Genome sequencing of model organisms, including the yeast S. cerevisiae [1] and the annotation of all the open-reading frames tremendously boosted our knowledge of the genetic landscape of many species. Then, systematic gene deletion efforts provided a comprehensive collections of gene-deletion mutants amenable to systematic phenotypic analyses [2], [3], [4], [5], [6], [7], [8], [9]. Thus,

in S. cerevisiae, the essential and non-essential genes were systematically characterized in haploid segregants of heterozygous mutant diploids, based on the viability or inviability phenotype conducted under standard, nutrient rich growth condition (YPD). It is now estimated that the reference S288C genome carries ~ 1,100 genes essential for cell viability (Yeast Deletion Project www-sequence.stanford.edu/group/yeast_deletion_project/deletions3.html). However, the growing use of various strain backgrounds by the yeast community encounters the complex problem of the genotype-phenotype relationships, exemplified by the varying phenotype of deletion mutants that turned-out to be essential or not, according to the strain background [10]. Being confronted with difficulties to build a previously reported viable gene deletion in a yeast strain, we encountered and explored the intriguing case of the background-specific viability/inviability of the POL32 gene, in close and distant natural S. cerevisiae isolates.

Pol32 is a subunit of the DNA polymerase Polδ complex which includes the evolutionary conserved Pol3, Pol31 and Pol32 subunits, respectively encoded by the POL3, POL31 and POL32 genes in S. cerevisiae. The human orthologous proteins are POLD1, POLD2, and POLD3, respectively [11]. POLD2 and POLD3 are commonly amplified in human tumors and both germline and sporadic mutations were found in POLD1 (reviewed in [12]). In the S. cerevisiae heterotrimeric Polδ complex, the largest subunit Pol3 (p125) is the essential catalytic subunit that elongates the discontinuous lagging strand with a proof-editing activity while Pol31 (p55) is an essential subunit of the holoenzyme with a role that remains to be elucidated. While intensively studied, the role of the tightly associated Pol32 subunit (p40) also remains to be elucidated.

Considering the central role of Pol δ in replication, it was not surprising that the inactivation of the POL3 and POL31 genes in S. cerevisiae was lethal [11]. However, it was rather unexpected that the deletion of the POL32 gene was fully viable in S. cerevisiae [11], a result later confirmed in the closely S288C related FY and BY S. cerevisiae backgrounds ([3], [13], present study). In contrast, the inactivation of the POL3, POL31 and POL32 orthologous genes in S. pombe [4],[14] ,mouse and human cells [15], [16], [17] were lethal, raising the case of a species-specific difference. Another evolutionary difference is that S. pombe and mammals encode a non essential fourth small subunit (p12) of unknown function but reported to stabilize the Polδ complex [18], [19]. Here, we describe how we could not delete the POL32 gene in the haploid Saccharomyces cerevisiae SK1 strain, raising the hypothesis of a yeast intra-species variability rather than an inter-species feature. Suspecting a yeast strain background effect, we conducted a comparative functional and genomic analysis to identify suppressor(s) of the pol32Δ mutant viability/inviability in close or distant S. cerevisiae strains. We found that single-nucleotide polymorphisms of Pol31 explain the dispensability or not of the evolutionary conserved Pol32 subunit in distinct yeast clades.