The science of molecular evolution has paved the way for understanding evolutionary processes in which genetic redundancy (the presence of two or more genes capable of serving the same functional role in an organism) is a major source of genetic novelty and robustness. In fact, a recent analysis of 106 bacterial genomes revealed that a significant number of genetic redundancies have persisted in individual genomes , and systematic gene deletion experiments have demonstrated that approximately 300 out of 4000 genes are indispensable for two bacterial species, Escherichia coli  and Bacillus subtilis , suggesting the presence of considerable redundancy in the bacterial genome. However, these findings also raise the question of how genetic redundancy is maintained within a genome, because functionally redundant genes are likely to be eliminated by selective pressure as shown in a large-scale analysis of protein-protein interactions in Saccharomyces cerevisiae, in which the redundant interactions due to duplicated genes generally are not persisted long after gene duplication . In order to explain the process, a theoretical model has been developed to provide insight into the retention of redundant genes, which is hypothesized to depend on their degree of functional redundancy , there have been no empirical studies of gene evolution to support this theoretical model directly. The study described here aimed to substantiate the model of evolution of genetic redundancy on the basis of the analysis of a ribonuclease family that has contributed to our understanding of some aspects of molecular evolution, such as adaptive evolution [6,7], positive Darwinian selection , and the origin of retroviruses with long terminal repeats (LTR) .
Ribonuclease H (RNase H; EC 126.96.36.199), one member of the ribonuclease family, is an enzyme that specifically degrades the RNA moiety of RNA-DNA hybrids . Because various studies have revealed the presence of RNase H in eukaryotes, prokaryotes, and retroviruses, this compound is considered to be one of the most widely conserved enzymes . Although the physiological functions of RNase H are not fully understood, this enzyme is thought to play several roles in DNA replication [12-14], DNA repair [15,16], and RNA transcription [17,18]. In terms of its medical importance, RNase H activity in retroviruses (including HIV-1) is necessary for their replication, and the enzyme has thus been regarded as one of the drug targets for AIDS chemotherapy . This enzyme is also suggested to be related to the antiviral immune response in humans, because mutations of the RNase H-encoding gene have been found in individuals affected by a human neurological disease, Aicardi-Goutières syndrome . Therefore, the accumulation of experimental data on the secondary structures and enzymatic features of RNase H from many studies of its biological significance has provided us with an opportunity to use this knowledge in the field of molecular evolution.
Unlike retroviruses, which possess a single RNase H gene, most prokaryotic and eukaryotic genomes contain multiple RNase H genes. According to the nomenclature for these enzymes, prokaryotic RNase H is generally classified into three groups: RNase HI, HII, and HIII. Eukaryotic RNase H is divided into RNase H1 and H2 . Phylogenetic analyses using RNase H sequences have proposed the following classification: Type 1 (prokaryotic RNase HI, eukaryotic RNase H1, and viral RNase H) and Type 2 (prokaryotic RNase HII and HIII, and eukaryotic RNase H2), and it is important to note that no prokaryotic species with a combination of RNase HI and HIII genes has yet been identified . Additionally, in contrast to eukaryotes, which tend to contain both RNase H1 and H2 genes, the combination of RNase H genes in prokaryotes varies among species, and the overall nature of this variation is poorly understood. Therefore, further study is required to clarify the presence or absence of RNase H genes in these species.
We conducted a comparative analysis of the complete genomes of 353 prokaryotes (326 bacteria and 27 archaea) and examined the combination of RNase H genes and the potential evolutionary processes that could explain the effects of functional redundancy on gene evolution, as described by a theoretical model of genetic redundancy . Our findings suggest that the RNase HI and HIII genes have evolved in a mutually exclusive manner owing to their functional similarities. This molecular evolution of RNase H genes is the first actual example of how the degree of functional redundancy has implications for changes in gene constitution during the course of evolution.