Divergent Evolution of Eukaryotic CC- and A-Adding Enzymes

Abstract

Synthesis of the CCA end of essential tRNAs is performed either by CCA-adding enzymes or as a collaboration between enzymes restricted to CC- and A-incorporation. While the occurrence of such tRNA nucleotidyltransferases with partial activities seemed to be restricted to Bacteria, the first example of such split CCA-adding activities was reported in Schizosaccharomyces pombe. Here, we demonstrate that the choanoflagellate Salpingoeca rosetta also carries CC- and A-adding enzymes. However, these enzymes have distinct evolutionary origins. Furthermore, the restricted activity of the eukaryotic CC-adding enzymes has evolved in a different way compared to their bacterial counterparts. Yet, the molecular basis is very similar, as highly conserved positions within a catalytically important flexible loop region are missing in the CC-adding enzymes. For both the CC-adding enzymes from S. rosetta as well as S. pombe, introduction of the loop elements from closely related enzymes with full activity was able to restore CCA-addition, corroborating the significance of this loop in the evolution of bacterial as well as eukaryotic tRNA nucleotidyltransferases. Our data demonstrate that partial CC- and A-adding activities in Bacteria and Eukaryotes are based on the same mechanistic principles but, surprisingly, originate from different evolutionary events.

Keywords: Salpingoeca rosetta; Schizosaccharomyces pombe; enzyme evolution; tRNA nucleotidyltransferase.

Figure 1

Phylogenetic distribution of eukaryotic species bearing genomes with more than one tRNA nucleotidyltransferase gene. (A) Schematic representation of class II tRNA nucleotidyltransferase organization. Motifs A to E of the catalytic core are indicated in green and blue. Additional motifs involved in A-addition are the basic/acidic motif (B/A; indicated as R (arginine) and D (aspartate), other basic/acidic combinations are also found) and the flexible loop, both located between motifs A and B. (B) Phylogenetic tree summarizing the data from 35 eukaryotic genomes analyzed in Betat et al. [8]. The tree topology represents all major eukaryotic clades. The total number of genomes analyzed per clade is given in brackets. Species names are given for genomes containing more than one candidate tRNA nucleotidyltransferase sequence. (C,D) Phylogenetic distribution of tRNA nucleotidyltransferases in Choanoflagellata and Taphrinomycotina, respectively. Boxes next to the species name indicate the type of candidate tRNA nucleotidyltransferase genes (green—alphaproteobacterial a-type; blue—ancestral eukaryotic e-type). Closed and open boxes indicate whether candidates are derived from genome or transcriptome data. The trees to the left of the species names are for illustrative purposes as the species phylogeny is not well established in some parts. Multifurcations, question marks, and dotted branches indicate uncertainty about the phylogenetic position.

Figure 2

Phylogenetic network of a- and e-type tRNA nucleotidyltransferase sequences from Choanoflagellata. On the right of the main split, a-type sequences of CCA-adding enzymes from Metazoa cluster with a-type sequences from Choanoflagellata (CC-adding enzyme in S. rosetta). On the left of the main split, e-type enzyme sequences from Fungi and Plants cluster with candidate e-type enzyme sequences of Choanoflagellata (A-adding enzyme in S. rosetta). An asterisk (*) marks sequences with experimentally verified function. Frames indicate the enzymes of S. rosetta that were analyzed in this study. Bootstrap values are indicated as percentage values, n.a. represents no shared split. Sequence names consist of the taxonomic group prefix (eu—Eukaryota, Unikonta; euh—Eukaryota, Holozoa; eb—Eukaryota, Bikonta; euf—Eukaryota, Fungi) as well as the first three letters of the genus and species name (see Supplementary Data).

Figure 3

Phylogenetic network of tRNA nucleotidyltransferase sequences from Taphrinomycotina in a greater context. The split in the center separates e-type (left) and a-type (right) sequences. The left side includes all sequences from genome-sequenced Taphrinomycotina. Schizosaccharomyces species are represented by both confirmed and putative CC-adding and A-adding enzyme sequences. Each of the corresponding subtrees of these Schizosaccharomyces enzymes resembles the species phylogeny in topology and branch length, indicating that both enzymes evolved at the same rate. These data exclude a horizontal gene transfer event and indicate subfunctionalization shortly after gene duplication. The cluster on the right side includes CCA-, CC-, and A-adding enzyme sequences of bacterial origin (including metazoan a-type sequences of alphaproteobacterial origin). An asterisk (*) marks sequences that have been experimentally verified for their function. The frames indicate the investigated S. pombe enzymes. Bootstrap values are given as percentage values, n.a. indicates no shared split. Sequence names consist of the taxonomic group prefix (eu—Eukaryota, Unikonta; euh—Eukaryota, Holozoa; eb—Eukaryota, Bikonta; euf—Eukaryota, Fungi) as well as the first three letters of the genus and species name (see Supplementary Data).

Figure 4

The S. rosetta enzymes are tRNA nucleotidyltransferases with complementing partial activities. (A) Alignment of S. rosetta a-type CC- and e-type A-adding enzymes with sequences of experimentally verified CCA-adding enzymes of C. elegans (a-type; accession number (AC): Q93795; [43]), H. sapiens (a-type; AC: Q96Q11; [44,45]), C. glabrata (e-type; AC: XP_449283.1; [46]) and A. thaliana (e-type; AC: Q94K06; [47]). The alignment is restricted to the N-terminal catalytic core elements. Highly conserved positions are indicated in blue. The alignment shows the high similarity of a-type CCA- and CC-adding enzymes as well as e-type CCA- and A-adding enzymes. The S. rosetta CC-adding enzyme shows strong deviations from the conserved flexible loop sequence of related a-type CCA-adding enzymes. (B) Radiolabeled tRNA^Phe (top) or tRNA^Phe-CC (bottom) was incubated with either S. rosetta a-type enzyme (left) or e-type enzyme (right). Nucleotide addition on radiolabeled tRNA^Phe (top) or tRNA^Phe-CC (bottom) identifies S. rosetta a-type and e-type enzymes as bona fide CC-and A-adding enzymes, respectively. In the presence of CTP only, the A-adding enzyme catalyzes a slight misincorporation of an additional C residue, a side reaction also observed for other tRNA nucleotidyltransferases [41,42]. (C) The flexible loop of S. rosetta CC-adding enzyme exhibits strong deviations from the consensus sequence “DGRxAxV” found in eukaryotic a-type CCA-adding enzymes. In the chimeric enzyme SHS, the corresponding region of H. sapiens CCA-adding enzyme (H) was transplanted into the CC-adding enzyme of S. rosetta (S). (D) Chimera SHS incorporates three nucleotides into a tRNA without CCA end and the terminal A residue into a tRNA carrying two C residues (tRNA-CC). M, mock incubation of the tRNA in the absence of enzymes; C, control tRNA^Phe with CCA end as size marker.

Figure 5

S. pombe performs CCA-addition with collaborating CC-and A-adding enzymes. (A) Alignment of the catalytic core of both S. pombe e-type enzymes with sequences of tested e-type CCA-adding enzymes of A. thaliana (accession number (AC): Q94K06; [47]), L. albus (AC: AAB03077.1; [48]), C. glabrata (AC: XP_449283.1; [46]), and S. cerevisiae (AC: P21269; [49]). Highly conserved residues are indicated in blue. Similar to the S. rosetta counterpart, the S. pombe CC-adding enzyme shows strong deviations from the conserved flexible loop sequence responsible for its functional restriction. (B) Both enzymes were incubated with labeled substrates tRNA^Phe (top) or tRNA^Phe-CC (bottom). The e-type enzyme 1 (left) exclusively adds two C residues on tRNA^Phe but does not accept tRNA^Phe-CC for A-addition and therefore represents a CC-adding enzyme. The e-type enzyme 2, on the other hand, accepts only tRNA^Phe-CC as a substrate and adds just a terminal A residue and is therefore identified as an A-adding enzyme. Similar to the S. rosetta enzyme (Figure 4B), this enzyme misincorporates an additional C residue if CTP only is offered. Taken together, these enzymes represent typical CC-and A-adding enzymes. (C) In the CC-adding enzyme, the flexible loop is lacking the conserved sequence “Yxxx(x)SRxP” found in eukaryotic e-type CCA-adding enzymes. For construction of an enzyme chimera carrying the flexible loop and the B/A motif of a closely related fungal CCA-adding enzyme, the corresponding sequence of T. flavorubra (T) was inserted into the CC-adding enzyme of S. pombe (S), generating the chimera STS (S, S. pombe CC-adding enzyme N-terminus; T, flexible loop of T. flavorubra CCA-adding enzyme; S, S. pombe CC-adding enzyme C-terminus). (D) On tRNA^Phe lacking the 3’-CCA terminus, chimera STS incorporates three nucleotides. While the addition of the terminal residue is not very efficient, one has to consider that the inserted regions in a chimeric protein are not perfectly adjusted to the context of the host protein. Hence, wild type-like activity cannot be expected. On tRNA^Phe-CC, the chimera efficiently adds a terminal A residue, clearly demonstrating that A-addition is restored to a considerable extent. M, mock incubation of the tRNA in the absence of enzymes; C, control tRNA^Phe with CCA end as size marker.