This thesis work applies Atomic Force Microscopy (AFM) for single-molecule imaging to understand the interactions between DNA and DNA-binding proteins. Flap Endonuclease (FEN) and ParB are the two proteins explored here, that bind DNA as an intermediate step in DNA replication and bacterial plasmid segregation respectively. With the objective to study FEN-flap DNA interactions, different immobilization methods, as well as protein and DNA modification methods, have been applied. The images of 100 bp flap DNA have revealed a Y-shaped flap DNA showing a thicker and higher double strand and a thinner less-elevated single strand attached to it. Helical resolution was obtained for DNA immobilized on Ca2+ treated mica and the molecules showed the major grooves of DNA. This was noteworthy as the helical resolution was difficult to resolve for such short DNA fragments and these observations have not been reported previously. However, there were no or very scarce ssDNA seen during flap DNA sample imaging, signifying the immobilization technique was not suitable for flap DNA. At a high salt concentration, salt precipitates were seen as noise on the mica surface. Ca2+ and Ni2+ was discovered to be more effective than Mg2+ for immobilization of dsDNA, with Ni2+ allowing the least amount of mobility of DNA in the consecutive frames. PLO treated mica proved to be the most suitable surface for DNA immobilization during imaging in buffer conditions, though it was not effective to visualize the short flaps. However, it allowed imaging with helical resolution where both the major and minor grooves of long DNA, like plasmids, was visualized. The imaging was simpler and more efficient if the length of the DNA strands was longer than 100 bp. Hence, the method of assembly of long DNA oligonucleotides to form 300 bp flap DNA with ~ 100 nt flap was applied such that the sample could be imaged with a good resolution and did not mobilize during consecutive scans.
Dynamic imaging of inactive FEN interaction with flap DNA showed the mobilisation of DNA on the mica surface, indicating that protein recognition and interaction affected the DNA conformation and disrupted the forces surrounding DNA that anchored it to the PLO treated mica surface. Active FEN could be seen intermittently mobilizing the DNA, threading it and cleaving the ssDNA branch. A method of motion tracking and quantification of the movement of sections of DNA was developed to allow the segmentation of DNA and tracking of each of the segments to recognise the DNA sections most affected by the protein. It was found that FEN caused the DNA to bend in the vicinity and time duration of its binding, and changes in DNA shape were witnessed in the short time span of the protein’s appearance close to DNA, after which the DNA anchored to the mica surface and immobilized. The results corroborated the previously reported theories that FEN binds the dsDNA, bends the ssDNA strand and makes it thread through the arch of the active site. However, we have proposed that FEN initially binds to DNA (both flap and overhang) not at the branch point, as previously proposed by crystallography studies, but anywhere on the DNA and is capable of ‘sliding’ on it to find the ssDNA junction to which it can anchor and initiate the nuclease activity step.
ParB interactions with parS were performed by imaging the reaction mixture immobilized on PLO treated mica and comparing the trace lengths of the DNA molecules. When the ParB protein was bound to DNA, the molecules displayed a ‘beadiness’ and the length of DNA was reduced in comparison with the length of the DNA negative control. The presence of CTP in the buffer caused the protein to bind to DNA in a sequence non-specific manner, hence caused the sliding of the protein off the DNA. These observations were consistent with the previously reported research, yet how ParB recognised the DNA in the first place in the absence of any parS site, remains unknown. Potential future experiments, building from this work, are also discussed.
