Single-Molecule Localization Microscopy Using Time-Lapse Imaging of Single-Antibody Labeling

Abstract

In single-molecule localization microscopy (SMLM), immunofluorescence (IF) staining affects the quality of the reconstructed superresolution images. However, optimizing IF staining remains challenging because IF staining is a one-step and irreversible process. Sample labeling through reversible binding presents an alternative strategy, but such techniques require significant technological advancements to enhance the dissociation of labels without sacrificing their binding specificity. In this work, we introduce time-lapse imaging of single-antibody labeling. Our versatile technique utilizes commercially available dye-conjugated antibodies. The method controls the antibody concentrations to capture single-antibody labeling of subcellular targets, thereby achieving SMLM through the labeling process. We further demonstrate dual-color single-antibody labeling to enhance the sample labeling density. The new approach allows the evaluation of antibody binding at the single-antibody level and within the cellular environment. This comprehensive guide offers step-by-step instructions for time-lapse imaging of single-antibody labeling experiments. The protocol enables the application of the single-antibody labeling technique to a wide range of targets.

Basic Protocol 1:

Sample preparation for single-antibody labeling

Basic Protocol 2:

Data acquisition for single-molecule localization microscopy

Alternative Protocol 1:

Dual-color single-antibody labeling using OptoSplit II equation

Basic Protocol 3:

Image analysis

INTRODUCTION

In single-molecule localization microscopy (SMLM), immunofluorescence (IF) staining remains one of the most widely used techniques for sample labeling. IF staining immobilizes the molecular target dye-conjugated primary or secondary antibodies. To this end, the antibody quality determines the quality of the superresolution images. Most common SMLM techniques utilize IF staining, including direct stochastic optical reconstruction microscopy (dSTORM) and DNA-based points accumulation for imaging in nanoscale topography (DNA-PAINT) (Heilemann et al., 2008; Jungmann et al., 2010, 2014; Rust et al., 2006; Schnitzbauer et al., 2017). The optimization of sample labeling often involves empirical adjustments to the antibody concentration and other IF staining conditions. However, evaluating antibody labeling is nearly impossible until after the superresolution image reconstruction is complete.

An alternative labeling strategy is through fluorescently labeled probes that transiently bind to the target. This approach offers several advantages, including stoichiometric labeling, reduced linkage error, flexibility in the choice of the fluorophore, enhanced labeling densities, and multiplexing imaging. Universal PAINT (u-PAINT) utilizes fluorescently labeled ligands and antibodies to label tagged and endogenous membrane proteins (Giannone et al., 2010). Protein-PAINT (pPAINT) employed signaling proteins to investigate T-cell signaling with multiplexing capability (Farrell et al., 2022). Another related approach is image reconstruction by integrating the exchangeable single-molecule localization (IRIS). IRIS utilizes rapidly exchanging probes to achieve high-density labeling and multiplexed imaging (Kiuchi et al., 2015). Recently, the IRIS technology has been extended to using fast-dissociating antibodies produced through hybridoma technology and antibody fragments (Miyoshi et al., 2021) and engineered fast-dissociating antibody fragments against molecular epitope tags generated through protein engineering (Zhang et al., 2022). Reversible binding has also been chemically enhanced in the superresolution census of molecular epitope tags (SR-COMET) technique (Gunasekara et al., 2022). While transient labeling represents significant technological advancements, there is still a need for a more versatile approach that uses readily available antibodies. To this end, we introduced an SMLM technique using time-lapse imaging of single-antibody labeling (Fig. 1; Gunasekara et al., 2023).

Figure 1: Overview of the concept of time-lapse imaging of single-antibody labeling.

(a) Schematic illustration of the single-molecule localization microscopy (SMLM) using single-antibody labeling on a fixed cell. (b) Time-lapse imaging with single-antibody labeling (optimization of antibody concentration, NII selection, and imaging). (c) Reconstructed superresolution image from single-antibody labeling.

This protocol presents a comprehensive step-by-step guide for single-antibody labeling, consisting of three Basic Protocols and an Alternate Protocol. Basic Protocol 1 focuses on the sample preparation steps for imaging microtubules and mitochondria in U2OS cells using dye-conjugated primary or secondary antibodies. The protocol can be modified and adapted for imaging other subcellular targets in different cells. Basic Protocol 2 covers time-lapse imaging with single-antibody labeling, multiplexed imaging with single-antibody labeling, and dual-color single-target imaging with single-antibody labeling. Basic Protocol 2 details optimizing antibody concentrations and non-illuminating intervals (NII), while Alternate Protocol 1 details dual-color imaging using OptoSplit II. Basic Protocol 3 provides steps for single-molecule analysis and superresolution image reconstruction using the ThunderSTORM ImageJ plug-in. Additionally, we provide critical tips for sample preparation and imaging procedures, a troubleshooting guide and timing information for planning the experiments.

BASIC PROTOCOL 1: SAMPLE PREPARATION FOR SINGLE-ANTIBODY LABELING

The sample preparation follows a standard IF staining protocol. This protocol provides a guide for preserving subcellular structures (microtubules and mitochondria) in U2OS cells and the sample preparation for two distinct single-antibody labeling approaches: 1) single-antibody labeling using dye-conjugated primary antibodies and 2) single-antibody labeling using dye-conjugated secondary antibodies on samples stained with unconjugated primary antibodies. For additional information on solution preparation, refer to the reagents and solution section. All the steps are conducted at room temperature unless otherwise specified.

CAUTION: Paraformaldehyde is a carcinogenic, toxic chemical. Wear protective gloves, protective clothing, and eye protection. Avoid breathing vapors. Wash skin thoroughly after handling.

CAUTION: Glutaraldehyde is a toxic chemical. Harmful if swallowed or inhaled. Causes skin burns and eye damage. Wear protective gloves, protective clothing, and eye protection. Avoid ingestion and inhalation. Wash skin thoroughly after handling.

CAUTION: Sodium azide is a hazardous substance. Fatal if ingested, comes into contact with skin if inhaled. Wear protective gloves, protective clothing, and eye protection. Wash skin thoroughly after handling.

Materials

U2OS cell line (ATCC, HTB-96)

U2OS cells line expressing HA-α-tubulin (Gunasekara et al., 2022)

DMEM (Fisher Scientific, 11960069, 500 mL)

Fetal Bovine Serum (MilliporeSigma, F0926, 500 mL)

Penicillin-streptomycin (Fisher Scientific, 15140–122, 100 mL, 10,000 units/mL penicillin and 10,000 μg/mL streptomycin)

L-glutamine (Fisher Scientific, 25030– 081, 100 mL)

Trypsin (Fisher Scientific, Trypsin-EDTA, 25200-056, 0.25%)

DPBS (Thermo Fisher Scientific, 14190-144, 500 mL)

Paraformaldehyde (Electron Microscopy Sciences, 15710, 16%)

Glutaraldehyde (Electron Microscopy Sciences, 16120, 10%)

Bovine serum albumin (BSA) (MilliporeSigma, A7906-100G)

Goat serum (Thermo Fisher Scientific, New Zealand origin, 16210072, 500 mL)

Unconjugated primary antibodies (Table 1)

Table 1:

Unconjugated antibodies used for IF staining.

Antibody	Target	Antibody concentration (nM)
ReadyTag anti-HA (12CA5) (Bio X Cell, RT0268, 25 mg)	HA tag on α-tubulin (microtubule)	67
Rabbit polyclonal Tom20 (Thermo Fisher Scientific, MA5-32148)	Tom20 (mitochondria)	27

MES (MilliporeSigma, M3671-50G, MW 195.24)

Potassium chloride (Fisher Scientific, P217–500, MW 74.55)

Magnesium chloride hexahydrate (ACROS Organics, 413415000, MW 203.31)

EGTA (MilliporeSigma, E3889-25G, MW 380.35)

Triton X-100 (MilliporeSigma, X100-1 L)

Sodium hydroxide (Fisher Chemical, S318-500)

Hydrochloric acid (Sigma Aldrich, 258148-500ML)

Sodium azide (ACROS Organics, UN 1687, MW 65.01)

Fixation buffer: paraformaldehyde (3.7%, v/v), glutaraldehyde (0.1%, v/v) in DPBS

Blocking buffer: BSA (5%, w/v) or goat serum (10%, v/v) containing Triton X-100 (0.3%, v/v) in DPBS

Post-fixation buffer: paraformaldehyde (3.7%, v/v) in DPBS

Tissue culture dishes (Falcon, 353003, 100x20 mm)

8-well chambered cover glasses (Thermo Fisher Scientific, 55411; Cellvis, C8-1-N)

Parafilm (Amcor, PM996)

Syringe filters (SyringeFilter.com, LLC, PVDF Filter, 30mm, 0.45 μm)

Serological pipettes (Fisherbrand, 2 mL, 13-678-11C; 5 mL, 13-678-11D; 10 mL, 13-678-11E)

Pipettes (Gilson Pipetman L P1000L; L P200L; L P120L; L P2L)

Biological safety cabinet (Thermo Fisher Scientific,1300 Series A2)

Incubator (Thermo Fisher Scientific, Series 8000 WJ)

Water purification system (Thermo Scientific, Barnstead GenPure xCAD Plus)

Analytical balance (Sartorius, Quntix Precision Balance 124-1S)

Water bath (Fisher Scientific, Isotemp GDP 10)

Orbital shaker (LE8S adjustable lab oscillator)

Tissue culture microscope (Nikon, ECLIPSE Ts2, Inverted microscope)

pH meter (Thermo Fisher Scientific, Orion Dual Star)

Protocol for cell seeding and fixation

1. Maintain the U2OS cells in a humidified atmosphere with 5% CO₂ at 37 °C. Subculture when the cells reach ~70% confluency.

   Culture U2OS cell line in DMEM supplemented with 10% fetal bovine serum, 2 mM L-glutamine, 100 units/mL penicillin, and 100 μg/mL streptomycin.

2. Seed ~5,000 cells per well using 400 μL of culture medium in an 8-well chambered cover glass. Let the cells grow at 37 °C and 5% CO₂ in an incubator. After 36 hr, the cells are ready to be fixed.

   After 36 hr, the cells achieve a confluency of ~40%.

3. Wash cells with 400 μL of DPBS once.

4. Add 200 μL of fixation buffer and incubate the sample for 15-20 min at room temperature.

5. Remove the fixation buffer and wash the sample two times with DPBS.

6. Add fresh DPBS. Fixed samples can be stored at 4 °C.

   (Optional) Add sodium azide solution (5 mM, in DPBS) to inhibit bacterial growth.

   Cover sample chambers with parafilm during storage to prevent the cells from drying out.

Sample preparation for single-antibody labeling using dye-conjugated primary antibodies

• 7)Add 200 μL of blocking buffer (BSA, 5% w/v or goat serum 10% v/v) containing permeabilization reagent (Triton X-100, 0.3% v/v) to the sample. Incubate for 0.5-2 hr at room temperature.

  The blocking buffer contains either BSA (5% w/v) or goat serum (10% v/v) if a goat-host secondary antibody is used. Permeabilization is done by adding Triton X-100 (0.3% w/v) to the blocking buffer. The permeabilization step should be skipped if labeling plasma membrane proteins. The permeabilization step should also be skipped if cell extraction was performed during fixation, i.e., use of the cytoskeletal buffer (see details in fixation buffer under REAGENTS AND SOLUTIONS).

• 8)(Optional) Add 200 μL of post-fixation buffer containing paraformaldehyde (3.7% v/v) in DPBS. Incubate the sample for 10 min at room temperature.

  This step preserves the sample blocking for single-antibody labeling.

  This step may be omitted if the antibody no longer effectively stains the sample or results in poor staining after post-fixation. During the initial stages of our protocol development, we observed that omitting this step resulted in a noticeable increase in non-specific binding events.

• 9)(Optional) Wash the sample two times with DPBS.
• 10)Add 200 μL of fresh DPBS. The sample is ready for time-lapse imaging with single-antibody labeling using dye-conjugated primary antibodies as described in Basic Protocol 2.

  (Optional) Add sodium azide (5 mM, in DPBS) to inhibit bacterial growth.

  Samples can be stored at 4 °C. Cover sample chambers with parafilm during storage to prevent the cells from drying out.

Sample preparation for single-antibody labeling using dye-conjugated secondary antibodies

• 11)Add 200 μL of blocking buffer (BSA, 5% w/v or goat serum, 10% v/v) containing permeabilization reagent (Triton X-100, 0.3% v/v) to the sample. Incubate for 0.5-2 hr at room temperature (same as Step 7).
• 12)Dilute the primary antibody in 200 μL of blocking buffer to a final concentration within the manufacturer-recommended range.

  In our study, we used concentrations of 4 μg/mL (27 nM) for full-length IgG (Table 1).

• 13)Replace the blocking buffer with the diluted primary antibody. Incubate the sample for 1-4 hr at room temperature.
• 14)Wash the sample three times with DPBS.
• 15)Add 200 μL of post-fixation buffer containing paraformaldehyde (3.7% v/v) in DPBS. Incubate the sample for 10 min at room temperature.

  This step preserves the primary antibody staining on the sample.

• 16)Wash the sample three times with DPBS. Add 200 μL of fresh DPBS after washing.

  The sample is ready for time-lapse imaging with single-antibody labeling using dye-conjugated secondary antibodies as described in Basic Protocol 2.

  Samples can be stored at 4 °C. Cover sample chambers with parafilm during storage to prevent the cells from drying out.

  (Optional) Add sodium azide solution (5 mM, in DPBS) to inhibit bacterial growth during storage.

BASIC PROTOCOL 2: DATA ACQUISITION FOR SINGLE-MOLECULE LOCALIZATION MICROSCOPY

The following protocol provides the steps for three implementations of single-antibody labeling: 1) time-lapse imaging with single-antibody labeling, 2) multiplexed imaging with single-antibody labeling, and 3) dual-color single-target imaging with single-antibody labeling.

The first section of the protocol outlines the steps for a basic single-antibody labeling experiment. This section includes the steps for initiating the imaging process, identifying the optimal antibody concentration, selecting the optimal NII, and setting up the imaging parameters for a subcellular target. Fiducial markers need to be added to the sample for drift correction. A short trial run must be performed to determine the optimal antibody concentration and NII before proceeding with extended acquisitions. This protocol is written for using an inverted microscope (Nikon Instruments, Eclipse Ti2E) with 100×/1.49 oil objective and a Prime 95B sCMOS camera. The laser power densities indicated in the steps are measured after the objective. Imaging is performed using a total-internal-reflection fluorescence (TIRF) microscopy setup. Data acquisition is performed using ND acquisition in Nikon NIS-Elements software (version 5.21.03).

The second section of the protocol outlines the extended steps for multiplexed single-antibody labeling using different antibodies. The third section of the protocol focuses on dual-color single-antibody labeling of a single target. The rationale for developing this assay is that single-antibody labeling can yield lower sample labeling densities than other SMLM techniques. Dual-color single-antibody labeling increases the labeling density while minimizing the spatial overlap between molecules. Additionally, an Alternate Protocol describes the use of OptoSplit II for dual-color imaging.

Materials

Chamber slide with a target of interest prepared according to Basic Protocol 1

Dye-conjugated antibodies (Table 2)

Table 2:

Dye-conjugated antibodies used for time-lapse imaging with single-antibody labeling.

Antibody	Target	Antibody concentration (nM)
ReadyTag anti-HA (12CA5) (Bio X Cell, RT0268, 25 mg)-Alexa Fluor 6471	HA tag on α-tubulin (microtubule)	0.5
F(ab′)2-Goat anti-Mouse IgG Alexa Fluor 647 (Thermo Fisher Scientific, A21237)	Anti-HA (12CA5)	0.063
α-tubulin monoclonal antibody (DMIA) Alexa Fluor 488 (Thermo Fisher Scientific, 53– 4502–82)	α-tubulin (microtubule)	0.003
Goat anti-rabbit Alexa Fluor 647 IgG (Thermo Fisher Scientific, Superclonal Recombinant, A27040, 1 mg)	Rabbit polyclonal Tom20	0.125

¹In-house dye conjugation with AF 647 (Gunasekara et al., 2022).

DPBS (Thermo Fisher Scientific,14190-144, 500 mL)

Fiducial markers (TetraSpeck microspheres, Thermo Fisher Scientific, T7280, 0.2 μm, or gold colloids, Ted Pella Inc., 15711–20, 100 nm)

Imaging buffer: Dye-conjugated antibodies diluted in DPBS (refer Table 2 for antibody concentrations used in our study)

Parafilm (Amcor, PM996)

Water purification system (Thermo Scientific, Barnstead GenPure xCAD Plus)

Oil-immersion objective (100x Apo TIRF objective, numerical aperture 1.49, working distance 0.12)

Immersion oil (Cargille Laboratories, 16245, refractive index, 1.5150)

Inverted fluorescence microscope (Nikon Instruments, Eclipse Ti2E with Perfect Focus System)

Quad-band set (Chroma Technology, TRF89901-ET-405/488/561/640nm)

Emission filters (Chroma Technology, ET450/50m, ET525/50m, ET600/50m, ET700/75m)

Laser unit (Nikon, LUN-F, serial number 3-476)

sCMOS camera (Prime 95B, serial number A18B203004)

Image acquisition software (Nikon, NIS Elements version 5.21.03)

Power meter (Thorlabs, PM100D)

Initiating data acquisition

1. Allow the sample to reach room temperature before proceeding if stored at 4 °C.

2. Turn on the microscope, camera, acquisition computer, and lasers.

   If the Prime 95B sCMOS camera is connected to the acquisition computer via Peripheral Component Interconnect Express (PCIe), ensure to turn on the camera before starting the computer.

3. Select the objective (e.g., 100x/1.49 oil objective) and apply a layer of index-matching immersion oil.

4. Position the sample on the microscope stage and raise the objective (e.g., 100x/1.49 oil objective) until the immersion oil contacts the sample.

   If the objective carries an adjustable correction collar ring, adjust it to match the thickness of the cover glass. For instance, the thickness of cover glass #1 is 0.13-0.16 mm and the correction ring needs to be set at the corresponding marking at 23 °C (white color marking) for image acquisition.

   In the Nikon Eclipse Ti2E system, the PFS (perfect focus system) maintains the focus during imaging. Raise the objective using the focus knob until the PFS is engaged (PFS indicator ‘on’) and mount the sample.

5. Open the acquisition software that controls the data acquisition.

6. Choose the optical configuration (excitation laser, quad-band set, and emission filter) for the fluorophore selected for single-antibody labeling.

   For instance, if an antibody is conjugated with Alexa Fluor (AF) 647, select the excitation laser (640 nm), quad-band set (TRF89901-ET), and emission filter (ET700/75m). Save the settings to optical configuration.

7. Select the laser power intensity level and set the camera parameters.

   A laser power density between 20-30 W/cm² and a camera exposure of 50 ms can be used initially. These parameters can be adjusted to improve the signal-to-noise ratio (SNR).

8. Add fiducial markers (TetraSpeck microspheres or gold colloids).

9. Sonicate gold colloids for 3 min or vortex the TetraSpeck microspheres for 30 s before use. Dilute TetraSpeck microspheres (size: 200 nm), 1:1000 in DPBS. Dilute Gold colloids (size: 100 nm), 1:5 dilution in Mili-Q water.

   Incubate TetraSpeck microspheres for 5 min and gold colloids for 10-20 min.

   Fiducials can also be added to the sample after the sample preparation before placing it on the microscope stage.

10. Use brightfield illumination, locate a cell, and focus on it.

11. Identify fiducial markers near the cell that can be used for drift correction (Fig. 2).

12. Replace the fiducial marker solution with 200 μL of DPBS. The sample is ready for image acquisition.

Figure 2: Fiducial markers for drift correction in imaging.

(a) Detection of TetraSpeck microspheres using low laser powers. (b) Visualization of gold colloids using brightfield illumination. Yellow arrowheads point to fiducial markers in each image. The dashed line contours the cell. Scale bars: 2 μm.

Time-lapse imaging with single-antibody labeling

• 13)The optimal antibody concentration (in the imaging buffer) and NII need to be adjusted to capture of sufficient single-molecule events without significant spatial overlap (Fig. 3a vs. b). Start with 0.003-0.5 nM antibody concentration and 1-20 s NII.

  The optimal concentration and NII differ significantly for different antibodies and targets. Our study (Gunasekara et al., 2023) conducted an NII scan, exploring NIIs 1, 3, 5, 10 and 20 s for different antibody systems. Through DBSCAN (density-based spatial clustering of applications with noise) characterization, we observed that increasing the NII led to increased capture of high-density events, likely associated with specific antibody binding events.

• 14)Prepare the imaging buffer by diluting the dye-conjugated antibody in DPBS.

  Refer to Table 2 for the dye-conjugated antibody concentration.

  When the concentration of dye-conjugated primary or secondary antibodies is provided in mg/mL, use the appropriate molecular weights of the antibodies (full-length IgG=150 kDa, F(ab’)₂=110 kDa) to convert the concentration to nM.

  If the stock solution is highly concentrated (100x or more), perform a serial dilution to minimize pipetting errors.

• 15)Add 300 μL of the imaging buffer to the mounted sample well.
• 16)Increase the laser power density (100-400 W/cm²) and set the exposure time (50 ms).

  The provided laser power and exposure time serve as a starting point. Adjust the parameters to capture more photons per frame while maintaining a low background signal.

• 17)Adjust the TIRF angle to achieve the maximum signal-to-noise ratio (SNR).

  An optimal TIRF angle restricts out-of-focus excitation and maximizes the SNR. In our study, we used a laser power density ranging between 100-400 W/cm² and an exposure time of 50 ms, resulting in an SNR of 3-5. The antibody concentration was maintained within the range of 0.003-0.5 nM, depending on the antibody. (Fig. 4 and Table 2).

• 18)Open NIS Elements software and right click to find acquisition controls. Go to ND Acquisition.

  In our study, we used ND Acquisition to set the NII and the frames. A similar acquisition control tab can be used for a different microscopy system to set up the imaging parameters and NII

• 19)In the ‘Time’ tab, input the NII in the ‘interval’ column. Input the frames in the ‘Loops’ column.
• 20)Select ‘Time’ tab and click ‘Run now’ to record a movie consisting of at least 50 frames.
• 21)Replay the recorded movie and evaluate the sparsity of the binding events.

  A single-molecule event density of 0.09 to 0.3 events/μm² serves as a starting point for new antibody-target systems (Gunasekara et al., 2023).

• 22)To determine the average number of single-molecule events per μm², select three representative regions (e.g., 15X15 μm) of the movie captured in step 21.
• 23)Count the number of single-molecule events in each frame and calculate the average number of events per μm².

  ThunderSTORM plug-in in ImageJ can be used for this purpose (Refer to Basic Protocol 3 for using ThunderSTORM plug-in).

  In our study, the counts of single-molecule events in a chosen area (e.g., 225 μm²), with a range of events (e.g., 20-70), is calculated to be between 0.09 to 0.3 events per μm².

• 24)Adjust the antibody concentration of the imaging buffer if necessary.

  If overlapping single-molecule events are observed right after the addition of the imaging buffer or an accumulation of events as the acquisition progresses, terminate the acquisition. Decrease the concentration of the antibody solution by 10-fold and reevaluate. If the concentration is still too high, decrease another 10-fold. If the concentration is too low, increase by 10-fold, reevaluate, etc. Use a new sample well and repeat steps 20-23.

  Increasing the antibody concentration beyond 1 nM may increase the background fluorescence from diffusing antibodies in the imaging buffer, decreasing the SNR.

• 25)Cover the sample well with a parafilm to prevent the imaging buffer from evaporating.
• 26)Select the number of frames in the ‘loops’ column of the ND acquisition.
• 27)Capture sufficient single-molecule localizations to resolve the underlying structure (Refer to Fig. 1i and Fig. 2c in Gunasekara et al., 2023).

  The optimal number of frames required may vary between antibody-target systems. We acquired 3000 frames for microtubules using a 20 s NII (Fig. 5).

  NIIs extend the image acquisition duration and increase the time needed to collect sufficient frames. Given a 20 s NII, acquiring 3000 frames result in ~17 hr acquisition time.

Figure 3: Single-molecule event density for single-antibody labeling.

Representative single-molecule events detected (a) before and (b) after the concentration optimization in three antibody-target systems in Table 2. The higher number of overlapping events in (a) renders single-molecule localization difficult. The antibody concentrations used are as follows. [12CA5-AF 647] =1 nM (a) and 0.5 nM (b); [DMIA-AF 488] = 1 nM (a) and 0.003 nM (b); [F(ab′)₂-AF 647] = 0.5 nM (a) and 0.063 nM (b). NII= 20 s. Scale bars: 2 μm.

Figure 4: Cross-sectional intensity profile across the indicated yellow line showing the SNR for three antibody-target systems in Table 2.

(a) Endogenous α-tubulin labeled with DMIA-AF 488. (b) HA tag on α-tubulin labeled with primary antibody anti-HA (12CA5) and secondary antibody F(ab′)₂-AF 647. (c) Tom20 in mitochondria labeled with primary antibody anti-Tom20 and secondary antibody goat anti-rabbit-AF 647. Scale bars: 2 μm.

Figure 5: Superresolution images of microtubules in U2OS cells using single-antibody labeling.

The comparison of reconstructed images after (a)500, (b)1000, (c)2000, and 3000 frames (F) is shown for microtubules. DMIA-AF 488 was used to label endogenous α-tubulin. [DMIA- AF 488] = 0.003 nM, NII=20 s and exposure time= 50 ms. Scale bars: 1 μm.

Multiplexed imaging with single-antibody labeling

• 28)Optimize the antibody concentrations for each target as described in steps 13-24.
• 29)Prepare the imaging buffer by diluting all antibodies. Add the imaging buffer to the sample well and set the NII.
• 30)Choose the optical configurations (excitation lasers, quad-band set, and emission filters) based on the fluorophores selected for multiplexed imaging. Set the laser power density (100-400 W/cm²) and exposure time (50 ms).

  Carefully adjust the TIRF angle and focus to obtain the highest SNR for different channels.

• 31)Set the number of frames and start the image acquisition.

  Cover the sample well with a parafilm to prevent the imaging buffer from evaporation.

  For multiplexed imaging, ND acquisition uses a separate tab marked as “λ”. Add the optical configuration for each channel to the ‘λ’ tab.

  Before starting the acquisition, select both ‘time’ and ‘λ’ tabs.

Dual-color single-target imaging with single-antibody labeling

This protocol outlines a strategy to enhance the sample labeling density. This strategy uses a mixture of the same antibody conjugated with two different dyes.

• 32)Prepare the imaging buffer by diluting the antibodies. Add the imaging buffer to the sample well and set the NII.

  Use the optimized antibody concentration in steps 13-24.

• 33)Select the optical configurations (excitation lasers, quad-band set, and emission filters) based on the fluorophores selected for multiplexed imaging. Set the laser power density (100-400 W/cm²) and exposure time (50 ms).

  Carefully adjust the TIRF angle. Then focus on the binding events to obtain the highest SNR for both channels.

• 34)Set the number of frames and start the image acquisition.

  Cover the sample well with a parafilm to prevent the imaging buffer from evaporation.

  For dual-color imaging, ND acquisition uses a separate tab marked as “λ”. Add the optical configuration for each channel to the ‘λ’ tab.

  Before the acquisition, select both ‘Time’ and ‘λ’.

Alternate Protocol 1: Dual-color single-antibody labeling using OptoSplit II

In this protocol, we utilize OptoSplit II to perform dual-color single-target imaging. OptoSplit II allows a single camera to simultaneously capture images from two distinct optical wavelengths. Identify the set of filters and beam splitter for your specific application. For detailed information on the installation and usage, refer to the manual provided by the manufacturer.

Materials (also see Basic Protocol 2)

OptoSplit II LS 1.0X (Cairn Research, OptoSplit II)

Cube to split 488 and 647 nm (Cairn Research)

• T565LPXR-UF2 Dichroic

• AT540/30M emission filter 25mm

• ET655LP long pass emission filter 25mm

Cube to split 561 and 647 nm (Cairn Research)

• T640LPXR-UF2 LP Dichroic 26x38x2mm

• ET615/40M emission filter 25mm

• ET655LP long pass emission filter 25mm

Acquisition software (Nikon, NIS-AR 5.42)

Software add-on for the OptoSplit II (Nikon, MQS41410: NIS-D SPLITTER Dual View)

1. Turn on the microscope, camera for OptoSplit II, acquisition computer, and lasers.

   If the Prime 95B sCMOS camera is connected to the acquisition computer via Peripheral Component Interconnect Express (PCIe), ensure the camera is turned on before starting the computer.

2. Open the acquisition software. Select the optical configuration (excitation lasers, quad-band set, and emission filters).

3. Prepare the imaging buffer by diluting the antibodies. Add the imaging buffer to the sample well and set the NII.

   Use the optimized antibody concentration in Basic Protocol 2 steps 13-24.

4. Set the laser power density (100-400 W/cm²) and the number of frames (500-3000).

5. Go to "Live" mode. The acquisition window displays spectrally resolved images side by side, representing the two channels used. The window may also show superimposed images, allowing for visual comparison of the combined signals from two channels.

6. Use split adjusters and aperture adjusters to visualize and adjust the position of the region of interest (ROI). Refer to OptoSplit II manual for detailed instructions on ROI adjustment.

7. Cover the sample well with a parafilm to prevent the imaging buffer from evaporation.

8. Select the ‘Time’ in the ND acquisition tab, input the NII in the ‘interval’ column. Input the frames in the ‘Loops’ column. A similar acquisition control tab can be used for a different acquisition software setup to adjust the imaging parameters and NII.

9. Start image acquisition.

BASIC PROTOCOL 3: IMAGE ANALYSIS

ThunderSTORM is an ImageJ plug-in designed for the analysis of fluorescence single molecule data and the reconstruction of superresolution images using SMLM.

Materials

Raw data file acquired according to Basic Protocol 2

ImageJ software: https://imagej.nih.gov/ij/download.html

Thunder_STORM.jar file: https://github.com/zitmen/thunderstorm

Initializing image reconstructions using ThunderSTORM plug-in

1. Install ImageJ and download the Thunder_STORM.jar file. Copy it into the plug-in subfolder of the ImageJ installation. Locate the ThunderSTORM plug-in in the Plugins menu of ImageJ to confirm the successful installation

2. Load the raw data file in ImageJ. Open the ThunderSTORM plug-in and select ‘Run analysis’. A window containing a set of processing parameters will appear.

3. Select ‘camera setup’ and set the parameters such as effective pixel size, photoelectrons per A/D count and the base level.

   Refer to ThunderSTORM user’s guide, version 1.3, processing 2D data and guidelines for the choice of parameters (Ovesný et al., 2014). In our study, we used the following parameters: pixel size 147 nm, photoelectrons per A/D count 0.98, and base level 100.

4. Set the other parameters such as image filtering, approximate localization of molecules, sub-pixel localization of molecules, and visualization of the results. Use the selected parameters and click ‘Preview’.

   Refer to ThunderSTORM user’s guide, version 1.3, processing 2D data and guidelines for the choice of parameters (Ovesný et al., 2014).

   In our study we used the following parameters as the starting point: Wavelet filter (B-Spline order 2 and scale 4) for image filtering, Local maximum with 8-neighbourhood connectivity for approximate localization of molecules, PSF: Integrated Gaussian and Maximum likelihood as fitting method for sub-pixel localization of molecules. For visualization of the results ‘Averaged shifted histograms’ or ‘Normalized Gaussian’ was used at a forced lateral uncertainty of 20 nm (Fig. 6a).

5. Visualize the detected molecules in the ‘Detections in frame 1’ result window. If the visualization indicates the lower or higher count of detected molecules, adjust the parameters in step 4 and redo the ‘Preview’.

Figure 6: Superresolution data analysis.

(a)Illustration of ThunderSTORM ‘Run analysis’ window highlighting parameters such as camera setup, image filtering, finding approximate localization of molecules, sub-pixel localization of molecules and visualization of the results. (b) Illustration of the result file and drift correction parameters (c) Images of a fiducial marker demonstrating (i) sub-optimal and (ii) optimal drift correction with their respective drift trajectories. (d) Superresolution image of microtubules (i) before and (ii) after drift correction. Scale bar: 500 nm.

Drift correction using a fiducial marker

• 6)While the file is open in ImageJ, use the square selection tool to identify a fiducial marker present during the entire acquisition.
• 7)Open the ThunderSTORM plug-in and select ‘Run analysis’ using the parameters set for image reconstruction.
• 8)In the result window, click the ‘Drift correction’ tab.
• 9)Check the ‘Fiducial marker’ and set parameters for drift correction.

  Set parameters such as max distance, min marker visibility ratio, and trajectory smoothing factor. Refer to ThunderSTORM user’s guide, version 1.3, processing 2D data and guidelines for the choice of parameters (Ovesný et al., 2014). In our study, we used the following parameters as starting point: max distance 1000, min marker visibility ratio 0.02 and trajectory smoothing factor 0.02 (Fig. 6b).

• 10)Check ‘Save to file’ and define a path to save the drift-corrected file. Click ‘Apply’. A sample x-y drift profile will display.

  After performing drift correction, the corrected fiducial marker becomes visible in the image. If the drift correction is sub-optimal (Fig. 6c), repeat step 9.

Analyzing raw single-molecule data and applying drift correction

• 11)Unselect the square selection from step 6. Select ‘Run analysis’ in ThunderSTORM and click ‘Ok’.

  Once the analysis is complete, a ‘results’ table and the corresponding reconstructed image will appear. The selected single molecule localizations on the raw data will appear as red crosses on the movie file.

• 12)Apply the drift correction.

  Click the ‘Drift correction’ tab in the result table.

  Check ‘Load from file’. By default, the information of the last saved drift correction file is kept. In this case, click ‘Apply’. The x-y drift profile appears.

  The drift-corrected result file and the reconstructed image appear.

• 13)In the ‘Visualization’ tab, set the visualization method, and magnification to render the final superresolution image.

  For the visualization of the results, we used ‘Averaged shifted histograms’ or ‘Normalized Gaussian’ at a forced lateral uncertainty of 20 nm. Magnification was set to 7.3 to have a pixel size of 20 nm in the reconstructed image.

  Additional post-processing steps can be used to optimize the image reconstruction. Refer to ThunderSTORM user’s guide, version 1.3, post-processing modules.

  (Optional) Change the ‘LUT’ using ‘Lookup Tables’ in ImageJ as intended. The default LUT from ThunderSTORM is ‘gray’.

Overlaying channels in multiplexed and dual-color single-antibody labeling

• 14)Open the reconstructed images of each channel separately in ImageJ.
• 15)Select ‘Image’ and go to ‘Color’. Click ‘Merge Channels’ from the drop-down menu. Assign colors for each channel using the ‘Merge Channels’ dialog box.

  For instance, if one image corresponds to the red channel, and the second image corresponds to the green channel, select 'Red' for the first channel and 'Green' for the second channel using the drop-down menu.

• 16)Select ‘Create Composite’ and click ‘OK’ to create the overlay. Use ‘Save As’ to save the overlay image as a new file.
• 17)Superimpose the images by aligning the fiducial markers from both channels. Select ‘Image’ and go to ‘Transform’. Click ‘Translate’ from the drop-down menu. Adjust X and Y offset.

REAGENTS AND SOLUTIONS

Blocking buffers

1. BSA (5%, w/v), Triton X-100 (0.3%, v/v) in DPBS

   • Mix:
   • 100 μL of BSA (from stock solution 10%, w/v)
   • 20 μL of Triton X-100 (from stock solution 3%, v/v)
   • 80 μL of DPBS to prepare 200 μL of blocking buffer
   • This buffer is freshly prepared before use
2. Goat serum (10%, v/v), Triton X-100 (0.3%, v/v) in DPBS

   • Mix:
   • 20 μL of goat serum (from stock solution 100%, v/v)
   • 20 μL of Triton X-100 (from stock solution 3%, v/v)
   • 160 μL of DPBS to prepare 200 μL of blocking buffer
   • This buffer is freshly prepared before use

Fixation buffer:

Paraformaldehyde (3.7%, v/v), glutaraldehyde (0.1%, v/v) in DPBS

• Mix:

• 46 μL of paraformaldehyde (Electron Microscopy Sciences, 15710, 16%)

• 2 μL of glutaraldehyde (Electron Microscopy Sciences, 16120, 10%)

• 152 μL of DPBS to prepare 200 μL of fixation buffer

• This buffer is freshly prepared before use

  The components of the extraction buffer used in our study are as follows: Paraformaldehyde (3.7%, v/v), glutaraldehyde (0.1%, v/v), and Triton X-100 (0.5%, v/v) in cytoskeleton buffer (Gunasekara et al., 2023). In this protocol, the permeabilization step is combined with the fixation step to preserves the morphology of the microtubules during chemical fixation. The components and the final concentrations of the cytoskeleton buffer are as follows. MES (10 mM, pH 6.1), potassium chloride (90 mM), magnesium chloride (3 mM), and EGTA (2 mM) (Kiuchi et al., 2015). This solution can be stored at room temperature for up to one month.

Post-fixation buffer:

Paraformaldehyde (3.7%, v/v) in DPBS

• Mix:

• 46 μL of paraformaldehyde (Electron Microscopy Sciences, 15710, 16%)

• 154 μL of DPBS to prepare 200 μL of post-fixation buffer

• This buffer is freshly prepared before use

COMMENTARY

Background Information:

IF staining is a critical step in sample preparation of SMLM, typically accomplished through immunolabeling using dye-conjugated primary or secondary antibodies. The quality of the resulting superresolution image relies on the quality and specificity of the antibodies utilized. Currently, the selection of antibodies for IF staining in SMLM largely relies on empirical methods, emphasizing the need for careful antibody validation and optimization.

The interactions between antibodies and epitope targets are often characterized by the equilibrium dissociation constant (K_D), typically falling within the nanomolar range. In conventional IF staining, primary antibodies are commonly used at concentrations ranging from 2-10 μg/ml or 13.3-66.7 nM (for full-length IgG, molecular weight ~150 kDa). These concentrations exceed the K_D values by one to two orders of magnitude. The large excess of antibodies labels the majority of the target at equilibrium. In our study, we employed low antibody concentrations below the nanomolar range. This process scales down the on-rate, enabling the capture of antibody labeling through single-molecule imaging. Thus, single-antibody labeling captures the progression of the immunolabeling process one antibody at a time.

Furthermore, we implemented a stroboscopic imaging scheme by incorporating an NII between consecutive frames to ensure that each frame captures an adequate number of single-molecule events without significant spatial overlap. As manifested by the extended acquisition time, the off-rate of the antibody is unmodulated and remains slow. This is a distinct feature compared to existing PAINT techniques that promote the off rate of the labeling probes. However, a noted advantage of our technique is the higher SNR due to multiple dyes per antibody. Signal amplification may also be achieved by multiple dye-conjugated secondary antibodies interacting with a single primary antibody. The extended image acquisition time represents a major limitation of our technique. Among approaches to accelerate the dissociation are antibody fragmentation and protein engineering methods (Gunasekara et al., 2022; Miyoshi et al., 2021; Zhang et al., 2022). However, promoting antibody dissociation without compromising binding specificity is a challenging task that highly relies on the specific interactions between the antibody and antigen. To that end, our dual-color single-target antibody labeling strategy provides an alternative approach to increasing labeling density within a given duration of acquisition. Our imaging method also enables the screening and evaluation of antibody labeling within the native cellular environment. Our approach provides valuable insights into the behavior and performance of antibodies in their cellular environment, augmenting existing assays to enhance the understanding of their performance for superresolution imaging applications.

Critical Parameters

A critical task in image acquisition is optimizing the dye-conjugated antibody concentration and NII to capture sufficient single-molecule events without significant spatial overlap (step 13-24, Basic Protocol 2). The optimal concentration and NII varies for different antibody-target systems. A single-molecule event density of 0.09 to 0.3 events/μm² serves as a starting point for optimization. We maintained the concentration within the range of 0.003-0.5 nM (Table 2) while maintaining an NII of 20 s.

When preparing the samples for imaging using dye-conjugated secondary antibodies, selecting the appropriate concentration of unconjugated primary antibody is important (step 12, Basic Protocol 1). As a starting point, use the manufacturer-suggested concentration. If the resulting superresolution image does not demonstrate the anticipated labeling density, consider increasing the primary antibody concentration to enhance the substrate abundance. For instance, in our study, we used a primary antibody (12CA5) concentration of 10 μg/mL (67 nM) for the labeling of HA tag on α-tubulin to image with secondary antibody F(ab′)₂-AF 647.

Troubleshooting

During each of the procedures outlined in Basic Protocols 1-3, it is anticipated that several challenges may arise. To address these issues, Table 3 presents a summary of the problems, potential causes, and solutions.

Table 3:

Troubleshooting table

Problem	Possible cause	Solution
Single-molecule events outside the cell (equal or significantly greater than within the cell)	Non-specific binding is high	Extend the blocking time (up to 2 hr) if using BSA in the blocking buffer or use serum from the species that corresponds to the secondary antibody in the blocking buffer Increase washing steps (five times) after primary antibody staining (step 14, Basic Protocol 1)
Uneven illumination observed following sample mounting	Bubbles in immersion oil	Wipe off the immersion oil completely and then reapply it to the objective Let the chamber slide reach room temperature before mounting
Low SNR (less than 3)	High background signal	Decrease the antibody concentration (by 10-100-fold) in the imaging buffer to reduce the background signal Adjust the TIRF angle to restrict out-of-focus excitation to maximize the SNR (step 17, Basic Protocol 2)
Overlapping single-molecule events	High antibody concentration in imaging buffer	Decrease the antibody concentration (by 10-100-fold) in the imaging buffer (step 24, Basic Protocol 2) Introduce a photobleaching step (with a maximum laser power density) every 20-50 frames in the image acquisition
Insufficient number of fiducial markers for drift correction	Cannot identify fiducial markers near the cell to be imaged	Sonicate (gold colloids) or vortex (TetraSpeck microspheres) the stock solutions to uniformly suspend the fiducials (step 9, Basic Protocol 2) Increase the concentration of fiducial markers (5-fold) or incubate the fiducial markers longer (until at least 3 or more settle down)
Diffusive appearance of the morphology (e.g., microtubules) after image reconstruction	Overlapping single-molecule events	Decrease the antibody concentration (by 10-100-fold) in the imaging buffer (step 24, Basic Protocol 2) Apply post-processing by selecting the parameter of choice (e.g., sigma) in ThunderSTORM analysis (Refer to ThunderSTORM user’s guide, version 1.3, post- processing modules)

Understanding the results

Our study used microtubules and mitochondria to demonstrate the imaging method. Basic Protocol 1 outlines the key steps for preserving subcellular structures in U2OS cells. Basic protocol 2 outlines the steps for using the NIS Elements software to perform single-antibody labeling. A comparable software package with time-lapse imaging capabilities can be used for this.

Tables 1 and 2 provide the antibody concentrations used in our study for sample preparation and time-lapse imaging with single-antibody labeling (at NII=20 s), respectively. We optimized the antibody concentration and NII to achieve a single-molecule event density ranging from 0.09 to 0.3 events per μm². Fig. 3 shows the single-molecule event density before and after optimization for three antibody-target systems. Fig. 4 illustrates cross-sectional intensity profiles showing SNR for different antibody-target systems. This was obtained using the laser power density of 100-400 W/cm² and exposure time of 50 ms.

We used the ThunderSTORM plug-in, available in ImageJ (Ovesný et al., 2014), for image processing. Open the raw data file (acquired according to Basic Protocol 2) in ImageJ and run ThunderSTORM analysis to obtain superresolution image. To process the data, set the camera parameters: effective pixel size, photoelectrons per A/D count and the base level. In our study, we used an effective pixel size of 147 nm, photoelectrons per A/D count of 0.98, and the base level of 100. Select the image filtering parameters and methods to identify the localization of molecules. In our study we used Wavelet filter, Local maximum and PSF: Integrated Gaussian (Fig. 6a). For visualization of the superresolution image during data analysis, select the visualization methods. In our study we used Averaged shifted histograms or Normalized Gaussian. Preview the molecular localizations with given parameters. If the visualization indicates the lower or higher count of detected molecules, consider adjusting the image filtering parameters (e.g., B-Spline order and scale if using Wavelet filter). After the analysis, the reconstructed superresolution images and the table of localized molecules (results table) are considered the final results. Then apply the drift correction (Fig. 6d). The post-processing methods can be applied to remove any poor-quality data points, further refining the obtained results and the ThunderSTORM user’s guide. In our study, events with a sigma value exceeding 170 exhibited a diffusive appearance and were removed in the post-processing. The superresolution images of microtubules obtained by single-antibody labeling is shown in Fig. 5.

To process the data from multiple channel acquisitions (e.g., multiplexed and dual-color imaging) begin by opening each channel individually in ImageJ. Perform image analysis according to the steps outlined in Basic Protocol 3. Once the reconstructed superresolution images are generated, proceed to create overlaid images using ImageJ (steps 14-17, Basic Protocol 3). Use the fiducials to align images generated from different channels. Representative superresolution images from multiplexed imaging of microtubules and mitochondria is shown in Fig. 7. Representative dual-color images of microtubule are shown in Fig. 8. We achieved 61-83 nm resolution for adjacent microtubules by single-antibody labeling (Gunasekara et al., 2023)

Figure 7: Multiplexed superresolution imaging with single-antibody labeling.

(a) Superresolution images of (a) microtubules and (b) mitochondria. (c) Dual color overlay of the superresolution images of (a) and (b). Microtubules are shown in magenta, and mitochondria are shown in green. For microtubule imaging, DMIA-AF 488 was used to label endogenous α-tubulin. [DMIA- AF 488] = 0.003 nM. For mitochondria imaging, Tom20 was labeled using primary antibody anti-Tom20 and imaged with secondary antibody goat anti-rabbit-AF 647. [Goat anti-rabbit-AF 647] = 0.125 nM. NII=20 s. The reconstructed image is from 3000 frames. Scale bars: 1 μm.

Figure 8: Dual-color single-antibody labeling.

HA tag on α-tubulin labeled with primary antibody anti-HA (12CA5) and secondary antibody F(ab′)₂-AF 647. Superresolution images of microtubules were obtained with an equal molar of 0.25 nM of (a) F(ab′)₂-AF 488 (green) and (b) F(ab′)₂-AF 647 (magenta). NII= 20 s. (c) Dual-color overlay of the superresolution images from (a) and (b). (d) Zoomed-in views of the regions indicated in panel c. Scale bars: 5 μm (a, b, c) and 250 nm (d).

Time considerations

The complete protocol outlined here requires ~5 days for completion. Table 4 summarizes the time consideration for each Basic Protocol.

Table 4:

Time considerations

Protocol	Time considerations
Basic Protocol 1	Day 1: Cell seeding (10-20 min) Day 3 (after 36 hr of seeding): Fixation (15-20 min) Day 3: Sample preparation for single-antibody labeling using dye-conjugated primary antibodies (1-2.5 hr)1 Day 3: Sample preparation for single-antibody labeling using dye-conjugated secondary antibodies (3-7 hr)1
Basic Protocol 2	Day 4: Initiating data acquisition (30-40 min) Day 4: Determination of optimal antibody concentration and NII (steps 13-27) (1-3 hr) Day 4/5: Time-lapse imaging with single-antibody labeling (1-17 hr)2 Day 4/5: Multiplexed imaging with single-antibody labeling (1-17 hr)2 Day 4/5: Dual-color single-target imaging with single-antibody labeling (1-17 hr)2
Basic Protocol 3	Day 5/6: Initializing image reconstructions (30 min) Day 5/6: Analyzing raw single-molecule data and applying drift correction (1 hr)

¹The timing depends on the antibody-target system.

²The timing depends on NII and the number of frames selected for the acquisition (For instance, with a 1 s NII, capturing 3000 frames takes 1 hr, while a 20 s NII, for 3000 frames extends to 17 hr).

Data Availability Statement

Images and movies relevant to this protocol have been published in our earlier manuscript (Gunasekara et al., 2023). New images have been incorporated as figures in this work. The data that support this protocol are available from the corresponding author upon reasonable request.