Conventional BODIPY Conjugates for Live-Cell Super-Resolution Microscopy and Single-Molecule Tracking

Summary

Conventional BODIPY conjugates can be used for live-cell single-molecule localization microscopy (SMLM) through exploitation of their transiently forming, red-shifted ground state dimers. We present an optimized SMLM protocol to track and resolve subcellular neutral lipids and fatty acids in living mammalian and yeast cells at the nanoscopic length scale.

Abstract

Single molecule localization microscopy (SMLM) techniques overcome the optical diffraction limit of conventional fluorescence microscopy and can resolve intracellular structures and the dynamics of biomolecules with ~20 nm precision. A prerequisite for SMLM are fluorophores that transition from a dark to a fluorescent state in order to avoid spatio-temporal overlap of their point spread functions in each of the thousands of data acquisition frames. BODIPYs are well-established dyes with numerous conjugates used in conventional microscopy. The transient formation of red-shifted BODIPY ground-state dimers (D_II) results in bright single molecule emission enabling single molecule localization microscopy (SMLM). Here we present a simple but versatile protocol for SMLM with conventional BODIPY conjugates in living yeast and mammalian cells. This procedure can be used to acquire super-resolution images and to track single BODIPY-D_II states to extract spatio-temporal information of BODIPY conjugates. We apply this procedure to resolve lipid droplets (LDs), fatty acids, and lysosomes in living yeast and mammalian cells at the nanoscopic length scale. Furthermore, we demonstrate the multi-color imaging capability with BODIPY dyes when used in conjunction with other fluorescent probes. Our representative results show the differential spatial distribution and mobility of BODIPY-fatty acids and neutral lipids in yeast under fed and fasted conditions. This optimized protocol for SMLM can be used with hundreds of commercially available BODIPY conjugates and is a useful resource to study biological processes at the nanoscale far beyond the applications of this work. 

Introduction

Single-molecule localization microscopy (SMLM) techniques such as stochastic optical reconstruction microscopy (STORM) and photo-activated localization microscopy (PALM) have emerged as methods for generating super-resolution images with information beyond Abbe’s optical diffraction limit^1,2 and for tracking the dynamics of single biomolecules^3,4. One of the requirements for probes compatible with SMLM is the ability to control the number of active fluorophores at any time to avoid spatial overlap of their point spread functions (PSF). In each of the thousands of data acquisition frames, the location of each fluorescent fluorophores is then determined with ~20 nm precision by fitting its corresponding point-spread function. Traditionally, the on-off blinking of fluorophores has been controlled through stochastic photoswitching^1,2,5 or chemically induced intrinsic blinking⁶. Other approaches include the induced activation of fluorogens upon transient binding to a fluorogen-activating protein^7,8 and the programmable binding-unbinding of labeled DNA oligomers in total internal reflection fluorescence (TIRF) or light sheet excitation⁹. Recently, we reported a novel and versatile labeling strategy for SMLM¹⁰ in which previously reported red-shifted dimeric (D_II) states of conventional boron di-pyromethane (BODIPY) conjugates^11,12,13 are transiently forming and become specifically excited and detected with red-shifted wavelengths.

BODIPYs are widely used dyes with hundreds of variants that specifically label sub-cellular compartments and biomolecules^14,15,16. Because of their ease of use and applicability in living cells, BODIPY variants are commercially available for conventional fluorescence microscopy. Here, we describe a detailed and optimized protocol on how the hundreds of commercially available BODIPY conjugates can be used for live-cell SMLM. By tuning the concentration of BODIPY monomers and by optimizing the excitation laser powers, imaging and data analysis parameters, high-quality super-resolution images and single molecule tracking data is obtained in living cells. When used at low concentration (25-100 nM), BODIPY conjugates can be simultaneously used for SMLM in the red-shifted channel and for correlative conventional fluorescence microscopy in the conventional emission channel. The obtained single molecule data can be analyzed to quantify the spatial organization of immobile structures and to extract the diffusive states of molecules in living cells¹⁷. The availability of BODIPY probes in both green and red forms allows for multi-color imaging when used in the right combination with other compatible fluorophores.

In this report, we provide an optimized protocol for acquiring and analyzing live-cell SMLM data using BODIPY-C₁₂, BODIPY (493/503), BODIPY-C₁₂ red and lysotracker-green in multiple colors. We resolve fatty acids and neutral lipids in living yeast and mammalian cells with ~30 nm resolution. We further demonstrate that yeast cells regulate the spatial distribution of externally added fatty acids depending on their metabolic state. We find that added BODIPY-fatty acids (FA) localize to the endoplasmic reticulum (ER) and lipid droplets (LDs) under fed conditions whereas BODIPY-FAs form non-LD clusters in the plasma membrane upon fasting. We further extend the application of this technique to image lysosomes and LDs in living mammalian cells. Our optimized protocol for SMLM using conventional BODIPY conjugates can be a useful resource to study biological processes at the nanoscale with the myriad available BODIPY conjugates.

Protocol

NOTE: For yeast cloning and endogenous tagging please refer to our recent publication¹⁰.

1. Preparation of yeast cell samples for imaging

1. Prepare a liquid overnight culture of a w303 yeast strain. Using a sterile wooden stick, spot a small amount of yeast cells from an agar plate containing yeast extract–peptone–dextrose into a culture tube with ~2 mL of synthetic complete dextrose (SCD) medium. Incubate the tube overnight in a shaking incubator at 270 rpm and 30 °C.
2. Perform a 1:50 morning dilution of the cells in SCD. Continue to culture the diluted cells for 4 h at 30 °C in a 270 rpm shaking incubator, allowing the cells to grow in exponential phase and to reach an optical density (OD) of ~0.6.
   NOTE: The procedure can vary here depending on which metabolic state is being studied. BODIPY conjugates do not require cells in the exponential growth phase. However, be cognizant of autofluorescence from dead cells during the stationary phase, as it can cause a background signal too strong to analyze single BODIPY-D_II emitters.
3. For studying fasting cells, grow the yeast culture for 2 days without exchange of media.
4. At ~30 min prior to plating the cells, incubate a chambered coverglass with 80 µL of 0.8 mg/mL sterile Concanavalin A (ConA) in deionized H₂O at room temperature. After 30 min, wash the coverglass three times with deionized H₂O.
5. At ~30 min before imaging, pipette the cells on the chambered coverglass, with the correct volume of fresh SCD to achieve an optical density of ~0.12 (typically 60 µL yeast culture at OD ~0.6 in 240 µL SCD). Let the cells settle and adhere to the ConA surface for 30 min.
6. Add the desired BODIPY conjugate directly to the chambered coverglass at a final concentration of ~100 nM.
   NOTE: A BODIPY concentration optimization experiment may be required depending on BODIPYs local density in a particular cellular compartment.

2. Preparation of mammalian cells for SMLM imaging

1. Maintain the mammalian U2OS cells in non-fluorescent DMEM with 10% fetal bovine serum, 4 mM glutamine, 1 mM sodium pyruvate and 1% penicillin-streptomycin antibiotics in a T25 flask.
   NOTE: Cells can also be maintained in DMEM with 10% fetal bovine serum and 1% penicillin-streptomycin antibiotics, however, the medium needs to be exchanged before imaging with a non-autofluorescent solution.
2. Split the cells at 70-80% confluency 1:5 in a single well of an 8-well plate. Culture the cells in the 8-well plate for 12 to 24 h before imaging.
3. Add BODIPY-C₁₂, LysoTracker Green or any other BODIPY conjugate at a final concentration of 100 nM (stock solutions in dimethyl sulfoxide [DMSO]) 10 min prior to imaging. This time can vary based on the desired experiment.
   NOTE: Imaging can be performed at ambient temperature (23 °C) using live cell imaging solution. However, imaging at 37 °C and 5% CO₂ with non-fluorescent DMEM mixed with 10% fetal bovine serum, 4 mM glutamine, 1 mM sodium pyruvate and 1% penicillin-streptomycin antibiotics is preferred to keep cells closer to physiological conditions and to make biological conclusions.

3. Equipment preparation

1. Mount the appropriate filter sets in the emission path based on the emission color of BODIPY being used.
   NOTE: A quad-band dichroic mirror (zt/405/488/561/640rdc) first separates the excitation from the emission light. The green emission (525 nm) and red emission (595 nm) are then split by a dichroic long-pass beam splitter (T562lpxr) followed by band-pass filters ET525/50 in the green channel and ET 595/50 in the red channel. The two channels are then projected to different areas of the same camera chip. Similarly, the red emission (595 nm) and the far-red emission are first split by a dichroic long-pass beam splitter (FF652-Di01) followed by band-pass filters ET 610/75 in the red and FF731/137 in the far-red channel.
2. Turn on the microscope, microscope stage, lasers (488 nm, 561 nm) and camera. Here an inverted microscope with a perfect focus system and an EMCCD camera cooled to -68 °C are used.
3. Add a drop of immersion oil on the microscope objective.
4. Open the Hal4000 software (see the Table of Materials) that controls the LED light for bright field imaging, laser powers, laser shutters and camera settings for imaging. Set the EMCCD gain to 30 and the camera temperature to -68 °C. Prepare the camera and corresponding software to record movies at 20 Hz.
   NOTE: This technique applies to any wide-field microscope capable of photo-activated localization microscopy (PALM) and stochastic optical reconstruction microscopy (STORM) imaging. Corresponding software may vary.
5. Turn on the microscope stage heater and set it to a temperature of 37 °C and to a CO₂ level of 5%. Adjust the objective correction collar accordingly.
6. Mount the sample on the microscope stage and focus until the focusing system engages. Move the stage using the stage controller until healthy cells appear in the field of view.
   NOTE: For imaging with yeast cells at room temperature, there is no need to turn on the heater or CO₂ control.
7. Turn on the appropriate lasers for the excitation of monomers as well as dimers. For BODIPY green or LysoTracker green, we use a 561 nm laser to excite D_II states for SMLM and a 488 nm laser to excite monomers for conventional fluorescence.
   NOTE: For BODIPY red, use a 640 nm laser to excite D_II states for SMLM and a 561 nm laser to excite the monomers for conventional fluorescence. For BODIPY red, adjust the 561 nm and 640 nm laser powers to visualize bulk fluorescence in the red channel and single molecule bursts in the far-red channel. The typical power for 561 nm is ~0.06 W/cm² and ~5 kW/cm² for 640 nm. For BODIPY green, expect to also see conventional fluorescence in the green emission channel under 561 nm excitation. For BODIPY red, expect to see conventional fluorescence in the red channel under 640 nm excitation. This signal arises from anti-Stokes emission, which becomes useful for monomer/dimer co-localization images with continuous laser excitation.

4. Data acquisition

1. Load laser shutter sequences for the excitation of monomers as well as dimers.
   NOTE: We typically use nine single molecule excitation frames at 561 nm followed by one conventional excitation frame at 488 nm. This offers a brighter conventional fluorescence signal and avoids leaking of another 488 nm excitable fluorophore such as green fluorescent protein (GFP) into the red single-molecule detection channel in multi-color imaging applications. Alternatively, turn on the 561 nm laser continuously and rely on the anti-Stokes emission for conventional images in the shorter wavelength channel.
2. Tune the laser powers such that single molecule fluorescence bursts are detected in the red-shifted emission channel under 561 nm excitation, and conventional fluorescence appears in the green emission channel with 488 nm excitation. Typical laser powers of the 561 nm laser will be around 0.8-1 kW/cm² for SMLM, and 0.035-0.07 W/cm² for the 488 nm laser in the conventional fluorescence imaging mode.
3. Choose a destination folder for movies and record 5,000-20,000 acquisition frames to collect enough localizations for reconstructing super-resolution images.
4. Move to different fields of view and repeat the steps above to collect data from more cells.

5. Data analysis and single-molecule tracking

1. Load the movie into a SMLM analysis software.
   NOTE: Any software¹⁸ can be used. We use INSIGHT (see the Table of Materials) and cross-validate the results using the ThunderSTORM¹⁹ plugin for imageJ (Fiji).
2. Visually screen the movie and adjust contrast settings such that single-molecule fluorescence blinking is visible. If needed restrict the region or the frame range for SMLM data analysis if parts of the sample are continuously fluorescing.
3. Set single molecule identification parameters for fitting with 2D Gaussian PSFs (ROI: 7 x 7 pixels with pixel size 160 nm, width 260-650 nm, height > 50 photons). Visually screen through some example frames to check the identification parameters and to reliably detect the distinct single molecule fluorescence bursts (see Figure 1C).
   NOTE: Certain identification parameters such as height and width can be slightly adjusted to optimize the recognition of visually perceived single molecule fluorescence signals.
4. Perform SMLM image analysis with the optimized identification parameters and then render each single molecule as a 2D Gaussian whose width is weighted by the inverse square root of the detected number of photons.
5. Assess the quality of the data. Use restricted frame ranges to observe single molecule distributions at more specific instances in time. This can account for organelle movement during data acquisition.
6. To further analyze the spatial distribution and dynamics of the molecule distributions, export the obtained molecule list containing the coordinates, frames of appearance, photons, widths and heights of the localizations. Import the molecule list in custom written analysis procedures.
7. For obtaining spatial information of the single molecule distribution, calculate the radial distribution function ρ(r), which represents the density of localizations as a function of the radial distance²⁰. To obtain ρ(r), calculate unique pair-wise distances of all localizations, construct the histogram with bins centered at r_i with height H(r_i) and with a width dr and divide by 2πr_i*dr; (ρ(r_i) = H(r_i)/π(r_i+dr)²-r_i²). The radial distribution function can then be used to quantify and compare the degree of clustering as well as the characteristic size of clusters.
8. To obtain dynamic information about the diffusion of molecules, link localizations that appear for example within 3 pixels (0.48 µm) in consecutive data acquisition frames to create single molecule traces.
   NOTE: The linking distance will depend on the diffusion of molecules and the density of localizations. The maximum linking distance can be estimated by analyzing the density of localizations in each frame^21,22. The average density was determined to be 0.043 localizations per µm²; thus a 0.48 µm radius was within a low enough density to ensure that different molecules were not linked together.
9. Average the displacements for different lag times Δt from multiple traces lasting at least three lag times to create a mean squared displacement (MSD) vs. Δt plot. Fit the MSD vs. Δt curve with the equation MSD = 4DΔt + σ² to calculate the average diffusion coefficient D³.

Results

Here, we present an optimized sample preparation, data acquisition and analysis procedure for SMLM using BODIPY conjugates based on the protocol above (Figure 1A). To demonstrate an example of the workflow for acquiring and analyzing SMLM data, we employ BODIPY (493/503) in yeast to resolve LDs below the optical diffraction limit (Figure 1B-F). Examples of the different multi-color imaging modes of BODIPY in conjunction with other probes such as...

Discussion

In this protocol, we demonstrated how conventional BODIPY conjugates can be used to obtain SMLM images with an order of magnitude improvement in spatial resolution. This method is based on exploiting previously reported, red-shifted D_II states of conventional BODIPY dyes, which transiently form through bi-molecular encounters. These states can be specifically excited and detected with red-shifted wavelengths and are sparse and short-lived enough for SMLM. By tuning the concentration of BODIPY monomers along wi...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
BODIPY C12	ThermoFisher	D3822	Green fatty acid analog
BODIPY C12 Red	ThermoFisher	D3835	Red fatty acid analog
BODIPY(493/503)	ThermoFisher	D3922	Neutral lipid marker
Concanavalin A	Sigma-Aldrich	C2010	Cell immobilization on glass surface
Drop-out Mix Complete w/o nitrogen base	US Biological	D9515	Amino acids for SCD
Dextrose	Sigma-Aldrich	G7021	Carbon source for SCD
Eight Well	Cellvis	C8-1.58-N	Chambered Coverglasses
Eight Well, Lb-Tek II	Sigma-Aldrich		Chambered Coverglasses
ET525/50	Chroma		Bandpass filter
ET595/50	Chroma		Bandpass filter
ET610/75	Chroma		Bandpass filter
Fetal Bovine Serum (FBS)	Gibco	26140079	Serum
FF652	Semrock		Beam splitter
FF731/137	Semrock		Bandpass filter
FluoroBrite DMEM	ThermoFisher	A1896701	Cell culture medium
Hal4000	Zhuang Lab, Harvard University		Data acquisition software
Ixon89Ultra DU-897U	Andor		EMCCD camera for photon detection
Laser 405, 488, 561, 640 nm	CW-OBIS		Lasers for excitation
Insight3	Zhuang Lab, Harvard University		Single molecule localization software
L-Glutamine	Gibco	25030-081	Amino acid required for cell culture
live-cell imaging solution	ThermoFisher	A14291DJ	Imaging buffer
Lysotracker Green	ThermoFisher	L7526	Bodipy based lysosome marker
Mammalian ATCC U2OS cells (Manassas, VA)	Dr. Jochen Mueller (University of Minnesota)
Nikon-CFI Apo 100 1.49 N.A	Nikon		Oil immersion objective
Penicillin streptomycin	Gibco	15140-122	Antibiotics
Sodium Pyruvate	Gibco	11360-070	Supplement for cell culture
T562lpxr	Chroma		Beam splitter
Trypsin-EDTA	Gibco	15400-054	Dissociation of adherent cell
W303 MATa strain	Horizon-Dharmacon	YSC1058	Parental yeast strain
Yeast Nitrogen Base	Sigma-Aldrich	Y1250	Nitrogen base without amino-acids
zt405/488/561/640rdc	Chroma		Quadband dichroic mirror