No orthologs could be found in our B oleracea
No orthologs could be found in our B. oleracea BAC library for AtCk1/2, which could mean that it is unique to Arabidopsis, or that we failed to detect it in our library. However, the second possibility is unlikely considering that Southern blotting analysis with four restriction enzymes disclosed fragments corresponding to no more than five copies for Ck1 in B. nigra. The AtCk1/2 sequence showed the lowest identity with any of the other sequences, including the one from Oryza, except for the truncated sequence BoCk1. Although phylogenetic analysis of nucleotide sequences (not shown) provided moderate support for a sister relationship between AtCk1/2 and AtCk1/1 plus BoCk1b, we suspect that this may have resulted from long-branch attraction, to which nucleotide sequence data may be more susceptible than amino Phenformin sequences. We hypothesize that AtCk1/2 is a relict sequence perhaps similar to the ancestral Ck1 gene in Brassicaceae. The possible events leading to the origin of different homologs can be inferred from Fig. 4. The Oryza Ck1 sequence has only 11 exons, while all Ck1 homologs from Arabidopsis and Brassica have either 13 or 14. Sequences from additional taxa within Brassicaceae and closely related eudicot families are needed to establish the ancestral number of exons for the family, but our results indicate that the presence of intron 11, and therefore of 14 exons, is the ancestral condition for Brassica and Arabidopsis Ck1 genes. Fig. 4 suggests the existence of an ancestral gene of 14 exons early in the evolution of these taxa, perhaps similar in sequence and structure to AtCk1/2. This gene might have undergone divergence following geographical isolation of founder species of n=4 or 5 chromosomes . Such an event would have resulted in the formation of two similar orthologs, one acting as a precursor for AtCk1/1 and BoCk1b in one branch, and the second one as precursor for the rest of the AtCk1 and BoCk1 genes in a second branch. This gene might have undergone segmental duplication resulting in one copy of the gene with 14 exons (which evolved into AtCk1/4 and BoCk1a) and a second copy of the gene carrying a splice site change resulting in the lack of intron 11. This mutant gene would have given rise to AtCk1 and BoCk1c in Arabidopsis and Brassica, respectively. The presence of segmental duplications in the genome of A. thaliana indicate that it has a polyploid origin , , . It has been proposed that this species might have originated from an ancestral genome of n=4 chromosomes undergoing polyploidization resulting in species of n=8 chromosomes. This assumption is based on the fact that most Arabidopsis species, except for A. thaliana, have a genome of n=8 chromosomes. A. thaliana might have derived its n=5 genomic number by subsequent chromosomal elimination , . Considering this hypothetical scenario, it is then possible to consider that A. thaliana acquired AtCk1/1 and AtCk1/4 from two different ancestral genomes by hybridization followed by amphiploidy. The genome carrying AtCk1/4 would have carried also the 13-exon paralog AtCk1/5. In Brassica species, a similar scenario could be expected, except that two additional copies of BoCk1c by another cycle of duplication-originated paralogs BoCk1d and BoCk1e, which might or might not be on the same chromosome. Our results suggest that of these three, only BoCk1e encodes a complete functional protein and the other two have apparently evolved into pseudogenes. The extensive divergence of AtCk1/2 from the rest of the genes might indicate that it is closer to the ancestral Ck1 gene in the Brassicaceae. The level of divergence of this gene is similar to that observed between Oryza Ck1 and Brassicaceae Ck1 genes, except that BoCk1/2 is structurally more similar (exon number) to the rest of its Brassica counterparts. A. thaliana might have conserved a copy of this relict gene that is now gone in Brassica species. The scenario outlined above provides a summary of our current understanding of the evolution of Ck1 homologs from Arabidopsis and Brassica. An alternative scenario invoking triplication of the Brassica genome from one similar to Arabidopsis cannot be discarded based on the data for a single gene family. However, in such case, one would expect to find 9–10 copies of Ck1 in Brassica. Further, these copies would be expected to cluster in trios. Such clustering was only observed for BoCk1c, BoCk1d, and BoCk1e, but their close affinity suggests that they are the product of recent segmental duplications. It is evident, however, that extensive sequence reshuffling has taken place in the evolution of the Brassicaceae, as can be seen from the 3′ end of these genes containing segments displaying identities to unrelated genes and even to a chloroplast gene. Additionally, our results indicate that chromosome number reduction  is not necessarily concomitant with a significant reduction of gene copy number, as observed for the Ck1 genes. Undoubtedly, the events leading to duplications are complex and thorough understanding of the evolution of these loci requires a full analysis of orthologous and paralogous sequences in order to determine their origins and relationships.