Genomduplikation und die Evolution von Genclustern in Teleostiern

The traditional view on the evolution of novel gene functions was divided in two groups, those who believed that genome duplication is more important and those who promoted the primary importance of regulatory evolution. Already before the genomic era, the idea emerged that new gene functions would only evolve after a duplication event, so that one copy continues with the original function while the other one changes. Some researchers claimed that major transitions in evolution have only been possible through an increase in gene number due to gene or genome duplications. The finding of a series of genome duplications that occurred during vertebrate evolution was taken as support for this hypothesis.Regulatory evolution is the change of regulatory elements by mutations followed by a changed expression pattern and possibly a new function. A new research field, evolutionary development (EvoDevo) that joins gene expression with evolutionary aspects was able to bring together gene duplication and regulatory evolution during recent years.Chapter 1 presents a study on the evolution of a biochemical pathway, glycolysis, and the duplication history of the genes involved. We investigated ten enzymes for which we obtained sequences from genomic and EST (expressed sequence tags) databases for different vertebrate species. Even though many of the genes involved follow a pattern of 2R/3R in vertebrates and show subfunctionalization in terms of tissue specialization, a general trend for the retention of paralogs could not be deduced.Among the most cited examples of duplicated genes, as well as of regulatory evolution, are Hox clusters - an ideal study object for comparative genomics. Hox genes establish positional information during embryonic development and in this way determine, which structure develops at which position. Therefore, a very precise regulation of those genes is essential. In Chapter 2, the advantages of the Hox clusters system for studying conserved non-coding sequences (CNS) is discussed, together with a review on recently published papers using a new method to identify CNS.In Chapter 3, another multigene family was studied, the KCNA genes that code for voltage-gated potassium channels (shaker-related). In tetrapods, eight genes were identified and of those, six are organized in two 3-gene-clusters. Teleost fish have four clusters with only one gene lost and an additional 2-3 genes that are located elsewhere in the genome.Apart from the Hox genes already mentioned, there are many other homeobox proteins that are also transcription factors and some of them are also arranged in conserved clusters. The ParaHox cluster consists of three genes (gsh, pdx/xlox, cdx) and this genomic region was also subject to a series of duplications. In contrast to the rather well conserved Hox clusters, the ParaHox clusters are often not recognizable as such anymore and only the neighboring receptor tyrosine kinases (RTKs) provide supporting evidence of large scale duplication events. While tetrapods have one cluster with three genes (and three additional genomic regions with one ParaHox gene each), the situation in fishes is even more complicated. Each of the eight genomic regions that resulted from genome duplications, the so-called paralogons, contain at maximum one ParaHox gene. To the contrary, RTKs exist in multiple copies. In Chapter 4, we compare the ParaHox paralogons of several fish species in terms of gene order and the size the clusters cover within a genome. Additionally, we characterize the C1-paralogon of Astatotilapia burtoni, an East-African cichlid. Some of the RTKs are known to be involved in coloration of fishes and these characteristics make the ParaHox paralogon highly interesting. We could show that duplicated ParaHox genes not only were eliminated from fish genomes that now have the same number of ParaHox genes as tetrapods, but that the gene order was rearranged several times in the case of the D paralogon. The surrounding of the RTKs contains only very few CNS, while we were able to detect conserved areas surrounding the ParaHox genes. In contrast to the Hox genes, the cluster arrangement in this case seems not to be required for correct gene expression.In Chapter 5 we present the genomic sequences of Hox clusters in the haplochromine cichlid Astatotilapia burtoni. One possible source for the amazing variation of cichlids is regulatory change of developmentally important genes. We compared the Hox cluster setup with other teleost fishes and found that relative high percentages (11-38%) of the intergenic regions are made up of CNS.Regulatory evolution offers an easy and fast possibility for genes expressed in a very specific pattern to evolve new functions, while for genes that are expressed more widely, at least in a temporal sense such as potassium channels, duplication and subsequent evolution of protein sequences might provide enough variety for adaptations