Phylogeny-wide analysis of social amoeba genomes highlights ancient origins for complex intercellular communication

Abstract

Dictyostelium discoideum (DD), an extensively studied model organism for cell and developmental biology, belongs to the most derived group 4 of social amoebas, a clade of altruistic multicellular organisms. To understand genome evolution over long time periods and the genetic basis of social evolution, we sequenced the genomes of Dictyostelium fasciculatum (DF) and Polysphondylium pallidum (PP), which represent the early diverging groups 1 and 2, respectively. In contrast to DD, PP and DF have conventional telomere organization and strongly reduced numbers of transposable elements. The number of protein-coding genes is similar between species, but only half of them comprise an identifiable set of orthologous genes. In general, genes involved in primary metabolism, cytoskeletal functions and signal transduction are conserved, while genes involved in secondary metabolism, export, and signal perception underwent large differential gene family expansions. This most likely signifies involvement of the conserved set in core cell and developmental mechanisms, and of the diverged set in niche- and species-specific adaptations for defense and food, mate, and kin selection. Phylogenetic dating using a concatenated data set and extensive loss of synteny indicate that DF, PP, and DD split from their last common ancestor at least 0.6 billion years ago.

Figure 1.

Phylogenetic position and life cycles of test species. Schematic representation of the SSU rRNA phylogeny of all social amoebas, with solitary amoebas as outgroup. DF and PP reside in groups 1 and 2, respectively. They differ from DD by generally forming small clustered and/or branched fruiting structures from a single aggregate, in which prespore cells differentiate first and the stalk is formed by dedifferentiation of prespore cells at the apex. In contrast, group 4 taxa form robust solitary unbranched fruiting structures and set aside appropiate proportions of prestalk cells after aggregation. They also differentiate into three additional cell types that support the stalk and spore head, and their sorogens can migrate extensively to seek optimal spots for fruiting body construction (Schaap et al. 2006; A Skiba and P Schaap, in prep.).

Figure 2.

Genome structures in social amoebae. Chromosomes are not drawn to scale. Telomere and centromere properties were deduced as described in Supplemental Material. Specific short sequence motifs are indicated as letters, identified common chromosomal regions as labeled areas. The polymorphism in the PP-specific sequence motif is indicated by the polymorphic nucleotides in brackets. Above this motif the number of observed repeat units is indicated.

Figure 3.

Phylogenetic tree of PKS KS domains of social amoebae. The domains were extracted from the PKS proteins using hmmer models of DD-specific domains (Zucko et al. 2007). A ClustalW alignment was fed in the phylogeny software puzzle (Schmidt and von Haeseler 2007) and confirmed by neighbor-joining analysis using PHYLIP (Felsenstein 1989).

Figure 4.

Architectural conservation of adenylate and guanylate cyclases. Proteins were identified by TBLASTN search with DD cyclase sequences and query of SACGB with the IPR001054 A/G_cyclase domain. Protein phylogenies of nucleotidyl cyclases were constructed using Bayesian inference (Ronquist and Huelsenbeck 2003) from alignments of the shared functional domains as described in the legend to Supplemental Figure S5.6. For proteins with two cyclase domains, a tandem alignment of both domains was used, with the single domains of ACB and ACG used twice. The tree is presented unrooted and decorated with the domain architectures of the proteins as analyzed by SMART (Schultz et al. 1998). The posterior probabilities (BIPP) of nodes are indicated by colored circles. Locus and gene ids: (ACB) DD: DDB_G0267376, PP: PPL_10642, DF: DFA_10857; (SGC) DD: DDB_G0276269, PP: PPL_05930, DF: DFA_12195; (GCA) DD: DDB_G0275009, PP: PPL_11010, DF: DFA_10832; (ACA1) DD: DDB_G0281545, PP: PPL_01657, DF: DFA_04345; (ACA2) PP: PPL_12370; (ACA3) PP: PPL_10658; (ACG) DD: DDB_G0274569, PP: PPL_06767, DF: DFA_08118; (Pia) DD:DDB_G0277399, PP: PPL_02311, DF: DFA_00057; (CRAC) DD: DDB_G0285161, PP: PPL_07100; (Rip3) DD: DDB_G0284611, PP: PPL_11127, DF: DFA_06425.

Figure 5.

(A) Ratio (d_S/d_N) of scaled synonymous (d_S) to scaled nonsynynonymous (d_N) base substitutions between species pairs. The summed number of proteins with a given value of d_S/d_N ratios is shown. Only true orthologs with a minimum sequence identity of 40% over at least 50% of the gene length were used for calculation. Protein pairs with a d_S of 0 are omitted. (Blue) DD/DF species pair; (red) DD/PP species pair; (green) DF/PP species pair. (B) Calibrated phylogenetic tree of major eukaryote lineages. A concatenated data set of 33 orthologous genes (Supplemental Table S6.1) was aligned using ClustalW. A neighbor-joining tree was reconstructed with MEGA4 (Pairwise gap deletion, Poisson correction, bootstrap test with 1000 replicates). The scale was calibrated using an estimated divergence time of 750 mya for the human/hydra split. The sequence of early evolutionary events cannot be resolved with this data set. A similar scale was obtained by using the moss/vascular plant split at 550 mya.