Computing fitness in each environment. From Fig. 1: Determining the fitness landscape of the yeast tRNA gene in multiple environments.

Organisms live in frequently changing environments, which affect their evolution. For instance, pathogens respond and evolve differently in different drug environments. Finding general rules governing the environmental impacts of mutations is thus important for understanding and predicting evolution.

Recent alumnus Chuan Li (U-M EEB Ph.D. 2017), postdoctoral fellow at Stanford University, and her former advisor, U-M EEB Professor Jianzhi “George” Zhang, published “Multi-environment fitness landscapes of a tRNA gene” in the journal Nature Ecology & Evolution, April 23, 2018.

Gene mutations are the raw materials of evolution. In a given environment, a mutation arising in an organism can benefit, harm or have no significant impact on the growth and reproduction of the organism carrying the mutation. Gaining insights into how random mutations affect the fitness of an organism provides valuable information about evolution. The fitness landscape is a conceptual tool used by researchers that allows them to visualize and predict evolution. The fitness landscape of a gene in a given environment describes how different mutations of the gene affect the fitness of an organism, which is usually represented by its reproductive rate.

Experimentally describing a fitness landscape is a formidable task because of the immense genotype space and difficulties of measuring fitness at a large scale. The genotype consists of vast numbers of genetic possibilities for any single organism. In addition, fitness of all these variants need to be quantified in an efficient, high-throughput and unbiased fashion.

To conquer these difficulties, a 2016 study published in Science by Li, Zhang and colleagues designed a powerful high-throughput method that combines precise gene replacement with next-generation sequencing to quantify the fitness landscape of a transfer RNA (tRNA) gene. The yeast gene they selected contains 72 nucleotides and holds the genetic instructions for making the tRNA molecule, which is part of the cellular machinery used to synthesize proteins from amino acids.

Their previous work successfully mapped a fitness landscape for this tRNA gene, where the researchers managed to assess the reproductive success associated with mutating each of the 72 nucleotides to A, T, G and C (the four nucleotides found in DNA), and 61 percent of all possible pair-wise combinations, as well as tens of thousands of variants carrying three or more mutations.

“We assessed one environment previously. However, organisms live in changing environments, so scientists are interested in describing the fitness landscape across multiple environments to better explain and predict evolution in the frequently changing natural world,” explained Li. “To elucidate this question, we applied this methodology to determine the fitness landscape of the yeast tRNA gene comprising over 23,000 variants in four different environments, including all possible single mutants, approximately 40 percent of all double mutants and a large number of variants carrying multiple mutations. The fitness of a genotype varies significantly across environments, but the pattern of change is quite simple. We proposed a piecewise linear model, a linear model with different slopes for beneficial and deleterious mutations, for predicting fitness landscape across environments, which enables researchers to infer fitness landscape in a new environment from an existing fitness landscape using only a few genotypes in the new environment.

“Our study reveals simple rules underlying a seemingly complex fitness landscape, opening the door to understanding and predicting fitness landscapes in general. We are also interested in how multiple mutations would interact to affect the fitness of the organism. Under each environment, we observe the prevalent pattern that having multiple mutations is worse than one would expect if mutations work independently. This pattern is called ‘negative epistasis.’ Interestingly, the magnitude of epistasis varies across environments.”

Li was recently named a Damon Runyon Fellow by the Damon Runyon Cancer Research Foundation. The recipients of this prestigious, four-year award are outstanding postdoctoral scientists conducting basic and translational cancer research in the laboratories of leading senior investigators across the country.