Origins of New Genes and Pseudogenes

Gene duplication was the first mechanism of gene generation to be suggested (Ohno, 1970), and this process does indeed appear to be the most common way of creating new genes. Duplications are typically classified according to the size of the portion of the genome that is duplicated; thus, a duplication may be described as involving an entire genome, large segments of a genome, individual genes, individual exons, or even specific parts of exons (Betrán & Long, 2002). The mechanisms that generate duplicate genes are diverse, and more details about these mechanisms are continually being discovered. These mechanisms include whole-genome duplications originating through nondisjunction, tandem duplications originating through unequal crossover, retropositions originating through the retrotranscription of an RNA intermediate, transpositions involving transposable elements (Jiang et al., 2004; Morgante et al., 2005), and duplications occurring after rearrangements and subsequent repair of staggered breaks (Ranz et al., 2007). Such duplications involve not only protein-coding genes, but also noncoding RNA genes. For example, a novel class of retroduplicates includes snoRNAs, which are a class of RNA genes that are involved in ribosomal RNA processing (Weber, 2006).

Much of the current excitement about gene duplication stems from the fact that with the number of sequenced genomes now available, researchers have more accurate estimates of how often genes duplicate, and these rates are extremely high. For example, more than 100 genes duplicate in the human genome per 1 million years (Hahn et al., 2007a). This means that the percentage of the genome affected by gene number differences (estimated to be 6%) contributes more to the differences between humans and chimpanzees than do single nucleotide differences between orthologous sequences (estimated to be 1.5% [Demuth et al., 2006]). High rates (17 genes per 1 million years) have also been estimated in flies (Hahn et al., 2007b). Additional excitement comes from the realization that duplications occur so often that individuals of the same species differ greatly in DNA content and gene number (i.e., many duplications are polymorphic and contribute to individual differences [Sebat et al., 2004]). It is estimated that, on average, two humans will differ by approximately 5 megabases of information.

Unexpectedly, several duplication trends have been described in genomes with respect to sex chromosome evolution. Many new male genes originate in species' Y chromosomes. Some of these male genes are organized in families that undergo gene conversion to avoid Y-chromosome degeneration. Male germ-line genes can also duplicate out of the X chromosome through retroposition (Betrán et al., 2002; Emerson et al., 2004; Lahn et al., 2001; Rozen et al., 2003). These findings reveal that genomic location and organization matter for gene origination and function.

Transposable elements (TEs) are so-called "selfish" segments of DNA that encode proteins that allow these segments to copy or move themselves within a genome. There are two types of TEs: DNA transposons and retrotransposons. DNA transposons are able to excise themselves out of the genome and be inserted somewhere else, whereas retrotransposons copy themselves through an RNA intermediate. Similar to viral insertions in the genome, TE insertions cause mutations and contribute to increased genome size, but they usually do not encode cellular proteins.

Interestingly, one way for a genome to acquire new genes is by recruiting transposable element proteins and using them as cellular proteins. Such events are called domestications of TE proteins. A recent review by Fechotte and Pritham (2007) compiled many examples of these events, and it revealed that some of the unexpected functions in which TE domesticated proteins play a role include functioning of the vertebrate immune system and light sensing in plants. Several examples of domestication have also been described in Drosophila (Casola et al., 2007). In one case, codomestication of the two proteins encoded by the PIF/Harbinger transposable element was observed. In this fascinating example, the two genes that the original TE contained were domesticated simultaneously; one of these genes encoded a transposase that binds and cuts DNA, while the other encoded a protein that contains a Myb/SANT domain believed to function in transcription, chromatin remodeling, and protein-protein interactions. More data are needed to reveal whether both of these proteins were domesticated to function in the same biological process.

Scientists use the term "lateral gene transfer" to refer to the case in which a gene does not have a vertical origin (i.e., direct inheritance from parent to offspring) but instead comes from an unrelated genome. It is well known that this sort of transfer occurs between bacteria, and that it also has taken place between the genomes of the cellular organelles (mitochondria and chloroplasts) and the nuclear genomes (Roger, 1999). However, more recent transfer events between organelles and/or endosymbiont bacteria continue to occur (Bergthorsson et al., 2003; Hotopp et al., 2007). For example, large-scale sequencing efforts have revealed that much of the genome of the intracellular endosymbiont Wolbachia pipentis was integrated into Drosophila species (Hotopp et al., 2007). However, the mechanism for these transfers remains largely unknown, and the functional consequences of some of these transfers have yet to be explored.

Existing genes can also fuse (i.e., two or more genes can become part of the same transcript) or undergo fission (i.e., a single transcript can break into two or more separate transcripts), thereby forming new genes. Interestingly, it has been observed that chimeric fusion genes sometimes involve two copies of the same gene (e.g., the alcohol dehydrogenase gene in Drosophila), and when that happens, the resulting genes undergo parallel evolution (Jones & Begun, 2005), in which they shift away from the functions of their parental genes.

Figure 1

Figure Detail

New genes can additionally originate de novo from noncoding regions of DNA. Indeed, several novel genes derived from noncoding DNA have recently been described in Drosophila (Begun et al., 2007; Levine et al., 2006). For these recently originated Drosophila genes with likely protein-coding abilities, there are no homologues in any other species. Note, however, that the de novo genes described in various species thus far include both protein-coding and noncoding genes. These new genes sometimes originate in the X chromosome, and they often have male germ-line functions.

The action of all the mechanisms described in the previous sections leads to exon shuffling (i.e., the observation that many genes share exons) (Gilbert et al., 1997; Li et al., 2001). In addition, analyses of "young" genes (genes that originated only a few million years ago) allow investigators to document all of the events that gave rise to these genes, because time has not eroded the footprints of these events. Using this approach, it has been inferred that combinations of several mechanisms and several events are often responsible for the generation of new genes (Betrán & Long, 2002). Two good examples of this are the jingwei gene (Long & Langley, 1993) and the SETMAR gene (Cordaux et al., 2006; Figure 1).