Horizontal gene transfer (HGT) is a form of genic inheritance that occurs between individuals in a population or between species rather than from parent to offspring. HGT is most prevalent in bacteria, where it is an important source of novel metabolic pathways, pathogenicity factors, and antibiotic resistance. In spite of the importance of HGT, we understand little about the evolutionary barriers to HGT. This project takes a systematic experimental approach to elucidate factors that select for or against horizontally transferred genes by quantifying intrinsic selection acting on newly transferred genes, by experimentally transferring and expressing several hundred genes across species boundaries. We will be able to systematically classify genes as resistant or permissive to transfer, examine the effect of the function and position in metabolic and regulatory networks on resistance to transfer, as well as identify any genes with substantial intrinsic benefits.
We have identified, by introducing genes from Salmonella to Escherichia coli, a number of important factors that act as selective barriers to horizontal gene transfer. The most important factors were gene dosage sensitivity, protein disordered regions and the growth environments. In the first instance we were able to show that genes sensitive to expression imbalances showed significant deleterious effects upon transfer and those insensitive to imbalances in expression were either neutral or only mildly deleterious. Secondly, we discovered that genes with more protein disordered regions were more likely to be deleterious. This results is interesting, as to our knowledge, this had never been suggested or demonstrated. However, it is still unclear why this might be. One suggestion, is that these disordered regions give give rise to miss-folding of proteins that then become toxic in the cell, although this has yet to be demonstrated. Lastly we endeavoured to understand the role of the environment in determining fitness effects of newly transferred genes, one aspect that might effect fitness but can not be detected bioinformatically but only through direct experiments. We were able to show that a gene's fitness effect upon transfer is significantly dependent on the environment. Interestingly the most volatile genes (i.e. those showing the largest changes in fitness between environments) where genes of intermediate fitness effects, with those at the extremes being less effected by the environment. In addition, we failed to find any effect of some commonly proposed genic properties, e.g. codon usage, gene function, and divergence. Our ability to test these hypotheses was the result of a fast high throughput protocol that was able to measure very small fitness effects (s ~ 0.00001) which was previously not possible experimentally. With this approach the possibility of measure thousands of genes or transfers between many species is now possible.
This research has important implications for understanding the spread of antibiotic resistance and pathogenicity factors that have an important impact on public health, agriculture, and our economy.
