Single-molecule microscopy of molecules tagged with GFP or RFP derivatives in mammalian cells using nanobody binders Academic research paper on "Chemical sciences"

Accepted Manuscript

Single-molecule microscopy of molecules tagged with GFP or mRFP derivatives in mammalian cells using nanobody binders

Evgenia Platonova, Christian M. Winterflood, Alexander Junemann, David Albrecht, Jan Faix, Helge Ewers

PII: DOI:

Reference: To appear in:

S1046-2023(15)30007-4 http://dx.doi.Org/10.1016/j.ymeth.2015.06.018 YMETH 3734

Methods

Received Date: Revised Date: Accepted Date:

26 March 2015 3 June 2015 24 June 2015

Please cite this article as: E. Platonova, C.M. Winterflood, A. Junemann, D. Albrecht, J. Faix, H. Ewers, Single-molecule microscopy of molecules tagged with GFP or mRFP derivatives in mammalian cells using nanobody binders, Methods (2015), doi: http://dx.doi.org/10.1016/j-ymeth.2015.06.018

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Single-molecule microscopy of molecules tagged with GFP or mRFP derivatives in mammalian cells using nanobody binders

Evgenia Platonova1'2, Christian M. Winterflood1'2, Alexander Junemann3, David Albrecht1'2, Jan Faix3 and Helge Ewers1'2'4*

'Randall Division of Cell and Molecular Biophysics, King's College London, SEI 1UL, United Kingdom

2Institute of Biochemistry, ETH Zurich, Zurich, Switzerland 3Medical Highschool Hannover, Hannover, Germany

4Institut für Biochemie und Chemie, Freie Unive rlin, Thielallee 63, 14195

Berlin, Germany

*Corresponding author: Helge Ewers

Institut für Biochemie und Chemie Freie Universität Berlin Thielallee 63 14195 Berlin Germany

Tel: +49 30 838 60644 Email: helge.ewers@fu-berlin.de

Abstract

With the recent development of single-molecule localization-based superresolution microscopy, the imaging of cellular structures at a resolution below the diffraction-limit of light has become a widespread technique. While single fluorescent molecules can be resolved in the nanometer range, the delivery of these molecules to the authentic structure in the cell via traditional antibody-meditaed techniques can add substantial error due to the size of the antibodies. Accurate and quantitative labeling of cellular molecules has thus become one of the bottlenecks in the race for highest resolution of target structures. Here we illustrate in detail how to use small, high affinity nanobody binders against GFP and mRFP family proteins for highly generic labeling of fusion constructs with bright organic dyes. We provide detailed protocols and examples for their application in super-resolution imaging and single-particle tracking and demonstrate advantages over conventional labeling approaches.

1. Introduction

In recent years the limit of optical resolution has been pushed down to a few nanometers with the development of several groundbreaking fluorescence superresolution imaging methods. Several of them rely on the light-induced temporal separation of the fluorescence emission of single fluorophores (PALM[1], STORM[2], dSTORM[3]). Such single-molecule localization microscopy (SMLM) techniques have enabled the resolution of fluorescence-labeled features in unprecedented detail and are becoming increasingly used to decipher the organization of sub-cellular structures. Traditionally, fluorescence microscopy is highly attractive for the ability to visualize locations and interactions of biomolecules due to relatively simple and quick sample preparation, outstanding contrast and high specificity. However, due to the nature of the SMLM technique, several problems that could be neglected in diffraction-limited microscopy pose significant problems in SMLM. Since SMLM depends on the resolution of thousands of single molecules, the limiting factor is the successful labeling of the target molecule. While existing bright fluorophores such as Alexa Fluor 647 (AF647) afford enough photons for a high

accuracy in the localization of the dye in the sample, whether the final image accurately represents the target structure largely depends on the efficiency and accuracy of its delivery. If the dye is not delivered into close proximity of the target structure or binds unspecifically to cellular structures, that leads to significant distortions.

Traditionally, bright fluorophores are delivered by indirect labeling with a primary target-specific antibody and a secondary antibody tagged with a bright dye (Figure 1A). There are two problems to this method: first the availability of specific primary antibodies, which stands even more acute when those antibodies should be from different species for successful multi-color imaging, and secondly the relatively

large size of IgG molecules, which is in the range of 10 nm. The latter might be overcome by directly labeling primary antibodies or Fab fragments. However, this requires additional procedures, such as antibody fragmentation, fluorophore conjugation and testing. This is generally time-consuming and expensive making it inflexible and suitable only in specific cases. Only for a limited number of molecules specific small binders exist, which can bring the dye into closest proximity possible to the target such as the actin-binding toxin phalloidin (Figure 1D). Efforts have been made to develop more generic small-sized binders which include DNA aptamers [4] or bicyclic peptides [5] created using in vitro selection, but these approaches have not found wide application in SMLM yet.

Other strategies to bring fluorophores close to the target structure for SMLM make use of genetically encoded labels, which comprise a broad category of tags generated in a field that has been very active due to the requirement of switchable fluorophores for SMLM [1]. These include recombinantly expressed photoactivatable and photoconvertible fluorescent fusion proteins such as mEos2 [6], PAmCherry [7] or mMaple [8], but their superior labeling specificity through the genetic fusion with the target molecule comes at the expense of a lower localization precision due to generally lower photon yield in comparison with organic fluorescent dyes. Another approach is to use special catalytic protein fusion tags, like SNAP [9] or Halo [1,10] which undergo selective covalent linkages with their synthetic chemical ligands Figure 1C). This requires the generation and characterization of new constructs. Since the fusion-proteins can only be made visible via their ligands, the estimation of transfection efficiency and verification of protein functionality in live cells becomes problematic. Furthermore, it is harder to access to what extent quantitative labeling can be achieved.

wh (Fi

Here we demonstrate a method that relies on the expression of recombinant constructs labeled with fluorescent proteins, which are used as epitopes for the binding of small high-affinity alpaca antibody-fragments, called nanobodies. Camelids such as llamas and alpacas have IgG molecules composed of two heavy chains only, without light chains [11]. As a consequence, the epitope-binding hypervariable domain consists of a single amino-acid chain in a small (~10-15 kDa) fold, which can easily be expressed recombinantly and has been named a nanobody. The small size and ease of production and genetic engineering of nanobodies has made them promising targets for biomedical imaging and as molecular therapeutics [12,13]. In biological research, while several nanobodies especially against cytoskeletal components have been developed [14], and specific nanobodies fused to GFP have been for intracellular detection as chromobodies [15]. Especially nanobodies against the green fluorescent protein (GFP) have been successful reagents for pulldown of tagged molecules [16] and in structured illumination [17] and single molecule localization superresolution microscopy [18]. The latter use nanobodies to deliver single organic dyes to single GFP proteins in cells is explained in detail in this manuscript.

Specifically, we make use of two nanobodies generated against GFP and RFP. The widespread usage and continuous development of fluorescent fusion proteins has led a wealth of well-characterized constructs. Many of the fluorescent proteins used today e derived from small successive modifications of only a few ancestral proteins, namely the green fluorescent protein from Aquorea victoria and the red fluorescent protein DsRed from Discosoma sp. Nanobodies against GFP and mRFP and many derivatives as detailed in Table1. The GFP and RFP nanobodies were use to deliver bright organic dyes which are highly suited for single molecule-based superresolution imaging and for single-molecule tracking applications [2,18,19]. In the

following, we will use data generated in our lab to illustrate the specific advantages of nanobodies over conventional labeling strategies for single and multicolor single molecule localization microscopy, as well as for single and multicolor single-particle tracking (SPT).

2. Application of nanobodies to study cellular ultrastructure

2.1 Increased labeling density and resolution in nanobody-labeled structures.

First, we demonstrate the performance of nanobody labeling of dense multicomponent assemblies in the crowded cellular environment. The nuclear pore complex (NPC) is highly packed cylindrical multi-protein structure with an outer diameter of ~120 nm. We used human U2OS cells stably expressing the nuclear pore protein Nup43 tagged with EGFP for SMLM imaging via AF647 labeled anti-GFP nanobody. We were able to visualize ring structures of ~100 nm in diameter on the lower surface of the nucleus of fixed cells, which is consistent with the predicted position of Nup43 (Szymborska et al., 2013) (Figure 2A).

In some cases even the rotational eightfold symmetry arrangement could be observed. We made a comparison with the commonly used indirect immunofluorescence method with primary anti-GFP antibody and secondary anti-mouse antibody labeled with AF647 to visualize Nup43-EGFP (Figure 2B). Though the reconstructed image contained localizations that could be attributed to the NPC, it was not possible to observe the full ring structure, likely owing to steric hindrance against labeling with relatively large antibodies in the dense environment of the nuclear pore complex. Therefore, we conclude that the small size of nanobodies provides superior accessibility of epitopes in the crowded cellular environment. We must note here, that generally it is possible to visualize the NPC structure in SMLM measurements with antibodies, at least against some epitopes [3,20,21], and we used this example to draw attention to how drastically epitope accessibility can affect the resulting superresolution image.

Finally, we demonstrate that the size of the nanobody improves the resolution due to a minimized linkage-error. We also show that the use of nanobodies supersedes the resolution obtainable using proteins directly tagged with common photoactivatable

fluorescent proteins. To do so, we labeled caveolae, which are flask-shaped invaginations of the plasma membrane 60-80 nm in diameter formed by the membrane associated protein Caveolin 1 (Cav1), using three different strategies. First, we used anti-Cav1 primary antibodies and AF647 tagged secondary antibodies to detect caveolae in rat epithelial NRK52E cells. When we performed the SMLM measurement we were able to visualize Cav1 assemblies on the membrane of fixed cells as rings with the average diameter of ~116 nm, resulting from the two-dimensional projection of a hollow cup-shaped caveolae in the plane of the membrane (Figure 2C). Next, we transiently overexpressed Cav1-EGFP in NRK52E cells and performed labeling with AF647-anti-GFP nanobodies. The caveolae seen in the reconstructed SMLM image had the same morphology, but exhibited a smaller diameter of ~94 nm (Figure 2D). Finally, we performed SMLM imaging of the caveolae in HeLa cells transiently transfected with Cav1-PAmCherry. In the reconstructed SMLM images caveolae were manifest as point-clouds of ~128 nm in diameter and a ring shape of the structure was detectable merely in few cases (Figure 2E). Although genetic tagging with a PAFP provides superior labeling efficiency and specificity with the smallest linkage error of the three compared labelling methods,

the photon yield of PAmCherry is 4-5 times lower than AF647 resulting in a roughly two-fold lower localization precision. Taken together, we conclude that nanobody labeling allows the resolution of finer details in SMLM measurements (Figure 2F).

2.2 Resolution in dual color nanobody-mediated single molecule localization microscopy.

We next used nanobodies for dual color SMLM via spectral demixing [4,22]. This method provides high resolution for both colors, minimizes cross-talk, and is free from chromatic aberrations and thus overall allows for very accurate colocalization

analysis in SMLM [5,23]. We here aimed to demonstrate the quality of labeling via nanobodies by accessing two distinct members of a small compact multiprotein structure. To that end, we performed dual color imaging of caveolae at the plasma membrane of fixed cells. Human U2OS cells were co-transfected with Cavl-EGFP

and Cavl-mCherry and labeling was performed with AF647-anti-GFP and CF680-

anti-RFP nanobodies. Reconstructed SMLM image showed that the two fusion proteins, as expected, colocalized on the scale of few tens of nanometers within caveolae revealing their detailed morphology (Figure 3A). In addition, we were able to resolve the hollow ring-shaped projection of the caveolae in both colors, as two peaks less than 50 nm apart are detectable in the lateral profiles of both channels (Figure 3B).

Next, we challenged our method to visualize two closely spaced, but morphologically different components of internalizing caveolae. Dynamin2 is a GTPase that is shown to associate with caveolae and operates as a fission collar around the neck of mature caveolae initiating their internalization from the plasma membrane to form cytosolic vesicles. The dominant negative form of dynamin2, Dyn2K44A [1,24], a mutant that cannot bind GTP, inhibits caveolar fission and arrests mature caveolae on the plasma membrane with Dyn2K44A accumulated around not completely constricted necks. When we transiently co-expressed Cavl-mCherry and Dyn2K44A-EGFP in U2OS

cells, fixed cells and labeled them with AF647-anti-GFP and CF680-anti-RFP nanobodies, we could clearly resolve by SMLM imaging the association of Dyn2K44A-EGFP localizations with Cavl-mCherry-positive structures and found the Dyn2K44A-EGFP staining to appear as a discrete spot in the center of the caveolae (Figure 3C). Consequently, in the lateral profile of a single caveolae, Dyn2K44A-EGFP staining appears as a single peak centered between two peaks for Cavl-mCherry (Figure 3D).

2.3 Single particle tracking

We next demonstrate how nanobodies can be used in the uPAINT [6,25] modality of SPT to study cellular membranes. Here, the protein of interest carrying an extracellular fluorescent tag is labeled with a nanobody bearing a bright and photostable dye whose spatial position can be determined with nanometer accuracy over time. In this way, the motion of transmembrane proteins can be quantitatively described on a single-molecule level.

We show here simple and elegant way of employing nanobodies for dynamic and nanoscopic investigation of small (less than 2 |m) mushroom-shaped protrusions from the neuronal dendritic shaft termed spines. Solute and membrane-bound diffusion into the peculiar-shaped spines is widely investigated [7,26,27] and it is still not entirely clear how spine access is mechanistically regulated for both types of influx.

In single spines of cultured hippocampal neurons we first performed fluorescence recovery after photobleaching (FRAP) experiments of membrane associated CD4-mRFP (Figure 4 IA-B). W e then complemented this ensemble diffusion measurement with experiments on the single-molecule level by additionally performing SPT on the same spine in vivo using Atto647N-anti-RFP nanobodies (Figure 4C). A welcome advantage being that SPT also provides the geometry of the spine with less than 15 nm resolution which is required to apply mathematical diffusion models. The combination of SPT and FRAP can help in overcoming artifacts inherent to each of the methods and to provide a more accurate description of membrane diffusion. Further, nanobodies could be successfully used to monitor the lateral mobility of two different types of membrane-associated molecules simultaneously [8,28]. We show an example of dual color SPT of L-mHoneydew-GT46 and GPI-EGFP in the plasma

membrane of live rat endothelial NRK52E cells. GPI is a glycosylphosphatidylinositol anchor that tethers in this case EGFP to the outer leaflet of the plasma membrane while GT46, a synthetic receptor, connects mHoneydew to the membrane via a single-spanning transmembrane domain. mHoneydew is a yellow-fluorescent derivative of dsRed which is recognized by anti-RFP nanobodies. As both EGFP and mHoneydew emit in the green-yellow region of the spectrum we could employ the orange-red and far-red dye pair Alexa Fluor 555 (AF555) and Atto647N. Both dyes have excellent quantum yield and high photostability and are thus ideally suited for SPT and their emission could be spectrally separated using a dichroic mirror. When AF555-anti-RFP and Atto647N-anti-GFP nanobodies were added to the imaging buffer in sub-nanomolar concentrations, we were able to trace their movement simultaneously on the basal membrane of the double-transfected cells (Figure 4D). This simple technique yields hundreds of trajectories for both molecules in a single experiment and clearly detects the difference in membrane motility of the lipid-anchored and the transmembrane molecule (Figure 4E-F). Our methods thus provides straight-forward access to dual color single particle tracking.

forward

3. Experimental procedures 3.1 Nanobody purification and labeling 3.1.1 Expression and purification of nanobodies

The coding sequence of the anti-GFP-nanobody [9,29,30] was obtained by reverse translation using Gene Designer 2.0 software (DNA 2.0 Inc., Menlo Park, USA) for optimal expression in E. coli. The designed sequence lacks the 3' end encoding the C-terminal His-tag, but carries an additional extension encoding the residues GKGSKGSKSK to markedly improve the efficiency of chemical labeling of the protein at Lysine residues. The novel protein is referred to as anti-GFP-nanobody-4K. The designed sequence was synthesized as BamHIISaH fragment by GenScript (Piscataway, USA), and ligated into the same sites of pGEX-6P-1 (GE Healthcare, USA).

GST-tagged anti-GFP-nanobodies were expressed in E. coli strain Rosetta (Promega) and purified from bacterial extracts on glutathione-conjugated agarose (Sigma-Aldrich, Germany) using standard procedures. The GST tag was subsequently cleaved by incubating the purified fusion protein with PreScission protease (GE Healthcare) in phosphate-buffered saline (PBS), pH 7.3, supplemented with 1 mM dithiothreitol (DTT) and 1 mM EDTA overnight at 4°C. After cleavage, the GST tag was separated from the anti-GFP-nanobodies by size exclusion chromatography using a preparative 26I60 Superdex G75 column (GE Healthcare) (Figure 5A). Anti-GFP-nanobody containing fractions were pooled, dialyzed against 150 mM KCl, 1 mM DTT, 60% glycerol and 20 mM imidazole (pH 7.4) and stored at -20°C for later use (Figure 5B).

3.1.2 Fluorescent labeling of nanobodies

nanobody ratio of ~ 1.5. Briefly, the nanobodies stored in PBS are dialyzed into 0.1 M

For the conjugation of fluorescent dyes to nanobodies, we use the most common technique for labeling peptides via free amines, N-hydroxysuccinimidyl ester chemistry. Anti-GFP- and anti-RFP nanobodies are labeled with N hydroxysuccinimidyl ester derivatives of AF647 (Life Technologies), CF680 (Biotium), AF555 (Life Technologies) and Atto647N (Atto-Tec) according to the manufacturer's protocol using a 5-fold molar excess of dye, resulting in a dye to

alyzed into 0.1

NaHCO3, pH 8.3 via 3 kDa MWCO desalting columns (Zeba Spin, Thermo Fisher). The lyophilized succinimidyl-esters dyes are dissolved in DMSO to 10 mg/ml, stored at -80 °C and for labeling added in 5-fold molar excess to the nanobodies. The reaction is incubated for 1-2 h in the dark at RT. The nanobody is purified from the excess of unreacted dye via buffer exchange into PBS by three consecutive times passages through desalting columns and stored at 4 °C. We noticed, that for anti-RFP nanobodies (Chromotek) addition of 15% (v/v) DMSO for labeling and further storage prevents their precipitation. The labeling ratio is determined by absorption spectrometry according to the manufacturer's labeling protocol and is in a range 1.0 1.5.

3.2 Sample preparation 3.2.1 Cell culture and transfection

ording to the

All cell lines are grown in low glucose DMEM without phenol red supplemented with

10% fetal bovine serum, penicillin, streptomycin and GlutaMAX (all Life Technologies) at 37 °C in a CO2-controlled humidified incubator. For U2OS cells stably expressing Nup43-GFP 500 |g/ml G418 (Invitogen) is added to the medium. Cells are transferred to round glass coverslips (Fisher Scientific) and transiently transfected using Lipofectamine 2000 (Life Technologies). We found that exchanging

the medium after 1-2 h of transfection greatly enhances cell viability with all cell lines used. We allow cells to grow 4-8 h after transfection for all caveolae imaging and 1848 h after transfection for all other experiments.

The preparation and cultivation of hippocampal neurons from E18 Sprague-Dawley rats is performed after standard procedures as described previously [31]. Neurons too are transfected with Lipofectamine 2000 reagent between 5 or 7 days in vitro (DIV) and imaged earliest at DIV 14.

3.2.2 Sample preparation and SMLM imaging

l vitro (DI

Generally, the sample preparation for imaging via nanobodies consists of several simple steps and can be performed in 2.5 h. This makes it very easy to perform SMLM measurements on the same day to avoid deterioration of the sample. We also note that nanobody-staining performed equally well after fixation with most common reagents: paraformaldehyde (this study, [32], glutaraldehyde [32], methanol [33], when the measures to reduce the specific to the fixative background are taken. This is important because some structures such as actin or microtubules require specific staining conditions and not all antibodies are compatible with glutaraldehyde fixation. The overall steps of the labeling protocol are visualized in a Diagram (Figure 6) and the detailed protocols for the staining procedures performed here are provided below. Before fixation for nanobody and antibody staining of nuclear pore complexes, U2OS cells stably expressing Nup43-EGFP are rinsed with warm PBS and then extracted with 0.1% (v/v) Triton X-100 in PBS for 3 min to remove the cytoplasmic pool of Nup43 molecules. Cells are rinsed once with warm PBS and fixed with 4% PFA / 2% sucrose in PBS at RT for 10 min. The fixation is stopped by rising coverslips three times in PBS, followed by 10 min incubation in PBS with 50 mM NH4Cl. To image caveolae, fixed cells are permeabilized for 5 min in

0.25% Triton X-100 in PBS. Subsequently, cells were washed and blocked with several drops of Image-IT FX (Invitrogen) for 30 min, and after that with 1% BSA / 4% horse serum / 0.004% NaN3 in PBS. The incubation with nanobodies, diluted in the 1% BSA / 0.004% NaN3 / PBS, was carried out at least for 1 h at RT or overnight at 4 °C. The incubation with primary antibodies was overnight at 4 °C and next the samples were washed and incubated with the appropriate secondary antibodies for at least 1 h at RT. This is followed by a final round of washing. For SMLM imaging, the switching buffer used contains 130 mM 2-Mercaptoethanol / 30 mM Mercaptothanolamine / 0.2 M Tris, pH 8.0 with 5 % (w/v) glucose, 0.25 mg/ml (50 U/ml) glucose-oxidase and 20 ^g/ml (600 U/ml) catalase. As nanobodies do not show an observable increased tendency of dissociating in the presence of reducing agents necessary for photoswitching it is possible to image for prolonged periods and multiple image series can be generated without any loss of localization density and image quality.

The intensity of the 643 nm imaging laser is ~2 kW/cm2 and 473 nm activation laser intensity is automatically adjusted to keep the average number of localizations per

frame constant (maximum intensity ~0.5 kW/cm2). We record a minimum of 30'000 frames with an exposure time of 25-55 ms.

3 Sample preparation for SPT measurements

live-cell single particle tracking the microscope stage temperature was stabilized to 35°C. We use physiological imaging buffer (145 mM NaCl, 5 mM KCl, 10 mM Glucose, 10 mM HEPES, 2 mM CaCl2, 1 mM MgCl2, 0.2% (w/v) BSA, 10 mM ascorbate[34], both for neurons and tissue culture cells. We add freshly diluted from a stock nanobodies to cells at a concentration of 25 pM immediately prior to image acquisition. Multiple image series are recorded, for U2OS cells typicaly 10-20 and for

neurons 5-10, of 500 frames with 25 ms exposure time and 5 ms laser illumination time.

For live-cell FRAP measurements the objective and sample holder stage of confocal microscope are heated to 37°C. For each color a prebleaching image sequence is recorded and a single 100 ms bleaching pulse applied to the region of the dendrit spine head. Post bleach imaging is performed for 200 frames with 20 ms an integration times for EGFP and mRFP respectively. Per measurement five to ten post bleach imaging sequences are recorded.

it live t

e to ten po

ritic 40 ms

3.3 Technical instrumentation and image analysis 3.3.1 dSTORM

SMLM is performed on a custom-built setup. In brief, a 473 nm laser (100 mW, Laserglow Technologies) is used for activation and a 643 nm laser (150 mW, Toptica Photonics) for imaging. Both lasers are focused onto the back-focal plane of an Olympus NA 1.49, 60x, TIRF-objective. A quad-edge dichroic beamsplitter (405/488/532/635 nm, Semrock) is used to separate fluorescence emission from excitation light. Emission light is filtered by two bandpass emission filters 700/75 nm (Chroma) and a 675/50 nm (AHF) and focused by 500 mm tube lense onto a back-illuminated EM-CCD chip (Evolve, Photometrics) which is water-cooled to -80 °C. Images are acquired with MicroManager [35].

data analysis is performed in MATLAB (Mathworks). Single positions are determined by Gaussian fitting based on a maximum likelihood estimator [36]. A previously described image-correlation based drift-correction is employed[33].

Ima All

3.3.2 SD-dSTORM

For dual color imaging a single 643 nm laser is used. Emission light is filtered by two 700/75 nm (Chroma) bandpass filters and split by long-pass dichroic beamsplitter (690 nm, AHF Analysetechnik) into the long and short wavelength channels, which are imaged onto separate parts of the CCD chip. The emission of AF647 is split roughly 45%/55% and CF680 80%/20% and produces a localization pair that can be identified in both channels. The assignment of each localization pair to either die is based on the intensity ratio. Reconstruction of the final two-color image sed on the localizations only from the long-wavelength channel.

3.3.3 SPT and FRAP

To perform SPT in one color the setup is used as described above for dSTORM. For two-color SPT measurements an additional 473 nm laser (100 mW, Laserglow Technologies) and a 556 nm laser (200 mW, Laserglow Technologies) are used. A quad-edge dichroic beamsplitter (405/488/532/635 nm, Semrock) is used to separate fluorescence emission from excitation light. Emission light is filtered by quad-band bandpass filter (446/523/600/677 nm, Semrock). A longpass dichroic beamsplitter (635 nm, Semrock) is used to separate AF555 fluorescence from Atto647N fluorescence and the separated emission beams are additionally filtered by bandpass filters 607/70 nm (Semrock) for AF555 and 700/75 nm (Chroma) for Atto647N. Two separate 500 nm tube lenses focus the emission light onto camera chip. We use multi-color fiduciary markers (100 nm diameter Tetraspeck beads, Life Technologies) for overlaying two channels in a resulting SPT image. A transformation matrix is calculated with MATLAB built-in routine cp2tform using affine transformation on a calibration image of the beads.

Molecules selected for particle tracking are required to have lateral localization precision better than 15 nm in each frame. Localizations with distance smaller than

~500 nm in consecutive frames are grouped into a track and only tracks with a minimum length of 20 consecutive frames are considered for analysis. For each track a Diffusion coefficient D was calculated from the mean square displacement (MSD) using the following relationship:

where r2 is (xj+i - x;)2 + (yi+1 - yi)2 in the z'-th frame, D is the diffusion c is the elapsed time between two successive frames. The diffusion coefficient D was calculated from the slope of a linear fit through the points z'=2 to z'=4 of the MSD plotted versus t. Nanobodies bound to the coverglass exhibited a D of < 10-4 ^m2/s

For live-cell FRAP measurements we use a custom-assembled spinning disk confocal fluorescence microscope based on an Olympus IX71 equipped with a 100 x Olympus (PlanSApo N, NA 1.40 oil) objective. Acquisition was controlled by MetaMorph software (MAG Biosystems) and the FRAP laser was controlled using the I-Las Version 1 protocol. Images are captured by an Evolve EM-CCD camera

MSD(t) = <r2> = 4-D-t

and were excluded from further analysis.

Conclusion

In conclusion, nanobody-mediated labeling provides a simple and generic approach to label virtually any of the widely available GFP- and RFP- derived fusion constructs for advanced single molecule imaging applications including single particle tracking and single molecule localization microscopy. The advantages include very precise and consistent labeling as well as superior accessibility in complex and dense samples due to the high affinity and small size of the nanobodies. For SMLM imaging applications, sample preparation is very robust and involves only a few simple steps. Imaging can be performed as early as two hours after fixation. Nanobodies are also insensitive to the choice of fixation reagent, expanding number of specimens that can be studied, and to the presence of reducing agents necessary for photo switching. Furthermore, the use of nanobodies against GFP and RFP simultaneously allows dual-color SMLM investigation of sub-cellular for any pair of functional GFP and RFP fusion constructs. This makes dual color SMLM imaging easily accessible to a myriad of biological problems. We show an example of resolving Cav1 and Dyn2K44A

proteins within small macromolecular complex of caveolae well enough to afford an \

insight into the nanoscopic geometry of their organization.

Finally, we demonstrate the application of nanobodies for SPT measurements in in combination with the ensemble technique fluorescence recovery after photobleaching, which can provide better insight into the global and local aspects of protein motility in the plasma membrane. For use in SPT, nanobodies provide better access to spatially confined membrane regions (e.g. the basal membrane or synapses [37]), while their monovalency assures the absence of crosslinking. Lastly, we demonstrate how nanobodies can be applied to simultaneous SPT of two different types of membrane molecules.

Acknowledgements

The authors thank Mike Heilemann for sharing U2OS cells stably expressing Nup43-EGFP. The authors acknowledge support from the NCCR Neural Plasticity and Repair, the Holcim Foundation, the Swiss National Fund, a Marie Curie Fellowship (CW), the NCCBI, and the SPP 1464 of the DFG (JF and AJ).

The authors declare no conflict of interest.

References

Betzig E, Patterson G H, Sougrat R, Lindwasser O W, Olenych S, Bonifacino J S, Davidson M W, Lippincott-Schwartz J and Hess H F 2006 Imaging Intracellular Fluorescent Proteins at Nanometer Resolution Science 313 1642-5

Rust M J, Bates M and Zhuang X 2006 Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM) Nature Methods 3 793-6

Heilemann M, van de Linde S, Schüttpelz M, Kasper R, Seefeldt B, Mukherjee A, Tinnefeld P and Sauer M 2008 Subdiffraction-Resolution Fluorescence Imaging with Conventional Fluorescent Probes Angew. Chem. Int. Ed. 47 6172-6

Ellington A D and Szostak J W 1990 In vitro selection of RNA molecules that bind specific ligands. Nature 346 818-22

Heinis C, Rutherford T, Freund S and Winter G 2009 Phage-encoded combinatorial chemical libraries based on bicyclic peptides Nature Methods 5 502-7

McKinney S A, Murphy C S, Hazelwood K L, Davidson M W and Looger L L 2009 A bright and photostable photoconvertible fluorescent protein Nature Methods 6 131-3

Subach F V, Patterson G H, Manley S, Gillette J M, Lippincott-Schwartz J and Verkhusha V V 2009 Photoactivatable mCherry for high-resolution two-color fluorescence microscopy Nature Methods 6 153-9

McEvoy A L, Hoi H, Bates M, Platonov; H, Liphardt J and Campbell R E 2012 use in multiple imaging modalities. PL

Cranfill P île: a photo

P J, Baird M A, Davidson M W, Ewers photoconvertible fluorescent protein for 7 e51314

Gautier A, Juillerat A, Heinis C, Correa I R Jr., Kindermann M, Beaufils F and Johnsson K 2008 An Engineered Protein Tag for Multiprotein Labeling in Living Cells Chemistry & Biology 15 128-36

Los G V, Encell L P, McDougall M G, Hartzell D D, Karassina N, Zimprich C, Wood M G, Learish R, Ohana R F, Urh M, Simpson D, Mendez J, Zimmerman K, Otto P, Vidugiris G, Zhu J, Darzins A, Klaubert D H, Bulleit R F and Wood K V 2008 HaloTag: a novel protein labeling technology for cell imaging and protein analysis. ACS Chem. Biol. 3 373-82

Hamers-Casterman C, Atarhouch T, Muyldermans S, Robinson G, Hamers C, Songa E B, Bendahman N and Hamers R 1993 Naturally occurring antibodies devoid of light chains. iNature 363 446-8

Van Impe K, Bethuyne J, Cool S, Impens F, Ruano-Gallego D, De Wever O, Vanloo B, Van Troys M, Lambein K, Boucherie C, Martens E, Zwaenepoel O, Hassanzadeh-Ghassabeh G, Vandekerckhove J, Gevaert K, Fernández L Á, Sanders N N and Gettemans J 2013 A nanobody targeting the F-actin capping protein CapG restrains breast cancer metastasis. Breast cancer research : BCR 15 R116

Oliveira S, Heukers R, Sornkom J, Kok R J and van Bergen en Henegouwen P M P 2013 Targeting tumors with nanobodies for cancer imaging and therapy. Journal of controlled release : official journal of the Controlled Release Society 172 607-17

Van Audenhove I, Van Impe K, Ruano-Gallego D, De Clercq S, De Muynck K, Vanloo B, Verstraete H, Fernández L Á and Gettemans J 2013 Mapping cytoskeletal protein function in cells by means of nanobodies. Cytoskeleton (Hoboken, NJ) 70 604-22

Rothbauer U, Zolghadr K, Tillib S, Nowak D, Schermelleh L, Gahl A, Backmann N, Conrath K, Muyldermans S, Cardoso M C and Leonhardt H 2006 Targeting and tracing antigens in live cells with fluorescent nanobodies Nature Methods 3 887-9

[16] Trinkle-Mulcahy L, Boulon S, Lam Y W, Urcia R, Boisvert F M, Vandermoere F, Morrice N A, Swift S, Rothbauer U, Leonhardt H and Lamond A 2008 Identifying specific protein interaction partners using quantitative mass spectrometry and bead proteomes The Journal of Cell Biology 183 223-39

[17] Schermelleh L, Carlton P M, Haase S, Shao L, Winoto L, Kner P, Burke B, Cardoso M C, Agard D A, Gustafsson M G L, Leonhardt H and Sedat J W 2008 Subdiffraction Multicolor Imaging of the Nuclear Periphery with 3D Structured Illumination Microscopy Science 320 1332-6

[18] Ries J, Kaplan C, Platonova E, Eghlidi H and Ewers H 2012 A simple, versatile method for GFP-based super-resolution microscopy via nanobodies. Nature Methods 9 582-4

[19] Albrecht D, Winterflood C M and Ewers H Dual color single particle tracking nanobodies Methods and Applications in Fluorescence in press

[20] Szymborska A, de Marco A, Daigle N, Cordes V C, Briggs J A G and Ellenberg J 2013 Nuclear Pore Scaffold Structure Analyzed by Super-Resolution Microscopy and Particle Averaging. Science

[21] Löschberger A, van de Linde S, Dabauvalle M-C, Rieger B, Heilemann M, Krohne G and Sauer M 2012 Super-resolution imaging visualizes the eightfold symmetry of gp210 proteins around the nuclear pore complex and resolves the central channel with nanometer resolution Journal of Cell Science 125 570-5

[22] Lampe A, Haucke V, Sigrist S J, Heilemann M and Schmoranzer J 2012 Multi-colour direct STORM with red emitting carbocyanines Biology of the Cell 104 229-37

[23] Platonova E, Winterflood C M and Ewers H A simple method for GFP- and RFP-based dual color single molecule localization microscopy ACS Chem. Biol. in press

[24] Damke H, Baba T, Warnock D E and Schmid S L 1994 Induction of mutant dynamin specifically blocks endocytic coated vesicle formation. The Journal of Cell Biology 127 91534

[25] Giannone G, Hosy E, Levet F, Constals A, Schulze K, Sobolevsky A I, Rosconi M P, Gouaux E, Tampe R, Choquet D and Cognet L 2010 Dynamic Superresolution Imaging of Endogenous Proteins on Living Cells at Ultra-High Density Biophysical Journal 99 1303-10

[26] T0nnesen J, Katona G, Rozsa B and Nägerl U V 2014 Spine neck plasticity regulates compartmentalization of synapses. Nat Neurosci

[27] Ewers H, Tada T, Petersen J D, Racz B, Sheng M and Choquet D 2014 A Septin-Dependent Diffusion Barrier at Dendritic Spine Necks. PLoS ONE 9 e113916

Albrecht D, Albrecht D, Winterflood C M, Winterflood C M, Ewers H and Ewers H Dual color single particle tracking via nanobodies Methods and Applications in Fluorescence in press

[29] Kirchhofer A, Helma J, Schmidthals K, Frauer C, Cui S, Karcher A, Pellis M, Muyldermans S, Casas-Delucchi C S, Cardoso M C, Leonhardt H, Hopfner K-P and Rothbauer U 2009 Modulation of protein properties in living cells using nanobodies Nature Structural & Molecular Biology 17 133-8

[30] Kubala M H, Kovtun O, Alexandrov K and Collins B M 2010 Structural and thermodynamic analysis of the GFP:GFP-nanobody complex Protein Science 19 2389-401

[31] Kaech S and Banker G 2007 Culturing hippocampal neurons Nat Protoc 1 2406-15

[32] Platonova E, Winterflood C M and Ewers H 2015 A Simple Method for GFP- and RFP-

based Dual Color Single-Molecule Localization Microscopy ACS Chem. Biol. 150401145430007

Ries J, Kaplan C, Platonova E, Eghlidi H and Ewers H 2012 A simple, versatile method for GFP-based super-resolution microscopy via nanobodies Nature Methods 9 582-4

Ewers H, Tada T, Petersen J D, Racz B, Sheng M and Choquet D A Septin-Dependent Diffusion Barrier at Dendritic Spine Necks journals.plos.org 9 el 13916

Edelstein A, Amodaj N, Hoover K, Vale R and Stuurman N 2001 Current Protocols in Molecular Biology ed F M Ausubel, R Brent, R E Kingston, D D Moore, J G Seidman, Smith and K Struhl (Hoboken, NJ, USA: John Wiley & Sons, Inc.)

Smith C S, Joseph N, Rieger B and Lidke K A 2010 Fast, single-molecule localization that achieves theoretically minimum uncertainty Nature Methods 7 373-5

Groc L, Lafourcade M, Heine M, Renner M, Racine V, Sibarita J B, Lounis B, Choquet D and Cognet L 2007 Surface Trafficking of Neurotransmitter Receptor: Comparison between Single-Molecule/Quantum Dot Strategies Journal of Neuroscience 27 12433-7

■e 27 12433

Figures and Tables

Table 1.

Fluorescent proteins recognized by nanobodies.

anti-GFP

anti-RFP

EGFP, wtGFP, GFP S65T mRFP

TagGFP, sfGFP, pHluorin mCherry

EYFP, YFP, Venus, Citrin mOrange

ECFP, CFP mPlum

mHoneyDew mRFPruby

mKate2

Not recognized:

TurboGFP

dsRed Ta

Figure Legends

Figure 1: Means to deliver an organic dye to a target structure. (A) Traditionally, fluorophores are delivered to target structures by sandwich labeling with a primary target-specific and a secondary labeled antibody. (B) The combination of gen< labeling with GFP and a highly specific anti-GFP nanobody delivers the dye i much closer proximity to the target structure. (C) Enzymatic labeling methods such as the SNAP-tag combine genetic labeling with covalent coupling of an organic dye. (D) For some molecules specific small molecule binders are available which bring the dye into closest proximity possible. (As an example here the interaction of phalloidin with actin). Scale bar is 4 nm.

Figure 2: Increased labeling density and resolution in nanobody-labeled structures. (A-B) Differences in epitope accessibility of nuclear pore complex components between nanobodies and antibodies. Images show the lower surface of the nucleus in fixed U2OS cells stably expressing Nup43-EGFP. (A) Nup43-EGFP labeled via AF647 conjugated anti-GFP nanobodies. (B) Nup43-EGFP labeled with primary anti-GFP antibodies and secondary AF647 conjugated antibody. (C-E) Superresolution images of caveolae on the basal membrane of fixed cells obtained using different labeling strategies. (C) Caveolae in NRK52E cells labeled via rabbit anti-Ca

av1 antibody and secondary AF647 conjugated antibody. (D) Cavl-EGFP transiently expressed in NRK52E cells labeled with AF647-anti-GFP nanobodies. (E) PALM image of Cavl-PAmCherry in HeLa cells. In the lower panels are schematics of labeling strategy and statistics on the number localization per caveolae and their apparent diameter.

Figure 3: Resolution in dual color nanobody-mediated single molecule localization microscopy. (A) Two color images of individual caveolae on the basal membrane of fixed U2OS cells containing Cav1-mCherry (red) and Cav1-EGFP (green) and a schematic view of the relative localization of EGFP (green) and mCherry (red) tagged Cav1 molecules on the caveolar membrane. (B) Intensity profiles of Cav1-mCherry (red) and Cav1-EGFP along the box shown in A. The lateral profile shows a double peak owed to the cup-shape of caveolae with similar radial distances from the center. (C) Two color images of individual caveolae containing Cav1-mCherry (green) and Dyn2K44A-EGFP (red) and a schematic view of the relative localization of Cav1-mCherry (green) on the caveolar membrane and Dyn2K44A-EGFP (red) accumulated around caveolar neck. (D) Intensity profiles of Cav1-mCherry (green) and Dyn2K44A-EGFP (red) along the box in C showing as a single peak centered inside the two peaks of Cav1.

Figure 4: Single particle tracking and fluorescence recovery after photobleaching. (A-C) Combined fluorescence recovery after photobleaching (FRAP) technique and single particle tracking measurement on a dendritic spine of a cultured hippocampal neuron expressing CD4-mRFP. (A) Individual frames of an image sequence recorded during a FRAP experiment. (B) Fluorescence recovery curves obtained from five consecutive FRAP experiments in color and their mean in black. (C) Single-molecule tracks obtained during SPT via Atto647N-anti-RFP nanobodies on the same dendritic spine as in (A). (D-E) Dual color single particle tracking experiment of L-mHoneydew-GT46 and GPI-EGFP in the plasma membrane of live NRK52E cells via Atto647N-anti-GFP and Atto555-anti-RFP nanobodies. (D) Trajectories of L-mHoneydew-GT46 (red) and GPI-EGFP (green) on the basal

membrane. (E) Histograms and (F) cumulative probability plots of diffusion coefficients.

Figure 5: Purification of recombinant anti-GFP nanobody. (A) Elution profile of anti-GFP nanobody 4K. (B) Coomassie brilliant blue staining of a 15% SDS PAGE gel of purified nanobodies.

Figure 6: Protocol for nanobody labeling. Overview of labeling prot GFP nanobodies for SMLM imaging of subcellular structures.

otocol with anti-

Supplementary Figure Legends:

Figure S1. Localization statistics for imaging of Cavl-EGFP with anti-GFP Alexa 647-labelled nanobodies shown in Figure 2D.

Figure S2. Localization statistics for imaging of Cavl-PAmCherry shown in Figure 2E.

Figure S3. Two-color localization statistics for imaging of Cavl-mCherry and Dyn2K44A-EGFP shown in Figure 3C. (Top row) Localization precision and photon statistics for Alexa Fluor 647. (Bottom row) Localization precision and photon statistics for CF680. (Right-most chart) Corresponding histogram of the normalized intensity ratio for spectral-demixing where the red line delimits the assignment of the localization pair to Alexa Fluor 647 (<0.25) and the green line to CF680 (>0.5). The fraction of disregarded localizations was below 1%.

Figure S4. Two-color localization statistics for live-cell single-particle tracking

shown in Figure 4. (Top row) Localization precision and photon counts for Atto647N. (Bottom row) Localization precision and photon counts for Alexa Fluor 555.

Two-color l in Figure 4. ( ottom row) Lo

NUP43-EGFP a-GFP nanobody staining

NUP43-EGFP a-GFP antibody staining

a-Cav1 AB

Cav1-PAmCherry

f • • % « • i •

I* 4 W«

9* • Ä * r

** 600nm

Diameter

* ( : «Cav1V-> Y Antibody • AF647 N localizations: 96 ± 20 Diameter: 116 ± 11 nm

M Caví ■ GFP I Nanobody

N localizations: 91 ± 27 Diameter: 94 ± 18 nm

% ^ ^ 1 * é. * % A-

** * «Í • * *

.ft 600nm

* % N localizations: 80 ± 26

n/ Diameter: 128 ± 26 nm

M Cav1

1 PAmCherry

— Cav1-EGFP —Cav1-mCherry

— Dyn2K44A-EGFP -Cav1-mCherry

Dyn2K44A-EGFP M Cav1-mCherry

Distance (nm)

CD4-mRFP

Time (s)

D L-mHoneydew-GT46 and GPI-eGFP

a-0 M—

m Lm-Honeydew-GT46 — GPI-eGFP

■= 1 ro 0.8

2 0.6 Q.

§ 0.4 0.2 m0

¿3 -4

-4 -3 -2 -1 0 -GPI-eGFP — L-mHoneydew-GT46

-3 -2 -1 0 1 Log(D) (Mm2/s)

1. Plate cells

2. Transfect with GFP fusion protein

3. Fix cells

4%PFA/2%Sucrose PBS 50mM NHCL PBS

4. Permeabilize

0.25% Triton X-100

5. Block

Image-It FX

4%Horse Serum/1%BSA PBS

6. Stain with nanobodies

60 min

250ng/ml anti-GFP nanobodies in 1%BSA PBS

60 min

SMLM imaging

nanobodies are the epitope-binding domains of single chain antibodies nanobodies can be recombinantly expressed in bacteria and are thus easily available

a number of nanobodies against specific proteins and tags are available nanobodies provide better sample penetration than antibodies nanobodies deliver organic dyes closer to the target structure than antibodies