This thesis is concerned with two distinct fundamental research questions that are both investigated using the E. coli lac system. In the first chapters we investigate what the shape of biological fitness landscapes look like. Chapter 2 reviews recent progress in measurement of empirical fitness landscapes, and introduces the open questions in evolution that they may answer, such as why particular evolutionary paths are taken. In this chapter, we also introduce the concept of epistasis as a useful description of the local shape of fitness landscapes. In chapter 3 we describe existing in vivo measurements on lac repressor and operator mutants and show how these can be used to construct a fitness landscape of lac regulation. Using computer simulations we simulate mutational pathways and reveal that new regulatory interactions can easily evolve. Chapter 4 deals with the local structure of the lac landscape. We determine that the landscape is multi-peaked and, consistent with earlier predictions, show the presence of reciprocal sign epistasis. We conclude our analysis of the lac landscape in chapter 5 with a more global analysis of its structure, focusing on which landscape features are important for evolution. This study reveals that the essential features of the lac landscape can be sufficiently captured by modeling the presence or absence of additivity between functional residues. In chapter 6 we turn to another fundamental research question: how do random molecular fluctuations in the number proteins in a single cell propagate to its growth? Again, we use the E. coli lac system to investigate this question. But whereas the first part of this thesis consists of theoretical simulations of lac regulation, here we perform laboratory experiments on E. coli cells that require use of their lac enzymes for growth. By means of automated and highly sensitive fluorescence microscopy, we measure both fluctuations in lac level and in growth rate in individual growing cells. These experiments show that fluctuations in the growth rate of single cells can be linked to protein fluctuations, but also reveal a intricate dynamic interdependency between these two properties.

Kavli Institute of Nanoscience Delft

Applied Sciences