Experimental evolution of a duplicated fluorescent protein.

A) E. coli cells carry a plasmid containing a duplicated gene coding for coGFP. B) Upper panel: Plasmid with two copies of the cogfp gene, both under control of independently inducible promoters (Ptet and Ptac); lower panel: control single-copy plasmid with only one active gene copy, the other copy is not fluorescent due to mutations engineered into the chromophore. C) Overview of the directed evolution experiment of the duplicated cogfp gene (see text for details).

Gene duplication increases mutational robustness.

The vertical axis shows mutational robustness, measured as the percentage of cells that maintain their fluorescence after mutagenesis, as a function of time (in generations of directed evolution) on the horizontal axis. Thick blue and red lines stand for the median fraction of fluorescent cells for single-copy and double-copy mutant libraries respectively, while dotted lines and circles indicate data from the three biological replicates. One tailed Mann-Whitney tests, * p≤0.05, n=3.

Gene duplication does not lead to significantly faster evolution of green fluorescence.

A) The vertical axis shows green fluorescence (log10) as a function of time (generations) on the horizontal axis. B) The vertical axis shows green fluorescence (log10) normalized to the fluorescence of its ancestral population as a function of time (generations) on the horizontal axis. Thick blue and red lines stand for the median fluorescence of single-copy and double-copy mutant populations respectively, while dotted lines and circles indicate data from the three biological replicates. One tailed Mann-Whitney tests, * p<0.05, n=3.

Double-copy populations accumulate more mutations per gene than single-copy populations during the first two generations.

The vertical axis shows the average number of non-synonymous mutations per cogfp gene, as a function of time (in generations of directed evolution) on the horizontal axis. Thick blue and red lines stand for the median numbers of single-copy and double-copy mutant populations respectively, while dotted lines and circles indicate data from the three biological replicates. Mann-Whitney tests * p<0.05, n=3.

Populations with two gene copies are showing increased genetic diversity and dN/dS ratios.

The horizontal axes of all panels show time in generations during selection for green fluorescence. A) Average pairwise amino acid distance for coGFP molecules in single-copy populations (blue: active copy, grey: inactive copy), B) average pairwise amino acid distance for coGFP molecules of the active copy in single-copy populations (blue) vs both copies in double-copy populations (pink). C) dN/dS ratio in single-copy populations (blue: active copy, grey: inactive copy), D) dN/dS ratio of the active copy in single-copy populations (blue) vs both copies in double-copy populations (pink). Thick lines represent the median over three replicate populations, while dotted lines indicate data from the individual biological replicates. * p ≤ 0.05 Mann-Whitney test, n=3.

Combination of beneficial mutations emerged earlier in double-copy populations.

The vertical axis shows the frequency of indicated mutations and combinations thereof in the populations under selection for green fluorescence, as a function of time (in generations of directed evolution) on the horizontal axis. Thick blue and red lines stand for the median frequencies for single-copy and double-copy populations, respectively, while dotted lines and circles indicate data from the three biological replicates. Detailed statistics are reported in Table S3.