The present analysis has been performed only with three genes and, consequently, no general conclusions about the evolution of U12-type introns can be extracted. However, several interesting facts of this history have been found.
Conservation of introns has only been found in vertebrates, although U12-type introns have been reported also in insects and plants before [3]. Some introns are conserved in less species than others, for instance: intron 13-14 of ERCC5 has only been found in three species, and intron 7-8 of KIFAP3 has no orthologous introns in the species analysed. The alignment found in insects for the second U12-type intron of KIFAP3 without a homologous intron suggests a possible loss of the intron in the insects lineage or an insertion of it in vertebrates.
The homologous introns usually share the terminal nucleotides and the sequences of the donor, acceptor and branch site are also well conserved, with the more similar ones in the closest species: M. musculus and R. norvegicus, which diverged forty milion years ago [5] and F. rubripes and T. nigroviridis. The different conservation of nucleotides within a site depends on its importance for the recognition by the spliceosome complex. Looking at the sequences, what it can be observed how the less important nucleotides vary along evolution, so they are more different between more divergent organisms.
In our analysis it was found that the distance between the branch point and the acceptor site is not very conserved (results), although it is between 10 and 20 nucleotides. In one case it is 21, which is not a very big difference, but indicates that when more U12-type introns are analysed, maybe the range of this distance will be enlarged. In addition, this distance does not show any tendencies to be bigger or smaller in any specific taxa, suggesting that, as it is within the intron, an insertion or deletion of a small number of nucleotides between the branch point and the acceptor site is accepted, as long as it does maintain the distance in the appropiate range. The 'A' in the branch point is important for the assembly of the spliceosome complex; three cases where found without an 'A' at that position, but in all of them, the nucleotide before it is an 'A', so it could be used for that assembly.
The donor site of the intron found in T. nigroviridis homologous to the first U12 intron of NHE6 was not found by geneid, although it is very similar to the one found in F. rubripes. The former has a 'T' in position +6, where the nucleotide usually is a 'C'; in the data set from Levine and Durbin [4] only six cases like this are found. Because of the version of geneid used is based on the information from that set it does not find the donor site mentioned, whose finding suggest that as more U12 introns are identified, more 'T's will be found at this position. This fact, along with the differences found in some branch sites, can led us to think that maybe the signal sites in U12 introns are not so constrained as it is considered and that searching sites according to the ones found before maybe does not allow us to find all the U12 introns that exists. The combination of these searches with phylogenetic comparisons as the one performed in the present analysis can help improving the search of new U12 introns.
One intron that must be considered in the present discussion is the intron homologous to the second one of KIFAP3 found in chicken. As it has been explained in the results the donor site could be one nucleotide moved. Then to mantain the frame after the intron, deletion of the 'GAG' codon is necessary, which is not a very rarely phenomena. The problem is that a stop codon exists in the exonic sequence upstream of the intron. It could led us to think that the gene is a pseudogene in chicken, but it has a very good branch site (TCCTTAAC) and the score of the acceptor site is also very good (intron 16-17 KIFAP3 results), which is strange if the sequence is not under selection. It is also curious that the part of the alignment between the stop codon and the donor site is not very good, although the rest is. The gene is a novel ensembl prediction, and the possible intron found is within the last intron according to the prediction, but maybe the annotation is not exactly correct. It would be interesting to further analyse this possible intron, which maybe can tell us some things about the history of U12 introns.
With regard to the first intron of KIFAP3, where no homologous introns were found, the alignments of the region upstream of the intron with the rat, mouse and xenopus sequence are good, indicating a possible intron loss or the insertion of the U12-type intron in the human lineage. For the X. tropicalis, a donor site (u12-type) is found in that place, so maybe it maintains an intron in there, but with a different sequence downstream, suggesting a possible insertion or deletion. Actually, the alignments provided by ensembl when orthologous genes are suggested show differences at this region (data not shown). The frog has a region that do not have the genes in the other especies; the three mammals show differences between them at that place. So, according to our results, this one is not a very conserved U12 intron.
The analysis of some paralogs, although not very exhaustive, showed that U12-type introns are conserved in the same gene family in some cases, as reported before [3], indicating that they can be mantained after a gene duplication.
To sum up, it was observed that U12-type introns are conserved in different eukaryotic taxa as well as in some gene families, although the degree of conservation is not the same for different introns. No subtype switching within U12 introns or conversion from U12 to U2 introns has been found in the present study.
Finally, as many questions about U12-type introns remain without an answer, further study will be needed. A more complete phylogenetic analysis could provide insights into the origin of both spliceosomes and their evolution, and it would help to understand the splicing process and the components that take part in it.