Molecular Evolution and Diversification of Proteins Involved in miRNA Maturation Pathway

Abstract

Small RNAs (smRNA, 19-25 nucleotides long), which are transcribed by RNA polymerase II, regulate the expression of genes involved in a multitude of processes in eukaryotes. miRNA biogenesis and the proteins involved in the biogenesis pathway differ across plant and animal lineages. The major proteins constituting the biogenesis pathway, namely, the Dicers (DCL/DCR) and Argonautes (AGOs), have been extensively studied. However, the accessory proteins (DAWDLE (DDL), SERRATE (SE), and TOUGH (TGH)) of the pathway that differs across the two lineages remain largely uncharacterized. We present the first detailed report on the molecular evolution and divergence of these proteins across eukaryotes. Although DDL is present in eukaryotes and prokaryotes, SE and TGH appear to be specific to eukaryotes. The addition/deletion of specific domains and/or domain-specific sequence divergence in the three proteins points to the observed functional divergence of these proteins across the two lineages, which correlates with the differences in miRNA length across the two lineages. Our data enhance the current understanding of the structure-function relationship of these proteins and reveals previous unexplored crucial residues in the three proteins that can be used as a basis for further functional characterization. The data presented here on the number of miRNAs in crown eukaryotic lineages are consistent with the notion of the expansion of the number of miRNA-coding genes in animal and plant lineages correlating with organismal complexity. Whether this difference in functionally correlates with the diversification (or presence/absence) of the three proteins studied here or the miRNA signaling in the plant and animal lineages is unclear. Based on our results of the three proteins studied here and previously available data concerning the evolution of miRNA genes in the plant and animal lineages, we believe that miRNAs probably evolved once in the ancestor to crown eukaryotes and have diversified independently in the eukaryotes.

Keywords: Dawdle (DDL); Tough (TGH), Serrate (SE/ARS2), Argonaute (AGO), Dicer-Like (DCR/DCL), evolution; phylogeny; small RNA (smRNAs).

Figure 1

Distribution of key proteins (Dicer/Dicer-like (DCR/DCL)) and Argonaute (AGO) involved in the miRNA pathway in animal and plant lineages and the corresponding scheme of miRNA biogenesis in these lineages. (A) Tree of life representing the distribution of major proteins of the miRNA pathway (Dicer/Dicer-like (DCR/DCL)) and Argonaute (AGO): The presence and absence of DCL and AGO in various lineages is shown (adapted from work in [8,9,10,11]). AGO and DCLs appear in most eukaryotic lineages. The pattern of evolution, however, is not known for the accessory proteins (DDL, SE, and TGH) involved in the smRNA machinery. (B) Schematic of the miRNA pathway in animals and plants: The proteins involved in various steps of the pathway and the cellular compartments corresponding to miRNA processing in both lineages are shown. As is evident from this figure, miRNA biogenesis differs in two major aspects between the two lineages, one of them being the methylation of pri-miRNAs and the other being the cytoplasmic processing of mature miRNA, both of which are specific to the plant lineage. In animals, methylation is absent except for in piwi-interacting RNAs (piRNAs) and the maturation of miRNAs occurs in the cytoplasm itself.

Figure 2

Domain architecture of DDL (A), SE (B), and TGH (C) proteins and their distribution across the tree of life (D). The distribution of domains for all three proteins across various lineages is shown. The index for the domain shapes is shown at the bottom of the figure. The interdomain region is shown in brown thick lines. The domains are not to scale and roughly represent the multiple sequence alignment. (D) A representative tree of life showing the lineages with the indicated presence and absence of proteins and the corresponding domain distribution.

Figure 3

Reconciled maximum likelihood (ML) and Bayesian phylogenetic trees of DDL (A), SE (B), and TGH (C). The well-supported clades (PP > 90 and bootstrap values >900) are shown by thick black lines in the tree. The lineage-specific clusters for each protein are marked in different colors, and the corresponding domain architecture for the three proteins in each lineage is shown as alphabetical codes. The index for the domain is shown in the lower left corner of the figure. IQ-Tree and Mr.Bayes (CIPRES cluster) were used to run the ML (1000 bootstrap replicates) and Bayesian trees, respectively. The fungi cluster containing the DUF1604 and the SWAP domain of TGH orthologs is comprised of Ascomycota lineage fungi alone. The other lineages of fungi including the Basidiomycota, Zygomycota, and Chytridomycota contain TGH orthologs with a distinct domain architecture, as shown in Figure 3C.

Figure 3

Reconciled maximum likelihood (ML) and Bayesian phylogenetic trees of DDL (A), SE (B), and TGH (C). The well-supported clades (PP > 90 and bootstrap values >900) are shown by thick black lines in the tree. The lineage-specific clusters for each protein are marked in different colors, and the corresponding domain architecture for the three proteins in each lineage is shown as alphabetical codes. The index for the domain is shown in the lower left corner of the figure. IQ-Tree and Mr.Bayes (CIPRES cluster) were used to run the ML (1000 bootstrap replicates) and Bayesian trees, respectively. The fungi cluster containing the DUF1604 and the SWAP domain of TGH orthologs is comprised of Ascomycota lineage fungi alone. The other lineages of fungi including the Basidiomycota, Zygomycota, and Chytridomycota contain TGH orthologs with a distinct domain architecture, as shown in Figure 3C.

Figure 4

Functional divergence test for DDL (A) and SE (B) protein sequences using DIVERGE 2.0. Amino acids substitution rates among different classes (type II divergence) are shown in red (in stick form) in DDL (A), these residues were identified in the phosphopeptide binding region (V21 (PP = 4.269792), P31V (PP = 4.269792), and K95G (4.269792)) (details in Table 1) and (B) ARS2 domain overlap of SE in humans, and A. thaliana (B) shows the sequence divergence-driven structural differences within the DUF3546 domain in the C-terminal region from Asn263 to Leu268 (Human). This region may contribute to the divergent functional roles of SE in the plant and animal lineage. The PDB IDs of the structures and the colors used for the analysis are mentioned in the figures. In panel B, the structural homology between the structures of the two SE orthologs is depicted by RMSD values. FATCAT server was used for structural superposition. All figures were prepared using PyMol.

Figure 5

Proposed evolutionary history of miRNA processing factors (A) and size distribution of the mature miRNAs in different clades (B). (A) The distribution and divergence of DDL, SE, and TGH orthologs are shown across the tree of life. The presence and absence of the three proteins and their domain organization in each lineage are shown. The index at the bottom of the figure provides details on the symbols used for marking different domains and neofunctionalization of the three proteins. The index for the symbols used is shown in the lower left corner of panel A. (B) The miRNA-length (X-axis) and the miRNA-frequency (Y-axis) in various lineages are color-coded. The index for the colors used for various lineages is shown on the right side of the graph. The lineages chordates and non-chordates (including the protostomes, echinoderms, and hemichordates) are the subgroups of Metazoa of unikonts, whereas the lineages eudicots, monocots, and chlorophytes constitute the Viridiplantae lineage of bikonts.