Analysis of SEC-SAXS data via EFA deconvolution and Scatter

Biological small-angle x-ray scattering provides structural measurements of macromolecule and macromolecular complexes. Ideally, the sample to be measured should be monodispersed. Although in some cases, size exclusion chromatography SAXS is not sufficient to produce monodispersity, a software-based deconvolution of the SAXS data can be performed to produce an idealized SAXS curve.

Following this protocol, a deconvolution program and the user-friendly Scatter program can be used to analyze the Vaccinia virus DNA polymerase exominus mutants. To perform a background subtraction of the size exclusion chromatography data, open the Java-based Scatter 4 program and open the SEC tab. Drag and drop the reduced data files into the drop data below window and click the output directory button and open to set the output directory in which the data will be saved.

Enter the sample name in the save as box and click trace. Click on the edit details button to edit the experimental details and fill out the appropriate fields. To select the buffer frames manually, click set buffers.

Left-click and drag to reselect the buffer as a flat on the trace curve before the void volume of the size exclusion chromatography column of approximately 100 frames and click set buffer and update to recalculate the SEC file. Left-click and drag to select a region of interest on the signal plot and left-click and drag the cross-hairs in the heat map plot to select the zoom and subset of frames that will be used for merging. With another left-click, highlight the frames in the heat map corresponding to the area of the bottom right of the cross-hairs.

Ideally, the frame should highlight a region with a predominantly cyan color and a stable radius of gyration. When satisfied with the selected frames, click merge to merge the subtracted frames and click the analysis tab to view the data. For deconvolution of the data, load the dataset in the deconvolution program.

In the control panel under files, use the folder symbol to locate the data and highlight all of the DAT files. A plot of integrated intensity versus frame number will be drawn in the series plot. Under the series tab, click to highlight the curve and to open the LC analysis popup window.

To select a suitable buffer region, click add region to select an area before the peak of the chromatogram and one after the solvent front and click set buffer. Popup windows will indicate that frames have not been selected for the background. To start the EFA, right-click on the highlighted file and select EFA from the menu.

In the popup window, the single value decomposition of the dataset should be observed. In the controls box, check the use frames box values to confirm that the whole peak area to deconvolute is covered in the intensity plot. The singular values plot will show the intensity of the singular values above the baseline.

If the left and right single vectors do not match, change the significant singular vector number to two and move the frames until the left and right singular vectors are similar. The EFA will be calculated, generating plots in the forward and backward directions for each vector and indicating when the components start and exit the solution profile for the selected size exclusion chromatography SAXS data. RAW will attempt to identify the ranges.

If necessary, use the arrows to change the ranges so that each circle is the start of an inflection point, rising from or falling to baseline and click next. To reduce or eliminate spikes, identify approximately which frame corresponds to the spike using the range control arrows to adjust the component range controls. When a minimum Chi square has been achieved, click back to perform a validation check.

If the original EFA plots still look valid, click next and save EFA data to save the plots. Then click done to close the EFA window. Then in the RAW window, open profiles in the plot panel to view the curves.

In the profiles tab of the control panel, right-click to save the curves as DAT files. For SAXS determination, open the scatter analysis tab and select G to select the manual Guinier analysis tool. In the plot, add or remove points such that the residuals do not have a smile or frown feature.

The selected data in the Guinier fit should not exceed the maximum queue multiplied by radius of gyration limit of 1.3. Click normalized Kratky. The resulting plot provides an assessment of the structural state of the macromolecule, globular, cylindrical, disordered, normalized for mass, and concentration.

Click volume of correlation. The total scattered intensity and an integrated area of the total scattered intensity as a function of Q plots will appear as a quick reference for validating the quality of the scattering curve. To start the flexibility analysis, click flexibility.

Each panel in the popup window will show a plot exploiting a power law relationship that exists between the compact and the elongated flexible biopolymers. Use the slider at the bottom of the box to change the view of the data until a plateau in one of the plots is reached. Immediately after performing a flexibility analysis, click volume.

A popup of three graphs will be generated. The Porod-Debye plot tracks where the slider from the flexibility plot was left and shows the plateaued area data. To calculate the volume of the particle, move the start and end points until the blue line on the plot fits the plateaued region.

For an unbiased result, the residuals in the Porod-Debye exponent power law fit should show no pattern. Under the P of R tab, the real space distribution and the scattering curve for the sample can be observed. Ideally, the distribution curve should be smooth with no waves and should just gently touch the x-axis.

Right-click on the sample name and click find DMAX to open a new window. The limits for the maximum dimension are preset with the suggested maximum Q range, the maximal data points to be used for the calculation, the lower and upper maximum dimension limits, and a lower and upper alpha score. Click start.

A composite distribution will be created along with a suggested maximum dimension and alpha level. If these data are acceptable, close the window to return to the P of R tab where the reciprocal space plot will now be cropped to match the suggested maximum Q range. Select the more model and click background to set the alpha level and maximum dimension to the suggested values from the popup box.

Click refine. A cross-validation plot will open, showing whether any points had to be rejected as indicated in red. If there are only a few rejected points and the distribution looks good, then the model is good.

You'll print a report in the analysis tab. Left-click to highlight the sample and right-click on the sample name. Select create report from the single dataset.

A text box will open to allow comments and a PDF document will be produced showing all of the figures and values generated. In this representative analysis, E9 DNA polymerase exonuclease minus mutant was bound to DNA and run using size exclusion chromatography small-angle x-ray scattering. Two peaks were observed.

The first large peak represents the E9 DNA polymerase exonuclease minus mutant DNA complex. And the second indicates the unbound state. While the classic approach of selecting frames provides a stable radius of gyration of the complex in the first peak, the second peak is clearly merged and the radius of gyration across the plot shows that the second peak of interest does not have a stable radius of gyration due to cross-peak contamination.

In this analysis, only five frames could be used that showed a semi-stable radius of gyration. When subtracted, they gave a radius of gyration of 36.3 angstroms. When the peaks were deconvoluted using EFA, the corresponding curve for the second peak was overlaid with the original, showing a clear decrease in signal to the noise and lower radius of gyration of 34.1 angstroms.

A Kratky plot of the data reveals that the complex with the convoluted peak is more globular as confirmed by the PR curve which gives a maximum dimension of 108.5 angstroms for the deconvoluted curve. The non-deconvoluted data for this analysis is more elongated with a maximum dimension of 120 angstroms, most likely due to the heterogeneity arising from the unbound E9 polymerase minus exonuclease mutant. The most critical steps are in selecting the singular values and the range of data used as these greatly affect the accuracy of the deconvolution.

The results should not be taken on their own, but further assessed using additional techniques such as analytical centrifugation or multi-angle laser light scattering to allow their biological interpretation. In-line column-coupled SAXS in combination with deconvolution and a user-friendly interface like the program Scatter provides is a powerful package to provide meaningful structural data even from intrinsically difficult systems.