Date Published: May 11, 2004
Publisher: Public Library of Science
Author(s): Philipp Khaitovich, Gunter Weiss, Michael Lachmann, Ines Hellmann, Wolfgang Enard, Bjoern Muetzel, Ute Wirkner, Wilhelm Ansorge, Svante Pääbo
Abstract: Microarray technologies allow the identification of large numbers of expression differences within and between species. Although environmental and physiological stimuli are clearly responsible for changes in the expression levels of many genes, it is not known whether the majority of changes of gene expression fixed during evolution between species and between various tissues within a species are caused by Darwinian selection or by stochastic processes. We find the following: (1) expression differences between species accumulate approximately linearly with time; (2) gene expression variation among individuals within a species correlates positively with expression divergence between species; (3) rates of expression divergence between species do not differ significantly between intact genes and expressed pseudogenes; (4) expression differences between brain regions within a species have accumulated approximately linearly with time since these regions emerged during evolution. These results suggest that the majority of expression differences observed between species are selectively neutral or nearly neutral and likely to be of little or no functional significance. Therefore, the identification of gene expression differences between species fixed by selection should be based on null hypotheses assuming functional neutrality. Furthermore, it may be possible to apply a molecular clock based on expression differences to infer the evolutionary history of tissues.
Partial Text: Advances in microarray technology have made the systematic study of expression levels of thousands of transcripts possible. This has been heralded as a major step forward in understanding the function of genomes, since transcript expression levels are expected to correlate with biological functions. Although this is clearly the case for many genes that change their expression in response to environmental stimuli (e.g., Spellman et al. 1998; Hughes et al. 2000; Miki et al. 2001), it is not known whether evolutionary changes in gene expression are determined primarily by Darwinian selection or by stochastic processes. Indeed, the extent to which natural selection has shaped the properties of organisms has been hotly debated ever since Charles Darwin proposed that organisms are adapted to their environment as a result of natural selection. At the molecular level, the view that most changes are due to Darwinian selection was challenged by Kimura’s neutral theory of molecular evolution (Kimura 1983). This theory states that the vast majority of differences seen in nucleotide and amino acid sequences within and between species have no or only minor selective effects. Consequently, their occurrence within a species and the fixation of differences between species are primarily the result of stochastic processes. Thus, it is believed today that the evolution of the overwhelming majority of synonymous nucleotide changes within protein-coding exons, as well as changes in noncoding parts of genomes, are determined by mutational processes and random genetic drift (Li 1997). In fact, even at the level of morphology, it has been argued that many features are not adaptive, but instead result from physical constraints or historical accidents (Gould and Lewontin 1979). However, since selection acts at the level of the phenotype while variation is generated at the level of the genotype, the proportion of changes caused by selection can be expected to be largest at the phenotypic level and smallest at the DNA sequence level. As a corollary, we may expect the proportion of selected changes to gradually decrease at the proteome and the transcriptome levels, since these are located progressively further from the phenotype. Consequently, a large proportion of transcriptome changes might be explained by historical accidents rather than by selective events.
We show that a neutral model of evolution can predict the main features of transcriptome evolution in the brains of primates and mice. A neutral model is also in agreement with published observations in Drosophila (Rifkin et al. 2003) and fish (Oleksiak et al. 2002). Although selective scenarios that explain some or even most of these observations can be found, the combined evidence presented leads us to conclude that a neutral model is the most adequate null model for transcriptome evolution. This suggests that the majority of gene expression differences within and between species are not functional adaptations, but selectively neutral or nearly neutral. The main challenge now is to develop a mathematical model of transcriptome evolution that allows quantitative predictions of transcriptome changes. Such a model, combined with experimental data estimating the normal variation of gene expression within a species and the relative contributions of genetic and environmental factors to this variation, should allow adaptive gene expression changes to be identified. Further work is also needed to reveal whether proteome evolution is also dominated by changes that are largely selectively neutral.