During the TRISCPOL project I attempted to shed light on the role of the FeS cluster in human Pol δ. I have experimentally validated FeS cluster binding by human Pol δ using radioactive iron incorporation assays and identified the key residues in the Cys B motif of the POLD1 subunit that are relevant for ligation of the FeS cluster. I have successfully developed and employed a codon-optimised system for the expression of human Pol δ in Sf9 cells. Using a combination of rational enzyme engineering and site-directed mutagenesis, I have generated a number of human POLD1 mutants, which were either completely or partially defective in binding of an FeS cluster. The ability to coordinate an FeS cluster by the engineered POLD1 variants was assessed using radioactive iron incorporation assays in Sf9 cells. One variant of human POLD1, which is almost completely unable to bind an FeS cluster, and two partially FeS cluster binding-deficient variants were selected for an in-detail characterisation in vitro using EMSA, thermal inactivation, primer extension and degradation assays.
In contrast to data from yeast, our data show only a partial destabilisation of human Pol δ in the absence of an FeS cluster. However, this partial structural defect translates in significant enzymatic deficiencies. Namely, the FeS cluster-deficient variant exhibits strongly reduced efficiency and processivity of DNA polymerisation. Additionally, this mutant replicase has modulated 3’-5’ DNA nucleolytic activity. Interestingly, instability of the FeS cluster mutant can be efficiently alleviated in the presence of the sliding clamp PCNA. PCNA also rescues the efficiency and processivity of DNA polymerisation, but not the defect of proof-reading of the FeS cluster-deficient human Pol δ. As revealed by thermal inactivation experiments, the FeS cluster-deficient variant of human Pol δ complexed with PCNA forms a less stable holo-enzyme and is inactivated more rapidly than wild-type Pol δ. Finally, the structural defect of the FeS cluster-deficient Pol δ in the presence of PCNA lowers the fidelity of DNA replication in vitro. Interestingly, the partially destabilised FeS cluster variants of POLD1 assemble into stable complexes. The thermal inactivation profile of these enzymes is very similar to the wild-type complex. Moreover, the DNA polymerisation activity of these variants seems to be unchanged. However, like the FeS cluster-deficient mutant, these enzymes had partial defects in their proof-reading abilities. Moreover, the fidelity measurements show clearly that both of the enzymes have strongly reduced accuracy of DNA synthesis. Intriguingly, the decrease in fidelity is almost the same as observed for the proofreading-deficient variant of wild-type Pol δ. In essence, our in vitro data confirm the previously suggested structural role of the FeS cluster in the stabilisation of human Pol δ. More importantly, we provide evidence that an FeS cluster-binding defect is linked to a decreased proof-reading efficiency. The proof-reading capability of human Pol δ is not only important to maintain high fidelity during unperturbed DNA synthesis, but is also a well-known kinetic barrier that limits the involvement of this replicative enzyme in translesion DNA synthesis. Therefore, the TRISCPOL in vitro data are potentially relevant for further understanding of the role of human Pol δ in DNA damage tolerance and mutagenesis. This in turn is highly relevant for resistance of cancer cells and may be utilised to develop new improved methods of treatment of this disease.
In parallel to our in vitro experiments, we have generated cell lines that inducibly express variants of POLD1 including the wild-type protein and several FeS cluster-binding mutants, and allow concomitant down-regulation of endogenous POLD1. The characterisation of these cell lines is currently in progress. We have performed clonogenic survival assays in conditions of DNA replication stress and observed an important role of the FeS cluster in cell survival. Moreover, DNA replication dynamics were compared between control cells and cells depleted of POLD1. Since depletion of POLD1, however, did not result in significantly altered replication dynamics, we have not pursued this approach any further. Instead we are currently setting up a FACS protocol, which will allow us to label for markers of DNA damage and investigate the efficiency of DNA replication.
