Structure Comparison
The primary goal of this project is to correlate design parameters of the different DNA origami nanostructures with their stability in certain environments, which could lead to the prediction of performance of future designs and help incorporate favorable qualities. Using the favorable qualities found, a new structure will be designed in cadnano that has an overall increased stability. To do this, the team decided to focus on four different parameters: scaffold routing, segment length, structure surface area, and structure cross section. The segment length, cross section, and surface areas affect the thermodynamic, magnesium cation and serum stability of each structure. Table R.1, below, shows the differences between structure segment length, structure surface area, and structure cross sections.
Table R.1: Structure Parameters
Figure R.1, below, compares the scaffold routing of the 18 helix bundle (hb) symmetrical, 18hb seamless, 18hb with 42 base pair (bp) crossovers, and the 18hb with nucleators. A(i) is the scaffold routing of the 18hb symmetrical and A(ii) is the scaffold routing of the 18hb seamless. The 18hb seamless is missing crossovers between the junctions in the scaffold routing. In A(iii) the 18hb with 42 bp crossovers, the scaffold routing contains crossovers every 42 bps, where the 18hb symmetrical contains crossovers every 21 bps. In A(iv), the 18hb with nucleators contains nuclease attachment sites in the routing. B(i) is the staple routing of the 18hb symmetrical, B(ii) is the staple routing for the 18hb with 42 bp crossovers, and B(iii) is the staple routing for the 18hb with nucleators.
Figure R.1: Parameter Evaluation
The thermodynamic, magnesium cation, and fetal bovine serum (FBS) stability were characterized for each of these structure design parameters as well and summarized below.
Thermodynamics and Kinetics of Structure Folding and Stability
Testing was performed to analyze the temperatures and time required for each structure to fold. For temperature analysis, a 60℃-40℃ rapid fold (refer to Methodology) was performed to determine the best temperature for each structure to fold (figure R.2A). In order to test thermodynamic stability, the folded structures were also tested using a quantitative Polymerase Chain Reaction (qPCR) machine to plot melting temperatures (figure R.2C). The structures were also rapid folded at their preferred annealing temperature at varying time points ranging from 1 minute to 1 hour in order to determine how kinetics affects the fold process (figure R.2B).
Figure R.2: (A)Electrophoresis gels displaying results of rapid folding the Flat 18hb (blue border) and the Symmetric 18hb (red border) at different temperatures. Both structures were held at the annealing temperature for 4 hrs. The temperatures are (from left to right): 60.0℃, 58.8℃, 56.5℃, 52.6℃, 47.7℃, 43.9℃, 41.2℃, 40.0℃. (B)Electrophoresis gels displaying results of rapid folding the Flat 18hb (blue border) and the Symmetric 18hb (red border) for different durations. The Flat 18hb was held at 47℃, and the Symmetric 18hb was held at 52℃. The times are (from left to right): 1min, 3min, 5min, 10min, 20min, 30min, 60min, 120min. (C)Fluorometer melt curves for the Flat 18hb and Symmetric 18hb using SYBR green as the fluorescent dye. (D)Table summarizing the maximum annealing temperature, minimum annealing time, and melting temperatures for each structure.
As seen in the summary data in Figure R.2D, the 18hb with 42 bp crossovers has the highest folding temperature as well as the quickest folding time. The structure has twice as many consecutively hybridized base pairs before a staple crossover and these longer staple sections potentially provide increased stability which could cause the higher folding temperature; this will be discussed further in the Thermodynamic Stability MATLAB section below. Further clarification of the differences between designs can be found in the Methodologypage, and full analysis of the individual structures can be found in the Supplemental page.The 18hb with nucleators also folded at a higher temperature than the other structures with the same scaffold lattice. The nucleating strands provides sites for folding to begin and the initial staple binding allows the structure to fold at a higher temperature. The Flat 18hb folded at a much lower temperature than the other 18hb structures and also required a significant amount of annealing time. It is therefore concluded that the shape of this structure is not ideal for increased stability. The rest of the 18hb structures that are designed with a honeycomb lattice fold between 52.6℃ and 56.5℃ and between 5 and 20 minutes.
Even though all of the 6hb structures begin to fold at 56.5℃, the differences are conveyed from the widely varying annealing times. The 6hb normal and 6hb normal v2 both folded within 10 minutes whereas the seamless and 6hb with overhangs require two hours to fold. A possible hypothesis for the time difference may be due to the missing seam in the 6hb seamless. Consequently, more time may be needed for the staples to bind along the path of the structure. Currently, it is unknown why the 6hb with overhangs requires more time and understanding why is one of the future goals of the team.
The melting temperatures of the structures all appeared similar between 60.5 and 71.0℃. The 18hb with 42 bp crossovers has the highest melting temperature along with the highest folding temperature and is the most stable structure tested from a thermodynamics standpoint. The structure with the lowest melting temperature was the 18hb with nucleators. One potential hypothesis is that the nucleators themselves are melting and, because of their importance to folding, the rest of the structure falls apart around them. All of the different structures melt at a higher temperature than the temperature they folded at initially, which is likely explained by the cooperativity in the folding process.
Thermodynamic MATLAB Simulations
In addition to performing experiments on the structures, MATLAB simulations were run to predict the thermodynamic stability of each structure based on their cadnano designs. The program calculates the melting temperatures of consecutive staple segments to determine when those segments will begin to bind during the rapid fold process. The MATLAB code then graphs the proportion of bases bound to the scaffold as a function of temperature (figure R.3A), the location of segments that have melting temperatures above the annealing temperature (figure R.3B), the length of each staple, and the melting temperature of each staple (supplemental figures).
Figure R.3: Structure design analysis using MATLAB. (A) Graph of % structure folded vs temperature. MATLAB calculates the annealing temperature of each continuous staple segment and plots of the cumulative proportion of annealed staple segments as a function of the annealing temperature. (B) Plot of staple segments along scaffold routing. Each staple segment is mapped to its corresponding section of the scaffold, and segments are divided into two groups based on their melting temperature. (C) Theoretical percent of structure bound at the empirically determined annealing temperature for each structure tested.
The outputs show that the 18hb with 42 bp crossovers and Square 18hb structures both have very high staple segment folding temperatures. The longer staple segments in the 18hb with 42 bp crossovers is the likely cause of the increased thermodynamic stability and we take advantage of this characteristic when designing a new structure in cadnano. Overall the proportion of bases required to fold the structure varies dramatically across the structures. The Seamless 18hb has only 1% of its staple segments bound when the structure is folded while the Square 18hb has over 20% of its segments bound before it folds. The one unknown is the 18hb with 42 bp crossovers as it is possible that the structure can fold at temperatures above 60℃. The pattern observed is that the 6hb structures require a low amount of segments to be bound (2-4%) while the 18hb structures, with the exception of the seamless, were all at 10% or higher. The 18hb seamless is able to fold at the same temperature as the 18hb symmetrical while having ten times less segments bound to the structure. The hypothesis is that once an initial segment is bound to the structure, cooperative binding is able to take place along the entire length of the structure, decreasing the looping energy. This idea is supported by the 6hb structures due to their longer scaffold sections, making cooperative binding more likely to occur. More testing is still needed, however, to confirm this theory. Removing the seam does appear to make the structure fold at temperatures higher than it would otherwise, therefore the new structure was designed to be seamless.
Animation Model R.1: Magnesium is chelated by EDTA causing the dsDNA to dissociate.
Magnesium Stability
Structure stability was tested in varying magnesium concentrations over a 24 hour period. Initially, blue loading dye containing EDTA was used when performing agarose gel electrophoresis to test the effects of varying salt concentrations (figure R.4A). However, the EDTA present in the dye chelated Mg2+ present in the buffer, causing structures to fall apart when MgCl2 concentration was below 10 mM as depicted in Animation Model R.1. It was expected that any structures containing 10 mM MgCl2 or less, would fall apart. However, it was observed that the 18hb with 42 bp crossovers and the Square 18hb both remained folded well below 10 mM MgCl2. The square lattice structure appears to provide increased stability based on preliminary results. A square lattice was taken into consideration when designing the new structure.
Figure R.4: MgCl2 Stability template: (A)Electrophoresis gels displaying results of rapid folding the Symmetrical 18 helix bundle (blue border) and the Flat 18 helix bundle (red border) at decreasing concentrations of MgCl2. The concentrations varied as follows: 20, 15, 10, 8, 5, 3, 1, 0 mM MgCl2. (B) The gel electrophoresis graphs were then run through a gel analysis code in MATLAB and the results were plotted in this graph. (C) Table summarizing the initial and complete structure breakdown in MgCl2 for each structure.
Figure R.5: MgCl2 Stability Figure R.5 shows the denaturation of the symmetrical 18 helix bundle in 3mM MgCl2 over time. This data was collected on a spectrophotometer and the initial absorbance reading was subtracted from each subsequent dataset to only represent the change from the initial state.
The following table shows the results for the rest of the panel of structures:
Table R.2: MgCl2 Stability Results
Due to the results of the blue loading dye gels, an EDTA-free dye was then used to eliminate chelating effects. The experiments were then repeated with the use of the EDTA-free dye and it was observed that all of the structures were stable in lower concentrations of MgCl2. Every structure was seen to be stable between 1-3 mM of MgCl2 solution, breaking down at lower concentrations. In order to observe the effects of lower MgCl2 concentrations, three structures - 18hb seamless, Flat 18hb, and the 18hb with 42 bp crossovers - were tested again with a finer concentration gradient between 0 - 3 mM MgCl2, showing little variability between the structures MgCl2 stability.
Figure R.6: MgCl2 Stability
Figure R.6 displays the gel intensity analysis MATLAB graph of the Flat 18hb in varying MgCl2. This gel was made using the EDTA-free dye. The gradient is as follows: 3, 2.5, 2, 1.5, 1, 0.5, 0 mM MgCl2.
FBS Stability
Each structure was tested in different concentrations of FBS to determine the stability of each structure in a physiological-like conditions. Figure R.7 below displays a template characterization and analysis of the panel of structures for the FBS stability experiments after 24 hours incubation.
Figure R.7: MgCl2 Stability template (A) Electrophoresis gels displaying results of rapid folding the Symmetrical 18 helix bundle (blue border) and the Flat 18 helix bundle (red border) at decreasing concentrations of FBS. The concentrations varied as follows: 100, 75, 50, 20, 10, 5, 1, 0% FBS. (B) The gel electrophoresis graphs were then run through a gel analysis code in MATLAB and the results were plotted in this graph. (C) Table summarizing the initial and complete structure breakdown in MgCl2 for each structure.
The following table shows the results for the rest of the panel of structures:
Table R.3: FBS Stability Results
Most of the structures fell apart between 50-75% FBS concentration, whereas a few structures were stable in 100% FBS. Due to the nature of nuclease breakdown in DNA origami, surface area was hypothesized to play a role in the stability of FBS. Four structures - the Flat 18hb, Square 18hb, Symmetrical 18hb, and Seamless 18hb - were compared to examine the correlation between surface area and stability in FBS solution. The surface areas of these particular structures are shown in Table 5, below:
Table R.4: Structure Surface Area
From Table R.4, the Flat 18hb has the largest surface area while the Square 18hb has the smallest surface area. The Flat 18hb initially breaks down in 50% FBS, where the Square 18hb initially breaks down right below 75% FBS solution, confirming the effect of surface area on FBS stability. The agarose gel images of each structure were then analyzed, using a MATLAB code, to quantify where structure breaks down in FBS solution.
Figure R.8: FBS Stability results
Figure R.8 and Table R.3 show that the Flat 18 structure breaks down under the least percentage of FBS while the Square 18hb structure was the most stable. Based on this and the surface area of the different structures, we concluded that FBS stability and structure surface area are likely to be inversely related. An increase in FBS stability is observed with smaller or decreased surface area, thus the newly designed structure’s accessible surface area is relatively small.
Animation Model R.2:Nuclease activity cleaves phosphate backbone of DNA
FBS Kinetics
The DNA structures were incubated at room temperature in 100% FBS for different time increments ranging from 1 to 24 hours. The intention was to see the effect of time on highly concentrated solution containing nucleases, the breakdown of each structure at a specific time frame was observed.
Figure R.9: 75% FBS Stability
Figure R.9 displays data collected on a spectrophotometer. The absorbance signal increases for the Symmetrical (orange) and the Flat (blue) 18hb in 75% FBS over time, matching the gel band shift in the FBS kinetics data.
The following table shows the results for the rest of the panel of structures:
Table R.5: 100% FBS Time Kinetics
Unfortunately, the results were inconclusive, with a structure gel band shift after just an hour in most cases. The team also repeated the experiment using a spectrophotometer, with 75% FBS as the buffer, and showed promising results as shown in figure R.9. The expected mechanism of nucleases in the FBS is that it cleaves the phosphate backbone of the DNA strands as seen in Animation Model R.2. However, in both the gel and spectrophotometer results we found that there is unfolding of DNA nanostructures prior to dsDNA breakdown. A new method for this particular test is needed in order to perform conclusive experiments to gain insight on the breakdown of structures in 100% FBS versus time. Potential options include adding salt to the solution, or analyzing the structural breakdown by Atomic Force Microscopy (AFM) or Transmission Electron Microscopy (TEM). As shown in figure R.10, even when there is a gel shift viable structures can still be easily accounted for using TEM, and even after 24 hours there is little to no visible difference in the nanostructures. Interestingly, even after PEG purification and gel purification, proteins are visible on the TEM images. This could make the contrast in images limited for analysis using this method, but it is still a feasible option.
