Fig 1.
Evolutionary adaptation to constitutive DNA replication stress in strains with different genomic features.
(A) Experimental scheme: Constitutive DNA replication stress was induced by removing Ctf4, a protein that coordinates replisome activities. Haploid (orange), diploids (blue) and cells with a severely decreased ability to perform homologous recombination (red, recombination-deficient), were subjected to 1000 generations of experimental evolution in the presence of constitutive DNA replication stress. (B) The fitness of the 6 ancestral strains (2 for each genome architecture) and of 24 evolved populations derived from them, relative to WT cells (s = 0). Cells were competed against reference strains with the same ploidy. Fitness data of haploid strains (orange) is from [17]. The semi-transparent gray dot represents a diploid strain that became haploid during the experiment and is thus excluded from subsequent analysis. Error bars represent standard deviations. a and α refer to the strains’ mating type (MAT locus). a/α indicates diploid strains. The P-values reported in figures are the result of t-tests assuming unequal variances (Welch’s test). The fitness values shown here are reported in S1 Data.
Fig 2.
Fitness increase over 1000 generations.
(A) Fitness change over the course of evolution for strains with ctf4Δ-induced replication stress. Average trajectories for populations with different genomic features are shown as thick lines. Thin traces represent individual populations’ trajectories. (B) Absolute fitness increase over 1000 generations for the evolved populations. Δs is the difference between the fitness of the evolved populations and the fitness of their ancestors, with both fitnesses measured relative to that of cells that contain Ctf4 (CTF4) and have the same ploidy. (C). Adaptation rates (measured as Δs/generation) during the early phase of evolution (0 to 100 generations) and the rest of the experiment (100 to 1000 generations). (D) Shows the correlation between the fitness increase during evolution (Δs) and the fitness defect of the ancestors relative to cells of the same architecture containing Ctf4 (the ancestor relative fitness). (E) Absolute fitness increase over 1000 generations for control strains that contained wild-type CTF4. Fitness data of haploid strains (orange and white) is from [17]. Error bars represent standard deviations. Grey line represents linear regression between the datapoints. The P-values reported in figures are the result of t-tests assuming unequal variances (Welch’s test). The data shown here are reported in or derived from S1 Data.
Fig 3.
Putative adaptive mutations in the evolved clones.
(A) Venn diagram representing the mutations (SNPs and small InDels 1-55bp) found in the evolved clones and populations. The area of the circles is proportional to the list of genes mutated at least once in the set of evolved clones for each genomic feature. The number within each sector represents number of genes which were mutated in one, two, or three genomic features. Tables include the genes whose mutations were significantly mutated in either global or feature-specific analysis. Symbols next to the genes highlight that they were statistically significantly mutated in strains with individual features (* refers to haploid, # to diploid and + to recombination-deficient features, e.g. SLD5 +# means that the gene is found mutated in strains with all three features, but is statistically significantly mutated only in diploid and recombination-deficient populations). The identities of all the genes in the Venn diagram intersections are reported in S5 Data (B) Copy number variations (CNVs) affecting different chromosomes (roman numbers) which appeared in at least two independent populations. Arabic numbers are the number of independent populations, in each genomic feature (highlighted by a different color), in which the CNVs were detected. Copy number change refers to the fragment’s gain or loss during the evolution experiment (i.e. +1 means that one copy was gained in haploid cells and that two copies were gained in diploid cells). Frames around the plots have the same color as the sectors in the Venn diagram, and encircle CNVs that have been found in one, two, or three genomic features. All the segmental amplifications detected are reported in S4 Table.
Fig 4.
Homozygous mutations in diploids evolved in the presence of constitutive DNA replication stress.
(A) Homozygous mutations (SNPs and small InDels of 1-55bp) detected in evolved ctf4Δ/ctf4Δ diploids clones distributed by chromosome (roman numbers). Dark blue diamonds represent loss of heterozygosity events (LOH). Light-blue triangles represent genes which contain two independent mutations (double hits) and the gray circles are the centromeres. The identities of the mutations shown here are reported in S6 Table. (B) DNA replication and DNA damage checkpoint genes with mutations in both copies in evolved ctf4Δ/ctf4Δ diploid clones. Gray lines represent evidence of genetic and physical interactions from the literature (https://string-db.org). Node diameter is proportional to the number of populations in which the gene was mutated. Dark blue represents LOH events, light blue represents double hits. Percentages refer to the average estimated frequency within populations of the clones carrying the respective mutation in homozygosity (LOH) or carrying both mutations (double hits).
Fig 5.
Mutations affecting sister chromatid cohesion, DNA replication and the DNA damage checkpoint.
(A) Heatmaps representing all the mutations (SNPs, small InDels of 1-55bp and segmental amplifications) causally implicated in these phenotypes in the 92 sequenced evolved clones. (B) Frequency of SNPs and small InDels (1-55bp) affecting genes (Open reading frames and associated regulatory regions). (C) Frequency of genes associated with the three processes that are present on segmental amplifications. The number of hits and the identities of the genes which were mutated are reported in S6 Data.
Fig 6.
Reconstructed adaptive mutations can produce evolved fitness increases.
(A) A subset of mutations affecting sister chromatid cohesion, DNA replication and DNA damage checkpoint were re-constructed in the the three ancestral strains that differed in their genomic features. The fitness of the re-constructed strains, relative to their respective ctf4Δ ancestor cells (s = 0) is depicted. For diploid strains, we replaced one (het) or two (hom) copies of the wild-type gene with the mutant allele. The fitness values shown here are reported in S7 Data. (B) Reconstructed strains carrying combinations of adaptive mutations affecting sister chromatid cohesion, DNA replication and DNA damage checkpoint recapitulate the fitness of their respective evolved populations after 1000 generations. The most frequently mutated gene affecting DNA replication was chosen for the reconstruction in individual features (e.g. IXR1 was mutated more often that SLD5 in haploids whereas the reverse was true in diploids). Fitness data of haploid strains (orange) is from [17]. The P-values reported in figures are the result of t-tests assuming unequal variances (Welch’s test). The fitness values shown here are reported in S7 Data.