Luque et al., (2014) In vivo evolution of metabolic pathways by homeologous recombination in mitotic cells.
So far, engineering of metabolic pathways has relied on in vitro techniques that can lead to the production of better enzymes, for example with more substrate specificity or better kinetic properties.
Thanks to directed protein evolution, it is possible to generate libraries of protein variants and then they will have to be screened for their functionality and desired features. An interesting approach is also a coupling of mutation and gene shuffling cycles but it can be applied to DNA sequences of maximum 5 kb, therefore it is not useful for in vitro shuffling of pathways, that are usually composed of gene clusters that overcome >10 kb.
Therefore, the research group thought about an in vivo approach in order to shuffle longer DNA sequences, such as a cluster of genes necessary for flavonoid production in yeast. It was already shown that in MMR deficient cells, homologous recombination between non-identical (homeologous) DNA sequences was possible, due to de fact that the cell is no longer able to reject DNA heteroduplexes. This property of the cell enhances the recombination frequencies between diverged homeologous sequences.
What the study wants to demonstrate is that it is possible to generate functional variable mosaic libraries of genes directly in vivo, integrate them correctly into the genome and to avoid therefore the problem of too long sequences.
What it should be highlighted of this research study is that they were able to generate in vivo, therefore under physiological conditions, different intragenic mosaic patterns of the flavonoid pathway genes. It is very interesting how the strains they sequenced did not have any indels or mutations in the chimeric genes and that the patterns were found only once.
I think one of the limitations of this approach is that it is possible to use this technique only for heterologous genes, at least if the aim is to integrate and engineer a whole cluster of genes. Engineering an endogenous pathway is not a feasible method because this way the recombination events could occur at the level of the endogenous genes and not among the homeologous of the cluster. Maybe, with further modifications this approach could find an application also for endogenous pathway engineering.
Considering the whole research study, I think this is a very good paper, especially for researchers involved in metabolic engineering studies.