High Precision FRET at Single-molecule Level for Biomolecule Structure Determination

1. PBS Buffer Preparation and Chamber Treatment

NOTE: Wear a laboratory coat and disposable gloves when performing wet chemical experiments. Use eye protection when aligning the laser.

1. PBS buffer preparation
   1. Dissolve 4.5 g of Na₂HPO₄, 0.44 g of NaH₂PO₄, and 3.5 g of NaCl in 400 mL of distilled water. Ensure a pH of 7.5 and sterilize the solution by autoclaving on a liquid cycle for 1 h (depending on the autoclave system).
   2. Take 15 mL of the PBS solution and mix it with 0.1 g of charcoal. Filter the mix by using a regular 20 mL syringe filter with a 0.2 µm pore size. Seal and store the PBS buffer at room temperature.
2. Microscope chambered cover glass treatment
   1. Add 500 µL of distilled water and 5 µL of polysorbate 20 nonionic surfactant (see the Materials List) to a chambered cover glass system (see the Materials List) and mix well. Let it soak for 30 min. Remove the polysorbate 20 solution and wash the chamber with distilled water twice. Let it dry.
      NOTE: The chamber is now ready to use.

2. DNA Sample Preparation

NOTE: Use designed labeled DNA strands (see the Materials List) for the creation of double-stranded DNA (dsDNA) standard samples. Designed oligos must not have dyes at the end of a polymer in order to avoid artifacts that can compromise the determined distance. The DNA sequence should be chosen to behave as a rigid body.

1. Add 1.5 µL of a donor labeled DNA strand and 4.5 µL of a complimentary (acceptor-labeled or non-labeled) DNA strand into a microfuge tube and mix them with 24 µL of nuclease-free water.
   NOTE: Depending on the mix of the selected oligonucleotides, the following samples will be generated: no-FRET, low-FRET, or high-FRET dsDNAs.
2. Hybridize the DNA using a thermal mixer and the following process: 95 °C for 10 min, 90 °C for 10 min, 80 °C for 10 min, 70 °C for 10 min, 60 °C for 10 min, 50 °C for 10 min, 40 °C for 10 min, 30 °C for 10 min, 20 °C for 10 min, 10 °C for 10 min, and holding at 5 °C.
   NOTE: The generated dsDNA standards can be kept in a -20 °C freezer for long-term storage or can be used immediately.

3. Protein Sample Preparation

Note: Starting with recombinant DNA for the expression of the protein of interest in bacterial systems, it is possible to mutate the residues from which the distances are to be measured into cysteines. To do so, use standard site-directed mutagenesis techniques⁴⁷. To facilitate protein purification, clone the recombinant and mutated DNA into a vector containing a purification tag (e.g., a His-tag). The glutamate subunit 1 ligand-binding domain (LBD) from the NMDA glutamate ionotropic receptor (GluN1) LBD (i.e., NMDA GluN1 LBD cloned into the pET-22b (+) vector) was used.

1. Protein expression
   1. Transform construct DNA plasmid into the expression system of choice⁴⁸.
      NOTE: The following steps will assume that the expression of a soluble protein is transformed in Escherichia coli. Purification from, for example, transfected mammalian cells⁴⁹ or transduced insect cells⁵⁰, is also possible, and detailed steps can be found elsewhere. Ensure that the E. coli strain selected is appropriate for the protein of interest. For example, the expression of proteins that contain disulfide bridges requires a strain of competent cells with a less-reducing intracellular compartment (see the Materials List).
   2. Inoculate a starter E. coli culture by using a sterile pipette tip to pick up a single transformed colony⁴⁸. Drop it into 100 mL of selective LB medium (see the Materials List) and allow the culture to grow overnight at 37 °C.
   3. Prepare LB broth (see the Materials List). Autoclave it to sterilize.
   4. Inoculate large-scale cultures of transformed E. coli by adding the overnight culture to 2 L of selective LB medium at a 1:500 ratio.
   5. Over the next few hours, assess the growth of the culture by monitoring the absorbance readings (Abs) of the culture at 600 nm, sometimes referred to as the optical density at 600 nm (OD₆₀₀), or by using a cell density meter. Note that the readings increase over time.
   6. Induce protein expression with a final concentration of 0.5 mM isopropyl-β-d-1-thiogalactopyranoside (IPTG) when the culture reaches an OD₆₀₀ of 0.7. Shake induced E. coli at 20 °C for 20-24 h.
   7. After protein induction, pellet the E. coli by spinning for 20 min at 3,000 x g and 4 °C. Discard the supernatant and store the E. coli pellet containing the intracellular protein at -80 °C until use.
2. Protein purification
   1. Lyse E. coli using a lysis method of choice (e.g., sonication, French press, nitrogen cavitation, etc.)⁵¹.
   2. Spin down the membrane and cell debris by centrifuging the lysate for 1 h at 185,000 x g and 4 °C.
   3. For a His-tagged protein, load the supernatant onto an equilibrated, nickel-charged, immobilized metal affinity chromatography (IMAC) column using a fast protein liquid chromatography (FPLC) system (see the Materials Table)⁵².
      NOTE: Equilibration buffer for the NMDA GluN1 LBD: 200 mM NaCl, 20 mM Tris, and 1 mM Glycine, pH 8. Elution buffer for NMDA GluN1 LBD: 200 mM NaCl, 20 mM Tris, 1 mM Glycine, and 400 mM Imidazole, pH 8.
      1. Wash the IMAC column with buffer containing a low amount (~12 mM) of imidazole.
      2. Elute the protein from the IMAC column using a linear gradient of imidazole from 12 mM to 400 mM.
   4. Dialyze the protein overnight in equilibration buffer without imidazole⁵³ by placing the eluate from step 3.2.3.2 into dialysis tubing and submerging it in the equilibration buffer under continuous stirring for 2-3 h. Repeat at least one more time.
      NOTE: Steps 3.2.3-3.2.6 assume the purification of a His-tagged protein. If purifying through some other method, adjust the protocol accordingly.
   5. Quantify the protein amount by taking the absorbance at 280 nm of the dialyzed protein and using Beer's Law (Absorbance unit = ε L c, where ε is the extinction coefficient (M^-1cm^-1), which can be obtained here⁵⁴; L is the light path length (cm); and c is the protein concentration (M)).
      NOTE: Various protein quantification assays are available, including the Bradford assay and bicinchoninic acid assay. Both provide accurate results.
3. Protein labeling
   1. Add donor (maleimide reactive cyan-green dye) and acceptor (maleimide reactive far-red dye) fluorophores to the purified protein at a 1:1:8 protein:donor:acceptor molar ratio.
   2. Incubate the protein and fluorophore mixture on ice for 30 min. Longer incubation times are possible.
   3. Pack a 0.5 mL Ni-Nitrilotriacetic acid (Ni-NTA) agarose column (see the Materials List) and equilibrate it using the same equilibration buffer as in step 3.2.3 while the protein is incubating.
      NOTE: Keep the amount of loaded protein in accordance with the resin binding capacity.
   4. After the 30 min incubation, load the protein/fluorophore mixture onto the column prepared in step 3.3.3 and purify by gravity flow.
   5. Wash off excess fluorophore with 5 mL of equilibration buffer.
   6. Elute the labeled protein by gravity from the column four times with 0.5 mL of elution buffer. Because the protein has already been purified from other proteins, no gradient is necessary.
   7. Check each eluate with a UV-Vis spectrometer to identify which fraction contains the labeled protein. Scan the absorbance from 230-700 nm to be able to ensure that the absorbance peaks from the protein (280 nm) and each fluorophore (493 nm for the cyan-green fluorophore and 651 nm for the far-red fluorophore) are visible in the eluate.
      NOTE: Typically, the protein will elute in fraction 2.
   8. Equilibrate a desalting column (see the Materials Table) with charcoal-treated PBS (step 1.1).
   9. Load the labeled protein onto the desalting column by gravity flow.
      NOTE: Keep the amount of loaded protein in accordance with the selected desalting column capacity⁵⁵.
   10. Elute using 3.5 mL of charcoal-treated PBS by gravity flow and collect 0.5 mL fractions of the eluate.
   11. Use a UV-Vis spectrometer to scan the absorbance of each eluate from 230-700 nm to identify which fraction contains labeled protein.
       NOTE: Steps 3.3.8-3.3.11 basically serve as a buffer exchange step. Other protocols that serve this purpose are also possible (e.g., extensive dialysis). Alternatively, one could go straight to step 3.3.8 from step 3.3.2.

4. Measurements Needed in Ensemble Conditions (in Cuvette)

1. Determination of the Förster constant
   1. Scan f_D, the fluorophore fluorescence emission (cps), in a fluorimeter by exciting the donor at 15 nm to its maximum absorbance wavelength in order to get the full emission spectrum. Monitor the emission beginning 5 nm after the excitation wavelength and ending 150 nm later. Use magic angle conditions by setting the emission polarizer to 54.7° and the excitation polarizers to 0°⁵⁶.
      NOTE: For the donor fluorophore used here, the Abs maximum occurs at 490 nm; 475 nm is used as the excitation wavelength, and the emission from 480-650 nm is monitored. For the acceptor, the Abs maximum occurs at 645 nm; 630 nm is used for the excitation wavelength, and the emission from 635-735 nm is monitored.
   2. Use the fluorimeter to perform an excitation scan of the acceptor fluorophore (Abs_A) ranging from 400-700 nm and use magic angle conditions by setting the emission polarizer to 54.7° and the excitation polarizers to 0°⁵⁶. Set the emission monochromator to 15 nm after the maximum-emission wavelength. Normalize to the maximum excitation value.
   3. On published tables, locate the extinction coefficient of the acceptor, ε_A (M^-1 cm^-1)⁵⁶, or use values provided by the manufacturer.
      NOTE: The value published for the acceptor fluorophore used in this manuscript is ε_A647 = 270,000 cm^-1 M^{-1 56}.
   4. Calculate the spectral overlap using J = Σf_D · (Abs_A · ε_A) · λ⁴, where f_D, Abs_A, and ε_A have been defined above λ and is the wavelength (nm). Use a worksheet to list in columns all the wavelength-dependent values obtained in steps 4.1.1-4.1.3. Align them according to wavelength. Perform the summation from the minimum wavelength of the donor emission (λ_min) to the maximum wavelength of the acceptor absorbance (λ_max).
   5. Calculate the Förster constant (R_o) using the following formula R_o⁶ = 8.79 x 10^-5 · J · κ² · Φ_F,D · n^-4, where J is the spectral overlap previously calculated in step 4.1.4, κ² is the orientation factor, Φ_F,D is the fluorescence quantum yield of the donor fluorophore, and n is the refractive index of the medium in which the fluorophore is situated.
      NOTE: Use n = 1.33 (if aqueous buffer is used) and κ² = 2/3.
   6. Locate the quantum yield value of the donor fluorophore (Φ_F,D, environment dependent) on published tables (see Reference 57) and use the value of the spectral overlap obtained in step 4.1.4 to calculate the final value of the Förster constant using the equation from step 4.1.5.
      NOTE: If the quantum yield is not available, follow step 4.2, below, to calculate it. In this case, use Φ_F,D = 0.8, which corresponds to a donor lifetime τ_D,r = 4.0 ns.
2. Determination of fluorescence quantum yield
   NOTE: The following procedure assumes only dynamic quenching. To consider static quenching, refer to Reference 56. However, PIE-MFD experiments are also useful in determining the quantum yield, even in the case of static quenching (see the Results).
   1. Select a reference fluorophore with similar absorbance and emission profiles for both the acceptor and the donor fluorophores for which the quantum yield (Φ_r) has been determined.
      NOTE: For the donor, Φ_r = 0.8 and τ_r = 4 ns, while for the acceptor, Φ_r = 0.32 and τ_r = 1.17 ns, which correspond to the Φ_r and τ_r for the cyan-green fluorophore- and the far-red fluorophore-labeled oligonucleotides, respectively⁵⁷.
   2. Measure the time-resolved fluorescence decay (f(t)) using the time-correlated single-photon counting (TCSPC) method at magic-angle conditions.
   3. Fit the fluorescence decay with a mono- or multi-exponential decay function in the form of f(t) = Σ_ix_ie^-t/τ_i, where x_i is the population fraction and τ_i is the population fluorescence lifetime.
   4. Calculate the species average lifetimes, 〈τ〉_x = Σx_iτ_i, where x_i is the population fraction and τ_i is the population fluorescence lifetime.
   5. Use the formula Φ_F,D = 〈τ_D〉_x * Φ_r / τ_r to calculate the fluorescence quantum yield of the donor fluorophore by plugging in the fluorescence lifetime and quantum yield of the reference, as well as the fluorescence lifetime of the donor fluorophore.
      NOTE: This method assumes dynamic quenching. For other Φ_F,D determinations, follow Lakowicz⁵⁶.

5. Experiment Alignment for PIE-MFD Single-molecule Detection (SMD)

NOTE: It is better to turn off the lights when taking measurements.

1. Equipment adjustment (Figure 1)
   NOTE: A home-built MFD setup depicted in Figure 1, with two pulsed lasers and 4 detection channels in an inverted microscope body, is used for this experiment. There are similar commercial systems.
   1. Turn on the 485-nm and 640-nm lasers and all detectors of the MFD setup. Open the software that controls the TCSPC acquisition and lasers. Make sure that the laser repetition rate is 40 MHz.
   2. Set the 485-nm pulsed laser power to 60 µW at an image plane of the 60X 1.2 N.A. water-immersion objective and the 640-nm pulsed laser power to 23 µW in pulsed interleaved excitation mode (PIE-MFD)⁴².
      NOTE: To set PIE-MFD, the two laser pulses are delayed in the laser controller software. For 485-nm laser excitation, the detection TCSPC channels (TAC channels) are 1-12,499 ("prompt" channel). For 640 nm laser excitation, the detection TCSPC channels (TAC channels) are 12,499-50,000 ("delay" channel).
   3. Add objective immersion liquid (a drop of double-distilled water) between the microscope objective lens and a cover glass slide. To ensure that the image plane is inside the solution and far from the glass surface, turn the adjustment knob one and a half turns after finding the second bright focal point due to the reflection of the lasers at the glass-liquid interface.
   4. Add 1 µL of 100 nM Rhodamine 110 to 50 µL of distilled water to the center of the cover glass. Ensure that the solution is also at the center of the microscope objective.
   5. Adjust the pinhole (size: 70 µm) positions (x and y direction one at a time) while monitoring the photon count rate on the acquisition software to maximize the number of photons detected.
2. Standard measurement SMD (work in a dark room)
   1. Use the sample from step 5.1.4 and record 120 s of the count rate by clicking the "Start" button on the time-tagged time-resolved (TTTR) control panel in "*.ht3" format⁵⁸ on the acquisition software.
   2. Compute offline fluorescence correlation spectroscopy (FCS)^59,60,61 (i.e., FCS measurement) to determine the characteristic time of diffusion, the number of molecules in the confocal volume, the triplet state kinetics, and the molecular brightness⁶².
      1. Open the software for FCS (Kristine, MFD suite). Select the experimental settings by clicking "Options" -> "Select Set up." Select a file with similar experimental settings and click "get parameters from file" to read the header information on the file.
      2. Select "Operate" -> "Correlate" to perform FCS.
         NOTE: Make sure that the channel numbers are properly specified and that the "TAC Gate" (TCSPC channels) is checked to select accordingly the prompt or delay channels.
      3. Select "Operate" -> "Global Fit of Correlation Curves" to open the fit routine. Use "equation #24" on the software and click "start."
         NOTE: Equation #24 on the software describes the autocorrelation function (G_c) of freely diffusing fluorescent molecules over a three-dimensional Gaussian illumination profile, such as⁶¹:
         
         where N is the mean number of molecules in the detection volume, x_T is the fraction of molecules exerting triplet-state kinetics with the characteristic time t_T, t_c, is the correlation time, t_diff is the diffusion time related to the geometrical parameter ω, and ω describes the Gaussian illumination profile. After the fit, take note of the diffusion time and the number of molecules in the confocal volume.
          
   3. Add 10 µL of 100 nM Rhodamine 101 into 50 µL of distilled water and mix well. Place this mix on top of the cover glass and ensure that the droplet is at the center of the objective lens. Click the "start" button on the TTTR control panel to record 120 s of data in TTTR format.
   4. Add 1 µL of 100 nM far-red fluorophore into 50 µL of distilled water and mix well. Place this mix at the center of the objective lens. Click the "Start" button and record 120 s of data in TTTR format.
   5. Place 50 µL of distilled water at the center of the objective lens. Click the "Start" button and record 300 s of data in TTTR format.
   6. Place 50 µL of PBS buffer at the center of the objective lens. Click the "Start" button and record 300 s of data in TTTR format.
   7. Take 1 µL of the mix from step 5.1.4 and mix with 50 µL of distilled water. Place this mix on the cover glass. First, click the "Start" button and collect 10 s of data in TTTR mode. Then, analyze the TTTR mode file using the Burst Integration Fluorescence Lifetime (BIFL) analysis software (Paris, MFD suite), as described in step 5.3.
      NOTE: Verify the number of bursts per second from the burst selection and analysis software^26,61. If the burst level is around 35 every 10 s, it is appropriate for single-molecule measurement.
   8. Continue recording the count rate in TTTR format for 1.5 h (i.e., TCSPC at SMD) to treat as a single-molecule measurement standard.
      NOTE: Due to the large file size, split raw "*.ht3" files into smaller-size files to load and process using BIFL.
3. Analysis of standard samples using BIFL
   1. Open the BIFL software (Paris).
   2. Select the setup for PIE in the "confirm set up" automatic pop-up window and read the header by selecting the file with similar experimental settings. Click "get parameters from file." Click "OK." Note that the pop-up window closes and is integrated into the Paris front-end. Click "OK" under "Next."
   3. Choose the files to analyze by clicking "Select" on "Data Path Array" to select the measurement to analyze.
      1. Click on "Green scatter" (for a water measurement), "Green BG" (for a buffer measurement), "Green thick" (for a 2 nM Rhodamine 110 measurement), "Red scatter" (for a water measurement), "Red BG" (for a buffer or water measurement), "Red thick" (for a 20-nM Rhodamine 101 measurement), "Yellow scatter" (for a water measurement), "Yellow BG" (for a buffer measurement), and "Yellow Thick" (for a 2 nM far-red fluorophore measurement).
      2. Click "OK" under "Next."
         NOTE: The "Green" channel corresponds to the signal of the green detectors in the "prompt" TCSPC channels. The "Red" channel corresponds to the signal of the red detectors in the "prompt" TCSCP channels. The "Yellow" channel corresponds to the signal of the red detectors in the delay TCSPC channels.
   4. Click "Adjust" next to "Data cut Burstwise" to adjust single-molecule selection parameters. In the new pop up window, select single-molecule events with two standard deviations from the mean interphoton arrival time ("dt") by changing the interphoton arrival time under "Threshold" and the minimum number of photons per single molecule event under "min. #." Click "Return" to close the pop-up window. Click "OK" under "Next."
      NOTE: The threshold, in ms units, depends on the background count rate. The typical minimum number of photons used is 60.
   5. Adjust the initial fluorescence lifetime "Color fit parameters" (e.g., from, to, and convolution) for the generated fluorescence decay parameters on the "Green", "Red," and "Yellow" colors. In the same window, adjust the "from" and "to" values for "Prompt" and "Delay." Click "Return" to close the pop-up window. Click "OK" under "Next."
      NOTE: Ensure that the 2-color excitation check-box is selected. "From" and "to" correspond to the initial and end bins on the fluorescence decay histogram (TAC channel number). If the initial fit parameters are selected properly, a fit function is added to the fluorescence decays on each channel.
   6. Select the location on the hard drive to which to save all processed ascii files in a parent folder.
      NOTE: Paris processes all selected bursts and creates multiple ascii output files than can be used by other programs for visualization (e.g., Margarita MFD suite).

6. dsDNA Standards and Sample Measurements

1. Add 500 µL of PBS buffer to a chambered cover glass and place a drop of distilled water between the chamber and the objective lens. Click the "Start" button on the control pedal and collect 5 min of data in TTTR mode to use for analysis.
2. Take a small amount (usually around 0.1 µL, concentration of around 1 µM) of dsDNA standard, add it to the PBS buffer, and mix well. First, collect 10 s of data by clicking "Start." Then, check the burst to get 35 bursts per 10 s (as in steps 5.2.7 and 5.3). Finally, collect >2 h of data in TTTR format, as described above.
3. Analyze the collected data for the dsDNA samples, as in step 5.3.3.
4. Visualize the burst histograms using the MFD suite (Margarita) and display the FRET efficiency versus 〈τ_D(A)〉_f or the F_D/F_A versus 〈τ_D(A)〉_f.
   1. Open the Margarita software and select "File" -> "Import all *.??4 and *.mti files." Select the parent folder containing various subfolders.
   2. Select the parameters to visualize by clicking next to the "X" (abscissa) of one of the parameters derived from Paris (e.g., tau green or 〈τ_D(A)〉_f); similarly, repeat this for the ordinate "Y" to select the desired parameter to visualize (e.g., FRET efficiency, F_D/F_A, or S_PIE PIE).
      NOTE: In this case, FRET efficiency, F_D/F_A, or S_PIE correct for proper background count rate in the green, red, and yellow channels; for quantum yields of the donor and acceptor; for the detection efficiency ratio (g_G/g_R); and for crosstalk (α). Here, g_G/g_R = 3.7 and α = 0.017, depending only on the instrument. Background count rates depend on the buffer used, and quantum yield values are previously determined.
   3. Add a FRET line by opening the "Overlay Equation" window by clicking "Display" -> "Overlay Equation." Select the static FRET line from the pop-up menu. Select the proper donor lifetimes and quantum yield parameters to generate the proper FRET line.
      NOTE: FRET lines for correlating various FRET indicators can be generated.
5. Determine the correction factor for the acceptor excitation by the donor excitation source (β) using the stoichiometry parameter by displaying FRET efficiency versus stoichiometry (S_PIE) in Margarita (Equation 1, below).
   NOTE: β is chosen such that the donor sample has a peak at S_PIE = 1.0 in the stoichiometry scale; the acceptor-only sample should have a stoichiometry of S_PIE = 0.0, and the dsDNA with both labels should have a stoichiometry of S_PIE ~0.5.
   NOTE: Instrument is now ready, and it is possible to measure FRET-labeled samples.
6. Measure and analyze FRET-labeled samples prepared in section 3 by following steps 6.3-6.4.