Figure 2. Most ATG genes duplicated as a result of the whole-genome duplication events near the ― 271 ―Within eukaryotes, duplication of the ATG genes is thought to be a major force driving the evolution of the autophagy pathway. One duplication hotspot was in vertebrates 2). However, the exact duplication timing, post-duplication evolutionary divergence, and how it relates to the functional differences between paralogs, have not been investigated in detail (Figure 1B).Experimental MethodsHomology search and phylogenetic analysis were used to identify the ancestral set of ATG genes at the root of Chordata, as well as the most likely timing of duplication and loss events. Using the pre-duplication species as proxies for the ancestral state, post-duplication divergence in the sequence and gene expression levels were compared between paralogs. The evolutionary fate of each paralog pair (i.e. how the ancestral function is carried out by the two paralogs after duplication) was classified by comparing the expression patterns of individual paralogs or the total expression of both paralogs to the pre-duplication level.Results and DiscussionThe common ancestor of Chordata only had one copy of ULK-1, ATG9, BECN, ATG2, WIPI, GABARAP, LC3, ATG4-1, ATG4-2, and ATG16, and most of them were duplicated in Gnathostomata (jawed vertebrates). GABARAP and BECN, however, were duplicated at later times (at the base of bony vertebrates and eutherians, respectively) (Figure 2). We also identified later, more lineage-specific duplication and loss events.root of vertebrates (1R and 2R), but GABARAP and BECN1 duplicated later.Having determined the most likely duplication timing of the ATG genes in vertebrates, we next compared the post-duplication evolutionary divergence between paralogs. The BECN, WIPI (WIPI1 and WIPI2), and ATG16 pairs display strong asymmetry at both the sequence and expression levels, where the less ancestor-like copy sequence-wise also has higher tissue specificity and lower gene expression level, the LC3 (LC3A and LC3B) pair shows no asymmetry, and the other pairs have partial asymmetry. Assuming that a similar expression pattern implies conserved function, we further categorized the duplicated pairs into different function-related evolutionary fate categories (i.e. how the ancestral function is carried out by the two paralogs after duplication; Figure 3). Consistent with the results above, within the pairs with strong asymmetry between paralogs, one paralog likely preserves the ancestral function, allowing
元のページ ../index.html#273