Artificial selection of microbial ecosystems for 3-chloroaniline biodegradation.

We present a method for selecting entire microbial ecosystems for bioremediation and other practical purposes. A population of ecosystems is established in the laboratory, each ecosystem is measured for a desired property (in our case, degradation of the environmental pollutant 3-chloroaniline), and the best ecosystems are used as 'parents' to inoculate a new generation of 'offspring' ecosystems. Over many generations of variation and selection, the ecosystems become increasingly well adapted to produce the desired property. The procedure is similar to standard artificial selection experiments except that whole ecosystems, rather than single individuals, are the units of selection. The procedure can also be understood in terms of complex system theory as a way of searching a vast combinatorial space (many thousands of microbial species and many thousands of genes within species) for combinations that are especially good at producing the desired property. Ecosystem-level selection can be performed without any specific knowledge of the species that comprise the ecosystems and can select ensembles of species that would be difficult to discover with more reductionistic methods. Once a 'designer ecosystem' has been created by ecosystem-level selection, reductionistic methods can be used to identify the component species and to discover how they interact to produce the desired effect.