2.3.1.1 The poplar rust case
A clue came when two poplar (Populus ) leaf rusts, caused by Melampsora medusae Thum. and M. larici-populina , reached Australia and thence New Zealand in 1973 (Van Krayenoord et al. 1974; Wilkinson and Spiers 1976; Spiers 1998). These fungal pathogens, between them, severely affected most of the planted poplar clones in both countries. While M. medusae can only affect poplar, M. larici-populina has a potentially complex life cycle, which in its most complex form is illustrated in Figure 1. While the poplar hosts are in leaf, both these rusts spread by urediniospores (asexual spore stage) , this spread being polycyclic with repeated cycles of infection within the season. At the start of the autumn, thick-walled, dikaryotic teliospores are produced by M. larici-populina, which overwinter on fallen poplar leaves and begin the sexual stage of the life cycle. Teliospores germinate in spring, producing basidiospores that infect the alternate, conifer host, generally larch (Larix sp.). Infection of poplar from mid-spring begins from aeciospores, produced on the conifer. This completes the sexual stage of the life cycle, starting afresh the polycyclic seasonal build-up of infection. Where the conifer host is not present, however, some overwintering of the uredinial stage, either on persisting poplar leaves in mild climates or possibly as mycelia within dormant leaf buds, can start the build-up (Walker et al. 1974; Wilkinson and Spiers 1976; Barrès et al. 2012; Albornoz et al. 2018). In a co-evolved host/pathogen relationship involving such leaf rusts, there should be a comparatively slow build-up of infection during the growing season. However, overwintering cycles of urediniospore production would accelerate the new seasonal build-up. In New Zealand, a semi-evergreen mutant clone (cv. Sempervirens) of the Lombardy poplar (Populus nigra L. cv. Italica) had been widely planted, being often favoured for orchard shelterbelts. This had originated in Chile as a somatic mutation. Its widespread presence, with the semi-evergreen habit and the inherent susceptibility to M. larici-populina (but not to M. medusae ), exacerbated the initial rust epidemic. Where winter conditions are suitably mild to allow retention of leaves, the all-year presence of infected leaves assured a continuing supply of urediniospores, favouring a far quicker build-up of infection in the new season than would be expected from deciduous material. Aided by this quick build-up, M. larici-populina spread rapidly and widely, defoliating poplars susceptible to both rusts, whileM. medusae unable to build up inoculum in the winter was outcompeted and remained restricted in distribution (Wilkinson and Spiers 1976). The semi-evergreen clone was very severely affected, and was soon almost eradicated, which presumably contributed to the rusts becoming less serious a few years after arriving. Other contributing factors would have included felling of trees of other severely affected cultivars, and a spontaneous proliferation of hyperparasitic and otherwise antagonistic microorganisms in phyllosphere communities on the leaf surfaces (Heather and Chandrashekar 1982).
The semi-evergreen cultivar shared the inherent susceptibility to infection with the original deciduous cultivar. That susceptibility, however, has exposed a potential danger of an evergreen habit in reducing protection from natural delays in the seasonal build-up of infection. Such a reduction can be expected to increase the level of resistance required for field fitness. That would increase selection pressures imposed by rusts, which operate alongside all the other selection pressures. Since increased selection pressure on any trait tends to reduce the possible selection intensity for other traits, that means a potential for ’selection overload’. Moreover, the poplar rust pathosystems are very complex (e.g. Barrès et al. 2012; Persoons et al. 2014; Albornoz et al. 2018), with a potential to be very dynamic and readily destabilised. Table 1 summarises the expected impacts of rust on the poplar host, in scenarios varying according to: severity of winter, presence of absence of alternate (conifer) hosts, and decidousness or evergreenness. Figure 2 outlines expected functional interplays, under a mild climate, between evergreen/semi-evergreen habit, growth potential, infection level and host performance,
Co-adaptation between rusts and poplars has tended to arise in climates that have severe or moderately severe winters. Thus, the mild climates of New Zealand and Chile are not the norm for the natural pathosystems. Yet these novel conditions outside climatic norms that provide an opportunity to examine the action of pathogens as evolutionary drivers influencing the persistence of deciduousness. Testing our hypothesis, and whether an evergreen poplar could ever be ecologically fit, however, faces major practical obstacles. Ideally, a population of evergreen genotypes, rather than just one, should be used for comparison with deciduous genotypes. However, there is no easy pathway to obtaining such a population. Even using the existing semi-evergreen clone would be suspect, because it might not match young seedlings, from which native poplars grow, for susceptibility to the pathogen within the growing season. Indeed, any matching of comparison material for inherent susceptibility would be challenging. Field testing would be wanted, where natural pathogens would be present. The impact of the pathogen(s) would need to be evaluated, most likely over several years as most impacts from rusts occur after repeated seasons of infection. Evaluating impact is likely to be complicated by the fact that presence of a susceptible evergreen host will likely influence disease development and impact on deciduous genotypes. Pathogen dispersal between stands could similarly complicate the picture. To manage these challenges, chemical control (or some effective but targeted biocontrol) of the pathogen(s) would need to be used to exclude rust impacts from a subset of trees. Support for our hypothesis would be obtained from the persistence of susceptible evergreen genotypes, and the relative success compared to susceptible deciduous genotypes, where the pathogen impacts are excluded from both, and the decline and eventual extinction of evergreen genotypes where pathogen impacts are not prevented.