Speciation in progress? A phylogeographic study among populations of Hemitrichia serpula (Myxomycetes)

Abstract

Myxomycetes (plasmodial slime molds, Amoebozoa) are often perceived as widely distributed, confounding to the "everything is everywhere" hypothesis. To test if gene flow within these spore-dispersed protists is restricted by geographical barriers, we chose the widespread but morphologically unmistakable species Hemitrichia serpula for a phylogeographic study. Partial sequences from nuclear ribosomal RNA genes (SSU) revealed 40 ribotypes among 135 specimens, belonging to three major clades. Each clade is dominated by specimens from a certain region and by one of two morphological varieties which can be differentiated by SEM micrographs. Partial sequences of the protein elongation factor 1 alpha (EF1A) showed each clade to possess a unique combination of SSU and EF1A genotypes. This pattern is best explained assuming the existence of several putative biospecies dominating in a particular geographical region. However, occasional mismatches between molecular data and morphological characters, but as well heterogeneous SSU and heterozygous EF1A sequences, point to ongoing speciation. Environmental niche models suggest that the putative biospecies are rather restricted by geographical barriers than by macroecological conditions. Like other protists, myxomycetes seem to follow the moderate endemicity hypothesis and are in active speciation, which is most likely shaped by limited gene flow and reproductive isolation.

Fig 1

Morphology of Hemitrichia serpula: a. mature fruiting body, b-c. capillitium of var. serpula (specimen M754) as seen in (b) light microscope and (c) scanning electron microscope (SEM), d. spore morphology of var. serpula, e-f. SEM micrograph of spores for (e) var. serpula (LE297865) and (f) var. parviverrucospora (sc28101), g. SEM of the internal linings of the peridium, var. parviverrucospora, h. close-up, showing the internal warts between the reticulations of var. parviverrucospora (sc28065).

Fig 2. Rooted consensus tree based on the 50% majority rule of Bayesian interference for partial SSU sequences for 40 ribotypes from 135 specimens of Hemitrichia serpula.

Shown are Bayesian posterior probabilities >0.70 and support values >50 for a corresponding tree calculated with RAxML (all branches indicated by a dot). Colored squares indicate the origin of the specimens. The ribotype number and the morphology of the respective specimens are indicated by smooth (var. serpula) and spiny circles (var. parviverrucospora). Question marks indicate the two specimens with undetermined spore morphology. Scale bars represent evolutionary distance as changes per site.

Fig 3. Phylogenetic analysis showing a mirrored image comparing tree topologies for partial SSU and EF1A sequences for 30 specimens.

Bayesian posterior probabilities >0.70 and RAxML support values >50 are indicated (nodes with dots). Scale bars represent evolutionary distance as changes per site. Ribotype (r) and EF1A genotype numbers (e) are shown for both markers. Dotted lines connect sequences from the same specimen, with symbols for var. serpula (smooth circles) and var. parviverrucospora (spiny circles) in the middle.

Fig 4. Statistical parsimony ribotype network representing genealogical relationships among 40 ribotypes estimated by TCS superimposed on a Bayesian interference tree.

Grey triangles are sized relative to the number of specimens per network. Line segments represent mutational steps between alleles. Circles are scaled in proportion to the number of sequences represented by each ribotype. Small circles between ribotypes indicate hypothetical transitional ribotypes. Colors designate the origin of the specimen. Morphotypes displayed by specimens showing the respective ribotype are indicated by smooth (var. serpula) or broken lines (var. parviverrucospora).

Fig 5. Hypothesized event-based (Vic = vicariance; Dis = dispersal) ancestral area reconstruction of H. serpula ribotypes as inferred by S-DIVA analysis of RASP.

Pie charts at the nodes give relative frequencies of the ancestral-area reconstruction. Grey triangles indicate ribotypes represented by multiple specimens.

Fig 6. Probability-based environmental niche models for the three major clades in a ribotype phylogeny of Hemitrichia serpula calculated with the MaxEnt algorithm.

Circles are located over areas where specimens were collected; their size is scaled according to the number of specimens collected in an area. Black filling of the pie diagrams indicates the proportion of specimens belonging to the respective clade. The underlying heat map shows the likelihood of occurrence for the respective clade.