Here we performed TEM and AFM imaging

Here, we performed TEM and AFM imaging to study APE1–DNA interactions. The results showed that APE1 can dynamically polymerize on DNA in an apparently sequence-independent manner. We propose that this protein polymerization allows for scanning of the structural properties of DNA with short residence time, in order to search for abasic sites. Cooperative oligomerization of KN-93 along a DNA duplex is well known for recombinases (RecA and Rad51) and for other proteins such as HNS, Stpa, LrpC, and Fur [[65], [92], [93]]. Such protein polymerization onto DNA is related to various functions and is referred to as “collaborative protein filaments” [92]. A DNA duplex containing an AP site manifests local flexibility, which is recognized by APE1. We demonstrated that APE1 can stabilize a kink, specific structure of the DNA duplex, as it was previously shown for PARP1 on a single-strand break [83] and for MC1 on its DNA-binding site [[84], [94]]. We propose that the APE1 oligomer functions as a sensor of conformational properties of DNA and this new function of APE1 allows for scanning of DNA and enables the search for local DNA structural perturbations such as an AP site and other DNA lesions. The finding that multiple APE1 proteins can be stably associated with an undamaged DNA duplex implies a decrease in the effective enzyme concentrations in solution, which in turn should lower enzymatic activities on long DNA substrates. In line with this notion, we observed a dramatic decrease in APE1-catalyzed cleavage activities on 64mer DNA substrates as compared to 17mer ones. Furthermore, the efficacy of AP site and αdA-DNA cleavage by APE1 can be ranked in the order 17mer > 22mer > 43mer > 64mer, strictly depending on the size of DNA fragments. Contrary to full-length APE1, the truncated APE1-NΔ61 protein manifested only a slight preference for short DNA duplexes as compared to long ones. Besides, E. coli AP endonuclease Nfo did not discriminate between DNA substrates of various length and cleaved 17–64mer duplexes with more or less similar efficacy. These data are in agreement with TEM and AFM imaging, which revealed that full-length APE1 tends to polymerize on DNA fragments, thus generating protein oligomers along DNA fragments. Altogether, these results suggest that the first 61 N-terminal residues of APE1 participate in protein polymerization along DNA and in formation of oligomerlike complexes by binding to undamaged DNA via electrostatic interactions between DNA phosphates and positively charged basic lysine and arginine residues. On the basis of our data, we can formulate a model in which APE1 binds to the DNA duplex termini in a stable manner and then recruits other APE1 molecules to bind nearby in a cooperative manner, which in turn initiates polymerization of the APE1 proteins along DNA, thus generating APE1-oligomers seen in TEM and AFM images (Fig. 10). Upon initial binding of DNA glycosylases to their substrates in the absence of APE1, the mutual “induced-fit” steps shift the conformational ensemble toward the transition state. APE1 oligomers mediate conformational changes in the DNA duplex, which may mimic the conformation of the transition state of the reaction, thereby lowering the entropic barrier and stabilizing the enzyme–substrate complex. In the equilibrium association of a DNA glycosylase with a damaged site in DNA, the rate constant of the forward reaction is close to a diffusion-controlled value, but when the DNA substrate is bound to APE1, it is preformed for the formation of a catalytically competent complex with DNA glycosylase. Consequently, the rate constant of the reverse reaction should be lower in comparison with the APE1-free case. By contrast, in the complex of DNA glycosylase with the end product (AP site), the dissociation rate constant is higher for the AP site associated with APE1 owing to distortion of DNA structure, which decreases the number of specific contacts between DNA glycosylase and AP site DNA. Overall, we can say that APE1-catalyzed stimulation of certain DNA glycosylases is based on the transformation of an “induced-fit” into a “conformational selection” mechanism. Finally, APE1-induced conformational changes in DNA helix stimulate (i) formation of a catalytically competent enzyme–substrate complex, (ii) disruption of the DNA glycosylase–end product complex, and (iii) probably the binding of redox-dependent transcription factors to DNA. We propose that the redox domain of an AP endonuclease, acquired during evolution of vertebrates, provides additional biological functions including redox and DNA glycosylase stimulation. In the process of natural evolution, these additional functions of APE1 have become essential for cellular proliferation and embryonic development in mammals.