|
内容記述 |
Living organisms have various characteristics that define lineages. The change in regulatory elements is thought to play a major role in development of these lineage specific characteristics. From the inception of molecular evolutionary studies, noncoding regions were suspected to be involved in gene regulation. Recent studies of genome comparisons among diverged species revealed that there are many highly conserved noncoding sequences (HCNSs) in vertebrates, and many of them actually contain regulatory elements. Based on the observations, one of the candidates for regulatory elements which contributed to the lineage specific evolution is the HCNSs conserved only in one lineage because these lineage specific HCNSs may have gained new functions during the evolution of the lineage. However, unlike the HCNSs conserved in the large lineage such as vertebrates, HCNSs conserved only in a small lineage comprised of closely related species such as primates and rodents have not been well studied. That is, identification of lineage specific HCNSs provides a new insight for the evolution of the corresponding lineage of organisms.<br/> I first analyzed human-marmoset and mouse-rat pairwise noncoding alignments, and determined to use the 100bp window which was a minimum length to detect conservations in the closely related species. The threshold for the conserved sequences in both human-marmoset and mouse-rat pairs was ≥98% identity. As the first filtering for identification of lineage specific HCNSs, I extracted conserved sequences with the thresholds as primate and rodent specific HCNS candidates from the human-marmoset and mouse-rat pairwise alignments. Using the extracted primate and rodent specific HCNS candidates as queries, I performed MegaBLAST search against 9 vertebrate genomes, and removed all HCNSs that were also conserved in non-primate or non-rodent vertebrate genomes. A total of 34,313 and 32,092 primate and rodent specific HCNSs were extracted. I further filtered these HCNSs by examining their lengths because longer conserved sequences were considered to be more strongly constrained. After these filtering processes, I finally obtained 223 primate- and 592 rodent-specific HCNSs.<br/> The SNP densities in primate and rodent specific HCNSs were significantly lower than those of genome averages. I therefore analyzed the derived allele frequency (DAF) within the primate specific HCNSs to measure the relative level of purifying selection acting on HCNSs. The level of DAF ≤0.1 within the HCNSs shows purifying selection signals in all human populations (Yoruba, Han Chinese + Japanese, and American of European Ancestor). This does not support the idea that the primate specific HCNSs are mutational cold spots and it is plausible that lineage specific HCNSs are under selective constraint. This suggests that lineage specific HCNSs tend to be under purifying selection, implying that primate and rodent specific HCNSs harbor important functions.<br/>  I also examined whether there is any differences in the distributions of lineage specific HCNSs and ultraconserved elements (UCEs) because the UCE is an extreme example of highly conserved vertebrate HCNSs. The distributions of primate and rodent specific HCNSs and vertebrate HCNSs were completely different in the genomes, suggesting that these lineage specific HCNSs and vertebrate HCNSs are independently evolved sets.<br/> To investigate the biological impact on the lineage specific HCNS on the evolution, I next examined the function of lineage specific HCNS-flanking genes (LHF genes). The statistically overrepresented functions of primate and rodent LHF genes were “anatomical development” and “transcriptional regulation”, which was consistent with the characteristics of known vertebrate HCNSs. Notably, the synonymous (dS) substitution of primate and rodent LHF genes were significantly smaller than those of genome wide genes, as well as the non-synonymous (dN) and dN/dS ratio. I also found that UCE-flanking genes showed significantly smaller dS values than those of genome wide genes. This indicates that there are stronger constraints on the LHF genes and UCE-flanking genes at nucleotide level compared to genes that are not associated with HCNSs. Indeed, orthologs of primate/rodent LHF genes in rodents/primates, the majority of which have no HCNSs, showed the same level of dS values with genome wide genes. This strongly suggests that there is a correlation between HCNSs and low dS genes. Given that the functions of LHF gene are important in development, the strong constraint on LHF genes at nucleotide level may be a result of tight regulation of the gene expression. For instance, many regulatory proteins bind to the LHF genes to regulate the gene expression by interacting with HCNSs.<br/> Interestingly, even though primate and rodent LHF genes showed similar functions to UCE-flanking genes, the majority of both LHF genes were different from the UCE-flanking genes. This suggests that independent sets of genes may have contributed to develop lineage specific characteristics. Conversely, the number of LHF genes which were shared by UCE-flanking genes was small but significantly larger than expected, and many of them were involved in nervous system development as transcriptional regulators. This suggests that certain groups of genes recruited new HCNSs in addition to old HCNSs which are conserved among vertebrates. <br/> Based on the results in this study, I propose a possibility that the lineage specific evolution occurred through the creation of new lineage specific HCNSs near two categories of genes. The first category is lineage specific sets of LHF genes. The creation of lineage specific HCNSs expands the set of LHF genes which are involved in development, but different from that of ancestral (vertebrate) HCNSs. The second category is particular groups of ancestral HCNS-flanking genes. One of the major gene groups are involved in nervous system development. The results in this study provide new insights into the lineage specific evolution through interactions between HCNSs and their LHF genes.<br/> |
要旨・審査要旨 (298.6 kB)
