Photoactivated Localization Microscopy with Bimolecular Fluorescence Complementation (BiFC-PALM)

Podsumowanie

Protein-protein interactions are visualized in cells with nanometer spatial resolution by combining bimolecular fluorescence complementation (BiFC) with photoactivated localization microscopy (PALM). Described here is the use of BiFC-PALM for imaging Ras-Raf interactions in U2OS cells for visualizing the nanoscale clustering and diffusion of individual Ras-Raf complexes.

Streszczenie

Protein-protein interactions (PPIs) are key molecular events to biology. However, it remains a challenge to visualize PPIs with sufficient resolution and sensitivity in cells because the resolution of conventional light microscopy is diffraction-limited to ~250 nm. By combining bimolecular fluorescence complementation (BiFC) with photoactivated localization microscopy (PALM), PPIs can be visualized in cells with single molecule sensitivity and nanometer spatial resolution. BiFC is a commonly used technique for visualizing PPIs with fluorescence contrast, which involves splitting of a fluorescent protein into two non-fluorescent fragments. PALM is a recent superresolution microscopy technique for imaging biological samples at the nanometer and single molecule scales, which uses phototransformable fluorescent probes such as photoactivatable fluorescent proteins (PA-FPs). BiFC-PALM was demonstrated by splitting PAmCherry1, a PA-FP compatible with PALM, for its monomeric nature, good single molecule brightness, high contrast ratio, and utility for stoichiometry measurements. When split between amino acids 159 and 160, PAmCherry1 can be made into a BiFC probe that reconstitutes efficiently at 37 °C with high specificity to PPIs and low non-specific reconstitution. Ras-Raf interaction is used as an example to show how BiFC-PALM helps to probe interactions at the nanometer scale and with single molecule resolution. Their diffusion can also be tracked in live cells using single molecule tracking (smt-) PALM. In this protocol, factors to consider when designing the fusion proteins for BiFC-PALM are discussed, sample preparation, image acquisition, and data analysis steps are explained, and a few exemplary results are showcased. Providing high spatial resolution, specificity, and sensitivity, BiFC-PALM is a useful tool for studying PPIs in intact biological samples.

Wprowadzenie

Protein-protein interactions (PPIs) are fundamental to biology¹ and are tightly regulated via spatiotemporal mechanisms across many time and length scales. Studies on cell signaling on the cell membrane, for example, have revealed dynamic, nanoscale spatial compartments that facilitate specific PPIs and cellular processes². Hence, the ability to probe PPIs in biological systems with sufficient spatial and temporal resolutions, high specificity, and high sensitivity is key to achieving a mechanistic understanding of biology.

Bimolecular fluorescence complementation (BiFC) is one of the few existing tools for visualizing PPIs in a cell with subcellular resolution and live-cell compatibility^3–5. The technique is relatively straightforward and involves splitting of a fluorescence protein into two non-fluorescent fragments; when genetically tagged to two interacting proteins and brought into proximity, the fragments can reconstitute to form a complete fluorescent protein, yielding fluorescent signal. When properly designed, BiFC probes should not spontaneously reconstitute in the absence of PPIs. As such, the fluorescence signal in a BiFC assay will only arise in the presence of PPIs, which enables direct visualization of PPIs with high specificity. Additional benefits of using fluorescence as the readout are high sensitivity, subcellular resolution, and compatibility with high throughput and high content screening assays, among others. For these benefits, a number of BiFC probes based on different parent fluorescent proteins have been developed. As in all other detection techniques based on conventional light microscopy, however, the spatial resolution of BiFC is limited to ~250 nm by the diffraction of light. This makes it a challenge to study the regulation of PPIs at the nanoscale, which, as alluded earlier and exemplified by lipid rafts⁶ and Ras nanoclusters⁷, is a critical length scale to understanding many cellular processes such as signaling.

BiFC has been combined with photoactivated localization microscopy (PALM)^8,9 to overcome this limit in spatial resolution for imaging PPIs¹⁰. PALM is a recent superresolution microscopy technique that circumvents the diffraction limit in fluorescence imaging through stochastic activation and subdiffractive localization of single fluorescent molecules. In each activation cycle, a fluorescent molecule emits a few hundred to a few thousand photons and gives rise to a single molecule image on the detector. While the image is diffraction-limited (~250 nm in width), its centroid can be determined with much higher precision, typically on the order of 10-50 nm depending on the number of photons detected. By activating and localizing each fluorescent molecule in the sample, a high resolution image can be reconstructed. Performed on living cells, single molecule tracking (smt-) PALM further permits acquisition of thousands of protein diffusion trajectories from a single cell¹¹. Importantly, PALM uses specialized fluorescent probes such as photoactivatable fluorescent proteins (PA-FPs) to achieve stochastic activation. Since both BiFC and PALM use fluorescent proteins, they were combined by splitting PAmCherry1, a commonly used PA-FP for PALM, into two fragments between amino acids 159 and 160.

The BiFC system based on split PAmCherry1 shows low background signal from spontaneous reconstitution of the two fragments. When genetically tagged to a pair of interacting proteins, the two fragments (RN = residues 1-159; RC = Met + residues 160-236) reconstituted efficiently to form complete PAmCherry1 proteins even at 37 °C and without incubating at lower temperatures, which is not the case for other BiFC pairs¹² such as the parent mCherry¹³. Furthermore, the reconstituted PAmCherry1 protein retained the photophysical properties of the parent PAmCherry1, such as high contrast ratio, medium photon yield, and fast photoactivation, among others, which are critical for accurate single molecule localization and high-resolution PALM imaging.

In this protocol, the use of BiFC-PALM for imaging Ras-Raf interactions in U2OS cells by using split PAmCherry1 (Figure 1A) is described. The first step is to design the constructs for expressing fusion proteins between PAmCherry1 fragments (i.e., RN and RC) and the proteins of interest. In theory, for each pair of candidate proteins (A and B), there are eight pairs of fusion proteins to be tested: RN-A/RC-B; RN-A/B-RC; RC-A/RN-B; RC-A/B-RN; A-RN/RC-B; A-RN/B-RC; A-RC/RN-B; and A-RC/B-RN. This process can often be simplified by taking into account the structural or biochemical properties of the candidate proteins. In the case of Ras, the protein is post-translationally modified at the C-terminal CAAX box (C=Cys; A=aliphatic; X=any), after which the AAX motif is cleaved off. Hence, RN or RC can only be fused to the N-terminus of Ras; this reduces the number of fusion protein pairs to four (Figure 1B). For Raf, the Ras-binding domain (RBD; residues 51-131) is used and can be tagged on either end. These four fusion configurations were generated: RN-KRas/RC-Raf RBD; RN-KRas/Raf RBD-RC; RC-KRas/RN-Raf RBD; and RC-KRas/Raf RBD-RN.

Additionally, the linker between the fragments and the proteins of interest will need to be considered. A flexible linker of about ten amino acids is often used as it provides sufficient freedom for complementation to occur. One such linker is (GGGGS)x2, although there are many others that have been successfully applied, including random sequences generated from a multiple cloning site (MCS)¹⁴. The length of the linkers may need to be optimized depending on the size of the proteins of interest and their orientations when interacting.

The PAmCherry1 fragments are contained in a small cloning backbone with flanking MCSs (see Materials List). Genes of interest can be inserted via the restriction sites or with a ligation-independent method. After cloning and sequence verification, the expression cassette is transferred to an expression vector using a recombinase reaction, a cloning process with high fidelity and high efficiency.

Next, the resulting expression constructs are transfected into the target cell line or, if a stable cell line is desired, packaged into lentivirus for infecting the target cell line. Transient transfection allows for quick validation of the BiFC configurations, but potential problems must be noted. Transfection using chemicals often stresses the cells, resulting in high autofluorescence; while the use of total internal reflection fluorescence (TIRF) microscopy can mitigate this background signal by limiting the illumination volume, TIRF ideal when the PPIs take place on the cell membrane. Additionally, transient transfections often lead to high levels of protein expression, far exceeding those of the endogenous proteins, which may cause artifacts in detecting the PPIs. Hence, it is recommended that stable cell lines be established after an appropriate BiFC configuration has been determined through initial testing. Stable cell lines also have the potential for tunable expression via doxycycline induction¹⁵.

After infection, cells are selected with antibiotics, typically puromycin and neomycin, one for each construct. The subsequent steps for sample preparation, image acquisition, and data analysis will be described in detail in the protocols.

With this approach, the formation of nanoscale clusters in BiFC-PALM images of Ras-Raf RBD are routinely observed (Figure 2A). Consistently, smt-PALM trajectories of Ras-Raf RBD show a heterogeneous distribution in the diffusion states (Figure 2B-D). These results suggest that Ras-Raf complexes exist in multiple states on the cell membrane, presumably as monomers and clusters, with potential biological implications. This work demonstrates the power of BiFC-PALM in selective imaging of PPIs in cells with nanometer spatial resolution and single molecule sensitivity, which would be difficult to obtain with conventional BiFC, fluorescence co-localization, or fluorescence resonance energy transfer (FRET).

When designing BiFC experiments and interpreting the results, it is important to keep in mind that the BiFC process is irreversible in most cases, including split PAmCherry1. Once the two fragments combine and form a complete PAmCherry1 protein, the linkage between the two fragments, and thus the PPI becomes permanent. This limits the use of BiFC and BiFC-PALM for monitoring the dynamics of the PPIs (i.e., binding kinetics and not the diffusion dynamics of the protein complex once it is formed) and at times could even lead to mislocalization of the PPI complexes.

An additional factor to consider when designing BiFC-PALM experiments is the delay in chromophore maturation that is inherent in fluorescent proteins. Once two proteins interact and the FP fragments come together, they typically refold on the order of seconds (half-time of 60 sec for EYFP³). However, subsequent chromophore maturation and fluorescent signal develop on the order of minutes. While fluorescence may be detectable within 10 min after reconstitution of split-Venus, a fast-folding YFP variant, the half-time for maturation in general is often around 60 min¹⁶. Split-PAmCherry1 was observed to have similar rates. Hence, BiFC and BiFC-PALM are currently poor choices for monitoring real-time PPI kinetics; other methods, such as FRET and a recently developed dimerization dependent fluorescent protein¹⁷, may be more suited for this purpose.

Protokół

1. Cloning

1. Determine the configurations to clone and choose a linker. Tag proteins with the fragments on the N- or C-terminus as described below so they do not disrupt their proper localization. Use a flexible linker such as (GGGGS)x2.
2. Genetically tag the proteins of interest to the PAmCherry1 fragments. As one option, use the cloning plasmids listed in the Materials List containing the fragments RN (PAmCherry1 residues 1-159) and RC (Met plus PAmCherry1 residues 160-236) with flanking MCS sequences.
   1. Linearize the plasmids containing the fragments RN and RC with inverse PCR (or restriction digest) at the point of insertion. Primers for inverse PCR are listed in Table 1. Below is an example PCR protocol used in the author’s lab (see Materials List for the exact materials used in this protocol).
      1. Mix the PCR reaction together, brought up to a final volume of 50 µl with water: to a thin-walled PCR tube, add the water, 25 µl of 2x master mix, 0.5 µl each of the forward and reverse primers (20 µM diluted stock, 0.2 µM final concentration), and 1 ng of template. Use the following cycling conditions: 98 °C for 30 sec, 30 cycles of 98 °C for 7 sec and 68 °C for 2 min, and a final extension at 72 °C for 10 min.
   2. PCR the genes of interest for insertion to the linearized plasmids containing RN and RC fragments. Perform the PCR as in step 1.2.1 with the extension time at 68 °C. Use primers with overhangs that contain a flexible linker sequence, such as GGAGGTGGAGGTAGTGGTGGAGGTGGAAGT for (GGGGS)x2, and 15 base pair homology to the ends of the linearized RN and RC plasmids for recombination.
      Note: The primer overhangs for the proteins of interest are presented in Table 2 if using the primers in Table 1 to linearize the plasmid backbone.
   3. Digest the PCR reaction with Dpn1 to eliminate the PCR template. Add 0.5 µl of Dpn1 (20 U/µl) directly to the 50 µl of PCR reaction. Incubate at 37 °C for 1 hr and heat inactivate at 80 °C for 20 min. If background template persists in later steps, increase the enzyme amount or lengthen the digestion time.
   4. Purify the PCR products via a spin column as described by the manufacturer.
   5. Perform a ligation-independent reaction to combine a linearized plasmid containing an RN or RC fragment with the gene of interest. Measure the concentration of the purified PCR products: combine 50 ng of linearized plasmid and 50 ng of the gene of interest with 1 µl of enzyme premix and bring the total volume to 5 µl with deionized water. Incubate at 50 °C for 15 min, place on ice, and add 20 µl of TE buffer.
   6. Transform recombinant plasmids into competent bacteria¹⁸.
   7. Select 2-4+ (as desired) colonies for O/N culture. Add 5 ml of LB broth and 5 µl of kanamycin (50 mg/ml stock) into a 14 ml polypropylene round-bottom tube, and add the selected bacteria colony with a sterilized inoculating loop. Incubate at 37 °C O/N while shaking at 225 rpm.
   8. Freeze 1 ml of O/N culture for glycerol stock¹⁹ and miniprep the remainder as described by the manufacturer.
   9. Perform sequencing to determine a good clone. If using the cloning plasmids in the Materials List, use M13 forward and reverse primers (in separate reactions) as necessary to obtain the complete sequence of the fusion construct. Compare with the expected sequence using an alignment tool such as BLAST from the National Center for Biotechnology Information (NCBI) - http://blast.ncbi.nlm.nih.gov/BlastAlign.cgi.
   10. Transfer the sequence-verified constructs to an expression plasmid via a recombinase reaction.
       1. Add 75 ng of the destination plasmid — such as pcDNA for transient transfections or pLenti for lentiviral packaging²⁰ — and 50 ng of the recombinant cloning plasmid to 1 µl of enzyme premix and bring up the total volume to 5 µl with deionized water. Incubate at 25 °C for 1 hr, then add 5 µl of TE buffer.
          Note: For lentiviral packaging and stable cell line generation, use a different destination plasmid for each construct, for example, one with puromycin resistance and the other neomycin. This allows for double selection of transduced cells. Also consider the promoter for each, such as the constitutive cytomegalovirus (CMV) promoter for high expression levels, or CMV with the TetO operator for tunable doxycycline-regulated expression.
       2. Transform 5 µl into competent bacteria and miniprep the plasmids as in steps 1.2.6 to 1.2.8, but using the appropriate antibiotic (typically ampicillin).
3. Determine the optimal BiFC configuration.
   1. Co-transfect expression plasmids into U2OS cells (or cell of choice).
      1. Add 350 µl of phenol red-free and antibiotic-free DMEM supplemented with 10% FBS into each well of an 8-well #1.5 glass bottom chamber slide. Plate about 7.5 x 10⁴ cells per well so cells are about 70%-90% confluent the next day.
      2. The day after plating, co-transfect expression plasmids with different configurations in each well using the preferred reagent and as described by the manufacturer.
         1. For each well, add 125 ng of each expression plasmid into 50 µl of serum-free reduced media in a 0.6 ml tube and pipette up and down to mix. Add 1 µl of the transfection reagent down the center of the tube and directly into the media. Agitate gently to mix and let the reaction sit for 30 min.
         2. Reduce the media in each well of the chamber slide by about half. Add the transfection mixture dropwise to the wells and shake gently to mix. Incubate cells at 37 °C and 5% CO₂ for 24-48 hr. Change the media the day after transfection.
   2. Image cells on a fluorescence microscope as in Protocol 2 beginning with step 2.1.4.
      Note: The sample may be imaged on a standard fluorescence microscope if the expression levels are high and the camera is sensitive enough for the relatively low photon output of PAmCherry1. Reconstituted PAmCherry1 molecules can be activated with an ultra violet filter set (405 nm) prior to imaging with a green excitation (561 nm) filter set.
4. Where applicable, create a negative control by, for example, introducing a point mutation into one of the proteins of interest that disrupts the interaction. Perform a site-directed mutagenesis²¹ on the cloning plasmid of the protein to be mutated and of the chosen configuration.
5. Generate a stable cell line by packaging the constructs into viral particles and infect U2OS cells (or cell line of choice). Consider an inducible (tetracycline) expression system¹⁵ so expression can be induced prior to imaging if prolonged irreversibility of BiFC is an issue, or if tunable expression is desired.
   CAUTION: Working with viruses is classified as Biosafety Level 2. While recent generations of viral packaging systems have greatly reduced the likelihood of producing replication-competent viruses, some risks are still present. References can be found at http://www.cdc.gov/biosafety/publications/.
   1. Repeat step 1.2.10 using a lenti- or retroviral destination vector²⁰. Increase the O/N culture to 100 ml for a midiprep and greater purity of the plasmids.
   2. On Day 1: Plate about 2 x 10⁶ 293T/17 cells into a 10 cm dish for each construct to be packaged.
   3. On Day 2: Prepare a transfection reaction using a transfection reagent of choice and as described by the manufacturer.
      1. Prepare a premix of packaging plasmids with final concentrations of 250 ng/µl for packaging plasmids 1 and 2, and 100 ng/µl for packaging plasmid 3 in sterile water. Add 10 µl of the packing plasmid premix and 2.5 µg of the target plasmid to 1 ml of serum-free reduced media in a 5 ml polypropylene round-bottom tube and pipette up and down to mix. Add 20 µl of transfection reagent down the center of the tube and directly into the media.
         1. Agitate gently to mix and incubate at RT for 30 min. Remove about half the media from the cells and add the transfection mixture in a dropwise manner. Shake gently to mix and incubate at 37 °C and 5% CO₂. Incubate 10 ml of media per plate in a tube with the cap loosened for temperature and CO₂ equilibration.
      2. On Day 3: Gently change the media with the pre-incubated media and return the cells to the incubator. Incubate another tube of media with the cap loosened.
         Note: Syncytia, or the formation of large multi-nucleated cells, can be a sign that the packaging is going well. However, cells are in a fragile state and can be easily detached. When changing media, tilt the pipettor so that it is horizontal and add the media dropwise into one edge of the plate.
      3. On Day 4: Harvest the virus by removing the media with a syringe and filtering through a 0.45 µm pore size filter into a clean 50 ml tube. Gently replace the media with the pre-incubated media.
      4. On Day 5: Harvest again as done previously in step 1.5.3.3 and combine common filtrate into the same 50 ml tube. Add 6 ml of virus concentrator to each tube and mix. Store at 4 °C for at least 24 hr until the virus precipitates.
         Note: Depending on the health of the cells, virus can be harvested a third time.
      5. Concentrate the virus by centrifuging at 1,500 x g for 15 min at 4 °C. Aspirate the supernatant and resuspend the pellet in 250 µl of media. Virus can be used immediately for infection or stored in 50 µl aliquots at -80 °C for up to one year.
      6. Plate about 3 x 10⁵ U2OS (or target) cells per well into a 6-well plate (2 ml of DMEM with 10% FBS per well) for infection, so cells are about 50% confluent at the time of infection.
      7. The following day, remove about half the media so about 1 ml remains. Coinfect with 50 µl of concentrated virus for each construct. Add 1 µl of polybrene (8 mg/ml) to each well. Incubate at 37 °C and 5% CO₂. Change media the next day and incubate for one more day before antibiotic selection.
         Note: The amount of virus to add can vary depending on the desired rate of infection. Viruses can be titrated using qRT-PCR.
      8. Add antibiotics to select for infected cells. Separate wells with uninfected cells that have not seen virus are required as canaries to test the efficacy of the selection. Incubate for 2-7 days, or until selection is complete.
         Note: Effective concentrations should be titrated initially, but they are typically 1.0 µg/ml for puromycin and 1.0 mg/ml for neomycin.
      9. Expand the cells into a larger 10 cm dish and proceed to Protocol 2.

2. Imaging Fixed Cells

1. Prepare a sample for imaging
   1. Clean an 8-well #1.5 glass-bottom chamber slide by adding 250 µl of 1 M NaOH for at least 1 hr and wash thoroughly with PBS.
      Note: Untreated chamber slides typically result in high background signal. Additionally, cells may be sensitive to residual NaOH and failure to thoroughly clean the glass surface (as many as 10x washes) may make cell adhesion difficult. If necessary, incubate the cleaned chamber slide with PBS O/N after washing. For sensitive cell lines, plating at a higher density (e.g., at 50%-60% initial confluency) may help.
   2. Plate about 5.5 x 10⁴ of the U2OS stable expression cells per well in 350 µl of phenol red-free DMEM with 10% FBS so that cells are healthy and not over confluent when imaging.
   3. Perform necessary treatments for the experiment, such as adding tetracycline to induce protein expression.
   4. Fix the cells immediately before imaging.
      1. Prior to fixation, prepare fresh paraformaldehyde (PFA) solution — 3.7% PFA in 1x PHEM with 0.1% glutaraldehyde (GA). For 10 ml PFA solution, weigh 0.37 g of PFA and transfer to a 1.5 ml microcentrifuge tube. Add 1 ml of distilled water and 30 µl of 1 M NaOH and vortex. Heat at 70 °C and vortex every 1-2 min until PFA is completely dissolved.
      2. Transfer dissolved PFA to a 15 ml conical tube and add 3.8 ml distilled water, 5 ml 2x PHEM buffer, 20 µl of 25% glutaraldehyde, and vortex to mix. Stock, 2x PHEM buffer is 120 mM PIPES, 50 mM HEPES, 20 mM EGTA, and 16 mM MgSO₄, with pH adjusted to 7.0 with 10 M KOH. This fixative solution can be stored at 4 °C for several days.
         CAUTION: PFA and GA are both hazardous. Prepare the solution in a fume hood and wear proper protective equipment when handling both chemicals. Avoid inhalation or direct skin contact.
      3. Remove the growth medium. Wash quickly with 500 µl of PBS and add 250 µl of the PFA fixative per well. Incubate the cells for 20 min at RT.
   5. Remove the PFA fixative from the sample and add 350 µl of PBS or imaging buffer (100 mM Tris with 30 mM NaCl and 20 mM MgCl₂, pH 8.5) per well.
   6. Vortex 100 nm gold particles to break up aggregates and add 35 µl per well (10x final dilution) for tracking stage drift during imaging.
2. Acquire images
   1. Turn on the microscope. Power on the 405 and 561 nm lasers, but keep the shutters (internal or external) closed at this point. Turn on the EMCCD camera and allow it to cool down. Ensure the 561 nm filters are in place.
   2. Open the acquisition software and set the exposure time to 100 msec and the EMCCD gain to 300 (range 1 – 1,000).
   3. Add immersion oil to the objective and secure the sample to the microscope stage.
   4. With either bright field or the 561 nm laser on (~1 kW/cm²), bring the sample into focus.
   5. For imaging Ras and other membrane proteins, use a 60X apochromat TIRF objective with 1.49 numerical aperture and bring the microscope into TIRF configuration. Adjust the excitation laser so that it is off-centered when hitting the back aperture of the TIRF objective; this causes the laser to deflect upon reaching the sample. Keep adjusting the laser until the critical angle is reached and the laser is being reflected back. Search for a cell to image with several gold particles in view. Set a region of interest that encloses the cell (or a region within the cell) and the gold particles.
      Note: TIRF decreases the penetration depth of the excitation laser and reduces the out of focus background signal. Additionally, drift correction is likely to be more accurate when more gold particles are in view.
   6. If available, engage the autofocus system to correct for sample drift in the z-direction. Otherwise, adjust the focus manually throughout the acquisition if the image goes out of focus.
   7. Begin acquisition with the 405 nm laser off and the 561 nm laser on (~1 kW/cm²) in case sufficient activation is occurring already (typically if expression levels are high). Otherwise, turn on the 405 nm laser at the lowest factory power setting (0.02 mW or 0.01 W/cm²) and increase gradually (0.1 mW at a time) as necessary until there are several tens of molecules per frame or so that single molecules are well separated.
   8. As data acquisition continues, gradually increase 405 nm laser power to keep the spot density roughly constant.
      Note: At lower 405 nm excitation power, some PAmCherry1 can already be activated. As these molecules photobleach, the population of remaining photoactivatable PAmCherry1 molecules decreases, requiring a higher photon flux to maintain the same density of molecules emitting fluorescence in each frame.
   9. Continue image acquisition until high 405 nm power (2.5 – 10 W/cm²) does not activate more activation events.
      Note: The total acquisition time depends on the expression level and efficiency of the complementation.
3. Process the images
   Note: There are many open source software options for image processing. Examples are ThunderSTORM (https://code.google.com/p/thunder-storm/) and QuickPALM (https://code.google.com/p/quickpalm/). The typical data processing procedures are briefly illustrated here using a custom Matlab package called wfiread for processing PALM data as an example. However, future revisions may not exactly follow what is described here.  For image processing with other software, please refer to the user documentations.
   1. Download the image processing software (Supplemental Code File) or the latest version at http://www.ohsu.edu/nan. Open Matlab and load the wfiread software (which is already put in the default path).
   2. View the raw image sequence and determine the region of interest. Select the area of the stack to be processed by left clicking and dragging a box around the desired area. If the region of interest was not reduced during acquisition (to about 256x256 pixels), select a smaller area to lessen the compute time. To deselect an area and processes the entire frame, right click anywhere in the image.
   3. Enter the range of frames to be processed.
   4. Select a gold particle (one that is isolated and uniform in shape) by left clicking on the image and dragging a small box around it. Under Particle Tracking, click the Track button. A graph will appear that shows the position of the selected gold particle across the stack of images, depicting the extent of the drift.
   5. Repeat this process to track as many gold particles as possible and determine the ones that track together (the particles will be color coded). Ideally the overall drift is on the order of one pixel at most.
   6. Individually select the gold particles that tracked together one at a time and click the Add Marker button under Particle Tracking. A green cross will appear signifying that it has been added as a marker and will be used to correct for drift.
   7. Adjust the Sigma Range, Smoothing Range and Threshold as necessary by changing the values slightly. Click the Find Particle button to test the settings. The particles that will be processed will be boxed and those that are not will be left out. The PAmCherry1 molecules, particles that are relatively round and bright, should be selected.
   8. Start processing the images by clicking the Make Coord File button and save the file.
      Note: The program will go through each frame within the defined frame range, find all fluorescent spots, and perform Gaussian fitting to find the centroid coordinates. A .cor file will be generated storing all coordinates and other fitting parameters of the processed single molecule images.
4. Post-process the images and render the PALM image
   Note: Other software may directly render the image after coordinate extraction.
   1. In Matlab, launch the ‘palm’ package and load the .cor file just created.
   2. Before rendering the PALM image using the coordinates, sort the individual coordinates (each from a ‘localization event’). The optimal sorting values may vary depending on the setup and fluorophore.
      1. For PAmCherry1, use the following values: Combining Frame - 8; Combining Distance (nm) - 100; Minimum RMS - 4; Minimum Fit Goodness - 0.25; and Max. Eccentricity - 1.4. See the Discussion section for additional explanation of these parameters.
   3. Click the Sort button to generate a new set of coordinates ready for rendering high resolution images.
   4. Enter values for proper rendering of the final image such as the raw pixel size (nm), the desired feature size (nm) in the rendered image, and the pixel size for the rendered image.
      Note: The raw pixel size must be determined for each setup. The rendered feature size can be set arbitrarily, but is typically set at a value slightly higher than the average localization precision. For PAmCherry1, the average localization precision is around 18 nm, so a feature size of 20 – 30 nm is appropriate. Of note, the localization precision was calculated by taking the standard deviation of the localizations of the same molecule across multiple frames²² and not by dividing the diffraction limit with the square root of detected number of photons. As for the pixel size of the rendered image, 10 nm is a good starting point; setting this value too low will result in unnecessarily large image files for unmagnified views.
   5. Click the Render button to generate the PALM image. Use the +/- magnifying icons in the tool bar of the figure window to zoom in and out.
   6. Optional: Perform cluster analysis of the reconstructed PALM image using either Ripley’s K test or other mechanisms such as simulation-aided DBSCAN (SAD)²³. Note: In the custom palm package, both Ripley’s K and SAD analysis are already built in; for coordinates generated from other software, the same analyses may be performed with additional custom scripts.

3. Single Molecule Tracking in Live Cells

1. Treat the tissue culture surface and plate the cells as described in Protocol 2. Since the temperature and/or CO₂ controlled stage may require a certain culture dish, ensure sure that the coverglass has the appropriate thickness (0.17 mm) for the microscopy setup.
   1. Transfect cells if doing a transient expression and perform any treatments needed for the experiment, such as tetracycline to induce expression, and incubate as necessary. This is also the same as described in Protocol 2.
   2. Place the culture dish on an on-stage incubator. Let the dish settle on the stage for a few minutes until the temperature and CO₂ concentration stabilize every time after mounting the culture dish or moving to a new region of interest. Block lasers at this step. If CO₂ control is not available, change to a CO₂ independent media right before imaging, such as Leibovitz’s L-15.
      Note: Phenol-red-free medium is recommended for depressing background fluorescence.
2. Acquire images as in Protocol 2. However, set an appropriate exposure time (typically 25-50 msec for PAmCherry1).
   Note: The molecule density must be lower than for fixed cells so that trajectories can be recorded without overlapping with each other. The specified few tens of molecules is typical for a 256x256 pixel acquisition at a pixel size of about 160 nm. The 561 nm laser power density was set to 0.8 kW/cm² to maintain a good signal-to-noise ratio while decreasing the phototoxicity to cells and increasing the average trajectory length.
3. Process the images.
   1. Extract coordinates of single molecules with the wfiread package as described in Protocol 2, or with another package.
   2. Reconstruct single particle trajectories using various single particle tracking (SPT) packages based on different algorithms found elsewhere²⁴. Note: Alternatively, this step may be accomplished using home built Matlab package (possibly available in future revisions at http://www.ohsu.edu/nan.
   3. Optional: After reconstruction, use an SPT package²⁴ to calculate displacement and diffusion constants from the trajectories. Additionally, use variational Bayes single-particle tracking (vbSPT)²⁵ to determine diffusion states of the target proteins. vbSPT Matlab package and documentation can be found at its official Sourceforge site: http://vbspt.sourceforge.net/.

Wyniki

The BiFC-PALM example shown is KRas G12D mutant interacting with the Ras binding domain (RBD) of CRaf (Figure 1A). As discussed, the RN or RC fragments were not tagged to the C-terminus of Ras because it would disrupt the membrane localization and hence the biological activity of Ras. This reduced the possible combinations from eight to four (Figure 1B). Each of these four combinations was introduced into U2OS cells using lentiviral infection. The cells were fixed as detailed in the prot...

Dyskusje

BiFC has been a commonly used technique for detecting and visualizing PPIs in cells, while PALM is a recent single molecule superresolution microscopy technique that enables nanoscale imaging of intact biological samples. The combination of BiFC with PALM achieved selective imaging of PPIs inside a cell with nanometer spatial resolution and single molecule sensitivity. BiFC-PALM extends the utility of both techniques, and as demonstrated in this work, shows great promise in revealing the molecular details of PPIs in thei...

Materiały

Name	Company	Catalog Number	Comments
Microscope frame	Nikon	Ti-U
60X oil immersion TIRF objective with 1.49 NA	Nikon	CFI Apo TIRF 60x H
EMCCD camera	Andor	iXon Ultra 897
561 nm laser	Opto Engine	MGL-FN-561
405 nm laser	Coherent	CUBE-405 50mW
561 nm dichroic mirror	Semrock	Di01-R405/488/561/635-25x36
561 nm filter	Semrock	FF01-525/45-25
405/561 nm notch filter	Semrock	NF01-405/488/568-25
Temperature and CO2 controlled stage	Tokai Hit	INUBG2ATW-PLAM
pENTR-D-TOPO-PAmCherry1_1-159-MCS	Addgene	60545
pENTR-D-TOPO-PAmCherry160-236-MCS	Addgene	60546
pcDNA3.2-DEST	Life Technologies	12489-019
pLenti-DEST	Addgene		http://www.addgene.org/Eric_Campeau/
Phusion High-Fidelity DNA Polymerase	Thermo Scientific	F-531
In-Fusion HD Cloning	Clontech	639649
LR Clonase	Life Technologies	11791
Vira Power Lentivirus Packaging	Life Technologies	K497500
X-tremeGENE Transfection Reagent	Roche	13873800
Lab-Tek II Chambered Coverglass	Thermo Scientific	155409	#1.5 glass bottom dishes
U2OS cells	ATCC	HTB-96
293T/17 cells	ATCC	CRL-11268
DMEM with phenol red	Life Technologies	11995
DMEM no phenol red	Life Technologies	21063
Fetal bovine serum	Life Technologies	10082
Leibovit's L-15, no phenol red	Life Technologies	21083-027
Reduced serum medium	Life Technologies	31985
Phosphate Buffered Saline	Life Technologies	14040
Syringe	BD Biosciences	309604
Syringe filter	Millipore	SLHV033RB
Lentiviral concentrator	Clontech	631231
Retroviral concentrator	Clontech	631455
10 cm culture dish	BD Biosciences	353003
6-well culture plate	BD Biosciences	353046
Polybrene	Sigma	107689
Puromycin	Life Technologies	A11138
G-418	Calbiochem	345812	Neomycin
Doxycyline	Fisher	BP2653
Tris base	Fisher	BP152
EDTA	Sigma	EDS
Sodium Hydroxide	Sigma	S5881
Paraformaldehyde	Sigma	158127
Glutaraldehyde	Sigma	G6257
PIPES	Sigma	P6757
HEPES	Sigma	H4034
EGTA	Sigma	3777
Magnesium Sulfate	Sigma	M2643
Potassium Hydroxide	Sigma	221473
Sodium chloride	Fisher	BP358
Magnesium chloride	Fisher	M33
100 nm gold particles	BBI Solutions	EM.GC100
Molecular grade water	Life Technologies	10977
Dpn1	New England Biolabs	R0176
PCR purification kit	Qiagen	28104
Miniprep kit	Qiagen	27104
Midiprep kit	Macherey-Nagel	740410
0.6 ml microcentrifuge tubes	Fisher	05-408-120
1.5 ml microcentrifuge tubes	Fisher	05-408-137
15 ml tubes	Fisher	05-539-12
5 ml polypropylene round-bottom tubes	BD Biosciences	352063
14 ml polypropylene round-bottom tubes	BD Biosciences	352059
50 ml tube	BD Biosciences	352070
PCR tube	GeneMate	C-3328-1
SOC medium	Life Technologies	15544
LB broth	BD Biosciences	244610
Kanamycin sulfate	Fisher	BP906
Competent cells	Life Technologies	C4040
Matlab	Mathworks