Structural Information from Single-molecule FRET Experiments Using the Fast Nano-positioning System

The overall goal of the Fast Nano-positioning system is to provide real-time structural information of biomolecules by combining single-molecule FRET experiments with rigorous statistical analysis. This method answers key questions in the field of structural biology such as the localization of flexible domains in macromolecules where standard structure biology tools cannot be applied. The major improvement of fast NPS compared to other FRET-based structural tools is that different dye models can be compared using not only the most likely positions, but a three-dimensional probability distribution.

The main advantage of this technique is that it combines structural data from the protein database with quantitative single-molecule fluorescence measurements. To begin the procedure, assemble a flow chamber and sample holder. To prepare the inlet and outlet screws, thread silicone tubing into hollow tab screws and cut the tubing straight on either end with a razor blade.

Screw the tab screws to the sample holder. Align quartz slide holes of the flow chamber with the sample holder threads. Gently tighten the glass holder screws to fix the flow chamber in place.

Thread 20-centimeter long pieces of silicone tubing into the inlet and outlet screws. Wash the flow chamber with 500 microliters of PBS. Then flush the flow chamber with 100 microliters of neutravidin solution in PBS, ensuring that a droplet forms at the end of the inlet tubing to prevent air from entering the sample chamber.

Clamp the inlet and outlet tubing shut after every manipulation to exclude air from the flow chamber. Incubate the chamber for 15 minutes at room temperature. Then, wash out the neutravidin solution with 500 microliters of PBS.

Add a drop of immersion oil on the prism, and screw the prism onto the sample chamber. The single-molecule FRET data is acquired using a total internal reflection fluorescence microscope. Start the camera software and the stage piezo-motor software.

Mount the chamber to the micrometer stage of the TIRF-microscope horizontally, as straight as possible, and in such a way that it intersects all three lasers. Add immersion liquid, and focus the microscope objective by checking the IR reflections. Set the camera parameters, the file path, and the auto-increment parameters.

Start the live camera feed and bleach background fluorescence using all three lasers with full intensities. Then, turn down the laser intensities. Then, load 100 microliters of a fluorescent bead solution into the chamber.

Wait 10 minutes for the beads to bind to the chamber surface. Record a movie with a field of view of 50 to 100 beads, and perform a batch analysis with a high peak finding threshold. Load the movie file, and choose two beads with distinct centers at opposite corners of the field of view.

Select the pixels of maximum intensity in those beads. Remove and clean the prism. Prepare a second sample chamber, and screw the prism holder onto the chamber.

Mount the chamber to the micrometer stage. Then, load 100 microliters of a 50 to 100 picomolar solution of biotinylated fluorescent-labeled sample into the chamber. Wait for the molecules to bind to the chamber surface.

Simultaneously click Take Signal to start recording and turn on the donor excitation laser. Adjust the laser power to ensure that at least 80%of the molecules in the field of view are bleached, Then, move the chamber to show an unbleached area, and record the next movie. Bleach the rest of the sample chamber in this way.

After the measurements at the TIRF-microscope, the acquired single-molecule FRET data needs to be analyzed to yield the mean FRET efficiency. Perform a batch analysis of the donor and acceptor movies of the sample. Then, load the batched movie files in the data analysis software.

Click on the button NOT Selected to indicate the individual FRET phases. Click on the beginning of the FRET period, then the time point when acceptor intensity decreases due to bleaching, and finally the time point when the donor intensity decreases due to bleaching. In the next window, select Yes to keep the selected FRET efficiency trace.

Once every molecule has been analyzed, save the traces and analyze the next movie. Once analysis is complete, run a script to combine the FRET results. Import the combined frame-wise FRET file into data analysis software using the default input option.

Plot the data as a histogram, and use the non-linear curve fitting tool to fit to a Gaussian. The exceed value is the main FRET efficiency. Use UV VIS and fluorescent spectroscopy to determine the isotropic Forster radius, and the steady-state anisotropy values for each dye.

The nano-positioning system was inspired by the well-known GPS, as we localize an unknown position, the antenna, with several known positions, the satellites. To begin the NPS analysis, launch the FastNPS software and create a new jobfile. To define the position priors of the labeled antennas and satellites, in the ModelDyePrior window, enter the spatial resolution of the position being defined.

Exclude the macromolecule interior by loading the corresponding centered PDB file. Enter the dye diameter and the skeletonization distance. Fill in the minimum and maximum position coordinates of the possible dye location.

If the position being defined is a satellite, select dye attachment via flexible linker, and enter the atom coordinates from the centered PDB file. Fill in the length and diameter of the linker. Click calculate accessible volume, save the position prior, and export it for visualization.

Repeat this process for all satellites and antennas. Once the positions have been defined, open the Define Measurement window, and create a new dye molecule. Load the position prior associated with that dye.

Fill in the fluorescence anisotropy, select an appropriate dye model, and activate the dye. Repeat this process until all of the dyes have been defined. Then, create a new measurement.

Select a FRET pair from the defined dyes. Enter the single-molecule FRET efficiency with error, and the isotropic Forster radius of this dye pair, and activate the measurement. Repeat this for all dye pairs.

Next, open the calculation window. If dyes have been assigned to different models, select User Defined. Otherwise, select the model to use for all dyes.

Perform the calculation, and open the results window. When using FastNPS, it is important to remember to adapt the models for all dyes in order to minimize the credible volumes, while of course staying consistent. If the consistency is lower than 90%click detailed consistency, and identify the dyes common to entries with less than 90%consistency.

Open the Define Measurement window, and change the model of the dye in question. Rerun the calculation in the mode User Defined. Export the credible volumes of the dyes singly, or as a batch.

Open the density files in the visualization software, and adjust credibility to the desired level. Single-molecule FRET efficiencies were measured between unknown antenna molecules and several known satellite molecules within an archaeal RNA polymerase open promotor complex. NPS analysis of this data determined the credible volumes of these dye molecules relative to the archaeal RNA polymerase open promotor complex.

Five different models where the dye molecules were used, and the results were compared. The credible volumes were consistent with the measured data when calculated with models that assumed that the dye can only rotate freely within a cone of unknown orientation. The classic model in particular is the most conservative, as it further assumes that the dye is in a fixed location.

This creates a relatively large credible volume, but the volume generally encloses the correct position. Higher precision models use the assumption that the dye is completely free to rotate within its fluorescence lifetime. The iso model, like the classic model, assumes that the dye is at a fixed position within the accessible volume.

The meanpos-iso model assumes dynamic averaging over all possible positions. While the credible volumes of the iso models are much smaller than those calculated by the non-iso models they are inconsistent with the measured data. We wanted to develop a technique that allows us to capture dynamic structural information of transient complexes, a problem that could not be addressed by existing structural tools.

Once the single-molecule FRET data has been recorded, the FastNPS analysis is done in a couple of hours, allowing for a quick comparison of different dye models. Following this procedure, other methods like MD simulations can be performed subsequently in order to obtain structured models that are consistent with experimental single-molecule FRET data.