Multiple infections, relatedness and virulence in the anther-smut fungus castrating Saponaria plants.

Multiple infections (co-occurrence of multiple pathogen genotypes within an individual host) can have important impacts on diseases. Relatedness among pathogens can affect the likelihood of multiple infections and their consequences through kin selection. Previous studies on the castrating anther-smut fungus Microbotryum lychnidis-dioicae have shown that multiple infections occur in its host plant Silene latifolia. Relatedness was high among fungal genotypes within plants, which could result from competitive exclusion between unrelated fungal genotypes, from population structure or from interactions between plant and fungal genotypes for infection ability. Here, we aimed at disentangling these hypotheses using M. saponariae and its host Saponaria officinalis, both experimentally tractable for these questions. By analysing populations using microsatellite markers, we also found frequent occurrence of multiple infections and high relatedness among strains within host plants. Infections resulting from experimental inoculations in the greenhouse also revealed high relatedness among strains co-infecting host plants, even in clonally replicated plant genotypes, indicating that high relatedness within plants did not result merely from plant x fungus interactions or population structure. Furthermore, hyphal growth in vitro was affected by the presence of a competitor growing nearby and by its genetic similarity, although this latter effect was strain-dependent. Altogether, our results support the hypothesis that relatedness-dependent competitive exclusion occurs in Microbotryum fungi within plants. These microorganisms can thus respond to competitors and to their level of relatedness.

Composition of epiphytic bacterial communities differs on petals and leaves.

The epiphytic bacterial communities colonising roots and leaves have been described for many plant species. In contrast, microbiologists have rarely considered flowers of naturally growing plants. We identified bacteria isolated from the surface of petals and leaves of two plant species, Saponaria officinalis (Caryophyllaceae) and Lotus corniculatus (Fabaceae). The bacterial diversity was much lower on petals than on leaves of the same plants. Moreover, the bacterial communities differed strongly in composition: while Pseudomonadaceae and Microbacteriaceae were the most abundant families on leaves, Enterobacteriaceae dominated the floral communities. We hypothesise that antibacterial floral volatiles trigger the low diversity on petals, which is supported by agar diffusion assays using substances emitted by flowers and leaves of S. officinalis. These results suggest that bacteria should be included in the interpretation of floral traits, and possible effects of bacteria on pollination are proposed and discussed.