Why do so many fungal open reading frames contain repeats?

Jan Schmid 1, Matt Wilkins2, Ningxin Zhang1, Rosie Bradshaw1, Murray Cox1, Richard Cannon3, Chris Schardl4
1Massey University, Palmerston North, New Zealand, 2Scientific Consulting, Palmerston North, New Zealand, 3University of Otago, Dunedin, New Zealand, 4University of Kentucky, Lexington, KY, USA

Tandem repeat-containing DNA mutates orders of magnitude faster than normal DNA by insertion and deletion of repeat units.  In microbes, hypermutable DNA repeats, when located in open reading frames (ORFs) or promoters in so-called contingency genes, can serve as an ‘insurance’. Their hypermutation constantly replenishes a pool of variants in the population, from which the fittest are chosen. If circumstances change suddenly, for instance if the host launches an immune response against the predominant allele of a contingency gene of a microbial pathogen, the very survival of the population may depend on the presence of rare variant alleles not targeted by this response. Hypermutability is a cost however when the environment is not changing, by constantly generating inferior alleles. Research by us and others has shown that the pathogenic yeast Candida albicans has 2600 hypermutable tandem repeat-containing ORFs (TR-ORFS), but that most of these may not be contingency genes used in short-term adaptation: Isolates of similar genetic background tend to have similar or identical TR-ORF alleles, even when isolated from different patient types and anatomical locations. Furthermore a genome- wide survey revealed that in the majority of TR-ORFs synonymous mutations reduce repetitiveness and thus mutability of the DNA, while retaining the encoded amino acid repeat. A survey of a set of other fungi (Saccharomyces Epichloë festucae, Cladosporium, Dothistroma), suggests that TR-ORFs are frequent in fungal genomes and that reduction of mutability by synonymous mutations is common. This suggests that TR-ORFs in fungi may often function to speed up adaptation by increasing the rate of evolution of novel proteins or for generating genetic background- specific alleles that work well under a variety of circumstances rather than as contingency genes. However, in C. albicans we observe that the association of specific alleles and specific genetic backgrounds often breaks down for a small number of TR-ORFs with high predicted mutability (VAR scores) and for which predicted mutability is not diminished by synonymous mutations, an indication that these TR-ORFs might be contingency genes.  The latter result also suggest that combining in silico predictions of mutability and the degree of its reduction by synonymous mutations might be a useful tool for detecting contingency genes in fungal genomes.