Convergence in ascospore discharge mechanism among pyrenomycete fungi based on 18S ribosomal RNA gene sequence

Fungi of the class Pyrenomycetes (Ascomycotina) form a morphological series ranging from those that shoot ascospores (sexual spores) forcibly from the ascus (spore sac) to fungi that ooze ascospores or have no obvious mechanism for ascospore release. Did forcible ascospore discharge evolve within these pyrenomycetes, or has it been lost in the group? We determined the sequences of the 18S ribosomal RNA gene from three fungi and used these, along with six sequences from our previous work and three sequences from GenBank, to infer the phylogeny of 12 ascomycetes with various ascospore discharge mechanisms. The 1720 base pairs of sequence data per fungus yielded 361 variable sites, 198 phylogenetically informative sites, and a single most parsimonious tree requiring 562 nucleotide changes. The tree shows that the capacity to shoot ascospores into the air has been lost or, less probably, gained repeatedly and independently. Species lacking forcible ascospore discharge are intercalated among three lineages of species with forcible discharge. In this tree, seven of the nine internal branches appeared in 95% or more of 500 bootstrap replicates. A tree uniting the fungi with forcible ascospore discharge into a monophyletic group required 45 additional steps and fit significantly less well with the data than the most parsimonious tree, based on a maximum likelihood test. Two of the fungi whose sequence we determined, Pseudallescheria boydii and Sporothrix schenckii, are not closely related to one another, even though both are human pathogens and both are from pyrenomycete lineages lacking forcible ascospore discharge. Using the well-resolved, most parsimonious tree, we inferred base substitution patterns in the 18S rRNA. The transition-to-transversion ratio was 1.9. Of all 12 possible substitutions, 29% were from U to C. At sites corresponding to yeast stem positions, A to G transitions were frequent, perhaps compensating for some of the U to C changes, and maintaining secondary structure base pairing (A to G:U to C = 3:4). In loop or bulge positions without secondary structure base pairing, U to C transitions were still frequent, but A to G transitions were rare (A to G:U to C = 1:5).