Experimental Fitness Landscape For Bacteriophage Infectivity

Yuuki Hayashi,Takuyo Aika,Tetsuya Yomo

2009-01-01

Abstract:In vitro molecular evolution is comprehended as the "adaptive walk" on the fitness landscape in sequence space, where the concept of the fitness is expanded to a measure of the physicochemical properties of biomolecules. Since the structure of the fitness landscape governs the evolvability of biomolecules, it is important issue for life science to depict statistical features of fitness landscapes. In our study, we focused on the evolvability of the infectivity of filamentous phage fd. The infection of the phage fd into E.coli host cell is initiated by attachment of the D2 domain in the g3p minor coated protein onto F-pilus of E.coli host cell. At the beginning of the evolution experiments, we made a variant phage, fd-RP, by replacing the D2 domain that includes a random polypeptide of 119 amino acids. Through 20 cycles of the random mutagenesis to the 13-130th sites in the "artificial D2 domain" and subsequent selection for the infectivity, the phage evolved to acquire a 1.7 x 10(4)-fold higher infectivity than the initial phage fd-RP. This results showed that a random sequence could acquire the infectivity through the iterative mutation and selection. We analyzed the experimental data based on the mathematical theory of the adaptive walk on the NK fitness landscape, which is a familiar mathematical model of the fitness landscape. As a result, we estimated several parameters that describe the structure of the infectivity landscape and depicted a rough view of the landscape. From the foot to the middle region, there are no local optima, while over the middle, many local optima appear drastically. This suggests that random sequences can easily evolve up to the middle region through iteration process of mutagenesis with only a few amino acid substitutions and subsequent selection.