Method Article

Single-Molecule Localization Microscopy of Membrane Proteins using Single-Antibody Labeling

1.2K views

⸱

DOI:

10.3791/69853

⸱

March 20th, 2026

 ,  ,  ,  , 
1Department of Chemistry, University of Illinois Chicago, 2University of Illinois Cancer Center, 3Department of Biomedical Engineering, University of Illinois Chicago, 4Department of Biochemistry and Molecular Genetics, College of Medicine, University of Illinois Chicago

Corresponding Authors: Hirushi Gunasekara <hgunas2@uic.edu>, Ying S. Hu <yshu@uic.edu>

In This Article

Summary

This protocol describes single-antibody labeling (SAL) to resolve the nanoscale spatial organization of plasma membrane proteins. By leveraging cumulative antibody-epitope interactions at the single-molecule level, membrane SAL (mSAL) maps local epitope distributions while simultaneously capturing antibody binding behavior in the native cellular environment.

Abstract

The plasma membrane defines the cell shape and serves as the interface that governs intercellular communication. Membrane proteins constitute a major class of therapeutic targets; therefore, super-resolving the cell membrane through its constitutive proteins holds great promise in advancing cell biology and antibody therapeutics. In this regard, single-molecule localization microscopy (SMLM) enables nanoscale visualization of protein organizations on biological structures. Despite its importance, applying SMLM to plasma membrane proteins poses unique challenges. In this protocol, we present an effective approach using time-lapse single-antibody labeling (SAL) termed membrane SAL (mSAL). We provide detailed step-by-step instructions, including optimization of the antibody concentration, laser power density, duration of non-illumination intervals, image reconstruction, and density-based cluster analysis, to resolve nanoscale membrane protein distribution and membrane morphology. We use the tetraspanin protein CD81 as the model membrane protein to demonstrate the capability of mSAL on both adherent and suspension mammalian cells. In addition to super-resolving the cell membrane and distributions of membrane proteins, our technique enables the investigation of the pharmacodynamics of therapeutic antibodies interacting with their membrane targets in the native membrane environment.

Introduction

The plasma membrane contains a diverse array of proteins that regulate cell mobility, adhesion, sensing, and intercellular communication. Understanding the nanoscale distribution of plasma membrane proteins is important, as protein function is often governed not only by expression levels but also by their spatial organization on the membrane1,2. Notably, many therapeutic antibodies target membrane proteins3,4, making nanoscale investigations essential for understanding how antibody interactions are influenced by the local membrane protein organization. However, precise mapping of membrane protein distribution, single-molecule profiling of their abundance, and real-time monitoring of their dynamic interactions with binding partners remain challenging.

Widely used single-molecule localization microscopy (SMLM) techniques, such as direct stochastic optical reconstruction microscopy (dSTORM)5,6 and DNA-based points accumulation for imaging in nanoscale topography (DNA-PAINT)7, use immunofluorescence (IF) staining for sample labeling. They achieve lateral resolutions of ~ 10 nm and have provided extensive insights into the nanoscale organization of therapeutic membrane targets, including human epidermal growth factor receptor 2 (HER2)8,9, CD2010, T cell receptors11,12,13, programmed death-ligand 114, B cell receptors15,16,17, and CD1917,18. For example, dSTORM studies using the therapeutic monoclonal antibody Trastuzumab to directly detect HER2, a key receptor implicated in breast cancer, have revealed distinct clustering states among HER2-positive patients8, while different therapeutic antibodies have been shown to modulate HER2 clustering9. DNA-PAINT combined with lattice light sheet microscopy has elucidated the oligomeric states and interaction profiles of CD20 in B-cells with therapeutic monoclonal antibodies rituximab, ofatumumab, and obinutuzumab10. Although conventional SMLM provides valuable insights into nanoscale organization, its full potential is limited by traditional IF staining approaches. The quality of the SMLM image is dependent on the IF staining, which in turn is governed by the specificity and affinity of the labeling antibody. Traditional SMLM uses high-affinity antibodies that typically achieve superior IF staining and image quality. However, most therapeutic antibodies interact with membrane receptors of live cells with moderate affinity to minimize off-target side effects19. For instance, bispecific T cell engagers (TCEs) exhibit a lower affinity for CD3, which prevents TCR internalization20,21,22. Such interactions cannot be accurately captured by conventional IF techniques.

We have previously developed a strategy that directly leverages single-molecule antibody-epitope interactions, termed single-antibody labeling (SAL). SAL enables high-density mapping of antibody-epitope interactions while preserving native binding equilibria23. While we have previously reported a protocol for intracellular cytoskeletal and mitochondrial proteins24, SAL on membrane proteins has not been demonstrated. In this protocol, we provide a step-by-step workflow for applying SAL to membrane proteins, demonstrated through the labeling of the tetraspanin CD81, a reported therapeutic target candidate25,26,27,28, in both suspension and adherent cells. Using fluorescently labeled anti-CD81 antibodies to directly probe the spatial distribution of CD81 on U2OS and Jurkat T cells, we demonstrate the application of SAL nanoscale studies of membrane proteins. We termed this strategy as membrane SAL (mSAL). Finally, we describe a clustering-based strategy to reduce the isolated single-molecule detections outside of the cell.

In essence, SAL uses time-lapse single-molecule labeling to capture individual antibody-epitope interactions, and subsequent single-emitter deconvolution yields a super-resolution image based on SMLM. By adjusting antibody concentration and selecting an appropriate non-illuminating interval (NII) between consecutive image frames, SAL matches the antibody on-rate to the single-molecule detection rate, enabling single-molecule localization. SAL requires longer NIIs due to the inherently slow binding kinetics of antibodies, which prolong the image acquisition durations (~3-17 h). SAL is best suited for applications that require precise nanoscale mapping of epitopes and the evaluation of individual antibody-epitope interaction dynamics. Despite its lower throughput, SAL offers unique insight into epitope organization and antibody engagement under near-physiological conditions, making it particularly valuable for characterizing therapeutic antibodies in situ.

Protocol

This protocol uses the established Jurkat E6.1 and U2OS cell lines and does not involve primary human or live animal materials.

1. Cell seeding/landing and fixation

NOTE: The sample preparation follows standard immunofluorescence (IF) staining protocols, optimized to preserve the subcellular organization of plasma membrane proteins across different cell types. If an established immunostaining protocol exists for the target of interest, users may follow that protocol up to the sample-blocking step 1.2.14. Single-antibody labeling (SAL) relies on the cumulative detection of fluorescently labeled antibodies as they bind to their antigen at the single-molecule level. Fluorescently labeled antibodies diluted in an imaging buffer are added to the samples on the microscope stage, and fluorescent labeling occurs dynamically as the image acquisition progresses. No fluorescent labeling is performed during the sample preparation phase.

  1. Prepare adherent cells, such as the U2OS cell line, as described in a previously published protocol24.
  2. Cell suspensions
    1. Maintain Jurkat T cells at a density of 1 x 105 to 1 x 106 cells/mL in 10 mL of complete RPMI media (RPMI 1640 supplemented with 10% (v/v) fetal bovine serum, 100 U/mL penicillin, and 100 µg/mL streptomycin) at 37 °C in a humidified incubator supplemented with 5% CO2.
      ​NOTE: The recommended 10 mL of complete media applies to T25 flasks; users should scale the volume accordingly when using different culture vessels.
    2. Prepare the glass surface by adding 200 µL of 0.01% Poly-L-Lysine (PLL) solution to the wells intended for use in an 8-well chambered cover glass. Incubate overnight at 4 °C or for 2 h at 37 °C.
    3. Count the cells from culture using a hemocytometer, then transfer 3.5 x 104 cells to a 1.5 mL microcentrifuge tube.
      ​NOTE: The cell count may need to be adjusted to achieve a good distribution of single cells. A seeding range of 30,000-60,000 Jurkat T cells achieve an ideal density where single cells are clearly separated from one another without being too sparsely distributed. The cell count may need to be adjusted for a new cell type.
    4. Centrifuge at 300 x g for 3 min to obtain a cell pellet. Discard the supernatant and resuspend the cell pellet in 200 µL of pre-warmed (37 °C) 1x Dulbecco's Phosphate Buffered Saline (DPBS). The resulting cell pellet will be visible as a small, white deposit at the bottom of the microcentrifuge tube.
      ​NOTE: Dispose of biohazardous waste according to the institution's guidelines and all applicable local, state, and federal regulations.
    5. Aspirate the PLL solution from the prepared chambered cover glass (step 1.2.2). If the PLL incubation takes place overnight at 4 °C, pre-warm the 8-well chambered cover glass to 37 °C before aspirating the solution.
    6. Add the cell suspension from step 1.2.4. onto the PLL-coated chamber well. Incubate in a humidified incubator at 37 °C supplemented with 5% CO2 for 30 min to allow cells to attach to the surface. Mix the cell suspension thoroughly before plating and pipette up and down to break up cell clumps and ensure even distribution across the well.
    7. Check the cell attachment to the glass surface using a light microscope. Under 10x magnification, focus the microscope onto the bottom of the cover glass.
      ​NOTE: Gently tapping the chambered cover glass while observing under the microscope may help assess whether the cells have attached to the surface or are still floating.
    8. Prepare 200 µL of the fixation buffer (4% paraformaldehyde and 0.1% glutaraldehyde in DPBS) and pre-warm it to 37 °C in a water bath.
      ​NOTE: The provided fixation buffer is optimized for CD81 and is generally suitable for preserving membrane proteins. For new membrane protein targets, the buffer composition may need to be adjusted to achieve optimal preservation. Users are encouraged to consult the literature for fixation conditions tailored to new targets.
      ​CAUTION: Paraformaldehyde is toxic and carcinogenic. Glutaraldehyde is toxic and corrosive. Wear appropriate personal protective equipment (PPE) when handling this reagent. In the event of skin contact, immediately wash the affected area thoroughly with soap and water. Dispose of the reagent according to the user's institution's guidelines and all applicable local, state, and federal regulations.
    9. Aspirate the DPBS from each sample well (from step 1.2.6) using a micropipette.
      ​NOTE: It is critical to perform all solution exchange steps slowly to avoid dislodging cells from the cover glass. Dispose of the biohazardous waste according to the user's institution's guidelines and all applicable local, state, and federal regulations.
    10. Add 200 µL of pre-warmed 37 °C fixation buffer to a corner of the well and incubate at room temperature for 15 min.
      ​CAUTION: Paraformaldehyde is toxic and carcinogenic. Glutaraldehyde is toxic and corrosive. Wear appropriate PPE when handling these reagents. In the event of skin contact, immediately wash the affected area thoroughly with soap and water. Dispose of the reagent according to the user's institution's guidelines and all applicable local, state, and federal regulations.
    11. Wash the fixed cells 3x with 500 µL of DPBS. Skip this step if proceeding with step 1.2.12.
    12. Optional: Slowly remove the fixation solution and add 200 µL of freshly prepared 0.1% sodium borohydride in DPBS, then incubate at room temperature for 7 min. Skip this step and proceed to step 1.2.14 if not using sodium borohydride reduction.
      NOTE: The reduction with sodium borohydride helps quench autofluorescence generated by glutaraldehyde and neutralizes unreacted aldehydes. The release of hydrogen gas as a result of sodium borohydride reacting with the aqueous solution will manifest as the release of bubbles. If the release of hydrogen gas is not observed, use a freshly made solution of sodium borohydride.
      CAUTION: Sodium borohydride is toxic and corrosive. Wear appropriate PPE when handling this reagent. In the event of skin contact, immediately wash the affected area thoroughly with soap and water. Dispose of the reagent according to the user's institution's guidelines and all applicable local, state, and federal regulations.
    13. Aspirate the sodium borohydride solution from each well, then wash three times with 500 µL of DPBS.
    14. Prepare 200 µL of the blocking buffer and add it to the well. Incubate at room temperature for 30 min. Do not use a permeabilization reagent (e.g., TritonX-100, saponin), as it may remove membrane proteins. The membrane SAL (mSAL) protocol presented here for CD81 did not include a blocking step.
      ​NOTE: A blocking buffer containing 5% (w/v) bovine serum albumin (BSA) in DPBS is recommended for mSAL with new targets. When using secondary antibodies, the corresponding serum is used for blocking. For example, 10% (v/v) goat serum in DPBS when using a goat-host secondary antibody. If an established blocking buffer recipe and a protocol exist for the target, users may follow that protocol. The blocking buffer can optionally be supplemented with isotype control antibodies or serum-derived antibodies to reduce off-target bindings of antibodies23,24. Only include a permeabilization reagent in the blocking buffer if the antibody targets an intracellular epitope of the membrane protein.
    15. For mSAL with dye-conjugated primary antibodies, directly proceed to step 1.2.20.
    16. Prepare 150 µL of the primary antibody staining solution (antibody at desired concentration diluted in the blocking buffer). Skip this step for mSAL with dye-conjugated primary antibodies.
      ​NOTE: For primary antibody staining, refer to the vendor's datasheet for recommended conditions. If concentration information is not provided, start with 4 µg/mL (27 nM) for full-length IgG23,24.
    17. Replace the blocking buffer with the primary antibody solution. Incubate at room temperature for 2 h or at 4 °C overnight in a humid chamber. Skip this step for mSAL with dye-conjugated primary antibodies.
    18. Wash the sample 2x with 400 µL of DPBS. Add 200 µL of blocking buffer to the well.
    19. Optional: Replace the blocking buffer with 200 µL of post-fixation buffer (4% paraformaldehyde) and incubate for 10 min at room temperature.
      ​NOTE: Post-fixation helps preserve sample blocking for SAL23,24. However, this step should be omitted if it reduces epitope accessibility. To assess the effect of post-fixation on epitope recognition by antibodies, perform standard immunofluorescence staining with and without post-fixation, and compare the staining quality.
    20. Optional: Wash the sample 2x with DPBS.
    21. Add 200 µL of DPBS into each well and cover the chamber with a transparent film to prevent evaporation. Samples can be stored at 4 °C for up to 1 month. If performing mSAL immediately after step 1.2.14, or 1.2.19 without post-fixation, keep the sample in the blocking buffer and proceed directly to step 2.2.
      NOTE: Maintaining the sample in DPBS supplemented with 5 mM sodium azide will prevent bacterial growth during storage.
      CAUTION: Sodium azide is acutely toxic and can be fatal if ingested. Wear appropriate PPE when handling this reagent. In the event of skin contact, immediately wash the affected area thoroughly with soap and water. Dispose of the reagent according to the user's institution's guidelines and all applicable local, state, and federal regulations.

2. Image acquisition

NOTE: The imaging procedure for both adherent and suspension cells is similar and is performed using total internal reflection fluorescence (TIRF) microscopy. However, imaging conditions should be optimized for each cell type.

  1. Allow stored samples (at 4 °C) to come to room temperature before imaging to prevent condensation on the glass. Skip this step if proceeding directly to imaging after sample preparation.
  2. Power on the microscope system: microscope, sCMOS camera, laser unit, and the computer with the image acquisition software. Allow the microscope system to stabilize for 5-10 minutes after powering on to allow the system's components to reach operating temperatures and minimize thermal drift.
    NOTE: mSAL can be performed using any TIRF microscope setup and compatible software. Power on the system according to the instrument's instructions.
  3. Apply immersion oil to a TIRF objective (e.g., 100x/1.49) and mount the chambered cover glass onto the microscope stage.
    NOTE: If the objective is equipped with a correction collar, ensure it is set to the correct thickness of the cover glass at 23 °C. The cover glass utilized in this study had a thickness of 0.13-0.17 mm.
  4. If equipped, engage the Perfect Focus System (PFS), then carefully raise the objective to the cover glass and bring the sample into focus.
  5. Prepare gold colloids as fiducial markers (1:2 dilution in ultra-pure water). Sonicate the gold colloids before dilution for 3 min to disperse aggregates.
  6. Gently aspirate the storage buffer (DPBS or blocking buffer) from the sample well and add 200 µL of the gold colloid solution to the sample. Incubate for 1 min and remove the solution.
  7. Use brightfield illumination to verify that enough gold colloids have been deposited before aspirating the solution. An ideal density consists of 5 to 10 gold colloids within the designated region of interest (ROI).
    NOTE: Alternatively, fluorescent microspheres (see Table of Materials) can be used at a dilution of 1:1,000 in DPBS.
  8. Wash 1x with 500 µL of ultra-pure water to remove excess gold colloids, then add 200 µL of DPBS.
  9. In the imaging software, create the optical configuration for the fluorophore (of the antibody probe), including the appropriate dichroic, laser excitation wavelength, and power density, camera integration time, emission filter, and camera pixel binning (if applicable). A pixel size of 150-160 nm is recommended to achieve optimal resolution.
  10. Create an optical configuration for a photobleaching step. A laser power of 380 W/cm2 or higher with a camera integration time of 1 s is recommended.
    NOTE: Photobleaching during acquisition helps minimize overlap between individual fluorescent antibody-binding events. The collected light can be directed through an emission filter with a non-matching wavelength to avoid camera saturation.
  11. Adjust the laser angle to establish TIRF illumination. Using brightfield illumination, identify a cell for mSAL. Ensure at least 5 gold colloids are visible within the ROI to correct stage drift during image reconstruction. Confirm autofluorescence from the gold colloids at the imaging laser power density.
  12. Set the imaging parameters: non-illuminating interval (NII), number of frames, and the frequency of photobleaching step (if required) in the image acquisition software. For mSAL of CD81 on U2OS and Jurkat T cells with CD81 antibody, optimal settings were an NII of 5 s, 2,000 total frames, and a photobleaching step every 5th frame.
    NOTE: These parameters must be optimized for each protein of interest, the antibody used for mSAL, and the cell type. Refer to step 3 for more details on optimization. The optimal number of frames may vary between the antibody-target as well as the cell type. For new target systems, capturing 2,000-3,000 frames is recommended as a starting point. NIIs extend image acquisition durations. For example, a 3,000-frame mSAL acquisition with a 20 s NII results in ~17 h of acquisition time.
  13. Prepare the imaging buffer by diluting the CD81 antibody to a concentration of 1 µg/mL in 200 µL of 5% BSA (w/v) in DPBS. For mSAL of CD81, the final antibody concentration was 0.5 nM for Jurkat T cells and 2 nM for U2OS cells.
    NOTE: The antibody concentration depends on the local substrate abundance and the affinity of the antibody. It is critical to optimize the antibody concentrations for each protein target, each antibody used for imaging, and each cell type. Refer to step 3.1 for more details on optimization.
  14. Add the prepared imaging buffer to the well and start the acquisition.

3. Imaging parameter optimization

NOTE: mSAL imaging parameters are influenced by the protein expression, its local organization, and the choice of antibody used for imaging. The objective of optimizing imaging parameters is to identify the antibody concentration and NII that produce spatially resolved single-molecule events while maintaining sufficiently high densities to maximize the epitope sampling.

  1. Antibody concentration
    1. Dilute the antibody to 0.5 nM for Jurkat cells and 2 nM for U2OS cells in 200 µL of 5% (w/v) BSA in DPBS and add the antibody solution to the well. For optimal single-antibody labeling, use an antibody concentration at or near its equilibrium dissociation constant (Kd); if Kd is unknown, starting at 1 nM is recommended.
    2. Use the live-view function in the imaging software configured for mSAL to monitor antibody binding.
    3. Adjust the TIRF angle and imaging parameters to achieve a signal-to-noise ratio (SNR) of ≥ 3.
      ​NOTE: An optimal TIRF angle improves the SNR by limiting out-of-focus fluorescence from diffusing antibodies in the imaging buffer. A laser power density of 100-200 W/cm2 and a 50 ms camera integration time can be used at the start. If the SNR is low, gradually increase laser power and/or the camera integration time and monitor the changes to the SNR until it reaches an optimal value.
    4. Configure the imaging parameters in the image acquisition software to obtain at least a 10-frame acquisition with an NII of 5 s (see step 2.12).
      ​NOTE: A 5 s NII is recommended as a starting point. An NII scan should be performed to identify the optimal NII for a new antibody-target system (step 3.2).
    5. Start the acquisition. Evaluate the sparsity of single-molecule localizations of the recorded movie.
      ​NOTE: Detailed steps on obtaining the localization density are described in a previous publication24. A single-molecule localization density between 0.02 and 0.15 events/µm2 was found to be optimal for mSAL.
    6. If multiple overlapping antibody-binding events are observed immediately after adding the imaging buffer or single-molecule localizations accumulate over time as the recording progresses, abort the acquisition. Reduce the antibody concentration by 2-fold and repeat the acquisition on a new sample. Repeat this process of reducing the antibody concentration by 2-fold and reevaluating the number of events per unit area until the desired single-molecule localization event density is achieved.
      ​NOTE: Overlapping events manifest as multiple large puncta that become brighter or larger upon the binding of a new antibody within the diffraction limit of the first antibody.
    7. If the event density is sparse, increase the antibody concentration in the imaging buffer by 2-fold and repeat the acquisition on a new sample. Repeat this process of increasing the antibody concentration by 2-fold and reevaluating the number of binding events per unit area until the desired single-molecule localization event density is achieved.
      ​NOTE: It is recommended to use a fresh sample for each concentration optimization trial to avoid potential interference from pre-bound antibodies in subsequent localization density evaluations.
  2. NII optimization
    1. Configure the imaging parameters in the image acquisition software to obtain a 50-frame acquisition (see step 2.12). Set the NII to 0 (no delay).
    2. Prepare the imaging buffer to the optimized antibody concentration from step 3.1 and add it to the sample well. Start the acquisition.
    3. Repeat steps 3.2.1-3.2.2, adjusting NIIs to 1, 3, 5, 10, or 20 s. Perform each acquisition on a new sample.
    4. Evaluate the density and the sparsity of single-molecule localizations of the recorded movie (see steps 3.1.5-3.1.7). The lowest NII with the optimal single molecule event density is determined as the optimal NII.
      NOTE: The antibody concentration may need to be optimized further to achieve optimal single-molecule localization density for the corresponding NII. Refer to step 3.1. Previous reports show that longer NIIs (10-20 s) were more selective towards higher-density binding events, but extending the NII beyond 20 s did not yield additional improvement23,24. If spatially overlapping binding events occur as the image acquisition progresses, introduce a photobleaching step at intervals sufficient to remove fluorescence from bound antibodies. For example, if overlapping events appear by the fifth frame, apply photobleaching every fifth frame. The appearance of spatially overlapping events is described in the NOTE of step 3.1.6.

4. Data analysis

NOTE: mSAL image reconstruction uses the free NIH software, Fiji29, with the ThunderSTORM30 plugin. Fiji version v1.54r and ThunderSTORM version 1.3 were utilized in this protocol. Detailed image analysis has been described in a previous protocol 24. In this section, we outline the analysis workflow used specifically for mSAL. Analysis presented in this protocol used a Prime95B sCMOS camera (serial number A18B203004) with photoelectrons per A/D of 0.98, and the camera base level of 100. The imaging pixel size set to 147 nm.

  1. Set the parameters for the reconstruction in the Run Analysis window. The parameters used for CD81 mSAL image reconstructions are described below. The Preview function should be used to evaluate the detected single-molecule binding events in the selected frame before running the full analysis.
    Image Filtering:
    Filter: Wavelet filter (B-spline)
    B-Spline order: 2
    B-Spline scale: 4
    Approximate localization of molecules:
    Method: Local Maximum
    Peak intensity threshold: std(Wave.F1)
    Connectivity: 8-neighbourhood
    Sub-pixel localization of molecules:
    Method: PSF: Integrated Gaussian
    Fitting radius [px]: 2
    Fitting method: Maximum likelihood
    Initial sigma [px]: 0.75
    Multi-emitter fitting analysis: not selected
    Visualization of the results:
    Method: Normalized Gaussian
    Magnification: 7.3
    Update frequency [frames]: 50
    Lateral uncertainty: 20
    NOTE: The pre-filtering parameters are used to enhance single-emitter detection over the background. They serve as starting points but should be optimized for each dataset. The approximate localization and detection method parameters are used for single-molecule binding events. Using local intensity maxima with 8-neighborhood connectivity is shown to provide reliable performance31. The integrated Gaussian method with maximum-likelihood fitting was selected as it is widely used and provides a reliable approximation of the point spread function (PSF) in single-molecule imaging applications32,33,34,35. The visualization method used in this protocol is Normalized Gaussian. Users may select the visualization method that best suits their data. The magnification was set to yield a reconstructed pixel size of 20 nm, which is appropriate for the expected localization precision achievable with antibody-based labeling, where the linkage error is on the order of ~10 nm. The magnification factor of 7.3 was calculated by dividing the diffraction-limited pixel size by the desired reconstructed pixel size (147 nm / 20 nm = 7.3). The update frequency determines how often the reconstructed image is refreshed during processing, but it does not influence the final result. The lateral uncertainty is set to 20 (but not forced) so that the reconstruction incorporates the calculated localization uncertainties. For additional details on available options, refer to the ThunderSTORM User Guide30 or click the question mark next to the input field.
  2. Start the ThunderSTORM analysis with an initial sigma of 1.6 pixels and a fitting radius of 3. Refine the parameters as needed. Multi-emitter fitting was not used, as event densities were kept non-overlapping in the diffraction-limited images.
  3. Click Ok to run the reconstruction. Once complete, the results window will automatically open.
  4. Use the Plot histogram function to find the peak sigma. Convert this value (nm) to pixels by dividing by the raw data pixel size, then rerun the ThunderSTORM analysis using the new value for initial sigma and the fitting radius set to an integer close to 3x initial sigma30.
  5. Determine the drift correction using the selected fiducial marker24.
  6. Perform image reconstruction over the full field of view and apply the calculated drift correction.
  7. For post-processing of the image, click the Filter tab of the results window. Apply filters to remove poorly localized events. In the displayed results for CD81, events with sigma greater than 200 nm and uncertainty greater than 20 nm were removed. Use the Plot histogram function to examine the sigma distribution and determine an appropriate cutoff.
    NOTE: Single-molecule localizations with large sigma values typically arise from dim fluorophores with low photon counts or from overlapping emitters. Localizations with an uncertainty greater than 20 nm were also removed, as an uncertainty exceeding the final pixel size indicates poor localization accuracy. For additional guidance on filtering parameters, refer to the ThunderSTORM User Guide30.
  8. Save the results table with single-molecule localization data for downstream analyses by selecting Export. The exported file will be in .csv file format and referred to as mSAL.csv in this protocol. Save the resulting Normalized Gaussian image.

5. Cluster analysis

NOTE: During epitope probing, antibodies interact with their intended targets (specific interactions) as well as off-target molecules (non-specific interactions). Because mSAL relies on the cumulative detection of single antibody-antigen binding events, some non-specific interactions inevitably contribute to the reconstructed image. Antibody-epitope interactions exhibit distinct kinetic fingerprints that can, in principle, distinguish specific from non-specific binding, where specific and non-specific events are differentiated by their local interaction densities, rather than by single-molecule dwell times36. Here, a density-based spatial clustering of applications with noise (DBSCAN)37 analysis is described to remove low-density events, retaining high-density events that are more likely to reflect specific antibody binding. In addition, density-based clustering can also provide quantitative insights into nanoscale spatial organization of proteins, particularly those known to naturally organize into clusters, such as CD81 into tetraspanin-enriched microdomains38,39,40,41,42. The provided DBSCAN analysis was performed using MATLAB 2025a. For users without access to MATLAB, performing similar DBSCAN analyses with open-source alternatives such as Python or R43,44,45 is recommended. The code utilized in this protocol can be obtained from: https://github.com/nrams12/DBSCAN_SAL_Clusterer/tree/main.

  1. Running the code
    1. Download the MATLAB file from the GitHub link above.
    2. Start MATLAB and open the code. On lines 37 and 38, set the minimum number of points (minPTS) and epsilon for the clustering parameters. The following parameters were utilized:
      Jurkat T cells:
      minPTS: 7
      Epsilon: 60
      U2OS cells:
      minPTS: 7
      Epsilon: 80
      NOTE: See step 5.2 for further details on optimizing these parameters.
    3. Set the filepath in the MATLAB software to the folder that contains the reconstructed mSAL.csv file.
    4. Run the code by clicking Run. The code will prompt the user to select a file. Select the mSAL.csv file and click Open.
    5. The clustered data is automatically saved into the folder named MATLAB Data in the filepath designated within the code. Open this file in ThunderSTORM by clicking Plugins > ThunderSTORM > Import/Export > Import results.
    6. Save the resulting single-molecule localization microscopy image.
  2. Optimization of the clustering inputs
    NOTE: The two input parameters for the DBSCAN algorithm are the minPTS and epsilon. Epsilon defines the radius around a point to establish its neighborhood, while minPTS specifies the minimum number of points within that neighborhood required to form a dense region. Points that satisfy this density requirement are classified as core points, and their neighborhoods are recursively added to the cluster. Points that do not meet the density requirement but fall within the neighborhood of a core point are labeled as border points, while all remaining points are considered noise points.
    1. Run the analysis using the inputs described above (step 5.1.2) as a starting point. Choose a minPTS value that is higher than the number of noise localizations but lower than the number of localizations on the cell. After selecting minPTS, adjust epsilon so that this number of localizations can be captured within the neighborhood radius for representative on-cell regions.
      NOTE: minPTS and epsilon can be adjusted independently of each other. A practical approach to selecting these parameters is to first examine the scatter plot of noise points outside the cell and compare them to the number of localizations observed within similar-sized areas on the cell.
    2. Evaluate the window labeled Figure 2 depicting the clustered data. If the data retains clusters on the cell containing the expected phenotype with minimized localizations outside the cell, the clustering inputs are adequate.
      NOTE: The clustered data should retain the structure of the data acquired using mSAL. If the clusters appear small, localized to only the very high-density regions of the original data, and exclude a significant number of binding events that correlate with the expected phenotype, the clustering parameters may be too restrictive. If the nonspecific binding outside the cell is not sufficiently reduced with large clusters, the clustering parameters may be too inclusive.
    3. If the clustering parameters are too restrictive, decrease the minPTS and increase the epsilon to include more points in the clustered data. If the clustering parameters are too inclusive, increase the minPTS and decrease the epsilon.
    4. After adjusting the minPTS and epsilon, rerun the code. Repeat steps 5.2.2-5.2.3 until optimal clustering is achieved.

6. Validation of antibody binding to its target

NOTE: For membrane proteins with no well-defined morphology, validating the accuracy of the resolved structures can be challenging. To validate the specificity of the antibodies, users may express a fluorescent protein (FP) fusion of the target protein. As an initial check, conventional immunofluorescence staining can be performed on cells expressing the FP fusion. For more precise validation when assessing new antibodies, mSAL can be performed on the FP-expressing cells to confirm that localized binding events correspond to the fluorescently tagged target. Here, a protocol for transfecting U2OS cells with CD81-mStayGold is described as an example, enabling confirmation that CD81 mSAL localizations correspond to the FP-labeled CD81 on the cell surface. Users may substitute any suitable fluorescent protein, provided its emission does not overlap with the fluorescence channel of the antibody being validated.

  1. Seed U2OS cells in an 8-well chamber slide and grow to ~70% confluency. Maintain cells in 200 µL of complete growth media (DMEM supplemented with 10% (v/v) fetal bovine serum, 100 U/mL penicillin, and 100 µg/mL streptomycin).
  2. Combine 100 ng of the DNA construct encoding the protein with a fluorescent reporter with 5 µL of Lipofectamine 2000. Vortex for 30 s and incubate for 30 min at room temperature.
    NOTE: This study utilized CD81 fused to mStayGold. Standard molecular cloning techniques were employed in generating DNA plasmids, though these are not described in this protocol. The DNA: Lipofectamine ratios provided are optimized for cells seeded in 8-well chamber slides and may need adjustment according to the manufacturer's instructions for other sample formats.
  3. Add 5 µL of DNA/Lipofectamine solution to cells. Incubate the chamber slide for 4 h at 37 °C, supplemented with 5% CO2 in a humid environment.
  4. Replace the media with pre-warmed complete growth media. Incubate for > 8 h at 37 °C supplemented with 5% CO2 in a humid environment to allow detectable levels of protein expression.
  5. Assess fluorescence expression levels on the cells using a fluorescence microscope or an imager. A transfection efficiency of ≥ 30% is recommended. If fluorescence expression levels are suboptimal, optimize the DNA: Lipofectamine ratio according to the manufacturer's guidelines.
    NOTE: Transiently expressing cells should be used for antibody validation within 24-48 h post-transfection.
  6. To fix the cells, please follow steps 1.2.8-1.2.13.
  7. Store the fixed samples in 200 µL of DPBS with 5 mM sodium azide until ready for imaging. Samples can be stored at 4 °C for up to 1 month.
    CAUTION: Sodium azide is acutely toxic and can be fatal if ingested. Wear appropriate PPE when handling this reagent. In the event of skin contact, immediately wash the affected area thoroughly with soap and water. Dispose of the reagent according to the user's institution's guidelines and all applicable local, state, and federal regulations.
  8. Follow steps 2.1-2.12 to set up the microscope and mount the sample. Identify a cell expressing the protein of interest using the fluorescence signal.
  9. For reference, capture a TIRF image of the cell. If a fluorescent protein in the same imaging channel as the mSAL antibody is used, perform photobleaching of the fluorescent protein. To photobleach the protein, utilize the maximum laser power for > 10 min until no signal is observed.
  10. Add the prepared imaging buffer to the well and start imaging. Analysis should be completed as described previously (step 4).
  11. Open both the reconstructed image and the TIRF image of the fluorescent protein acquired prior to the mSAL acquisition. Assess mSAL antibody binding locations by comparing the appearance and localizations of the expressed protein, depicted in the diffraction-limited TIRF image, to the mSAL image.

Results

The mSAL technique described in this protocol enables visualization of single antibody binding events to their corresponding plasma membrane protein targets, producing a super-resolution map of membrane protein distribution. CD81 is used here as a representative example, though the approach is readily adaptable to other membrane proteins.

Immobilizing Jurkat T cells with sufficient spacing between cells ensures membrane epitope accessibility by antibodies and glass surface for fiducials to settle (Figure 1A). In addition, pipetting techniques can directly affect the preservation of delicate membrane structures. Therefore, it is encouraged to pipette gently during solution exchange steps. Pipetting the solution into a corner of the well while holding the sample chamber at an angle helps prevent the solution flow from dislodging cells and disrupting the delicate membrane structures (Figure 1B).

Gold colloids are added as fiducial markers to correct for any lateral stage drift during the image reconstruction. The perfect focus system (PFS) on the microscope system maintains the sample in focus during long-term imaging by automatically correcting for focus drift using near-infrared light to track the coverslip surface and continuously adjusting the objective height. If using antibodies conjugated to a near-infrared-exciting fluorophore, the PFS function should be disabled. It is necessary to confirm the presence of sufficient gold colloids in the field of view before imaging. Figure 2A, B displays a U2OS cell in brightfield (Figure 2A) and 488 nm (Figure 2B) channels, with white arrows indicating the locations of the gold colloids. Gold colloids exhibit autofluorescence under laser illumination. It is important to note that while gold colloids may appear in the brightfield view, not all colloids exhibit visible autofluorescence in the fluorescent channels. After confirming a sufficient number of gold colloids in the region of interest, the illumination angle must be adjusted to or right below the critical angle for total internal reflection fluorescence (TIRF) or highly inclined and laminated optical sheet (HILO) illumination. If the incident angle of laser illumination is suboptimal, it may reduce the signal-to-noise ratio (SNR) of the single-molecule events by increasing the background signal (Figure 2C,D).

Imaging parameters such as antibody concentration, NII, and the frequency of photobleaching steps significantly affect the quality of the reconstructed image. Thus, optimizing these parameters is crucial to maintaining sufficient single-molecule densities in each frame without spatial overlap. Spatial overlap in single-molecule binding events reduces the accuracy of single-emitter fitting during downstream analysis. An NII scan can be performed to identify the optimal time-lapse interval, where several NIIs are tested to monitor the antibody binding kinetics at a given antibody concentration (Figure 3A). While stage drift is observed during mSAL acquisitions, it does not influence the evaluation of single-molecule localization density for NII optimization, because the emitter sparsity is assessed within individual frames and reflects the antibody binding rate in a single frame rather than the accumulation of events across the acquisition. Our previous work demonstrates that increasing concentration and decreasing the NII to maximize the throughput increases the amount of nonspecific binding captured during imaging23. For CD81, 1 nM antibody concentration and an NII of 0 or 1 s only captured minimal binding events between the first and tenth frames. Upon an increase of the NII to 20 s, several binding events occurred between the first and tenth frames, which manifested as multiple large puncta, suggesting high spatial overlap. However, an NII of 5 s revealed several binding events between the first and tenth frames, with minimal spatial overlap. With this observation, we proceeded with an NII of 5 s. Thereafter, we reduced the antibody concentration to 0.5 nM to further minimize the spatially overlapping events between images and monitored accumulation over several frames (Figure 3B). At frame 5, several binding events are observed that are not spatially overlapping. By frame 10, some larger puncta can be observed, indicating the presence of spatially overlapping events. Therefore, a photobleaching step was implemented after every fifth frame during the acquisition.

For drift correction, the selected fiducial marker should be visible within the selected region through the entire acquisition. The reconstruction parameters may need adjustments between the analysis of the gold colloid and the full image reconstruction. After saving the drift correction file, reconstruct the entire field of view (Figure 4A) and apply the drift correction and filtering (Figure 4B). The resulting mSAL image and individual localizations can then be saved. The resulting mSAL image provides a histogram of the interactions between the antibody and its epitopes on the cell.

Antibodies can exhibit off-target binding46. In SAL, off-target events may manifest as low-density interactions23. Depending on the membrane protein of interest, it may naturally exist in a clustered organization, such as the tetraspanin CD81, known to associate with itself and form tetraspanin-enriched microdomains38,39,40,41,42. A clustering algorithm can be employed to discern high-density binding events from low-density, nonspecific binding. To this end, a density-based spatial clustering of applications with noise (DBSCAN) implementation, compiled in MATLAB, was used to cluster the single-molecule binding events based on local epitope density. Figure 5A displays a scatter plot of the CD81 mSAL localizations on a Jurkat T cell prior to clustering, and Figure 5B displays the corresponding normalized Gaussian rendering. Figure 5C displays the scatter plot of the clustered mSAL localizations after DBSCAN. The colors are assigned to individual clusters identified by the algorithm. A .csv file is exported containing the clustered single-molecule localizations. The .csv file can then be imported back into ThunderSTORM, and a normalized Gaussian image of the clustered localizations can then be created (Figure 5D).

Unlike larger organelles and cellular structures, such as mitochondria or microtubules, membrane proteins typically lack readily identifiable structures when labeled with fluorescent markers. Therefore, the antibody specificity may need to be validated by non-imaging methods. Nevertheless, a fluorescent protein (FP)-based approach enables a visual comparison between mSAL localizations and the known distribution of the tagged protein, providing a visual assessment of the membrane structure on which the target protein is distributed (Figure 6A). After the mSAL image is reconstructed (Figure 6B) and processed using DBSCAN (Figure 6C), the binding locations of the antibodies can be compared to the areas showing fluorescence intensity in the diffraction-limited image. The reconstructed super-resolution image can also reveal fine membrane features, such as the microvilli (Figure 6D). As a negative control for mSAL, dye-conjugated isotype control antibodies may be used. Overall, the combination of mSAL, density-based clustering analysis, and antibody binding validation establishes a workflow capable of uncovering the spatial distribution of membrane protein targets, such as CD81, and the nanoscale membrane topology, such as the microvilli.

Microscopy of bacterial colonies; pipetting in PCR workstation for DNA analysis experiment.
Figure 1: Preparation of Jurkat T cells for mSAL. (A) Brightfield image with 10x magnification of Jurkat T cells after landing on a poly-L-lysine surface for 30 min. (B) Image showing the recommended angle for the slow addition of the fixation solution to best preserve fine membrane structures. Please click here to view a larger version of this figure.

Cell imaging comparing TIRF and HILO microscopy techniques; fluorescence microscopy results.
Figure 2: Fiducial markers and single-molecule images between total-internal-reflection fluorescence (TIRF) and highly inclined laminated optical sheet (HILO) for mSAL acquisition. (A) Brightfield image of a U2OS cell. Arrows indicate the location of gold colloids. (B) TIRF image of the same U2OS cell with 488 nm laser excitation. Arrows indicate the location of visible gold colloids. (C) TIRF and (D) HILO images of the same U2OS cell during mSAL imaging. Scale bars: 10 µm. Please click here to view a larger version of this figure.

Fluorescence microscopy time-lapse; particle dynamics analysis over 20 seconds; scientific study.
Figure 3: Optimization of the NII and photobleaching step. (A) TIRF images of the same cell imaged with NIIs of 0 s, 1 s, 5 s, and 20 s, showing the accumulation of antibody binding events between the first and tenth frame with an antibody concentration of 1 nM. (B) TIRF images of a Jurkat T cell at different frames displaying the accumulation of antibody binding events with a 5 s NII and an antibody concentration of 1 nM. Scale bars: 2 µm. Please click here to view a larger version of this figure.

Fluorescence microscopy images of cellular structures; high-resolution details in insets.
Figure 4: mSAL data image reconstruction. (A) Reconstructed image of the mSAL acquisition prior to drift correction. A magnified view of the boxed region is shown to the right. (B) Reconstructed image of the mSAL acquisition after drift correction and filtering. A magnified view of the boxed region is shown to the right. Scale bars: 2 µm (A whole cell, B whole cell) and 500 nm (A magnified, B magnified). Please click here to view a larger version of this figure.

Cellular microscopy analysis; diagrams illustrating particle tracking and density visualization.
Figure 5: Density-based cluster analysis minimizes nonspecific binding background in mSAL data for a Jurkat T cell. (A) Scatter plot of mSAL data. Magnified views of the boxed regions are shown below. (B) Normalized Gaussian image of the unclustered mSAL data. Magnified views of the boxed regions are shown below. (C) Scatter plot of clustered mSAL data using DBSCAN. Magnified views of the boxed regions are shown below. Colors represent distinct clusters. (D) Normalized Gaussian image of the clustered mSAL data using DBSCAN. Magnified views of the boxed regions are shown below. Scale bars: 2 µm (whole cell), 500 nm (medium magnified), and 100 nm (fully magnified). Please click here to view a larger version of this figure.

CD81 protein imaging; microscopy, DBSACN-filtering; cell structure analysis; intensity graph results.
Figure 6: Validation of mSAL with CD81-mStayGold in a U2OS cell. (A) TIRF image of CD81-mStayGold. Magnified views of the boxes are shown to the right. (B) mSAL reconstructed image. Magnified views of the boxes are shown to the right. (C) mSAL reconstructed image after DBSCAN filtering. Magnified views of the boxes are shown to the right. (D) Line profiles of the lines in the magnified views in panels (A-C). Scale bars: 10 µm (whole cell), 1 µm (middle magnified), and 250 nm (right magnified). Please click here to view a larger version of this figure.

Discussion

The organization of plasma membrane proteins plays a vital role in regulating cell function. Conventional fluorescence microscopy techniques are limited by the diffraction limit of light, masking their nanoscale organization. SMLM overcomes this limitation by achieving spatial resolutions of 10-20 nm47, enabling the characterization of their nanoscale assemblies10,48,49. Clustering artifacts in SMLM arise from repeat detections of the same fluorophore50, which can give the appearance of protein clusters. The single-molecule detections depend on the photoswitching properties of the fluorophore and therefore do not fully represent the local epitope concentration51,52,53,54,55. PAINT-based techniques address this limitation by relying on single-molecule detections that occur through the transient binding and unbinding of probes to their targets56,57,58,59. mSAL is a PAINT-based technique that records single antibody-antigen interactions that constitute the single-molecule detections, thereby providing a readout of relative local epitope density. The direct probing of epitopes by antibodies can be leveraged to obtain semi-quantitative analysis of the local epitope organizations. SAL and mSAL offer several advantages over existing SMLM techniques. It can be performed on a standard TIRF microscope using any commercially available antibody and is compatible with both dye-conjugated primary or secondary antibodies. As mSAL does not depend on the photophysical properties of fluorophores, it supports the use of virtually any existing fluorophore and enables flexible multiplexing. The presence of multiple fluorophores per antibody also provides higher SNR for single-molecule events. However, mSAL is limited in throughput, and the slow binding kinetics of antibodies lead to longer NII and extended imaging times compared to standard SMLM techniques. Depending on the NII, mSAL could be 3-17 times slower than dSTORM and DNA-PAINT. A typical mSAL acquisition needs ~2,000 frames. At a 5 s NII, the total imaging time is ~3 h, and at a 20 s NII, the total imaging time is ~12 h. In contrast, typical dSTORM and DNA-PAINT acquisitions require 50,000-100,000 frames, but can be completed within 45 min to 1 h, owing to the rapid photoswitching kinetics of fluorophores in dSTORM and the fast DNA-oligo hybridization in DNA-PAINT. In addition, slow off-rates of certain antibodies may reduce achievable labeling densities, and labeling efficiencies depend strongly on the antibody and cell type used. Nonetheless, with optimized imaging parameters, our previous publications23,24 demonstrate that SAL achieves labeling densities that reflect local epitope distributions and yield super-resolution comparable to dSTORM and DNA-PAINT. As such, mSAL is best suited to applications requiring accurate representation of local nanoscale epitope arrangements and evaluating individual antibody-epitope interaction dynamics, whereas dSTORM and DNA-PAINT remain preferable for experiments prioritizing faster super-resolution imaging. In addition, sample preparation using the immunostaining process involves several buffer exchanges and washing steps, which can inadvertently disrupt delicate membrane structures, such as retraction fibers or microvilli. mSAL with dye-conjugated antibodies overcomes this challenge by eliminating the need for additional wash steps, as sample staining occurs simultaneously with image acquisition.

Time-lapse mSAL facilitates the evaluation of antibody labeling directly in cellular environments for super-resolution microscopy. This technique constitutes a robust approach for multiplexed SMLM23. The successful implementation of this strategy requires careful optimization of both dye-conjugated antibody concentration and NII. Appropriate antibody concentrations are necessary to achieve sufficient labeling density for single-molecule detection, while avoiding excessive signal overlap that diminishes spatial resolution. Likewise, optimization of the NII reduces temporal overlaps between fluorescent events and enhances selectivity for high-density binding interactions, thereby improving signal discrimination.

Collectively, these optimizations ensure accurate single-molecule localization and enhance the fidelity of super-resolution image reconstruction. The optimal antibody concentration for the CD81 antibody was 0.5 nM in Jurkat T cells and 2 nM in U2OS cells, paired with a 5 s NII. The photobleaching step was added after every 5 frames to reduce the capturing of accumulated single-molecule events. After acquiring 2,000 frames, the reconstructed images revealed the super-resolved distribution of CD81 on the cell membrane and microvilli.

Previous work on SAL23,24 demonstrated that specific binding events are localized as areas of high binding density, dependent on the local epitope concentration. Whereas conventional IF staining eliminates some nonspecific binding during washing steps, mSAL retains both specific and nonspecific events. To identify high-density regions, the clustering algorithm DBSCAN was applied to the mSAL antibody binding events. This approach is particularly useful for membrane proteins that are naturally organized into clusters, such as tetraspanins like CD8138,39,40,41,42. Applying a clustering algorithm also helps to remove the nonspecific binding. Although some nonspecific events outside the cell remained, DBSCAN removed a substantial amount of isolated, low-density single-molecule detections that were likely nonspecific. While complete removal of low-density events outside the cell is expected, it should be noted that, as with any clustering approach, there is an inherent compromise between noise reduction and data preservation in DBSCAN to ensure that the reconstructed image accurately reflects the underlying cluster. The clustering process may exclude some true binding events; however, the reduction in nonspecific background often outweighs the loss of a small fraction of specific detections. Although systematically quantifying the exclusion rate of specific events would be ideal, it is difficult to determine whether any individual event is truly specific without an independent reference. Given the complexity of protein-protein interactions, such benchmarking remains challenging. As a practical alternative, users can either assess the exclusion rate of events occurring outside the cell boundary or use a negative control mSAL acquisition with a relevant dye-conjugated isotype control antibody, where all detections are unequivocally nonspecific. While clustering-based filtering may slightly underestimate absolute event counts, the relative epitope distributions and comparative analyses in high-density areas remain reliable, as the filtering removes low-density (likely non-specific) events rather than high-density (likely specific) events, preserving the semi-quantitative nature of mSAL.

Common challenges in applying the mSAL technique include suboptimal antibody concentrations relative to epitope density, slow antibody binding kinetics, suboptimal NII values, improper reconstruction parameters, and unsuitable DBSCAN inputs. Selecting an appropriate NII and antibody concentration combination is critical, and we encourage users to carefully review steps 3.1 and 3.2 of the protocol. Multiple iterations are often necessary to identify suitable conditions. Accurate reconstruction is equally essential, as inappropriate ThunderSTORM parameters can noticeably degrade image quality. Users are therefore encouraged to understand how each parameter influences localization detection and fitting before performing mSAL reconstruction. Lastly, the DBSCAN input parameters are crucial for achieving meaningful clustering of the single-molecule binding events. We recommend that users first inspect a scatter plot of noise localizations outside the cell or localizations of a negative control mSAL (with a corresponding isotype control) and compare it with representative localizations in the cell. The chosen minPTS within the epsilon should be greater than the noise localizations but below the localizations of true on-cell events, and epsilon should be large enough to encompass at least the minPTS threshold for on-cell clusters.

In addition to CD81, mSAL can be readily extended to a wide range of membrane proteins, including clinically relevant immunotherapeutic targets on the plasma membrane60,61,62. Therapeutic antibodies act through direct interactions with extracellular epitopes. Therefore, visualizing how these antibodies interact with their targets at the nanoscale can provide further insight into their target interaction efficiency as well as the off-target interaction tendencies of a therapeutic antibody. mSAL using therapeutic antibodies will provide insights into the spatial distribution and accessibility of their epitopes, as well as their interaction dynamics in the native cellular environment, which can guide the rational design of new therapeutic antibodies.

In conclusion, we present a technique for capturing and mapping the nanoscale organization of membrane proteins by probing their distribution with low concentrations of antibody. The advantages of this technique include reduced solution exchange to preserve fine membrane structures, semi-quantitative labeling of target proteins, and the versatility to be applied to any antibody-epitope system, including therapeutic antibodies and their corresponding membrane protein receptors.

Disclosures

All authors declare that they have no conflicts of interest to disclose.

Acknowledgements

Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health under Award Number R35GM146786 and the College of Liberal Arts and Sciences, University of Illinois Chicago. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Materials

List of materials used in this article
NameCompanyCatalog NumberComments
anti-CD81 Alexa Fluor 488 AntibodyThermo Fisher ScientificMA5-44132Imaging antibody
Benchtop centrifugeEppendorf5424
Bovine serum albuminMilliporeSigmaA7906-100G
Chambered cover glass, 8-wellCellvisC8-1-N
DMEM mediaThermo Fisher Scientific11960069
Dulbecco's Phosphate Buffered SalineThermo Fisher Scientific14190-144
Fetal bovine serumMilliporeSigmaF0926-500ML
Glutaraldehyde, 10 %Electron Microscopy Sciences16120
Gold colloids, 100 nmTed Pella15711−20Fiducial marker
Immersion oil Cargille Laboratories16245
Inverted microscope equipped with TIRF moduleNikon InstrumentsEclipse Ti2-EMicroscope
Jurkat E6.1 cell lineMilliporeSigma88042803-1VLSuspension cells
Laser Launch, 6 lineOxxiusL6CC-CS8-1511Excitation laser
Lipofectamine 2000Thermo Fisher Scientific11668019Transfection reagent
Microcentrifuge tube, 1.5 mL VWR89000-028
Microscopy imaging softwareNikon InstrumentsNIS-Elements Advanced ResearchImage acqisition software
Objective, 100x/1.49 CFI APO TIRF Nikon InstrumentsMRD01991Microscope objective
Paraformaldehyde, 16 %Electron Microscopy Sciences15710
Penicillin-StreptomycinThermo Fisher Scientific15140-122Antibiotics
Poly-L-Lysine solution, 0.01% MilliporeSigmaP4707-50ML
Prime 95B sCMOS CameraTeledyne Vision Solutions01-PRIME-95B-R-M-16-CCamera
RPMI 1640 mediaThermo Fisher Scientific11875-093Media for suspension cells
Sodium azideThermo Fisher Scientific19038-1000
Sodium borohydrideMilliporeSigma213462-25G
T25 culture flasksThermo Fisher Scientific169900
TetraSpeck microspheresFiducial marker
Tissue culture dish, 100 mm Corning353003
U2OS cell lineATCCHTB-96Adherent cells

References

  1. Zulueta Diaz, Y., de las, M., Arnspang, E. C. Super-resolution microscopy to study membrane nanodomains and transport mechanisms in the plasma membrane. Front Mol Biosci. 11, 1455153(2024).
  2. Levental, I., Lyman, E. Regulation of membrane protein structure and function by their paralipidomes. Nat Rev Mol Cell Biol. 24 (2), 107(2022).
  3. Stephens, A. D., Wilkinson, T. Discovery of Therapeutic Antibodies Targeting Complex Multi-Spanning Membrane Proteins. Biodrugs. 38 (6), 769(2024).
  4. Chan, A. C., Martyn, G. D., Carter, P. J. Fifty years of monoclonals: the past, present and future of antibody therapeutics. Nat Rev Immunol. 25, 745-765 (2025).
  5. Rust, M. J., Bates, M., Zhuang, X. Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM). Nat Methods. 3 (10), 793-796 (2006).
  6. Heilemann, M., et al. Subdiffraction-resolution fluorescence imaging with conventional fluorescent probes. Angew Chem Int Ed Engl. 47 (33), 6172-6176 (2008).
  7. Jungmann, R., et al. Multiplexed 3D cellular super-resolution imaging with DNA-PAINT and Exchange-PAINT. Nat Methods. 11 (3), 313-318 (2014).
  8. Tobin, S. J., et al. Single molecule localization microscopy coupled with touch preparation for the quantification of trastuzumab-bound HER2. Sci Rep. 8 (1), 15154(2018).
  9. Wakefield, D. L., et al. Using quantitative single molecule localization microscopy to optimize multivalent HER2-targeting ligands. Front Med. 10, 1064242(2023).
  10. Ghosh, A., et al. Decoding the molecular interplay of CD20 and therapeutic antibodies with fast volumetric nanoscopy. Science. 387 (6730), eadq4510(2025).
  11. Lillemeier, B. F., et al. TCR and Lat are expressed on separate protein islands on T cell membranes and concatenate during activation. Nat Immunol. 11 (1), 90-96 (2010).
  12. Hu, Y. S., Cang, H., Lillemeier, B. F. Superresolution imaging reveals nanometer- and micrometer-scale spatial distributions of T-cell receptors in lymph nodes. Proc Natl Acad Sci U S A. 113 (26), 7201-7206 (2016).
  13. Rossboth, B., et al. TCRs are randomly distributed on the plasma membrane of resting antigen-experienced T cells. Nat Immunol. 19 (8), 821-827 (2018).
  14. Civita, S., et al. Nanoscale engagement of programmed death ligand 1 (PD-L1) in membrane lipid raft domains of cancer cells. FEBS J. 292 (22), 5979-5997 (2025).
  15. Lee, J., Sengupta, P., Brzostowski, J., Lippincott-Schwartz, J., Pierce, S. K. The nanoscale spatial organization of B-cell receptors on immunoglobulin M- and G-expressing human B-cells. Mol Biol Cell. 28 (4), 511(2017).
  16. Ferapontov, A., et al. Antigen footprint governs activation of the B cell receptor. Nat Commun. 14 (1), 976(2023).
  17. Mattila, P. K., et al. The actin and tetraspanin networks organize receptor nanoclusters to regulate B cell receptor-mediated signaling. Immunity. 38 (3), 461-474 (2013).
  18. Nerreter, T., et al. Super-resolution microscopy reveals ultra-low CD19 expression on myeloma cells that triggers elimination by CD19 CAR-T. Nat Commun. 10 (1), 3137(2019).
  19. Slaga, D., et al. Avidity-based binding to HER2 results in selective killing of HER2-overexpressing cells by anti-HER2/CD3. Sci Transl Med. 10 (463), eaat5775(2018).
  20. Haber, L., et al. Generation of T-cell-redirecting bispecific antibodies with differentiated profiles of cytokine release and biodistribution by CD3 affinity tuning. Sci Rep. 11 (1), 14397(2021).
  21. Abdelmotaleb, O., Schneider, A., Gassner, C., Märsch, S., Klein, C. The impact of CD3 affinity-attenuation on T cell engaging bispecific antibodies: is it really that simple. Expert Opin Drug Discov. 20 (8), 943-949 (2025).
  22. Dang, K., et al. Attenuating CD3 affinity in a PSMAxCD3 bispecific antibody enables killing of prostate tumor cells with reduced cytokine release. J Immunother Cancer. 9 (6), e002488(2021).
  23. Gunasekara, H., et al. Superresolution Imaging with Single-Antibody Labeling. Bioconjugate Chem. 34 (5), 825-833 (2023).
  24. Perera, T., Gunasekara, H., Hu, Y. S. Single-Molecule Localization Microscopy Using Time-Lapse Imaging of Single-Antibody Labeling. Curr Protoc. 3 (10), e908(2023).
  25. Küppers, R. CD81 as target for B cell lymphomas. J Exp Med. 216 (7), 1469-1470 (2019).
  26. Vences-Catalán, F., et al. CD81 is a novel immunotherapeutic target for B cell lymphoma. J Exp Med. 216 (7), 1497-1508 (2019).
  27. Vences-Catalán, F., et al. Targeting the tetraspanin CD81 reduces cancer invasion and metastasis. Proc Natl Acad Sci U S A. 118 (24), e2018961118(2021).
  28. Bailly, C., Thuru, X. Targeting of Tetraspanin CD81 with Monoclonal Antibodies and Small Molecules to Combat Cancers and Viral Diseases. Cancers. 15 (7), 2186(2023).
  29. Schindelin, J., et al. Fiji: an open-source platform for biological-image analysis. Nat Methods. 9 (7), 676-682 (2012).
  30. Ovesný, M., Křížek, P., Borkovec, J., Svindrych, Z., Hagen, G. M. ThunderSTORM: a comprehensive ImageJ plug-in for PALM and STORM data analysis and super-resolution imaging. Bioinformatics. 30 (16), 2389-2390 (2014).
  31. Ovesný, M., Křížek, P., Borkovec, J., Svindrych, Z., Hagen, G. M. Image analysis for single-molecule localization microscopy. In: Super-Resolution Imaging in Biomedicine. , 79-94 (2016).
  32. Thompson, R. E., Larson, D. R., Webb, W. W. Precise nanometer localization analysis for individual fluorescent probes. Biophys J. 82 (5), 2775-2783 (2002).
  33. Stallinga, S., Rieger, B. Accuracy of the Gaussian point spread function model in 2D localization microscopy. Opt Express. 18 (24), 24461-24476 (2010).
  34. Huang, F., Schwartz, S. L., Byars, J. M., Lidke, K. A. Simultaneous multiple-emitter fitting for single molecule super-resolution imaging. Biomed Opt Express. 2 (5), 1377-1393 (2011).
  35. Smith, C. S., Joseph, N., Rieger, B., Lidke, K. A. Fast, single-molecule localization that achieves theoretically minimum uncertainty. Nat Methods. 7 (5), 373-375 (2010).
  36. Chatterjee, T., et al. Direct Kinetic Fingerprinting and Digital Counting of Single Protein Molecules. Proc Natl Acad Sci U S A. 117 (37), 22815-22822 (2020).
  37. A density-based algorithm for discovering clusters in large spatial databases with noise. Ester, M., Kriegel, H. P., Sander, J., Xu, X. Int Conf Knowledge Disc Data Mining, 240, 226-231 (1996).
  38. Termini, C. M., Gillette, J. M. Tetraspanins Function as Regulators of Cellular Signaling. Front Cell Dev Biol. 5, 242379(2017).
  39. Levy, S., Shoham, T. Protein-protein interactions in the tetraspanin web. Physiology. 20, 218-224 (2005).
  40. Rubinstein, E., et al. CD9, CD63, CD81, and CD82 are components of a surface tetraspan network connected to HLA-DR and VLA integrins. Eur J Immunol. 26 (11), 2657-2665 (1996).
  41. Schmidt, S. C., Massenberg, A., Homsi, Y., Sons, D., Lang, T. Microscopic clusters feature the composition of biochemical tetraspanin-assemblies and constitute building-blocks of tetraspanin enriched domains. Sci Rep. 14 (1), 2093(2024).
  42. Zuidscherwoude, M., et al. The tetraspanin web revisited by super-resolution microscopy. Sci Rep. 5, 12201(2015).
  43. Comparative performance analysis of K-means and DBSCAN clustering algorithms on various platforms. Shahriar, N., Akib Al Faisal, S. M., Pinjor, M. M., Sharif Zobayer Rafi, M. A., Rahman Sarkar, A. Proceedings of the 22nd International Conference on Computer and Information Technology (ICCIT), , 1-6 (2019).
  44. Hahsler, M., Piekenbrock, M., Doran, D. dbscan: Fast Density-Based Clustering with R. J Stat Softw. 91 (1), 1-30 (2019).
  45. Pedregosa, F., et al. Machine Learning with PyTorch and Scikit-Learn: Develop machine learning and deep learning models with Python. J Mach Learn Res. 12, 2825-2830 (2011).
  46. Buchwalow, I., Samoilova, V., Boecker, W., Tiemann, M. Non-specific binding of antibodies in immunohistochemistry: fallacies and facts. Sci Rep. 1, 28(2011).
  47. Bond, C., Santiago-Ruiz, A. N., Tang, Q., Lakadamyali, M. Technological advances in super-resolution microscopy to study cellular processes. Mol Cell. 82 (2), 315-332 (2022).
  48. Garcia-Parajo, M. F., Cambi, A., Torreno-Pina, J. A., Thompson, N., Jacobson, K. Nanoclustering as a dominant feature of plasma membrane organization. J Cell Sci. 127 (23), 4995-5005 (2014).
  49. Tan, H. L., Bungert-Plümke, S., Kortzak, D., Fahlke, C., Stölting, G. Determination of oligomeric states of proteins via dual-color colocalization with single molecule localization microscopy. eLife. 11, e76631(2022).
  50. Annibale, P., Vanni, S., Scarselli, M., Rothlisberger, U., Radenovic, A. Identification of clustering artifacts in photoactivated localization microscopy. Nat Methods. 8 (7), 527-528 (2011).
  51. Dempsey, G. T., Vaughan, J. C., Chen, K. H., Bates, M., Zhuang, X. Evaluation of fluorophores for optimal performance in localization-based super-resolution imaging. Nat Methods. 8 (12), 1027-1036 (2011).
  52. Shivanandan, A., Deschout, H., Scarselli, M., Radenovic, A. Challenges in quantitative single molecule localization microscopy. FEBS Lett. 588 (19), 3595-3602 (2014).
  53. Endesfelder, U., Heilemann, M. Art and artifacts in single-molecule localization microscopy: beyond attractive images. Nat Methods. 11 (3), 235-238 (2014).
  54. Lambert, T. J., Waters, J. C. Navigating challenges in the application of superresolution microscopy. J Cell Biol. 216 (1), 53-63 (2017).
  55. Annibale, P., Vanni, S., Scarselli, M., Rothlisberger, U., Radenovic, A. Quantitative photo activated localization microscopy: unraveling the effects of photoblinking. PloS One. 6 (7), e22678(2011).
  56. Jungmann, R., et al. Quantitative super-resolution imaging with qPAINT. Nat Methods. 13 (5), 439-442 (2016).
  57. Unterauer, E. M., Jungmann, R. Quantitative Imaging With DNA-PAINT for Applications in Synaptic Neuroscience. Front Synaptic Neurosci. 13, 798267(2021).
  58. Stein, J., Stehr, F., Jungmann, R., Schwille, P. Calibration-free counting of low molecular copy numbers in single DNA-PAINT localization clusters. Biophys Rep. 1 (2), 100032(2021).
  59. Youn, Y., et al. Quantitative DNA-PAINT imaging of AMPA receptors in live neurons. Cell Rep Methods. 3 (2), 100408(2023).
  60. Yin, H., Flynn, A. D. Drugging Membrane Protein Interactions. Annu Rev Biomed Eng. 18, 51-76 (2016).
  61. Jin, S., et al. Emerging new therapeutic antibody derivatives for cancer treatment. Signal Transduct Target Ther. 7 (1), 39(2022).
  62. To'a Salazar, G., Huang, Z., Zhang, N., Zhang, X. -G., An, Z. Antibody therapies targeting complex membrane proteins. Engineering. 7 (11), 1541-1551 (2021).

Reprints and Permissions

Tags

Plasma MembraneSuper Resolution MicroscopyCD81 ProteinDensity Based ClusteringAntibody BindingDrift CorrectionTherapeutic Antibodies
Single-Molecule Localization Microscopy of Membrane Proteins using Single-Antibody Labeling
Contact Us
Recommend to Library

Privacy

Terms of Use

Policies

Copyright © 2026 MyJoVE Corporation. All rights reserved.