Experimental
Keywords: Basics of Single-Molecule Localization Microscopy, Imaging Buffer, Image Aquisition, Image Reconstruction, Labelling of Target Proteins, Sample Preparation, Microscopy Systems
Basics of Single-Molecule Localization Microscopy
The quality of a super-resolution image generated with single-molecule localization microscopy depends on experimental parameters such as the labelling density of the structural feature, the signal-to-noise ratio, the photoswitching of the fluorophore and the thresholding parameters used for image reconstruction.
Photoactivatable fluorescent proteins and organic fluorophores are the two main classes of fluorophores used for single-molecule localization microscopy (SMLM). They differ in the photophysical mechanism of photoactivation, photon emission yield and conjugation chemistry. Photoactivatable fluorescent proteins (paFPs) can be photoswitched in aqueous buffers. In contrast, organic fluorophores are often photoswitched by employing redox-chemistry induced transitions into non-fluorescent sates, and thus require specific chemical buffers. Often, 10-100 mM of a thiol compound at pH values rainging from 7.2 to 8.0 is used. Eventually, the oxygen concentration is reduced using enzymatic or chemical reactions in order to reduce photobleaching (e.g. for the carbocyanine fluorophores Cy5 and Alexa Fluor 647).
Using only paFPs for dual color experiments does not need a buffer change to achieve reducing conditions. Aqueous buffers (like PBS) or immobilization media can be used during data acquisition. In contrast, the challenge of dual-color experiments with organic dyes is to find matching buffer conditions, under which the two used fluorophores still show optimal photoswitching. Without optimized redox conditions, thiols interfere with photoswitchable proteins and can induce blinking as well as permanent bleaching of the fluorophores. In the following, some examples are described of how dual-color experiments were realized in theory.
Multi-color SMLM can be carried out either using two photoactivatable fluorescent proteins, two photoswitchable organic fluorophores, or a “hybrid” approach using one of each. Imaging is performed in a sequential way, starting with the far-red dye first. A suitable combination of two organiy fluorophores are Alexa Fluor 647 (or Cy5) and Alexa Fluor 532, in a PBS- or tris-buffered solution with 100mM mercaptoethylamine and a pH of 8.0 (see below for protocopls). The combination of Alexa Fluor 647 (or Cy5) and mEos2 in a PBS buffer at pH 7.2 and with 50 mM MEA is found to work well for the “hybrid” approach.
Imaging Buffer
The imaging buffer for SMLM experiments depend on the fluorophores used. Three stock solutions have to be prepared in advance, and can be kept in the freezer (note that poor photoswitching might occur if the thiol stock has aged; a fresh solution of MEA might be a solution to this).
(1) 100 mg/mL stock solution of glucose
100 mg glucose
900 µL PBS
100 µL glycerol
(2) Glucose oxidase and catalase stock solution
1 mg glucose oxidase (Sigma G2133-50KU)
20 µL Tris-HCl pH 7.5 (1M Stock)
25 µL KCl (1M Stock)
2 µL catalase from bovine liver (Sigma C100-50MG)
4 µL TCEP (from 1M stock) or DTT at same concentration.
0.5 mL glycerol
0.45 mL water
(3) 1M stock solution of mercaptoethylamin in PBS
114 mg cysteamine hydrochloride (MEA) (Sigma M6500-25G)
1 mL PBS
The imaging buffers for various fluorophores are listed below. Depending on the type of sample chamber (8-well chamber, microfluidic channel), a sufficient volume is added to the sample and eventually sealed with silicon.
Imaging buffer for imaging of Alexa Fluor 647/Cy5
400 µL of buffer (1)
50 µL of buffer (2)
50 µL of buffer (3)
450 µL of PBS
5 µL of 1M NaOH
Final pH should be around 7.4
Imaging buffer for imaging of Alexa Fluor 532
400 µL of (1)
50 µL of (2)
100 µL of (3)
400 µL of PBS
15 µL of 1M NaOH
Final pH should be 7.6 – 8.0
Imaging buffer for two-color imaging of Alexa647/Cy5 and mEos2
400 µL of buffer (1)
50 µL of buffer (2)
50 µL of buffer (3)
450 µL of PBS
5 µL of 1M NaOH
Final pH should be around 7.4
Imaging buffer for two-color imaging of Alexa647/Cy5 and Alexa532
400 µL of (1)
50 µL of (2)
100 µL of (3)
400 µL of PBS
15 µL of 1M NaOH
Final pH should be 7.6 – 8.0
Image Aquisition
Imaging is performed either in widefield, TIRF or inclided illumination mode. The integration time of the camera (iXon DV 887, Andor, Ireland) should be adjusted such that a blinking event can be captured on one or two frames on average (typically 10 - 200 ms). Specific settings of the two microscopes managed by our group are listed below.
- Integration time: 30 - 50 ms (Alexa Fluor 647/Cy5, Alexa Fluor 532), 100 - 200 ms (mEos2)
- Pre amplifier gain: 1
- EM gain: 200
- Frame transfer: on
- Read out rate: 10MHz at 14 bit
- Number of images: 5000 (very minimum) to 50000 frames
Custom-built setup
This setup is equipped with an argon/krypton ion laser (Innova70C, Coherent, Germany); excitation intensities were measured behind the objective.
- Alexa Fluor 647/Cy5: 647 nm at 35 to 50 mW (ca. 5 kW/cm2)
- mEos2: 568 nm at 25 to 30 mW (ca. 3 kW/cm2)
- Alexa Fluor 532: 514 nm at 100 mW (ca. 10 kW/cm2)
SMLM setup based on an Olympus TIRF 2
This setup is equipped with five laser modules emitting at 640 nm (LAS/640, Olympus, Japan), 561 nm (Sapphire561-FP, Coherent, Germany), 532 nm (EO-PS-II, Eksma Optics, Lithuania), 491 nm (LAS/491, Olympus, Japan) and 405 nm (Cube, Coherent, Germany).
- Alexa Fluor 647/Cy5: 640 nm at 35 mW (ca. 4 kW/cm²)
- mEos2: 561 nm at 60 mW (ca. 5 kW/cm²)
- Alexa Fluor 532: 532nm at 300 mW (ca. 20 kW/cm²)
Reactivation of Alexa Fluor 647/Cy5 in the absence of mEos2 and Alexa Fluor 532 was realized by low irradiation intensities ar 514 (532) nm (µW range); mEos2 was photoactivated at low irradiation intensities at 405 nm.
Image Reconstruction
The intensity pattern of a single fluorophore is described by a Bessel function. For practical reasons, this pattern is often approximated by a Gaussian function. We use custom-written software to analyze single-molecule movies and to generate a super-resolution image, e.g. rapidSTORM, QuickPALM (Plug-in for ImageJ/FIJI) or SimpleSTORM.
Different to a conventional microscopy image that contains intensity information, an SMLM image is build from a list of coordinates that were determined from single-molecule emission patterns. This offers the opportunity to post-analyze SMLM images using various coordinate-based algorithms and to determine cluster sizes, to quantify protein copy numbers or to estimate colocalization parameters. A suitable software we use for this purpose is LAMA.
Labelling of Target Proteins
The labelling density is a key parameter in super-resolution microscopy. In order to resolve a (continuous) structure, the fluorophore has to be placed at half the distance of the desired resolution (sampling theorem, Nyquist-Shannon criterion). This can be challenging for dense structures of high dimensionality: not only that a large number of fluorophores needs to be introduced; at the same time, the photoswitching rates have to be set accordingly, in order to avoid multiple emitters imaged at the same time.
Proteins can be labeled by immunochemistry employing a pair of primary and secondary antibodies. Tagged to a fluorescent dye, the secondary antibody is imaged with a fluorescence microscope. This implies that the fluorophore can be spaced 10 to 20 nm away from the target protein. As antibodies can also bind unspecifically, background signal occurs. Proteins can also be labeled genetically by introducing a photoactivatable fluorescent protein. This allows for stoichiometric labeling, and a suitable design of an experiment even allows protein couting. Typically, fluorescent proteins emit a smaller number of photons, which reduces the localization precision (which itself affects the spatial resolution).
Protein tags such as CLIP/SNAP represent an alternative approach: a mutant of the O6-alkylguanine-DNA alkyltransferase (AGT) is fused to the protein of interest, which itself bind O6-benzylguanine (BG) derivatives. The BG derivatives themselves are tagged to a fluorophore. Other approaches include the Halo-tag and the concepts of click chemistry.
Sample Preparation
The following section describes a possible experimental procedure for sample preparation for an SMLM experiment. Specific experiments might need some adaptation in some points. All volumes and amounts in the described steps refer to the preparation of a single well of a 8-well chambered coverslip (approximate volume of 800 µL; Labtek I, Thermo Scientific Inc.).
Surface cleaning
NaOH (200 µL, 2 M, AppliChem) was pipetted into the chamber and incubated for 20 minutes. After washing with PBS (3 x 5 minutes, 200 µL, AppliChem) poly-L-lysine (150 µL, Sigma) was added and incubated for 10 minutes and washed with PBS (3 x, 200 µL).
Seeding of HeLa cells
Cell handling was performed under sterile conditions. The growth medium was taken off the cell culture stock (HeLa) and the culture flask was washed once with PBS (1 mL). Trypsin (1.5 mL, Sigma) was added and the cell culture flask was swung for an equal distribution. The trypsin was taken off and the cell culture flask was incubated for two minutes at 37°C and 5% CO2 in the incubator (Nuaire DHD Autoflow, IBS Integra BioScience). The detached cells were taken up in growth medium (5 mL, DMEM, 10% FCS, 5% PenStrep, 5% L-glutamine) and diluted to a concentration of 37.5 cells per µL. The diluted cell suspension (400 µL) was pipetted onto the previously prepared sample chamber (15000 cells total). The cells were incubated for 24 hours at 37°C and 5% CO2.
Transfection with actin-mEos2
The plasmid (0.12 µL, 857 ng/L, actin-mEos2) was pipetted into serum-free medium (39.7 µL, DMEM, 5% PenStrep, 5% L-glutamine). Transfection reagent (0.2 µL, TurboFect, Fermentas) was added and the mixture was incubated for 15 minutes at room temperature. The transfection mixture was added to the growth medium in the chamber of the cell culture. The chamber was swung carefully and incubated for 12-24 hours at 37°C and 5% CO2.
Fixation and Immunostaining
Growth medium was removed from the sample chamber and an aqueous formaldehyde solution (200 µL, 3 v/v% in PBS, Polyscience) was added. After 15 min of incubation at room temperature, the solution was removed and the chamber was washed with PBS (3 x 200 µL, 5 min). A solution of fluorescent fiducial markers (200 µL, TetraSpeck, Invitrogen, 0.1 µm in diameter, diluted 1:500 in PBS containing 5% BSA) was added and incubated for 30 minutes (optionally; or 1:1000 dilution within the dSTORM imaging buffer). After washing with PBS (3 x 200 µL, 5 min), a solution with primary antibody (150 µL, 2 µg/mL) was added and incubated for 60 minutes at room temperature. The cells were washed with PBS (3 x 200 µL, 5 min) and a solution of secondary antibody (150 µL, 5 µg/mL) was added. After 60 minutes of incubation at room temperature, the cells were washed with PBS (3 x 200 µL, 5min) and post fixed with PFA (200 µL, 3 v/v% in PBS, Polyscience; this step is crucial, since antibodies detach from their target in the imaging buffer) for 15 minutes at room temperature. After a final round of washing with PBS (3 x 200 µL, 5 min), cells are ready to image or can be stored at 4°C for later imaging.
Microscopy Systems
SMLM can be performed on any standard widefield or TIRF microscope with some additional capabilities such as a powerful laser source, an objective with a high numerical aperture capable of TIRF illumination, and a single-photon sensitive camera such as an electron-multiplied charge coupled device (EMCCD) or a scientific complementary metal oxyde semiconductor (sCMOS) camera. We operate two different setups that are described below.
Custom-built microscope
The microscope body is based on an inverted microscope (Olympus IX71, Olympus, Japan) and placed on an air-damped optical table. The system is equipped with an immersion oil objective suitable for TIR/HILO imaging (Plan ApoN 60x / 1.45 oil, Olympus, Japan).
The excitation light is provided by a multi-line Ar/Kr-ion laser (InnovaI70C, Coherent, USA). An acousto-optical tunable filter (AOTF) (Pegasus) selects the desired excitation wavelength. A clean-up filter is used to block any other wavelengths. A second laser generates light of 405 nm (Cube, Coherent, USA). Both lasers are overlaid by dichroic and silver coated mirrors (suitable for the visible spectrum of light). The combined lasers are focused onto the back-focal plane of the objective via a lens telescope. A manually movable mirror can be used to direct the beam into TIR/HILO illumination. Emission light is collected by the objective and separated from the excitation light by a dichroic mirror. A pair of long path and band path filter is used to filter out any (non emission) background light. The emitted light is then focussed by the tube lens and a telescope (2.5 x magnification) onto the chip of an EMCCD camera (iXon DV 887, Andor, Ireland). The physical pixel size is 16 x 16 µm2, we operate various lenses that lead to an image pixel size of 107 x 107 nm2.
This system can be used for dual-color imaging of Alexa Fluor 647 (or Cy5)/Alexa Fluor 532 and Alexa Fluor 647 (or Cy5)/mEos2 (or mMaple) in 2D mode.
Enhanced OlympusIX81 TIRF Microscope
The OlympusIX81 (Olympus, Japan) is a fully-automated widefield fluorescence microscope, equipped with an oil immersion objective (UApo N 150 x 1.45, Olympus, Japan) and an EMCCD camera (iXon DV897DC, Andor, Ireland). The image pixel size is 107 x 107 nm2. The system is equipped with five laser sources emitting at 405, 491, 532, 561 and 640 nm (405 nm: Cube, Coherent, Germany; 488 nm: LAS/491, Olympus, Japan; 561 nm: Sapphire561-FP, Coherent, Germany; 640 nm: Las/640, Olympus, Japan). Except for the 532 nm laser (EO-PS-II, Eksma Optics, Lithuania), all lasers are coupled into the fully-automated TIRF unit via optical fibres (cell^TIRF motorized multicolor TIRF, Olympus, Japan). The TIRF unit enables motorized adjustment of the illumination angle (widefield or TIRF/HILO). The 532 nm laser is freely coupled into the microscope via a port of the TIRF unit, which was originally designed for only widefield epi-fluorescence imaging with a metal halide lamp. A telescope focuses the beam onto the back focal plane of the objective. A manual switch at the TIRF unit can change between epi (532 nm) and TIR/HILO (405, 491, 561, 640 nm) illumination.
This system can be used for dual-color imaging of Alexa Fluor 647 (or Cy5)/Alexa Fluor 532 and Alexa Fluor 647 (or Cy5)/mEos2 (or mMaple) in both 2D and 3D (astigmatism) mode.
The above paragraphs were adapted from the PhD thesis of B. Flottmann 2014.