Evolution of SET-domain protein families in the unicellular and multicellular Ascomycota fungi

Abstract

Background: The evolution of multicellularity is accompanied by the occurrence of differentiated tissues, of organismal developmental programs, and of mechanisms keeping the balance between proliferation and differentiation. Initially, the SET-domain proteins were associated exclusively with regulation of developmental genes in metazoa. However, finding of SET-domain genes in the unicellular yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe suggested that SET-domain proteins regulate a much broader variety of biological programs. Intuitively, it is expected that the numbers, types, and biochemical specificity of SET-domain proteins of multicellular versus unicellular forms would reflect the differences in their biology. However, comparisons across the unicellular and multicellular domains of life are complicated by the lack of knowledge of the ancestral SET-domain genes. Even within the crown group, different biological systems might use the epigenetic 'code' differently, adapting it to organism-specific needs. Simplifying the model, we undertook a systematic phylogenetic analysis of one monophyletic fungal group (Ascomycetes) containing unicellular yeasts, Saccharomycotina (hemiascomycetes), and a filamentous fungal group, Pezizomycotina (euascomycetes).

Results: Systematic analysis of the SET-domain genes across an entire eukaryotic phylum has outlined clear distinctions in the SET-domain gene collections in the unicellular and in the multicellular (filamentous) relatives; diversification of SET-domain gene families has increased further with the expansion and elaboration of multicellularity in animal and plant systems. We found several ascomycota-specific SET-domain gene groups; each was unique to either Saccharomycotina or Pezizomycotina fungi. Our analysis revealed that the numbers and types of SET-domain genes in the Saccharomycotina did not reflect the habitats, pathogenicity, mechanisms of sexuality, or the ability to undergo morphogenic transformations. However, novel genes have appeared for functions associated with the transition to multicellularity. Descendents of most of the SET-domain gene families found in the filamentous fungi could be traced in the genomes of extant animals and plants, albeit as more complex structural forms.

Conclusion: SET-domain genes found in the filamentous species but absent from the unicellular sister group reflect two alternative evolutionary events: deletion from the yeast genomes or appearance of novel structures in filamentous fungal groups. There were no Ascomycota-specific SET-domain gene families (i.e., absent from animal and plant genomes); however, plants and animals share SET-domain gene subfamilies that do not exist in the fungi. Phylogenetic and gene-structure analyses defined several animal and plant SET-domain genes as sister groups while those of fungal origin were basal to them. Plants and animals also share SET-domain subfamilies that do not exist in fungi.

Figure 1

Maximum likelihood phylogeny of 113 representative SET-domain sequences. Bootstrap values for the major SET-domain families that are higher than 60% by either of the maximum likelihood (ML) or the maximum parsimony (MP) methods are shown at the node (the two % values are ML/MP). Internal branches supporting the major SET-domain families with higher than 80% ML bootstrap values are also indicated by thick lines. Within the major SET-domain protein groups, bootstrap values by the ML analysis greater than 60%, 70%, 80%, and 90% are indicated by stars (*), filled circles (●), filled squares (■), and filled triangles (▲), respectively. SET-domain protein subgroups discussed in the main text are indicated by numbers. Representative domain names are shown for the major SET-domain protein families (see Additional file 3 for the domain names). Species abbreviations are given in Additional file 1. For this ML phylogeny, the gamma shape parameter and the proportion of invariant sites were estimated to be 0.779 and 0, respectively.

Figure 2

Distribution of SET-domain genes in the three kingdoms. Saccharomycotina and Pezizomycotina fungi are shown as "Sac" and "Pez", respectively, among the fungal kingdom. Species or group-specific duplication events of SET-domain genes are illustrated by dashed arrows. SET5/6 related filamentous fungi are shown as "SET-fil". Arrows marked with * indicate that such relationship is not significant and inconclusive. SET-MYND and SUV3-9 are found in S. pombe as well as Pezizomycotina fungi, but missing from Saccharomycotina fungi. This is indicated by the arrows marked with "(-Sac)".

Figure 3

Distribution of SET-domain families in fourteen genomes. Shaded in yellow are genes found in all tested species; peach-colored genes were found only in the Saccharomycotina. Genes found in tested genomes except in the Saccharomycotina are shaded in turquoise, while those found only in multicellular species are shown in pink. Genes specific for the filamentous fungi are shown in grey. A related gene found in Arabidopsis is shaded in grey as well. Metazoa-specific genes are shaded in Bordeaux red. Footnotes: ^a genome size; ^b approximate numbers of predicted open reading frames; ^c the Su(var)3-9 gene in S. pombe is known as the Clr4; ^d one of the two copies is known as the DIM-5 gene in N. crassa and the second shows a weak similarity to G9a; ^e filamentous fungal genes belonging to the large SET5/6 family; ^f "unknown" genes have no significant support to cluster with any SET-domain family identified in this study; ^g numbers in parentheses indicate numbers of total SET-domain genes found in these animal/plant genomes [20,21].

Figure 4

Domain architecture of the SET1/TRX family (a: the SET1 subfamily, b: the TRX subfamily). The divergent RNA recognition motifs (RRM) found in Arabidopsis and yeast (not identified in domain databases) are indicated by the pale color in the figure. Domains are not drawn to scale. For the TRX subfamily, representatives of two major animal subgroups differing by the positioning of the FYRN-FYRC (DAST) domains are shown. Note the presence of conserved domains between the animal and the plant representatives. For more structures see Additional file 3.

Figure 5

Domain architecture of the SET2/ASH1 family (a: the SET2 subfamilies, b: the ASH-1 subfamily). Domains are not drawn to scale. For more structures see Additional file 3.

Figure 6

Maximum likelihood phylogeny of the SET3/4 families among ten Saccharomycotina species. Bootstrap values greater than 60% are shown. The genomic position is indicated with accession numbers for the sequences from unannotated genomes. The gamma-shape parameter and the proportion of invariant sites were estimated to be 0.879 and 0, respectively.

Figure 7

Domain architecture of the SUV39/G9a family (a: the SUV39 subfamily, b: the G9a subfamily). Domains are not drawn to scale. For more structures see Additional file 3.