The aim of this work was to describe the selenoproteome of the newly sequenced species wolf-eel (Anarrhichtys ocellatus). To do so, we compared the homology between this species genome and zebrafish, our reference species. We carried out a comparative computational analysis with each selenoprotein, machinery elements and homologous proteins with cysteine from zebrafish to wolf-eel.

Here we report a comprehensive analysis of each of these proteins that reveal novel insights into the characterization of Sec in this organism. We also created this webpage in html language to show our results and actualized the wikipedia webpage for the wolf-eel.

We characterized Anarrhichtys ocellatus selenoproteome by searching for all known selenoproteins in Danio rerio. The search was supplemented with the analysis of SECIS elements via SEBLASTIAN, and with the subsequent phylogenetic analysis of proteins belonging to the same family. Overall, the data obtained showed 27 selenoproteins, 6 Cys-homologues and 6 Cys-homolog proteins related to Sec synthesis. It is important to note that one selenoprotein (GPx8) had acquired a Sec residue that did not appear in our reference species. By contrast, 2 selenoproteins were not found in our species, which stands to reason that they were lost in evolution (SELENOI and SELENOM).

One process contributing to the reduction of the selenoproteome is the conversion of Sec to Cys, which is a specific process to selenoproteins and can be achieved by a single point mutation that transforms a Sec UGA to Cys. However, Sec and Cys are not functionally equivalent [5]. In our species, we found 12 Cys-homologues that were the same in zebrafish, but in other proteins we were not able to determine if there was this kind of conversion, because the proteins that did not contain the Sec residue had poor alignments that did not fill up with zebrafish Sec residues.

We observed that some selenoproteins had been lost during evolution in Anarrhichtys ocellatus genome. This is the case of SELENOI and SELENOM. SELENOI is a selenoprotein in zebrafish, but we did not obtain any Sec residue nor SECIS elements in the comparative analysis with our species. These results suggest that the protein may had been lost or that the sequence had been missanotated, which led to a poor prediction of the protein. In the case of SELENOM, we did not obtain any Sec residue either, but Seblastian predicted a SECIS element. On account of that, we believe that Anarrhichtys ocellatus has lost the Sec in SELENOM more recently than SELENOI, and thus, the SECIS element remains in the sequence. Nevertheless, those sequences presented a significant amount of gaps, which indicate low quality annotation and prediction.

Additionally, several selenoproteins genes are found duplicated in Danio rerio, probably due to the whole genome duplication in the early evolutions of fishes. In zebrafish we found duplications in SelO, SelT, SelW, GPx1 and GPx4 [5]. However, in our species the duplicated protein for these families was not observed. The program predicted the same sequence for both proteins that were different in zebrafish, indicating that these families consist in only one protein per family.
However, some duplications present in zebrafish, remain in our species of interest. Some of them are SELENOU1a-SELENOU1b, GPx3a-GPx3b and DIO3a-DIO3b.

Our study suggest that Anarrhichtys ocellatus selenoproteome is quite conserved compared with the one of Danio rerio. We obtained high identity percentages between sequences, suggesting that the number of mutational changes that cause a change in the amino acid sequence had not been highly significant, evidencing high homology between the two selenoproteomes.

We also came out with some limitations. First, the fact that some proteins were missing a large number of amino acids made it difficult to obtain a high quality prediction. It is highly probable that, on the one hand, the annotation of Anarrhichtys ocellatus was not complete and that, on the other hand, it was fractionated in different scaffolds. Another limitation that we found was that Seblastian did not always coincide with our T-coffee alignment, which made us decide weather to consider our results correct or not. Despite all these limitations, we obtained significant results which can be useful for further studies.

For the first time, the selenoproteome of Anarrhichtys ocellatus has been computationally characterized using a comparative analysis with Danio rerio.