Patterns of Conservation and Diversification in the Fungal Polarization Network

Abstract

The combined actions of proteins in networks underlie all fundamental cellular functions. Deeper insights into the dynamics of network composition across species and their functional consequences are crucial to fully understand protein network evolution. Large-scale comparative studies with high phylogenetic resolution are now feasible through the recent rise in available genomic data sets of both model and nonmodel species. Here, we focus on the polarity network, which is universally essential for cell proliferation and studied in great detail in the model organism, Saccharomyces cerevisiae. We examine 42 proteins, directly related to cell polarization, across 298 fungal strains/species to determine the composition of the network and patterns of conservation and diversification. We observe strong protein conservation for a group of 23 core proteins: >95% of all examined strains/species possess at least 14 of these core proteins, albeit in varying compositions, and non of the individual core proteins is 100% conserved. We find high levels of variation in prevalence and sequence identity in the remaining 19 proteins, resulting in distinct lineage-specific compositions of the network in the majority of strains/species. We show that the observed diversification in network composition correlates with lineage, lifestyle, and genetic distance. Yeast, filamentous and basal unicellular fungi, form distinctive groups based on these analyses, with substantial differences to their polarization network. Our study shows that the fungal polarization network is highly dynamic, even between closely related species, and that functional conservation appears to be achieved by varying the specific components of the fungal polarization repertoire.

Fig. 1.

—The central part of budding yeast’s polarization protein network. (A) Polarity establishment and subsequent budding takes place at one location of the budding yeast cell membrane (cartoon). Insert: schematic overview of the 35 proteins selected from budding yeast and their functional groupings based on SGD (http://www.yeastgenome.org/) (Drees et al. 2001; Chang and Peter 2003; Madhani 2007; Martin and Arkowitz 2014). Cdc42 cycles between an active membrane-bound state (GTP) and an inactive cytosolic state (GDP). Depicted are the Cdc42 regulators and effectors, the Bem1/Cdc42 protein complex, and several downstream steps (color coded). Nrp1 has a presumed function in polarity establishment (see Laan et al. 2015), Ubi4 has a described epistatic interaction with Cdc42 (BioGRID; thebiogrid.org). (B) Matrix of the genetic (in red) and physical (in black) interactions between the 35 selected polarization proteins based on SGD protein data. Proteins are color coded with the functional groupings from the (A) panel. Protein abundance following Kulak et al. (2014) is displayed in the most right panel. Note that for Gic1 and Ubi4 no expression data were available.

Fig. 2.

—Correlation between genome quality and number of retrieved proteins. The top panel (Count) shows the distribution of strains/species for the number of retrieved proteins. The top center panel shows a statistically significant positive correlation between the N50 of scaffolds of the genome and the number of retrieved proteins. The bottom center plot shows a statistically significant negative correlation between L50 (scaffolds) and the number of proteins. The bottom panel shows statistically significant negative correlation between the number of scaffolds in the genome and the number of retrieved proteins.

Fig. 3.

—Phylogenetic relationships between the 298 fungal strains/species and the protein matrix for the 42 selected polarization proteins. The phylogeny is based on 242 protein sequences (613783 aa) and the approximately maximum likelihood method and the JTT model of amino acid evolution. Support values are almost exclusively >0.9, except when shown on the tree (11 instances). The tree includes the phyla: Microsporidia (in beige), Blastocladiomycota, and Chytridiomycota (in orange), Neocallimastigomycota (in yellow), Zoopagomycota (in dark green), Mucoromycota (in light green), Basidiomycota (in blue), and the Ascomycota (in purple). Subphyla are shades of the same phylum color.

Fig. 3.

—Phylogenetic relationships between the 298 fungal strains/species and the protein matrix for the 42 selected polarization proteins. The phylogeny is based on 242 protein sequences (613783 aa) and the approximately maximum likelihood method and the JTT model of amino acid evolution. Support values are almost exclusively >0.9, except when shown on the tree (11 instances). The tree includes the phyla: Microsporidia (in beige), Blastocladiomycota, and Chytridiomycota (in orange), Neocallimastigomycota (in yellow), Zoopagomycota (in dark green), Mucoromycota (in light green), Basidiomycota (in blue), and the Ascomycota (in purple). Subphyla are shades of the same phylum color.

Fig. 4.

—Protein matrix of the 43 species with highest genome quality. The matrix displays the similarity scores of the iterative ggMatch approach for species of the Mucoromycota (in green), Basidiomycota (in blue), and the Ascomycota (in purple). Proteins are ordered and color coded following figures 1A and 3. Essential proteins (in Saccharomyces cerevisiae) and short single domain proteins are labeled with green and yellow bullets at the bottom of the matrix. The life styles of the species are depicted in the far right column.

Fig. 5.

—Polarization proteins prevalence. Prevalence of the 42 polarization proteins for all examined fungal species (black circles), the non-Dikarya species (excluding the Microsporidia; yellow circles), the Basidiomycota species (blue circles), and the Ascomycota species (purple circles). Proteins are ordered based on their overall prevalence in all examined strains/species. The 70% criterion is marked by a horizontal red dotted line. Shading in the bottom part reflects grouping of proteins with < 20% prevalence in the all Fungi group (light gray; left), proteins with prevalence 20% < 70% (gray; center), proteins with >70% prevalence in all examined groups (i.e., core proteins; dark gray; right). Difference in prevalence between the Ascomycota and Basidiomycota is presented in the top panel (pink diamonds).

Fig. 6.

—The number of core proteins and strains/species. Depicted is the percentage of strains/species and the number of core proteins. Dotted horizontal lines represent the 85%, 90%, 95%, and 100% of strains/species levels.

Fig. 7.

—Comparison between budding yeast’s core and noncore proteins. (A) Significant difference in the observed prevalence of the core and noncore proteins (P value < 0.0001). (B) Number of genetic interactions (in red) and physical interactions (in black) of the 35 examined polarization proteins. No difference was observed between the core and noncore proteins in the number of genetic or physical interactions. (C) Significant difference in protein abundance between the two groups. Core proteins have higher protein abundance (P value = 0.005). Note that data for Gic1 (noncore) and Ubi4 (core) were unavailable. Core proteins are depicted as circles, while noncore proteins are depicted as squares. Black lines depict medians.

Fig. 8.

—Multiple factor analysis and correlations. (A) Multiple factor analysis of the number of polarization proteins, lineage, lifestyle, genomic quality, and genetic distances. The 298 strains/species are plotted and color-coded according to their phylogenetic lineage as in figure 3. Dimension 1 explains 17.02% of the observed variation and the following four factors constitute to its construction (in order of importance): lineage, lifestyle, genetic distance, number of observed proteins. Dimension 2 explains 10.94% of the variation in the data and is based on the variables lineage, lifestyle, and genetic distance. Main areas occupied by specific lineages are labeled accordingly for clarity. A clear distinction can be made between yeast-like fungi (left top corner), filamentous fungi (lower part), and unicellular nonyeast like fungi (right). (B) Cartoon depicting the topology of the major clades. The length of branches do not represent observed branch lengths. See figure 3 for full phylogeny. (C) The distribution of lifestyles (in percentages) for the twelve different phylogenetic lineages. The number of strains/species per lineages is given. Lifestyles are color-coded as in legend at the bottom left of the figure. The 298 strains/species are classified as unicellular, yeast, filamentous, dimorphic (either yeast/filamentous or yeast/pseudohyphal), and trimorphic following figure 3. (D) Pie plot depicting the percentage of variation explained by the three main dimensions. The two dimensions account for 28% of the observed variation, leaving 72% undefined. (E) The number of observed proteins in twelve different phylogenetic lineages. Groups are color-coded per lineage as in figure 3. Medians are given as black lines. (F) The number observed proteins in the different lifestyles. Black lines represent medians. (G) The number of observed proteins plotted versus the genetic distance (in respect to Saccharomyces cerevisiae, Schizosaccharomyces pombe, Ustilago maydis, Sporisorium reilianum). Strains/species are color-coded according to their lifestyle morphology.