. Pycnia develop and undertake a sexual stage as pycniospores. The following season, surviving plants re-emerge exhibiting aecia formed mostly on the stem, giving the typical “cluster cup” appearance, which produce aeciospores. Aeciospores germinate on the host tissue and germ tubes either directly penetrate tissue or grow through natural openings such as stomata. Telia, bearing teliospores on localised mycelium are then produced (Wilson & Henderson 1966) completing the cycle.
This paper investigates the hypothesis, presented by Silvertown et al. (2006), that P. hysterium has a regulatory influence on its host population dynamics and that this relationship is density-dependent (Silvertown et al. 2002). This assessment is undertaken by presenting data relating to the long-term population dynamics of the system in the PGE and attempting to obtain information on pathogen-induced mortality through the use of inoculation experiments. Biological characteristics of a biennial host – systemic pathogen system are then incorporated into an epidemiological model which is used to determine conditions under which population dynamics similar to those observed for T. pratensis infected by P. hysterium in the PGE are simulated. Specifically the epidemiological model takes into account the biennial nature of the host plant T. pratensis and the systemic characteristics of the rust, P. hysterium. The model is of SIR form, in which host categories are defined as Susceptible, Infectious and Removed (Anderson & May 1979), and models non-continuous host generations using discrete time. Such compartmentalised models of disease dynamics have been used previously to represent plant–fungal pathogen dynamics in continuous time (Gilligan 2002). The model developed here is an example where a discrete-time approach (Allen, 1994, Switkes, 2003) provides a more appropriate representation of the host–pathogen system, given the biennial host characteristics, the unusual yet key systemic nature of the rust pathogen, and the infection process linking second and first year plants. Expanded forms of the discrete SIR model have been used to describe gene frequency and disease spread in plant populations (Kesinger et al. 2001).
The model was developed to be sufficiently flexible to include the impact of pathogen-induced mortality on the ability of first year plants to overwinter and to re-emerge and grow in the second growing season. Pathogen-induced mortality – often termed virulence in human and animal epidemiology, see Antonovics (2005) – has been reported for a wide range of plant foliar diseases (e.g. Mycosphaerella laricinia on Pinaceae), systemic diseases (e.g. Urocystis trientalis on Trientalis europaea), as well as a range of cankers, wilts and butt rots (Gilbert 2002). Thus we incorporated within, as well as between, growing season pathogen-induced mortality. Specifically this type of pathogen-induced mortality has been reported with late infections of Puccinia, e.g. S. vulgaris infection by P. lagenophorae; at the beginning of a growing season, infected host plants have a lower probability of survival, but also increase their chances of mortality in the overwintering phase (Frantzen & Müller-Schärer 1999). Within our model, pathogen-induced mortality of individual plants has an effect on the population dynamics of the host, not simply by altering the number of seedlings through partial sterilisation of infectious individuals, but also by reducing the numbers of infected first year plant individuals which potentially become infectious when there is additional mortality between seasons. Additionally, within season pathogen-induced mortality reduces the number of second year infectious individuals that can infect first year susceptible plants in the same season. Within the modelling framework we consider both constant and variable pathogen-induced mortality.
Section snippets
Field observations
Vegetative and flowering T. pratensis individuals and signs of rust infection were recorded in each of the subplots prior to the annual hay cut at the PGE in late-Jun. from the years 1995–1998 to 2000–2004, using one 10 × 2 m quadrat placed in the centre of each of the 97 subplots from the 24 plots. Each of the plots varied in size and had a range of fertiliser and pH treatments applied (Silvertown et al. 2006). These counts may be an underestimation due to the difficulty of identifying first
Survey data
T. pratensis was found in all but one of the plots and 15 subplots between 1995 and 2008, across a range of fertiliser and pH treatments. The survey data demonstrate the outbreak nature of infected second year plants (Fig 2A) in plots under differing treatment types and of healthy second year plants across all plots in the PGE (Fig 2B), in both cases followed by a low but stable population size. Although no data were collected in 1999, 2000 and 2005 these data were interpolated using polynomial
Interpretation and discussion
The aim of this paper was to examine the hypothesis that T. pratensis, described as an outbreak species in the Park Grass Experiment, is regulated by the autoecious, demicyclic rust pathogen P. hysterium (Dodd et al. 1995). This was done by reference to field observations for T. pratensis and P. hysterium, developing a generic epidemiological framework appropriate for the life history characteristics associated with this host and pathogen system, and obtaining parameter values from small scale
Acknowledgements
The work presented here was funded by the BBSRC (BBSSA200412470). Particular thanks to Rothamsted Research for providing access to the Park Grass Experiment and its continued upkeep. Further thanks to two anonymous reviewers.
References (51)
Some discrete-time SI, SIR, and SIS epidemic models
Mathematical Biosciences
(1994)An epidemiological framework for disease management
Advances in Botanical Research
(2002)- et al.
Discrete-time models for gene frequencies and population densities in plant pathosystems
Nonlinear Analysis
(2001) - et al.
The supression of reproduction of Tragopogon pratensis by the rust fungus Puccinia hysterium
Fungal Ecology
(2010) Dynamics of plant–pathogen interactions in natural plant communities
Fungal pathogens and the structure of plant populations and communities
- et al.
Disease spread and population dynamics of anther-smut infection of Silene alba caused by the fungus Ustilago violacea
Journal of Ecology
(1988) - et al.
Population biology of infectious-diseases: 1
Nature
(1979) Plant venereal diseases: insights from a messy metaphor
New Phytologist
(2005)- et al.
Spatial and temporal heterogeneity explain disease dynamics in a spatially explicit network model
The American Naturalist
(2008)
The effect of fungal pathogens on plant communities
Pattern and patchiness in plant–pathogen interactions – causes and consequences
Annual Review of Ecology and Systematics
The current and future dynamics of disease in plant communities
Annual Review of Phytopathology
Determinants of species richness in the Park Grass experiment
American Naturalist
Does transmission of the rust pathogen, Puccinia punctiformis, require stem mining vectors?
Biocontrol Science and Technology
On Competition
Plant disease in natural populations of wild barley (Hordeum spontaneum)
Ecology of Infectious Diseases in Natural Populations
Community stability – a 60-year record of trends and outbreaks in the occurrence of species in the Park Grass Experiment
Journal of Ecology
An epidemiologic-study if Puccinia punctiformis (str) Rohl as a steppingstone to the biological control of Cirsium arvense (L) Scop
New Phytologist
Epidemiology and Plant Ecology: principles and applications
Wintering of the biotrophic fungus Puccinia lagenophorae within the annual plant Senecio vulgaris: implications for biological weed control
Plant Pathology
Spread of organisms: can travelling and dispersive waves be distinguished?
Basic and Applied Ecology
Evolutionary ecology of plant diseases in natural ecosystems
Annual Review of Phytopathology
Cited by (8)
The Unique Contribution of Rothamsted to Ecological Research at Large Temporal Scales
2016, Advances in Ecological ResearchCitation Excerpt :It is likely that similar processes are operating on Park Grass to maintain long-term community stability, for example, the functionally similar grasses Arrhenatherum elatius and Alopecurus pratensis tend to be found on the same plots but one tends to dominate in any given year (unpublished data). Finally, the potential importance of interactions of pathogens in maintaining taxonomic and genetic diversity of plant communities should not be forgotten and has been the subject of a number of recent papers using Park Grass as a model system (Jeger et al., 2014; Salama et al., 2010, 2012). As well as being valuable for studying the stability of ecosystems over time, the Rothamsted experiments are also an important resource for studying the resilience of ecosystems in response to a disturbance and have been used to investigate resilience of soil function.
The impact of climate change on disease in wild plant populations and communities
2022, Plant PathologyA long-term study of burning effects on a plant pathogen in tallgrass prairie
2017, Plant PathologyEcology of fungal plant pathogens
2017, The Fungal KingdomEcology of fungal plant pathogens
2016, Microbiology Spectrum