Normally lethal amino acid substitutions suppress an ultramutator DNA Polymerase δ variant

Abstract

In yeast, the pol3-01,L612M double mutant allele, which causes defects in DNA polymerase delta (Pol δ) proofreading (pol3-01) and nucleotide selectivity (pol3-L612M), confers an "ultramutator" phenotype that rapidly drives extinction of haploid and diploid MMR-proficient cells. Here, we investigate antimutator mutations that encode amino acid substitutions in Pol δ that suppress this lethal phenotype. We find that most of the antimutator mutations individually suppress the pol3-01 and pol3-L612M mutator phenotypes. The locations of many of the amino acid substitutions in Pol δ resemble those of previously identified antimutator substitutions; however, two novel mutations encode substitutions (R674G and Q697R) of amino acids in the fingers domain that coordinate the incoming dNTP. These mutations are lethal without pol3-L612M and markedly change the mutation spectra produced by the pol3-01,L612M mutator allele, suggesting that they alter nucleotide selection to offset the pol3-L612M mutator phenotype. Consistent with this hypothesis, mutations and drug treatments that perturb dNTP pool levels disproportionately influence the viability of pol3-L612M,R674G and pol3-L612M,Q697R cells. Taken together, our findings suggest that mutation rate can evolve through genetic changes that alter the balance of dNTP binding and dissociation from DNA polymerases.

Figure 1. Viability and mutation rates conferred by eex alleles.

(A) Top. Plasmid shuffling scheme. Diploid plasmid shuffling strains with chromosomal deletions of POL3 (lines with triangles) complemented by a POL3–URA3 plasmid (blue) were transformed with LEU2 plasmids (red) expressing various POL3 alleles (pol3-LEU2) on plates lacking leucine and uracil (-Leu -Ura). Bottom. The above pol3-LEU2 transformants were then plated on FOA media in 10-fold serial dilutions to assess viability. FOA-resistant colonies were used as replica cultures in a fluctuation assay to determine mutation rates conferred by (B) pol3-eex alleles, (C) pol3-01,eex, (D) pol3-L612M,eex and (E) pol3-01,L612M,eex (reprinted from Herr et al.32). Black bars indicate MMR proficient cells (BP8001). White bars indicate msh2Δ/msh2Δ cells (BP9101). “X” indicates no growth.

Figure 2. Mutation Spectra of pol3-01,L612M,eex cells.

We clonally isolated cells expressing mutator alleles after 17–20 generations of growth and performed whole genome sequencing. (A) Mutation frequencies. The frequencies of transitions and transversions were determined by dividing the number of observations of each class by the number of scored sites in the genome that could give rise to that mutation. The frequency of each class of insertion/deletion mutation (indels) were determined by dividing the number of observations by the total number of scored sites. (B) Mutation Fractions. The fraction of each class of mutation was determined by dividing the number of observation of each class by the total number of observed mutations. The error bars for both (A) and (B) represent the standard deviations of the mutation frequencies and fractions obtained from independent isolates of the same genotype.

Figure 3. Sensitivity of pol3-L612M,R674G and pol3-L612M,Q697R strains to conditions that modify dNTP pool levels.

(A) Regulation of dNTP synthesis. Phosphorylated Dun1 increases dNTP synthesis by negatively regulating Dif1 (Damage-regulated Import facilitator 1) Crt1, (Constitutive RNR Transcription 1), and Sml1. These three proteins repress either RNR gene expression or RNR assembly. (B) Genetic interactions with dun1Δ and sml1Δ. Two independent plasmid shuffling strains per genotype were plated onto FOA media in ten-fold serial dilutions. (C) Sensitivity to HU. Two independently shuffled strains per genotype were plated in ten-fold serial dilutions onto media containing increasing concentrations of HU.

Figure 4. Locations of eex amino acid substitutions in the S. cerevisiae Pol δ structure.

The S. cerevisiae Pol δ structure (Protein database accession code 3IAY) is shown as a schematic diagram of the α-carbon backbone (rendered in PyMol). Structural domains are color coded as follows: amino (gray), exo (red), palm (purple), fingers (blue) and thumb (green). Yellow spheres correspond to eex mutations from the present study and are labeled to indicate the amino acid substitution; light blue spheres correspond to antimutator mutations from a previous study. The incoming dNTP is denoted by green CPK sticks and the template nucleotide by orange CPK sticks. The primer DNA is represented by tan sticks and the template DNA, by brown sticks. Important non-mutated residues proximal to the antimutator substitutions are shown as space-filling spheres and color coded according to the domain. Active-site carboxylate side chains are gray CPK sticks coming out of the purple (palm) or red (exo) ribbons. Metal ions are small black spheres. (A) Overall distribution of eex amino acid substitutions. The α-carbons are indicated by color coded spheres. (B) eex substitutions affecting an interaction between the amino and palm domains. (C) eex substitutions with structural roles in exonuclease domain. The last three nucleotides of the DNA primer strand from the RB69 Pol Editing structure (Protein database accession code 1CLQ) were placed in the Pol δ exo active site (tan sticks) by aligning the conserved exo domain motifs of RB69 Pol and Pol δ in Pymol. (D) Interaction of the β-hairpin of the exonuclease domain with the primer strand. (E) G818C and the KKR motif implicated in binding of duplex DNA in the minor groove. (F) Amino acid substitutions affecting the polymerase active site. The ribose of the incoming dNTP stacks on Y613 and L612 (pink spheres). The γ-phosphate of the incoming dNTP contacts R674 and the primary amino group of S611. Q697 forms a hydrogen bond with R674, while K559N forms a hydrogen bond with S703.