Host Jumps and Radiation, Not Co-Divergence Drives Diversification of Obligate Pathogens. A Case Study in Downy Mildews and Asteraceae

Abstract

Even though the microevolution of plant hosts and pathogens has been intensely studied, knowledge regarding macro-evolutionary patterns is limited. Having the highest species diversity and host-specificity among Oomycetes, downy mildews are a useful a model for investigating long-term host-pathogen coevolution. We show that phylogenies of Bremia and Asteraceae are significantly congruent. The accepted hypothesis is that pathogens have diverged contemporarily with their hosts. But maximum clade age estimation and sequence divergence comparison reveal that congruence is not due to long-term coevolution but rather due to host-shift driven speciation (pseudo-cospeciation). This pattern results from parasite radiation in related hosts, long after radiation and speciation of the hosts. As large host shifts free pathogens from hosts with effector triggered immunity subsequent radiation and diversification in related hosts with similar innate immunity may follow, resulting in a pattern mimicking true co-divergence, which is probably limited to the terminal nodes in many pathogen groups.

Fig 1. Maximum likelihood tree for Bremia, inferred from the combined sequences of cox2, cox1, the spacer region between cox2 and cox1 genes, 18S, 28S, ITS1, and RxLR11 loci.

ML, MP, ME bootstraps, and MCMC posterior probability more than 70% are shown above or below the branch. Asterisk (*) marks the samples selected for cophylogenetic analyses.

Fig 2. Maximum likelihood tree for Asteraceae, inferred from the combined sequences of matK and ITS loci.

ML, MP, ME bootstraps, and MCMC posterior probability more than 70% are shown above or below the branch. Asterisk (*) marks the samples selected for cophylogenetic analyses.

Fig 3. Tanglegram for the Asteraceae (left) and Bremia (right).

Each tree is a consensus tree of ME, ML, MCMC, MP trees inferred from the combined sequences of matK and ITS genes for Asteraceae and cox2, cox1, the spacer region between cox2 and cox1 genes, 18S, 28S, ITS1, RxLR11 loci for Bremia. Each leaf of Bremia is linked to its host. Bold lines indicate significant links between taxa (ParaFit, P < 0.05), and dashed lines indicate marginally significant links between taxa (ParaFit, P > 0.05). Tanglegram was produced in TreeMap 3b (https://sites.google.com/site/cophylogeny/).

Fig 4. A potentially optimal reconciled tree between Asteraceae and Bremia inferred from CoRe-PA (A) and the simplified tree (B).

One of several best-fit scenarios proposed by the program is shown, which includes 25 codivergence, 26 host-switches, 2 duplication, and 13 sorting events. Green and red lines represent the host and parasite, respectively. Four evolutionary events are denoted: codivergences (black circle), duplication (white circle), sorting events (blue circle), and host-switches (red line with arrowhead). Dashed red lines indicate distant host jumps across different subfamilies of Asteraceae.

Fig 5. Bivariate plot of corresponding sequence divergences of pathogens and hosts at the level of species of Asteraceae.

The inlay illustrates the evolution of Bremia by host jumps as inferred from cophylogenetic analyses, genetic divergence, and maximum clade ages.

Fig 6. Number of divergence events counted in every 3 million years in each molecular clock tree of Asteraceae and Bremia.