Validation of the TLS assay. (A) Kinetics of replication of a lesion-free pSP plasmid in S. pombe. At various time points, total DNA was extracted from aliquots of the culture and subjected or not to DpnI digestion. We monitor the kinetics of appearance of plasmid DNA replicated in S. pombe cells by assessing the amount of colonies formed in E. coli. The plasmid replicated in S. pombe is represented by DpnI-treated sample (black bar) and the total amount of plasmid extracted (no DpnI treatment: grey bar). At time point 46 h, essentially all plasmids extracted from S. pombe are resistant to DpnI digestion, proving that they have been replicated in the yeast cells. (B) S. pombe transformation efficiency with control, TT-CPD and TT(6-4) plasmids in the parental numWT strain. (C) TLS assay in the parental numWT strain with lesion-free control plasmid (CONTROL) and plasmids carrying a single TT-CPD or TT(6-4) lesion. The relative amounts of white and blue colonies are determined in E. coli after replication in S. pombe during 72 h. The proportions expressed are per cent (%) are indicated on top of each bar. Error bars result from at least three independent experiments. As expected, with the lesion-free control plasmid both strands yield a similar number of plasmid progeny. We have no explanation for the slight bias (55:45) that is observed. With the lesion-containing constructs, the proportion of blue colonies dramatically decreases, reflecting the replication-blocking potential of the lesions (see text).

Genetic control of TLS pathways across TT(6-4) in S. pombe. The RTE is monitored in various strains as described in Figure 3. Error bars result from at least three independent experiments. The number of colonies recovered after transformation of each S.pombe strain is shown in Supplementary Table S1. (A) Error-free TLS pathway across TT(6-4). (B) Mutagenic TLS pathway across TT(6-4).

Outline of the TLS assay in S. pombe. (A) Scheme of the pSP-TT plasmid construct. The shuttle plasmid contains the ColE1 origin, the ampR resistance cassette, the bi-directional S. pombe ARSDblet replication origin and the URA4 marker for propagation and selection in E. coli and S. pombe, respectively. The single TT lesion, either a cyclobutane (CPD) or 6-4 photoproduct, is located at the beginning of the lacZ′ reporter gene opposite a short sequence heterology. Replication of the lesion-containing strand (DAMAGED strand) by TLS yields a functional lacZ′ gene whatever nucleotides are inserted opposite the TT lesion whereas the NON-DAMAGED strand carries a +1 frameshift that prevents lacZ expression. (B) Flow chart of the experimental design. The covalently closed circular plasmid carrying the single lesion is introduced into S. pombe cells by transformation. After a 72 h period of replication in the yeast cells, the plasmid is extracted, and the DpnI-resistant plasmid replication products are transformed into E. coli cells and plated on ampicillin X-gal plates. The RTE across a given lesion (TLS %) is defined as the proportion of blue colonies (LacZ+) divided by the total number of colonies multiplied by two (see Results paragraph). The TLS events (blue colonies) are further analysed by sequencing to determine the respective proportions of error-free and mutagenic TLS events.

Genetic control of the error-free TLS pathways across TT-CPD in S. pombe. The RTE is monitored in various strains carrying mutations in genes coding for DNA polymerases specialized in TLS or in genes affecting the ubiquitination status of PCNA or combination mutants. For sake of clarity, the strains are grouped with respect to the status of PCNA ubiquitination. Error bars result from at least three independent experiments. For TT-CPD essentially all TLS events (>99%) are error-free in the parental strain. The number of colonies recovered after transformation of each S.pombe strain is shown in Supplementary Table S1.

The ‘hanging tool-belt' model: recruitment of multiple TLS polymerases by PCNA poly-ubiquitination in S. pombe. The major TLS pathways across TT-CPD (A) or TT(6-4) (B) lesions in S. pombe are error-free and involve distinct sets of TLS polymerases: Polη and Polκ for TT-CPD, Polη, Rev1 and Polζ for TT(6-4). These pathways strictly depend on ubiquitination at residue K164 of PCNA through the Rad6/Rad18 complex. Moreover, these TLS pathways require a functional Rad8^Rad5/Ubc13–Mms2 complex that is known to attach K63-linked poly-ubiquitin chains to mono-ubiquitinated PCNA. We propose that in addition to the known interactions of these polymerases with PCNA (or among themselves), their interaction with the K63-linked poly-ubiquitin chain of PCNA facilitates their assembly into multi-polymerase complexes involved in TLS. In S. pombe, Polη and Polκ possess one UBZ domain, whereas Rev1 contains three UBM domains. Several minor TLS pathways are also inferred from careful analysis of the genetics data.