Final Report Summary - HFEN1DYNAMICS (Defining the Role of Flap Endonuclease 1 Conformational Dynamics in Catalysis)
The process of replicating our genetic material is awe-inspiring in its complexity, as it involves copying billions of nucleotides with exceptional speed and accuracy. This amazing feat is performed by a group of proteins that together form a 'replication machine'. Understanding the function of each of these proteins is critical since failure of any one of them may result in a life threatening disease. In addition to the problem of accurate replication, DNA is under continual onslaught from environmental chemicals and radiation (mainly the sun's rays). These alter the DNA bases so that they no longer pair up, as they should, again a life-threatening situation. Biological systems have evolved a number of enzymes (biological catalysts) that are involved in repairing damaged DNA. An enzyme involved in both DNA replication and repair is Flap endonuclease. Without flap endonucleases and enzymes like them (known as a family of proteins) life cannot exist. Understanding how flap endonuclease and its other family members function at the molecular level is the principle aim of this work.
Furthermore, the information garnered here would be invaluable for those who seek to develop new cancer therapeutics by creating inhibitors of flap endonucleases, as large amounts of flap endonucleases are present in cancer cells and its abundance in cancer cells allows cancers to progress more rapidly. Flap endonucleases are specifically involved in the removal of DNA 'whiskers' that form in two specialized processes; one in replication and one in repair. These whiskers of DNA have to be removed with extraordinary precision; otherwise, DNA replication and repair will be faulty and could result in cancer. Flap endonucleases are the molecular scissors that remove the whiskers with the necessary precision. Recently, we have obtained 'snap-shots' of the protein in various conformations or 3D structures (i.e. a picture of scissors in an open and a closed state), and this information has greatly increased our understanding of flap endonucleases and its family members.
Unfortunately, snap-shots of various conformations do not tell the entire story. Just as scissors must move to be an effective tool for cutting, flap endonucleases must be able to move its parts to cut the DNA whiskers. We proposed to use a specialized technique called nuclear magnetic resonance, which can probe the magnetic properties of atoms, to quantify the motions of FEN1. An understanding of the motions of flap endonucleases will solidify our understanding of flap endonucleases and its family members and help guide pharmaceutical drug discover and design programs.
Based on the newest information provided by the DSF studies as described in the midterm report, which showed that hFEN1 at NMR conditions is mono-disperse at pH 8.5 we continued our investigation into hFEN1-336 stability at NMR conditions. To ensure that the aggregation was not caused by aspects of the purification and concentration procedures, we modified our procedures by increasing the pH of the certain buffers and only concentrated the samples using ultrafiltration to avoid concentration gradients at the membrane surface. Except for the disappearance of certain peaks in the HSQC spectrum due to the increased exchange rate of the amide protons, the spectra hFEN1-336 at pH 8.5 indicated that the protein was still folded and in a conformation similar to that at pH 7.5.
The sample did, however, remain soluble longer showing little to no precipitation. Unfortunately, the loss of peaks in the HSQC due to the increased amide exchange rate at high pH prevented the continued use of this condition for NMR studies as assignment was even more difficult with the lack of long stretches of sequential connectivity. Because the protein is known to form additional secondary structure elements when it binds substrate or product DNA, we rationalized that the addition of DNA to the sample could result in the ability to assign the protein at this high pH as secondary structure elements slow the rate of amide proton exchange. To that end, we titrated substrate and product DNA into hFEN1-336 samples and monitored changes in the spectrum using TROSY-HSQC. Simple addition of lyophilized aliquots of DNA (1:1.5 protein/DNA ratio) to the protein did result in changes in the HSQC spectra; however, the quality of the spectra were poor. Furthermore, precipitation of the sample was observed even at pH 8.5.
Because hFEN1-336 is a structure specific rather than sequence specific, we rationalized that the problem associated with poor HSQC quality of the complex could be due to non-specific DNA induced aggregation at the high concentrations used for NMR. To determine if the method of making the complex was the cause of the poor spectral quality and precipitation, we opted to prepare the DNA complex at pH 8.5 under dilute conditions and moderate salt concentrations with subsequent concentration and exchange into the appropriate buffer. Doing this resulted in far better quality spectra and far less precipitation of the sample. Thus, we have devised a procedure for the preparation or hFEN1-336-DNA complexes for NMR studies. Despite the procedure, assignment of the complex was still difficult due to the lack of long sequential tracks due to the high exchange rate for the amide protons.
Thus, we returned to our attempts to stabilize the sample at pH 7.5 by dialyzing protein purified and concentrated at pH 8.5 into pH 7.5 buffer containing 5 mM MEGA-8 detergent. The addition of a detergent with a high critical micelle concentration was tried because the dynamic light scattering data reported in the midterm report hinted that the aggregation was due to a hydrophobic interaction. The spectra of the sample at pH 7.5 in the presence of 5 mM MEGA-8 resulted in the re-appearance of peaks lost at high pH. Furthermore, the sample did not flocculate as was seen previously as long as it was in the presence of MEGA-8 detergent. Despite solving this problem, we uncovered another problem in that some of the resonances in the HSQC spectrum under these conditions were doubled. In addition, some new peaks were observed suggesting a second conformer. Re-analysis of the assignment data from last year showed that some of the same resonances were doubled in assignment spectra and that the degree of doubling correlated with the age of the sample. As peak doubling for some resonances is usually seen at dimer interfaces, we hypothesized that the doubled peaks could be due to artefactual dimer formation at the high hFEN1-336 concentrations used for NMR.
In support of this, one recent structure of hFEN1 in complex with substrate DNA showed that the complex crystallized as a dimer. To test whether dimer formation was the cause of the peak doubling, we prepared a sample of hFEN1-336 at one tenth the concentration of the original sample and compared HSQC spectra of the concentrated and diluted samples. To achieve the same signal to noise ratio in both samples, the length of time to collect the spectrum of the dilute sample was much longer. The spectrum of the dilute samples was identical to the concentrated sample, thereby showing that the localized peak doubling was not due to dimer formation.
As we already suspect that various portions of the protein are dynamics, it is possible that he peak doubling is simply a region of the protein in slow exchange. Although we did not expect the problems we have encountered, we have solved some of them only to identify new problems. It is interesting that the peak doubling is not observed initially in the spectra and become increasingly more noticeable as the sample ages in the NMR tube. Initially, we did not notice the peak doubling as the sample precipitated throughout the course of experiments. Now that we have conditions that keep the protein in solutions, we would like to investigate further this peak doubling.
Such peak doubling phenomena as the sample ages has been observed before. In these cases, proline cis-trans isomerisation was the cause of localized peak doubling. Proline residues are unique in their ability to adopt both trans and cis conformer as all other amino acid residues in proteins exclusively adopt the trans conformer due to the fact that it is significantly more stable than cis conformer. Cis-Trans prolyl isomerisation has been shown to occur on a slow time scale of days and weeks due to the high activation necessary for the uncatalyzed reaction. Because all proline residues are initially in the trans conformation when translated, such isomerizations occur slowly over time. There are 22 proline residues in the hFEN1-336. Our current plan is to mutate these to determine if mutation to an alanine, which can only adopt the trans conformer, of any one proline results in the loss of peak doubling. Rather than randomly choose mutations, we have opted to mutate and prepare those proline residues that we expect from the assignment data to be the culprit. One tantalizing prospect is proline 188, which is C-terminal to a known phosphorylation site (i.e. serine 187).
The crystal structures of hFEN1-336 and hFEN1-354 show proline 188 in the trans conformation. In these structures, the serine 187 hydroxyl group is not surfaced exposed and is hydrogen bonded to a carboxylate of a sheet secondary structure. How such a residue can be phosphorylated without destroying the structure of the protein is puzzling. However, it is possible that cis-trans prolyl isomerization would expose the residue as movement about the serine 187-Prolien 188 C-N bond could result in exposure of the residue. Thus, we have begun our site directed mutagenesis on this residue. Due to the preliminary results obtained through the Marie Curie funding, we have recently secured funding to continue the project.
In addition to the work involving the NMR, Dr Finger's presence during the past two years has also resulted in the development of a new circular dichroism assay hFEN1-induced DNA unpairing and further exploration of several hFEN1 mutants to explore the mechanism of hFEN1 reaction.