Weird gene in a weird mammal: comparative and functional analysis of the highly divergent sand rat Pdx1 gene

Deleterious mutations in highly conserved genes are expected to be removed by natural selection. However, when these genes are trapped in dynamic regions of the genome, for example mutational hotspots, it becomes difficult to predict how the conflict between nucleotide changes and natural selection will be resolved. In this thesis, I investigate highly divergent genes in the fat sand rat (Psammomys obesus) associated with GC-skew. To understand how the tug-of-war between GC-skew and natural selection is resolved, I focus on a highly representative example of an extremely conserved gene trapped in an extreme GCrich region of the sand rat genome: pancreatic duodenum homeobox 1 (Pdx1). In Chapter 2, I hypothesize that the highly divergent sand rat PDX1 protein may be a result of adaptive selection, and that similar amino acid changes may be observed in other arid-living rodents. Using genomic and transcriptomic data gathered from 33 rodent species, I show that there is near 100% conservation of the PDX1 homeodomain sequence in non-gerbil species, including 15 rodent species that reside in arid or semi-arid habitats. The highly divergent Pdx1 gene observed in the sand rat is only found in three gerbil species, suggesting that adaptive or relaxed selection in an arid environment may not be the key factor affecting Pdx1 gene evolution in this lineage. Instead, other factors such as GC composition bias may be the main factor driving Pdx1 gene divergence in the gerbil subfamily. In Chapter 3, I hypothesize that natural selection may have only managed to purge extremely deleterious mutations while tolerating the existence of mildly deleterious mutations in these genes. Using mammalian cell transfection and pulse-chase labelling, I reveal loss of a conserved ubiquitination site and compensation from a unique, gerbil-specific ubiquitination site. This provides evidence for loss of a conserved amino acid with important function associated with GC-skew and links the presence of a gerbil-unique lysine residue with a specific function. In Chapter 4, I use qPCR to demonstrate significantly altered regulatory effects of the sand rat PDX1 protein on key target genes in rat pancreatic cells, Ins1 and Ins2. I also present establishment of CRISPR/Cas9 edited knock-in cell lines with the aim of teasing apart whether these observed differences were caused by changes in sand rat PDX1 binding site recognition. I conclude that the highly divergent sand rat PDX1 has likely lost conserved functional residues and has significantly reduced ability to initiate proper biological response to glucose stimulation in a rat cell environment. At the same time, results in Chapter 3 show the presence of gerbil-specific residues with specific function, indicating possible compensation by natural selection. I use genome wide comparison methods in Chapter 5 to demonstrate that extreme GC-skew has affected multiple highly conserved genes with assorted function. Pdx1 is not the only aberrantly divergent gene in the gerbil lineage, and these aberrantly divergent genes are linked to GC-skew and high GC regions in the gerbil genome. Four divergent genes, Pdx1, Insr, Medag and Spp1 are likely associated with obesity and abnormal metabolism observed in gerbils on a standard laboratory diet. Overall, this research provides further characterization of GC-rich genome regions in the gerbil lineage and functional analysis of a highly divergent sand rat gene, Pdx1, that is affected by GC-skew but also has significant functional importance. This work suggests that when extreme changes in genome nucleotide content occurs, natural selection may be incapable of removing some maladaptive mutations in highly conserved genes that are otherwise abolished in other lineages.