Letter pubs.acs.org/NanoLett
Instant Live-Cell Super-Resolution Imaging of Cellular Structures by Nanoinjection of Fluorescent Probes Simon Hennig,*,† Sebastian van de Linde,‡ Martina Lummer,§ Matthias Simonis,† Thomas Huser,† and Markus Sauer‡ †
Biomolecular Photonics, Department of Physics and §Department of Molecular Cell Physiology, Bielefeld University, Universitätsstr. 25, 33615 Bielefeld, Germany ‡ Department of Biotechnology & Biophysics, Julius-Maximilians-University, Am Hubland, 97074 Würzburg, Germany S Supporting Information *
ABSTRACT: Labeling internal structures within living cells with standard fluorescent probes is a challenging problem. Here, we introduce a novel intracellular staining method that enables us to carefully control the labeling process and provides instant access to the inner structures of living cells. Using a hollow glass capillary with a diameter of 7.0, it can be delivered by nanoinjection, by applying a negative voltage of −0.7 to −1.0 V to the electrodes. The staining procedure was carried out in PBS at a temperature of 37 °C with a single-barrel pipet with a diameter of 100 nm filled with ∼10−5 M paclitaxel−Oregon Green (Supplementary Video 1). Time-lapse fluorescence imaging enables direct monitoring of the labeling process, its efficiency, and demonstrates that paclitaxel−Oregon Green binds specifically to the tubulin structure within seconds to minutes upon intracellular delivery (Supplementary Figure 6a−e). Tubulin staining of the entire cell is completed within 15 min after start of the experiment. Next, we injected MitoTracker Red, a widely used stain for the visualization of mitochondria in living and fixed cells. Although MitoTracker Red is cell permeable and readily enables labeling of living cells, it is impossible to control
relative to the cell, this approach uses changes in ion conductance to control the approach and penetration of the tip and subsequent delivery of charged molecules, based on scanning ion conductance microscopy.20−22 Nanoinjection provides fast, minimally invasive and traceable staining with various fluorescent probes suitable for high resolution fluorescence microscopy of living cells. To reveal the power of this approach, we labeled intracellular structures in different cell types under conducting and nonconducting buffer conditions with conventional cell permeable and nonpermeable fluorescent probes while directly monitoring the labeling process. Live-cell wide-field fluorescence and super-resolution imaging by direct stochastic optical reconstruction microscopy (dSTORM)23,24 demonstrate the effectiveness of the technique to gain an instant view of the cellular labeling process even with subdiffraction-resolution. Notably, this technique is also applicable for the efficient sequential or simultaneous delivery of multiple different targets to the same cell. Comparison of labeling densities and structural details visible in living and fixed cells indicates that live-cell nanoninjection-based delivery of fluorescent probes perfectly preserves cellular target structures and epitope multiplicity. To label living cells by nanoinjection, the nanopipette needs to carefully penetrate the plasma membrane of a cell of interest. For this purpose, the approach of the tip to the cell is monitored using the ion flow through the nanopipette by measuring the corresponding conductance at a predefined voltage of typically 60 mV between the nanopipette and a bath electrode. After penetration of the cell with the nanopipette, fluorescent probes are delivered by adjusting the voltage applied to the electrodes to adapt to the net charge of the molecules using the electrophoretic mobility of the molecules (Figure 1 and Supplementary Figures 1−4). To drive charged molecules through a glass nanopipette by an electric field, we have to consider that the SiO2 groups at the 1375
DOI: 10.1021/nl504660t Nano Lett. 2015, 15, 1374−1381
Letter
Nano Letters
Figure 2. Directed staining of mitochondria in a single living HeLa cell surrounded by other cells by nanoinjection of MitoTracker Red. (a) At the beginning of the staining process only the fluorescence response of the nanopipette is visible. (b−e) During the first 30−120 s, mitochondria in the vicinity of the injection point are gradually stained. (f) After 150 s no further expansion of fluorescence in lateral direction is observed, indicating that the fluorescent probe has spread throughout the cytosol. (g, h) With continued injection of MitoTracker Red more diffuse background signals become visible, due to staining of mitochondria in the out-of-focus axial direction. (i) 240 s after the start of the staining process the fluorescence signal did not increase any further, indicating that the staining is saturated. Fluorescence images were taken at 15 W/cm2 at 568 nm with an integration time of 150 ms exciting with a highly inclined and laminated optical sheet (HILO).26 Scale bar, 10 μm.
Figure 3. Super-resolution imaging of nanoinjection-based labeling of actin filaments. (a) Representative wide-field fluorescence image of the actin cytoskeleton in living U2OS cells, taken immediately after nanoinjection of ATTO 655−phalloidin at a concentration of 10−5 M inside the pipet. The actin skeleton of a single living cell is visible within 20−46 s after starting the nanoinjection process. (b) dSTORM image of the actin structure of the living U2OS cell shown in (a). The dSTORM image of the actin structure demonstrates a moderately increased spatial resolution. Due to the dynamics of actin filaments in living cells, the resolution of the dSTORM images is limited compared to what can be achieved in fixed cells. dSTORM images were reconstructed from 15,000 frames recorded within 12 min. Scale bar, 5 μm.
proteins.27,28 ATTO 655−phalloidin (Supplementary Figure 5) exhibits a neutral charge at neutral pH and can thus assumedly be delivered by electroosmosis by applying a positive voltage of +200 to +500 mV for several seconds. Labeling of the actin skeleton of the entire cell was finished within just a few seconds (Supplementary Video 2a and 2b). Fluorescence micrographs demonstrate that stress fibers, the fine actin filament meshwork, and even filopodia can readily be visualized (Figure 3a). This finding confirms the efficiency of nanoinjection-based live-cell labeling and corroborates rapid diffusion and binding of ATTO 655-phalloidin to actin inside cells. The reductive environment inside living cells9 enables us to perform live-cell super-resolution imaging, which we demonstrate by imaging the actin cytoskeleton by single molecule localization microscopy (dSTORM).8,29 dSTORM imaging of
the labeling density of a single cell. MitoTracker Red exhibits a net positive charge at neutral pH (Supplementary Figure 5) and can thus be delivered by electroosmosis and its electrophoretic mobility by applying a positive voltage of +1.0 V. In these experiments the stock solution (1 mM) dissolved in DMSO was filled directly into the nanopipette. Following intracellular delivery, the first mitochondria become clearly visible already within 30 s (Figure 2). Notably, nanoinjection-based labeling offers a convenient way to stain selected cells with MitoTracker and control the labeling density simply by applying different injection times (Figure 2a−i). To fluorescently label actin structures in living U2OS cells by nanoinjection, we used ATTO 655−phalloidin. Since ATTO 655−phalloidin is cell-impermeable and prevents depolymerization of the actin skeleton, live-cell imaging of actin is typically performed using fluorescent lifeact fusion 1376
DOI: 10.1021/nl504660t Nano Lett. 2015, 15, 1374−1381
Letter
Nano Letters
Figure 4. Nanoinjection-based DNA labeling of the nucleus of a single living U2OS cell. The nanopipette was loaded with 10−3 M Sytox Green in PBS, pH 7.4, and converged toward the cytoplasm of the cell. (a) Image series of the staining sequence after injection of Sytox Green. The circle in the bright-field image indicates the position of the nanopipette inside the cytoplasm. The corresponding fluorescence image shows the fluorescent spot of the nanopipette. Application of a positive voltage of +1.0 V results in a fluorescence increase inside the cytoplasm after 15 s. After 25 s fluorescence appears in the nucleus, and within 120 s most of the Sytox molecules are located inside the nucleus indicating typical staining behavior of nucleic acids. At 300 s the entire nucleus is stained, and most of the fluorescence inside the cytoplasm has vanished (Supplementary Video 5). (b) Fluorescence intensity vs time recorded in the red and blue areas in (a). (c) Time dependence of the fluorescence intensity integrated over the entire cell as an indicator of the amount of molecules injected into the cell. After reaching the maximum after ∼50 s, photobleaching decreases the fluorescence intensity. This indicates that the molecules are not diffusing out of the nucleus as the intensity time trace in (b) might suggest by its decreasing fluorescence intensity after 170 s. The inset in (c) shows a detailed view of (b) for better comparison. Images were recorded at an integration time of 600 ms and an irradiation intensity of 3 W/cm2. Scale bar, 10 μm.
living U2OS cells at 37 °C was started immediately after the delivery of ATTO 655−phalloidin and removal of the nanopipette from the injected cell. In our experiments we used 10−5 M ATTO 655−phalloidin and completed the injection within 46 s at a voltage of +200 mV. By also irradiating the sample with a low laser power at 405 nm, we achieved a constant blinking rate of ATTO 655 over the entire acquisition time of ∼12 min.24,30 Due to the long acquisition time and the residual mobility of phalloidin-labeled actin filaments, the resulting dSTORM images show only moderate resolution improvement (Figure 3b and Supplementary Video 3). Furthermore, it has to be considered that the dSTORM recording started immediately after the nanoinjection process. Any unbound labels will thus limit the ultimate improvement in resolution that can be achieved. For live-cell labeling of lysosomes, we injected undiluted 15 mM LysoTracker stock solution (dissolved in DMSO) by electroosmotic forces to living HeLa cells. Labeling was stopped 71 s after starting the injection of molecules. The lysosome trafficking could easily be followed by time-lapsed fluorescence imaging (Supplementary Figure 7 and Supplementary Video 4). As a final proof of concept, we tested the possibility to specifically label nucleic acids in living cells by a cell impermeable DNA intercalator. We used Sytox Green, which shows an ∼500-fold increase in fluorescence intensity upon binding to double-stranded DNA. Therefore, Sytox Green is commonly used as a dead cell marker in flow cytometry.31
Upon application of a positive voltage of +200 mV to a nanopipette filled with 30 μM Sytox Green in PBS, the intercalator is dispersed into the cytoplasm, diffuses rapidly toward the nucleus, penetrates the nuclear membrane, and binds efficiently to DNA. Time-lapsed imaging of the fluorescence intensity in the cytoplasm and nucleus demonstrates that staining of the nuclear DNA is completed within 200 s (Figure 4 and Supplementary Video 5). Remarkably, direct nanoinjection of the intercalator dye into the nucleus of a living cell shows reduced background fluorescence in the cytoplasm and stable high fluorescence in the nucleus. The signal-to-noise ratio is substantially higher than typically observed following recommended staining protocols for live-cell and fixed intercalating dyes, respectively (Supplementary Figure 8). As demonstrated above, the EOF enables nanoinjection of neutral or positively charged fluorescent probes. This implies that nanoinjection should also provide a powerful tool for rapid multitarget staining of living cells. Here, even small differences in electrophoretic mobility can be used advantageously to control the delivery of different molecules (Figure 5). To stain the tubulin and actin cytoskeletons as well as the nucleus of a living cell in a single nanoinjection experiment, we used a mixture of paclitaxel−Oregon Green (50 μM), ATTO 655−phalloidin (5 μM), and DAPI (15 μM) and loaded 10 μL of this solution into a single-barrel nanopipette. We approached the cell by monitoring the ion conductance with a negative 1377
DOI: 10.1021/nl504660t Nano Lett. 2015, 15, 1374−1381
Letter
Nano Letters
Figure 5. Multitarget nanoinjection. (a, b) A single living U2OS cell was labeled with three different fluorescent probes delivered from a single barrel nanopipette. The labeled cellular structures were visualized by 3D fluorescence imaging. (c) Actin was visualized by ATTO 655−phalloidin (2 μL, 10−5 M), (d) β-tubulin by paclitaxel−Oregon Green (4 μL, 10−4 M), and (e) the nucleus was stained with DAPI (4 μL, 10−4 M). Staining of tubulin was carried out by applying a negative voltage of −1.0 V for 5 min; the nucleus and actin were stained by applying +1.0 V for 4 min. (a) XZ-overlay of the three different fluorescent probes with the nanopipette remaining inside the cell at the injection position, indicated by the arrow. (b) Zprojection of the cell. The dashed line shown in (a) indicates the plane of the XZ-image. (c−e) Z-projections of the single channels. The wide-field 3D-images were reconstructed from 2D images acquired in axial step sizes of 200 nm and an integration time of 500 ms for Oregon green, 250 ms for ATTO 655, and 150 ms for DAPI. The total acquisition time for the whole image was 15 min. Scale bar, 10 μm.
voltage of −60 mV. Under these conditions the electrophoretic mobility of the negatively charged paclitaxel−Oregon Green is low enough to minimize the flow of molecules out of the pipet to a value that was below our measurement sensitivity. DAPI and ATTO 655−phalloidin are both ejected by applying a positive voltage and are not diffusing out of the pipet during the approach process. After approaching and penetrating the cell, we increased the negative voltage to −1.0 V to induce paclitaxel−Oregon Green delivery and first stained the tubulin structure inside the cell. The staining procedure was completed within 3 min. Subsequently, the voltage was decreased, and 3D fluorescence imaging of tubulin was performed for the next 4 min. During fluorescence imaging, the nanopipette was kept at its position inside the cell. Afterward the voltage was changed to +1 V for simultaneous staining of actin filaments and DNA. Staining of the DNA was finished after 1 min confirming the efficient release of DAPI from the nanopipette, rapid cytoplasmic diffusion, nuclear penetration, and efficient binding to double-stranded DNA. Fluorescence imaging of DNA was conducted by decreasing the applied voltage to +250 mV to continue the delivery of ATTO 655−phalloidin. After DNA imaging, staining of the actin skeleton was completed, and imaging of actin filaments was started. Overall, the nanoinjection-based staining and 3D imaging of three different cellular structures in a single living cell required 15 min, which could be shortened, e.g., by performing simultaneous multichannel imaging (Figure 5). Of note, the xz-image clearly
depicts the location of the nanopipette at the delivery position within the cell. Our data indicate that other combinations of fluorescent probescell permeable as well as cell impermeable stainscan also be used for nanoinjection-based controlled multitarget staining of living cells. Although nanoinjection provides fast and minimally invasive staining of living cells and cellular compartments, it remains an invasive technique which can cause artifacts by the process of physically penetrating the cell. To investigate the extent of potential damage to the cells as evidenced by structural artifacts in the fluorescence images, we stained the actin skeleton of a single living U2OS cell with ATTO 655−phalloidin. We assumed that the nanopipette itself has a minor effect on the cellular structure since it has an outer diameter of only ∼100 nm, which is significantly smaller than, e.g., a typical microinjection tip. To image the dense actin network of U2OS cells at the basal cell membrane, we performed imaging in TIRF-mode. This allows the investigation of even very small changes in the lateral cellular structure and to subsequently determine the damaged area caused by the nanopipette. To obtain a final “damaged cell volume”, we assumed a cylindrical shape of the nanopipette. Typically, a cell is penetrated by the nanopipette to some hundred nanometers beyond the apical membrane if labeling of intracellular structures is intended. Staining of nuclear structures requires deeper penetration of the nanopipette into the cell. To determine the lateral damage area, we penetrated a U2OS cell to approximately 50 nm above the 1378
DOI: 10.1021/nl504660t Nano Lett. 2015, 15, 1374−1381
Letter
Nano Letters
Figure 6. Structural artifacts caused by nanoinjection. (a) TIRF image of actin filaments inside a single living U2OS cell stained with ATTO 655− phalloidin during the nanoinjection process. The bright spot indicates the position of the nanopipette inside the fine structure of the actin skeleton. (b) The same cell shortly after retraction of the nanopipette. Structural damage of the fine actin meshwork is clearly visible. (c, d) Detailed views of the areas highlighted in (a) and (b). (e) To appraise the damaged area caused by the penetration of nanopipette, we placed an intensity plot through this area including the pipet and after retraction of the pipet. The diameter of the affected area was determined to 2.5 μm. TIRF images were taken at 300 ms integration time at a laser intensity of 20 W/cm2. Scale bars, 10 μm (a, b), 2 μm (c, d).
Figure 7. Fluorescence image (z-projection after 3D deconvolution) of the actin skeleton of a single nanoinjected U2OS cell. (a) Directly after the nanoinjection, process actin filaments outside the cell membrane are clearly visible. (b) After fixation with PFA for 10 min, these filaments are no longer visible. Furthermore, a strong decrease in fluorescence intensity was observed. (c) 5 min after treatment with Triton X-100 for permeabilization enhanced fluorescence inside of the nucleus appears (highlighted area). Nanoinjection was performed with a double barrel pipet under DMEM conditions at 37 °C. Fluorescence imaging was carried out using widefield epi-fluorescence microscopy at 20 W/cm2 with an integration time of 150 ms/plane with z-planes acquired every 200 nm. Deconvolution was carried out as described in the Materials and Methods (SI). Scale bar, 10 μm.
basal membrane by first approaching the nanopipette to the surface and then subsequently lifting it to a height of 50 nm. After approaching and adjusting the height of the pipet we began to stain the actin network. Once this process was completed, we imaged the lower actin structure of the entire cell in TIRF-mode (Figure 6a). Afterward, we retracted the nanopipette and imaged the same area again under identical conditions (Figure 6b). After
retraction of the nanopipette, a small area at the former position of the pipet remains dark (devoid of fluorescence). It is very likely that the fine structure of the actin filaments was destroyed in this area. The width of the area was determined to be 2.5 μm, according to a lateral damage area of ∼4.9 μm2, assuming a circularly shaped damaged area. Figure 5 shows a penetration depth for the nanopipette of approximately 3.5 μm starting from the cell membrane to the inside of the cell. 1379
DOI: 10.1021/nl504660t Nano Lett. 2015, 15, 1374−1381
Letter
Nano Letters Assuming a cylindrical shape within the first few micrometers of the nanopipette, this means a volume of 11.9 μm3 will likely have been disrupted. Compared to the cell volume of approximately 4000 μm3 of the entire U2OS cell used for the injection process, only ∼0.3% of the entire cell volume is directly disturbed during the penetration process. Obtaining structural information about cells often includes fixation and permeabilization of cells to preserve them for further studies. Typical staining protocols utilize aldehydecontaining reagents, such as paraformaldehyde (PFA) or glutaraldehyde. PFA as one of the most widely applied fixation reagents effectively “freezes” the cell by cross-linking its proteins. Structural changes due to this process compared to the living cell are unavoidable. Furthermore, following fixation subsequent permeabilization (e.g., with Triton X-100) of the cellular membrane is often required to enable functionalized fluorescent probes to reach their destination. A typical staining artifact arising from this fixation method reported in the literature is increased background fluorescence if reactive aldehyde groups are not blocked.32,33 Nanoinjection enables investigation of structural changes that occur inside cells before and after the fixation process. For this we investigated the structural deformation of actin filaments by nanoinjection of ATTO 655−phalloidin. Labeling of the actin structure was carried out with a double-barrel nanopipette (see Supplementary Figure 4) using DMEM medium as described in Materials and Methods. The staining process was finished within 3 min. Subsequent imaging of the labeled structure with low laser intensity allowed us to take a quick view at the actin skeleton (Figure 7a), followed by fixation with 4% PFA dissolved in prewarmed PBS for 10 min. After the fixation process, we imaged the structure again and found a locally strong decrease in fluorescence intensity as well as a highly increased fluorescence background (Figure 7b). Additionally, we noticed that fine structures of the actin network that we could image well before fixation now appear blurred. Furthermore, filaments within filopodia at the outside of the cell membrane partially disappeared. Lastly, shrinkage of the cell’s area in the lateral direction due to the fixation process was observed (Supplementary Figure 9).34 Treating the fixed cell with Triton X-100 for 5 min revealed no further noticeable changes in the cellular structure nor in the total fluorescence signal (Figure 7c). The most notable effect of the permeating agent Triton X-100 is that after its application fluorescence is now also observed inside the nucleus (Figure 7c (highlighted area)). Since we found that the application of PFA reduces the detected fluorescence intensity (Supplementary Figure 10) and that the fluorescence response from single fluorophores is relatively stable under different buffer conditions (Supplementary Figure 11), we assume that the structural change in the actin skeleton is induced by PFA treatment. To verify this, we subsequently photobleached the entire fluorescence signal inside the investigated cell and completed the labeling protocol by adding more ATTO 655-phalloidin after permeabilization. Interestingly, we noticed that the previously injected and bleached cell exhibited no new fluorescence (Supplementary Figure 10c). This indicates that all actin binding sites for phalloidin are still occupied and binding of “fresh” fluorescent ATTO 655-phalloidin is prevented. To corroborate that PFA does not destroy the ATTO 655−phalloidin bond, we performed fluorescence correlations spectroscopy (FCS) measurements with freely diffusing ATTO 655−phalloidin under PBS and PFA conditions which resulted in no significant
differences in diffusion time (Supplementary Figure 12). In addition, nanoinjection-based staining of cells with ATTO 655−phalloidin in 4% PFA−PBS resulted in an unaltered labeling pattern (Supplementary Figure 13 and Supplementary Video 6). In accordance with the findings of Small et al.,35 PFA disrupts at least partially the actin structure, as demonstrated here at the single cell level. Consequently, the actin structure appears blurred and less detailed compared to the structure measured in living cells directly after nanoinjection. We demonstrate the electrokinetic delivery of a number of functionalized fluorescent probes to different types of living cells using nanopipettes with a diameter of ∼100 nm. Nanoinjection-based probe delivery provides a fast and easy way for efficient labeling of entire cells and cellular compartments simply by applying a voltage between the electrodes. The staining procedure takes between a few seconds to some 10 min and depends solely on the charge of the fluorescent probe and its concentration. Direct monitoring of the staining process by fluorescence microscopy enables fine and immediate control over the labeling density. We were able to inject cell permeable fluorescent probes into individual living cells as well as nonpermeable probes. Nanoinjection was found to be applicable to several conducting and nonconducting buffer environments by using a single-barrel or a double-barrel pipet. The direct intracellular staining enabled successful superresolution imaging of living cells by dSTORM. In these experiments an acquisition time of 13 min between the start of the staining procedure and the completed image acquisition was sufficient for the reconstruction of a super-resolved image. The entire process can, however, also be accelerated using smaller image areas and faster EMCCD or sCMOS cameras. Thus, nanoinjection-based labeling enables instant live-cell super-resolution imaging of intracellular structures. In addition, the technique allows the simultaneous delivery of at least three different types of fluorescent probes to a single cell with a single-barrel pipet by adjusting the concentration and applying positive and negative voltages to the molecules inside the pipet. Because nanoinjection-based delivery of fluorescent probes to living cells is minimally destructive, it provides a versatile and efficient new tool for the investigation of single cells under nearly native conditions.
■
ASSOCIATED CONTENT
S Supporting Information *
Detailed explanation of the nanoinjection principle, Materials and Methods, and additional figures. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank Fabian Humpert for performing FCS measurements as well as Saskia Bannister for additional preparations of U2OS cells.
■
REFERENCES
(1) Betzig, E.; et al. Imaging intracellular fluorescent proteins at nanometer resolution. Science 2006, 313 (5793), 1642−5.
1380
DOI: 10.1021/nl504660t Nano Lett. 2015, 15, 1374−1381
Letter
Nano Letters (2) Gustafsson, M. G. Surpassing the lateral resolution limit by a factor of two using structured illumination microscopy. J. Microsc 2000, 198 (Pt 2), 82−7. (3) Heilemann, M.; et al. Subdiffraction-resolution fluorescence imaging with conventional fluorescent probes. Angew. Chem., Int. Ed. Engl. 2008, 47 (33), 6172−6. (4) Hell, S. W. Microscopy and its focal switch. Nat. Methods 2009, 6 (1), 24−32. (5) Hennig, S.; et al. Quantum dot triexciton imaging with threedimensional subdiffraction resolution. Nano Lett. 2009, 9 (6), 2466− 70. (6) Rust, M. J.; Bates, M.; Zhuang, X. Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM). Nat. Methods 2006, 3 (10), 793−5. (7) Schoen, I.; et al. Binding-activated localization microscopy of DNA structures. Nano Lett. 2011, 11 (9), 4008−11. (8) Klein, T.; et al. Live-cell dSTORM with SNAP-tag fusion proteins. Nat. Methods 2011, 8 (1), 7−9. (9) van de Linde, S.; Heilemann, M.; Sauer, M. Live-cell superresolution imaging with synthetic fluorophores. Annu. Rev. Phys. Chem. 2012, 63, 519−40. (10) Shannon, C. E. Communication in the Presence of Noise. Proc. Inst. Radio Eng. 1949, 37 (1), 10−21. (11) Shroff, H.; et al. Live-cell photoactivated localization microscopy of nanoscale adhesion dynamics. Nat. Methods 2008, 5 (5), 417−23. (12) Barber, K.; et al. Delivery of membrane-impermeant fluorescent probes into living neural cell populations by lipotransfer. Neurosci. Lett. 1996, 207 (1), 17−20. (13) Laforge, F. O.; et al. Electrochemical attosyringe. Proc. Natl. Acad. Sci. U.S.A. 2007, 104 (29), 11895−900. (14) Neumann, E.; et al. Gene transfer into mouse lyoma cells by electroporation in high electric fields. EMBO J. 1982, 1 (7), 841−5. (15) Bruckbauer, A.; et al. Nanopipette delivery of individual molecules to cellular compartments for single-molecule fluorescence tracking. Biophys. J. 2007, 93 (9), 3120−31. (16) Viigipuu, K.; Kallio, P. Microinjection of living adherent cells by using a semi-automatic microinjection system. Altern. Lab Anim. 2004, 32 (4), 417−23. (17) Delehanty, J. B.; et al. Delivering quantum dot-peptide bioconjugates to the cellular cytosol: escaping from the endolysosomal system. Integr. Biol. (Camb.) 2010, 2 (5−6), 265−77. (18) Xu, J.; et al. Nanoblade delivery and incorporation of quantum dot conjugates into tubulin networks in live cells. Nano Lett. 2012, 12 (11), 5669−72. (19) Actis, P.; et al. Voltage-controlled metal binding on polyelectrolyte-functionalized nanopores. Langmuir 2011, 27 (10), 6528−33. (20) Chen, C. C.; Zhou, Y.; Baker, L. A. Scanning ion conductance microscopy. Annu. Rev. Anal. Chem. (Palo Alto, Calif.) 2012, 5, 207− 28. (21) Hansma, P. K.; et al. The scanning ion-conductance microscope. Science 1989, 243 (4891), 641−3. (22) Novak, P.; et al. Nanoscale live-cell imaging using hopping probe ion conductance microscopy. Nat. Methods 2009, 6 (4), 279− 81. (23) Heilemann, M.; et al. Super-resolution imaging with small organic fluorophores. Angew. Chem., Int. Ed. Engl. 2009, 48 (37), 6903−8. (24) van de Linde, S.; et al. Direct stochastic optical reconstruction microscopy with standard fluorescent probes. Nat. Protoc. 2011, 6 (7), 991−1009. (25) Chong-Woo Park, J.-P. K.; Marmé, N.; Möller, M.; Spatz, J.; Wolfrum, J.; Sauer, M. Accurate Delivery of single biomolecules by polyethylene glycol coated sucmicrometer pipettes. Chem. Phys. 2004, 301 (1), 105−110. (26) Tokunaga, M.; Imamoto, N.; Sakata-Sogawa, K. Highly inclined thin illumination enables clear single-molecule imaging in cells. Nat. Methods 2008, 5 (2), 159−61.
(27) Riedl, J.; et al. Lifeact: a versatile marker to visualize F-actin. Nat. Methods 2008, 5 (7), 605−7. (28) Riedl, J.; et al. Lifeact mice for studying F-actin dynamics. Nat. Methods 2010, 7 (3), 168−9. (29) Wombacher, R.; et al. Live-cell super-resolution imaging with trimethoprim conjugates. Nat. Methods 2010, 7 (9), 717−9. (30) van de Linde, S.; et al. Photoinduced formation of reversible dye radicals and their impact on super-resolution imaging. Photochem. Photobiol. Sci. 2011, 10 (4), 499−506. (31) Roth, B. L.; et al. Bacterial viability and antibiotic susceptibility testing with SYTOX green nucleic acid stain. Appl. Environ. Microbiol. 1997, 63 (6), 2421−31. (32) Schnell, U.; et al. Immunolabeling artifacts and the need for livecell imaging. Nat. Methods 2012, 9 (2), 152−8. (33) Weber, K.; Rathke, P. C.; Osborn, M. Cytoplasmic microtubular images in glutaraldehyde-fixed tissue culture cells by electron microscopy and by immunofluorescence microscopy. Proc. Natl. Acad. Sci. U.S.A. 1978, 75 (4), 1820−4. (34) Stadler, C.; et al. A single fixation protocol for proteome-wide immunofluorescence localization studies. J. Proteomics 2010, 73 (6), 1067−78. (35) Small, J.; et al. Visualising the actin cytoskeleton. Microsc. Res. Tech. 1999, 47 (1), 3−17.
1381
DOI: 10.1021/nl504660t Nano Lett. 2015, 15, 1374−1381
Instant Live-Cell Super-Resolution Imaging of Cellular Structures by Nanoinjection of Fluorescent Probes Simon Hennig,*,† Sebastian van de Linde,‡ Martina Lummer,§ Matthias Simonis,† Thomas Huser,† and Markus Sauer‡ †
Biomolecular Photonics, Department of Physics and §Department of Molecular Cell Physiology, Bielefeld University, Universitätsstr. 25, 33615 Bielefeld, Germany ‡ Department of Biotechnology & Biophysics, Julius-Maximilians-University, Am Hubland, 97074 Würzburg, Germany S Supporting Information *
ABSTRACT: Labeling internal structures within living cells with standard fluorescent probes is a challenging problem. Here, we introduce a novel intracellular staining method that enables us to carefully control the labeling process and provides instant access to the inner structures of living cells. Using a hollow glass capillary with a diameter of 7.0, it can be delivered by nanoinjection, by applying a negative voltage of −0.7 to −1.0 V to the electrodes. The staining procedure was carried out in PBS at a temperature of 37 °C with a single-barrel pipet with a diameter of 100 nm filled with ∼10−5 M paclitaxel−Oregon Green (Supplementary Video 1). Time-lapse fluorescence imaging enables direct monitoring of the labeling process, its efficiency, and demonstrates that paclitaxel−Oregon Green binds specifically to the tubulin structure within seconds to minutes upon intracellular delivery (Supplementary Figure 6a−e). Tubulin staining of the entire cell is completed within 15 min after start of the experiment. Next, we injected MitoTracker Red, a widely used stain for the visualization of mitochondria in living and fixed cells. Although MitoTracker Red is cell permeable and readily enables labeling of living cells, it is impossible to control
relative to the cell, this approach uses changes in ion conductance to control the approach and penetration of the tip and subsequent delivery of charged molecules, based on scanning ion conductance microscopy.20−22 Nanoinjection provides fast, minimally invasive and traceable staining with various fluorescent probes suitable for high resolution fluorescence microscopy of living cells. To reveal the power of this approach, we labeled intracellular structures in different cell types under conducting and nonconducting buffer conditions with conventional cell permeable and nonpermeable fluorescent probes while directly monitoring the labeling process. Live-cell wide-field fluorescence and super-resolution imaging by direct stochastic optical reconstruction microscopy (dSTORM)23,24 demonstrate the effectiveness of the technique to gain an instant view of the cellular labeling process even with subdiffraction-resolution. Notably, this technique is also applicable for the efficient sequential or simultaneous delivery of multiple different targets to the same cell. Comparison of labeling densities and structural details visible in living and fixed cells indicates that live-cell nanoninjection-based delivery of fluorescent probes perfectly preserves cellular target structures and epitope multiplicity. To label living cells by nanoinjection, the nanopipette needs to carefully penetrate the plasma membrane of a cell of interest. For this purpose, the approach of the tip to the cell is monitored using the ion flow through the nanopipette by measuring the corresponding conductance at a predefined voltage of typically 60 mV between the nanopipette and a bath electrode. After penetration of the cell with the nanopipette, fluorescent probes are delivered by adjusting the voltage applied to the electrodes to adapt to the net charge of the molecules using the electrophoretic mobility of the molecules (Figure 1 and Supplementary Figures 1−4). To drive charged molecules through a glass nanopipette by an electric field, we have to consider that the SiO2 groups at the 1375
DOI: 10.1021/nl504660t Nano Lett. 2015, 15, 1374−1381
Letter
Nano Letters
Figure 2. Directed staining of mitochondria in a single living HeLa cell surrounded by other cells by nanoinjection of MitoTracker Red. (a) At the beginning of the staining process only the fluorescence response of the nanopipette is visible. (b−e) During the first 30−120 s, mitochondria in the vicinity of the injection point are gradually stained. (f) After 150 s no further expansion of fluorescence in lateral direction is observed, indicating that the fluorescent probe has spread throughout the cytosol. (g, h) With continued injection of MitoTracker Red more diffuse background signals become visible, due to staining of mitochondria in the out-of-focus axial direction. (i) 240 s after the start of the staining process the fluorescence signal did not increase any further, indicating that the staining is saturated. Fluorescence images were taken at 15 W/cm2 at 568 nm with an integration time of 150 ms exciting with a highly inclined and laminated optical sheet (HILO).26 Scale bar, 10 μm.
Figure 3. Super-resolution imaging of nanoinjection-based labeling of actin filaments. (a) Representative wide-field fluorescence image of the actin cytoskeleton in living U2OS cells, taken immediately after nanoinjection of ATTO 655−phalloidin at a concentration of 10−5 M inside the pipet. The actin skeleton of a single living cell is visible within 20−46 s after starting the nanoinjection process. (b) dSTORM image of the actin structure of the living U2OS cell shown in (a). The dSTORM image of the actin structure demonstrates a moderately increased spatial resolution. Due to the dynamics of actin filaments in living cells, the resolution of the dSTORM images is limited compared to what can be achieved in fixed cells. dSTORM images were reconstructed from 15,000 frames recorded within 12 min. Scale bar, 5 μm.
proteins.27,28 ATTO 655−phalloidin (Supplementary Figure 5) exhibits a neutral charge at neutral pH and can thus assumedly be delivered by electroosmosis by applying a positive voltage of +200 to +500 mV for several seconds. Labeling of the actin skeleton of the entire cell was finished within just a few seconds (Supplementary Video 2a and 2b). Fluorescence micrographs demonstrate that stress fibers, the fine actin filament meshwork, and even filopodia can readily be visualized (Figure 3a). This finding confirms the efficiency of nanoinjection-based live-cell labeling and corroborates rapid diffusion and binding of ATTO 655-phalloidin to actin inside cells. The reductive environment inside living cells9 enables us to perform live-cell super-resolution imaging, which we demonstrate by imaging the actin cytoskeleton by single molecule localization microscopy (dSTORM).8,29 dSTORM imaging of
the labeling density of a single cell. MitoTracker Red exhibits a net positive charge at neutral pH (Supplementary Figure 5) and can thus be delivered by electroosmosis and its electrophoretic mobility by applying a positive voltage of +1.0 V. In these experiments the stock solution (1 mM) dissolved in DMSO was filled directly into the nanopipette. Following intracellular delivery, the first mitochondria become clearly visible already within 30 s (Figure 2). Notably, nanoinjection-based labeling offers a convenient way to stain selected cells with MitoTracker and control the labeling density simply by applying different injection times (Figure 2a−i). To fluorescently label actin structures in living U2OS cells by nanoinjection, we used ATTO 655−phalloidin. Since ATTO 655−phalloidin is cell-impermeable and prevents depolymerization of the actin skeleton, live-cell imaging of actin is typically performed using fluorescent lifeact fusion 1376
DOI: 10.1021/nl504660t Nano Lett. 2015, 15, 1374−1381
Letter
Nano Letters
Figure 4. Nanoinjection-based DNA labeling of the nucleus of a single living U2OS cell. The nanopipette was loaded with 10−3 M Sytox Green in PBS, pH 7.4, and converged toward the cytoplasm of the cell. (a) Image series of the staining sequence after injection of Sytox Green. The circle in the bright-field image indicates the position of the nanopipette inside the cytoplasm. The corresponding fluorescence image shows the fluorescent spot of the nanopipette. Application of a positive voltage of +1.0 V results in a fluorescence increase inside the cytoplasm after 15 s. After 25 s fluorescence appears in the nucleus, and within 120 s most of the Sytox molecules are located inside the nucleus indicating typical staining behavior of nucleic acids. At 300 s the entire nucleus is stained, and most of the fluorescence inside the cytoplasm has vanished (Supplementary Video 5). (b) Fluorescence intensity vs time recorded in the red and blue areas in (a). (c) Time dependence of the fluorescence intensity integrated over the entire cell as an indicator of the amount of molecules injected into the cell. After reaching the maximum after ∼50 s, photobleaching decreases the fluorescence intensity. This indicates that the molecules are not diffusing out of the nucleus as the intensity time trace in (b) might suggest by its decreasing fluorescence intensity after 170 s. The inset in (c) shows a detailed view of (b) for better comparison. Images were recorded at an integration time of 600 ms and an irradiation intensity of 3 W/cm2. Scale bar, 10 μm.
living U2OS cells at 37 °C was started immediately after the delivery of ATTO 655−phalloidin and removal of the nanopipette from the injected cell. In our experiments we used 10−5 M ATTO 655−phalloidin and completed the injection within 46 s at a voltage of +200 mV. By also irradiating the sample with a low laser power at 405 nm, we achieved a constant blinking rate of ATTO 655 over the entire acquisition time of ∼12 min.24,30 Due to the long acquisition time and the residual mobility of phalloidin-labeled actin filaments, the resulting dSTORM images show only moderate resolution improvement (Figure 3b and Supplementary Video 3). Furthermore, it has to be considered that the dSTORM recording started immediately after the nanoinjection process. Any unbound labels will thus limit the ultimate improvement in resolution that can be achieved. For live-cell labeling of lysosomes, we injected undiluted 15 mM LysoTracker stock solution (dissolved in DMSO) by electroosmotic forces to living HeLa cells. Labeling was stopped 71 s after starting the injection of molecules. The lysosome trafficking could easily be followed by time-lapsed fluorescence imaging (Supplementary Figure 7 and Supplementary Video 4). As a final proof of concept, we tested the possibility to specifically label nucleic acids in living cells by a cell impermeable DNA intercalator. We used Sytox Green, which shows an ∼500-fold increase in fluorescence intensity upon binding to double-stranded DNA. Therefore, Sytox Green is commonly used as a dead cell marker in flow cytometry.31
Upon application of a positive voltage of +200 mV to a nanopipette filled with 30 μM Sytox Green in PBS, the intercalator is dispersed into the cytoplasm, diffuses rapidly toward the nucleus, penetrates the nuclear membrane, and binds efficiently to DNA. Time-lapsed imaging of the fluorescence intensity in the cytoplasm and nucleus demonstrates that staining of the nuclear DNA is completed within 200 s (Figure 4 and Supplementary Video 5). Remarkably, direct nanoinjection of the intercalator dye into the nucleus of a living cell shows reduced background fluorescence in the cytoplasm and stable high fluorescence in the nucleus. The signal-to-noise ratio is substantially higher than typically observed following recommended staining protocols for live-cell and fixed intercalating dyes, respectively (Supplementary Figure 8). As demonstrated above, the EOF enables nanoinjection of neutral or positively charged fluorescent probes. This implies that nanoinjection should also provide a powerful tool for rapid multitarget staining of living cells. Here, even small differences in electrophoretic mobility can be used advantageously to control the delivery of different molecules (Figure 5). To stain the tubulin and actin cytoskeletons as well as the nucleus of a living cell in a single nanoinjection experiment, we used a mixture of paclitaxel−Oregon Green (50 μM), ATTO 655−phalloidin (5 μM), and DAPI (15 μM) and loaded 10 μL of this solution into a single-barrel nanopipette. We approached the cell by monitoring the ion conductance with a negative 1377
DOI: 10.1021/nl504660t Nano Lett. 2015, 15, 1374−1381
Letter
Nano Letters
Figure 5. Multitarget nanoinjection. (a, b) A single living U2OS cell was labeled with three different fluorescent probes delivered from a single barrel nanopipette. The labeled cellular structures were visualized by 3D fluorescence imaging. (c) Actin was visualized by ATTO 655−phalloidin (2 μL, 10−5 M), (d) β-tubulin by paclitaxel−Oregon Green (4 μL, 10−4 M), and (e) the nucleus was stained with DAPI (4 μL, 10−4 M). Staining of tubulin was carried out by applying a negative voltage of −1.0 V for 5 min; the nucleus and actin were stained by applying +1.0 V for 4 min. (a) XZ-overlay of the three different fluorescent probes with the nanopipette remaining inside the cell at the injection position, indicated by the arrow. (b) Zprojection of the cell. The dashed line shown in (a) indicates the plane of the XZ-image. (c−e) Z-projections of the single channels. The wide-field 3D-images were reconstructed from 2D images acquired in axial step sizes of 200 nm and an integration time of 500 ms for Oregon green, 250 ms for ATTO 655, and 150 ms for DAPI. The total acquisition time for the whole image was 15 min. Scale bar, 10 μm.
voltage of −60 mV. Under these conditions the electrophoretic mobility of the negatively charged paclitaxel−Oregon Green is low enough to minimize the flow of molecules out of the pipet to a value that was below our measurement sensitivity. DAPI and ATTO 655−phalloidin are both ejected by applying a positive voltage and are not diffusing out of the pipet during the approach process. After approaching and penetrating the cell, we increased the negative voltage to −1.0 V to induce paclitaxel−Oregon Green delivery and first stained the tubulin structure inside the cell. The staining procedure was completed within 3 min. Subsequently, the voltage was decreased, and 3D fluorescence imaging of tubulin was performed for the next 4 min. During fluorescence imaging, the nanopipette was kept at its position inside the cell. Afterward the voltage was changed to +1 V for simultaneous staining of actin filaments and DNA. Staining of the DNA was finished after 1 min confirming the efficient release of DAPI from the nanopipette, rapid cytoplasmic diffusion, nuclear penetration, and efficient binding to double-stranded DNA. Fluorescence imaging of DNA was conducted by decreasing the applied voltage to +250 mV to continue the delivery of ATTO 655−phalloidin. After DNA imaging, staining of the actin skeleton was completed, and imaging of actin filaments was started. Overall, the nanoinjection-based staining and 3D imaging of three different cellular structures in a single living cell required 15 min, which could be shortened, e.g., by performing simultaneous multichannel imaging (Figure 5). Of note, the xz-image clearly
depicts the location of the nanopipette at the delivery position within the cell. Our data indicate that other combinations of fluorescent probescell permeable as well as cell impermeable stainscan also be used for nanoinjection-based controlled multitarget staining of living cells. Although nanoinjection provides fast and minimally invasive staining of living cells and cellular compartments, it remains an invasive technique which can cause artifacts by the process of physically penetrating the cell. To investigate the extent of potential damage to the cells as evidenced by structural artifacts in the fluorescence images, we stained the actin skeleton of a single living U2OS cell with ATTO 655−phalloidin. We assumed that the nanopipette itself has a minor effect on the cellular structure since it has an outer diameter of only ∼100 nm, which is significantly smaller than, e.g., a typical microinjection tip. To image the dense actin network of U2OS cells at the basal cell membrane, we performed imaging in TIRF-mode. This allows the investigation of even very small changes in the lateral cellular structure and to subsequently determine the damaged area caused by the nanopipette. To obtain a final “damaged cell volume”, we assumed a cylindrical shape of the nanopipette. Typically, a cell is penetrated by the nanopipette to some hundred nanometers beyond the apical membrane if labeling of intracellular structures is intended. Staining of nuclear structures requires deeper penetration of the nanopipette into the cell. To determine the lateral damage area, we penetrated a U2OS cell to approximately 50 nm above the 1378
DOI: 10.1021/nl504660t Nano Lett. 2015, 15, 1374−1381
Letter
Nano Letters
Figure 6. Structural artifacts caused by nanoinjection. (a) TIRF image of actin filaments inside a single living U2OS cell stained with ATTO 655− phalloidin during the nanoinjection process. The bright spot indicates the position of the nanopipette inside the fine structure of the actin skeleton. (b) The same cell shortly after retraction of the nanopipette. Structural damage of the fine actin meshwork is clearly visible. (c, d) Detailed views of the areas highlighted in (a) and (b). (e) To appraise the damaged area caused by the penetration of nanopipette, we placed an intensity plot through this area including the pipet and after retraction of the pipet. The diameter of the affected area was determined to 2.5 μm. TIRF images were taken at 300 ms integration time at a laser intensity of 20 W/cm2. Scale bars, 10 μm (a, b), 2 μm (c, d).
Figure 7. Fluorescence image (z-projection after 3D deconvolution) of the actin skeleton of a single nanoinjected U2OS cell. (a) Directly after the nanoinjection, process actin filaments outside the cell membrane are clearly visible. (b) After fixation with PFA for 10 min, these filaments are no longer visible. Furthermore, a strong decrease in fluorescence intensity was observed. (c) 5 min after treatment with Triton X-100 for permeabilization enhanced fluorescence inside of the nucleus appears (highlighted area). Nanoinjection was performed with a double barrel pipet under DMEM conditions at 37 °C. Fluorescence imaging was carried out using widefield epi-fluorescence microscopy at 20 W/cm2 with an integration time of 150 ms/plane with z-planes acquired every 200 nm. Deconvolution was carried out as described in the Materials and Methods (SI). Scale bar, 10 μm.
basal membrane by first approaching the nanopipette to the surface and then subsequently lifting it to a height of 50 nm. After approaching and adjusting the height of the pipet we began to stain the actin network. Once this process was completed, we imaged the lower actin structure of the entire cell in TIRF-mode (Figure 6a). Afterward, we retracted the nanopipette and imaged the same area again under identical conditions (Figure 6b). After
retraction of the nanopipette, a small area at the former position of the pipet remains dark (devoid of fluorescence). It is very likely that the fine structure of the actin filaments was destroyed in this area. The width of the area was determined to be 2.5 μm, according to a lateral damage area of ∼4.9 μm2, assuming a circularly shaped damaged area. Figure 5 shows a penetration depth for the nanopipette of approximately 3.5 μm starting from the cell membrane to the inside of the cell. 1379
DOI: 10.1021/nl504660t Nano Lett. 2015, 15, 1374−1381
Letter
Nano Letters Assuming a cylindrical shape within the first few micrometers of the nanopipette, this means a volume of 11.9 μm3 will likely have been disrupted. Compared to the cell volume of approximately 4000 μm3 of the entire U2OS cell used for the injection process, only ∼0.3% of the entire cell volume is directly disturbed during the penetration process. Obtaining structural information about cells often includes fixation and permeabilization of cells to preserve them for further studies. Typical staining protocols utilize aldehydecontaining reagents, such as paraformaldehyde (PFA) or glutaraldehyde. PFA as one of the most widely applied fixation reagents effectively “freezes” the cell by cross-linking its proteins. Structural changes due to this process compared to the living cell are unavoidable. Furthermore, following fixation subsequent permeabilization (e.g., with Triton X-100) of the cellular membrane is often required to enable functionalized fluorescent probes to reach their destination. A typical staining artifact arising from this fixation method reported in the literature is increased background fluorescence if reactive aldehyde groups are not blocked.32,33 Nanoinjection enables investigation of structural changes that occur inside cells before and after the fixation process. For this we investigated the structural deformation of actin filaments by nanoinjection of ATTO 655−phalloidin. Labeling of the actin structure was carried out with a double-barrel nanopipette (see Supplementary Figure 4) using DMEM medium as described in Materials and Methods. The staining process was finished within 3 min. Subsequent imaging of the labeled structure with low laser intensity allowed us to take a quick view at the actin skeleton (Figure 7a), followed by fixation with 4% PFA dissolved in prewarmed PBS for 10 min. After the fixation process, we imaged the structure again and found a locally strong decrease in fluorescence intensity as well as a highly increased fluorescence background (Figure 7b). Additionally, we noticed that fine structures of the actin network that we could image well before fixation now appear blurred. Furthermore, filaments within filopodia at the outside of the cell membrane partially disappeared. Lastly, shrinkage of the cell’s area in the lateral direction due to the fixation process was observed (Supplementary Figure 9).34 Treating the fixed cell with Triton X-100 for 5 min revealed no further noticeable changes in the cellular structure nor in the total fluorescence signal (Figure 7c). The most notable effect of the permeating agent Triton X-100 is that after its application fluorescence is now also observed inside the nucleus (Figure 7c (highlighted area)). Since we found that the application of PFA reduces the detected fluorescence intensity (Supplementary Figure 10) and that the fluorescence response from single fluorophores is relatively stable under different buffer conditions (Supplementary Figure 11), we assume that the structural change in the actin skeleton is induced by PFA treatment. To verify this, we subsequently photobleached the entire fluorescence signal inside the investigated cell and completed the labeling protocol by adding more ATTO 655-phalloidin after permeabilization. Interestingly, we noticed that the previously injected and bleached cell exhibited no new fluorescence (Supplementary Figure 10c). This indicates that all actin binding sites for phalloidin are still occupied and binding of “fresh” fluorescent ATTO 655-phalloidin is prevented. To corroborate that PFA does not destroy the ATTO 655−phalloidin bond, we performed fluorescence correlations spectroscopy (FCS) measurements with freely diffusing ATTO 655−phalloidin under PBS and PFA conditions which resulted in no significant
differences in diffusion time (Supplementary Figure 12). In addition, nanoinjection-based staining of cells with ATTO 655−phalloidin in 4% PFA−PBS resulted in an unaltered labeling pattern (Supplementary Figure 13 and Supplementary Video 6). In accordance with the findings of Small et al.,35 PFA disrupts at least partially the actin structure, as demonstrated here at the single cell level. Consequently, the actin structure appears blurred and less detailed compared to the structure measured in living cells directly after nanoinjection. We demonstrate the electrokinetic delivery of a number of functionalized fluorescent probes to different types of living cells using nanopipettes with a diameter of ∼100 nm. Nanoinjection-based probe delivery provides a fast and easy way for efficient labeling of entire cells and cellular compartments simply by applying a voltage between the electrodes. The staining procedure takes between a few seconds to some 10 min and depends solely on the charge of the fluorescent probe and its concentration. Direct monitoring of the staining process by fluorescence microscopy enables fine and immediate control over the labeling density. We were able to inject cell permeable fluorescent probes into individual living cells as well as nonpermeable probes. Nanoinjection was found to be applicable to several conducting and nonconducting buffer environments by using a single-barrel or a double-barrel pipet. The direct intracellular staining enabled successful superresolution imaging of living cells by dSTORM. In these experiments an acquisition time of 13 min between the start of the staining procedure and the completed image acquisition was sufficient for the reconstruction of a super-resolved image. The entire process can, however, also be accelerated using smaller image areas and faster EMCCD or sCMOS cameras. Thus, nanoinjection-based labeling enables instant live-cell super-resolution imaging of intracellular structures. In addition, the technique allows the simultaneous delivery of at least three different types of fluorescent probes to a single cell with a single-barrel pipet by adjusting the concentration and applying positive and negative voltages to the molecules inside the pipet. Because nanoinjection-based delivery of fluorescent probes to living cells is minimally destructive, it provides a versatile and efficient new tool for the investigation of single cells under nearly native conditions.
■
ASSOCIATED CONTENT
S Supporting Information *
Detailed explanation of the nanoinjection principle, Materials and Methods, and additional figures. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail: [email protected]. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank Fabian Humpert for performing FCS measurements as well as Saskia Bannister for additional preparations of U2OS cells.
■
REFERENCES
(1) Betzig, E.; et al. Imaging intracellular fluorescent proteins at nanometer resolution. Science 2006, 313 (5793), 1642−5.
1380
DOI: 10.1021/nl504660t Nano Lett. 2015, 15, 1374−1381
Letter
Nano Letters (2) Gustafsson, M. G. Surpassing the lateral resolution limit by a factor of two using structured illumination microscopy. J. Microsc 2000, 198 (Pt 2), 82−7. (3) Heilemann, M.; et al. Subdiffraction-resolution fluorescence imaging with conventional fluorescent probes. Angew. Chem., Int. Ed. Engl. 2008, 47 (33), 6172−6. (4) Hell, S. W. Microscopy and its focal switch. Nat. Methods 2009, 6 (1), 24−32. (5) Hennig, S.; et al. Quantum dot triexciton imaging with threedimensional subdiffraction resolution. Nano Lett. 2009, 9 (6), 2466− 70. (6) Rust, M. J.; Bates, M.; Zhuang, X. Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM). Nat. Methods 2006, 3 (10), 793−5. (7) Schoen, I.; et al. Binding-activated localization microscopy of DNA structures. Nano Lett. 2011, 11 (9), 4008−11. (8) Klein, T.; et al. Live-cell dSTORM with SNAP-tag fusion proteins. Nat. Methods 2011, 8 (1), 7−9. (9) van de Linde, S.; Heilemann, M.; Sauer, M. Live-cell superresolution imaging with synthetic fluorophores. Annu. Rev. Phys. Chem. 2012, 63, 519−40. (10) Shannon, C. E. Communication in the Presence of Noise. Proc. Inst. Radio Eng. 1949, 37 (1), 10−21. (11) Shroff, H.; et al. Live-cell photoactivated localization microscopy of nanoscale adhesion dynamics. Nat. Methods 2008, 5 (5), 417−23. (12) Barber, K.; et al. Delivery of membrane-impermeant fluorescent probes into living neural cell populations by lipotransfer. Neurosci. Lett. 1996, 207 (1), 17−20. (13) Laforge, F. O.; et al. Electrochemical attosyringe. Proc. Natl. Acad. Sci. U.S.A. 2007, 104 (29), 11895−900. (14) Neumann, E.; et al. Gene transfer into mouse lyoma cells by electroporation in high electric fields. EMBO J. 1982, 1 (7), 841−5. (15) Bruckbauer, A.; et al. Nanopipette delivery of individual molecules to cellular compartments for single-molecule fluorescence tracking. Biophys. J. 2007, 93 (9), 3120−31. (16) Viigipuu, K.; Kallio, P. Microinjection of living adherent cells by using a semi-automatic microinjection system. Altern. Lab Anim. 2004, 32 (4), 417−23. (17) Delehanty, J. B.; et al. Delivering quantum dot-peptide bioconjugates to the cellular cytosol: escaping from the endolysosomal system. Integr. Biol. (Camb.) 2010, 2 (5−6), 265−77. (18) Xu, J.; et al. Nanoblade delivery and incorporation of quantum dot conjugates into tubulin networks in live cells. Nano Lett. 2012, 12 (11), 5669−72. (19) Actis, P.; et al. Voltage-controlled metal binding on polyelectrolyte-functionalized nanopores. Langmuir 2011, 27 (10), 6528−33. (20) Chen, C. C.; Zhou, Y.; Baker, L. A. Scanning ion conductance microscopy. Annu. Rev. Anal. Chem. (Palo Alto, Calif.) 2012, 5, 207− 28. (21) Hansma, P. K.; et al. The scanning ion-conductance microscope. Science 1989, 243 (4891), 641−3. (22) Novak, P.; et al. Nanoscale live-cell imaging using hopping probe ion conductance microscopy. Nat. Methods 2009, 6 (4), 279− 81. (23) Heilemann, M.; et al. Super-resolution imaging with small organic fluorophores. Angew. Chem., Int. Ed. Engl. 2009, 48 (37), 6903−8. (24) van de Linde, S.; et al. Direct stochastic optical reconstruction microscopy with standard fluorescent probes. Nat. Protoc. 2011, 6 (7), 991−1009. (25) Chong-Woo Park, J.-P. K.; Marmé, N.; Möller, M.; Spatz, J.; Wolfrum, J.; Sauer, M. Accurate Delivery of single biomolecules by polyethylene glycol coated sucmicrometer pipettes. Chem. Phys. 2004, 301 (1), 105−110. (26) Tokunaga, M.; Imamoto, N.; Sakata-Sogawa, K. Highly inclined thin illumination enables clear single-molecule imaging in cells. Nat. Methods 2008, 5 (2), 159−61.
(27) Riedl, J.; et al. Lifeact: a versatile marker to visualize F-actin. Nat. Methods 2008, 5 (7), 605−7. (28) Riedl, J.; et al. Lifeact mice for studying F-actin dynamics. Nat. Methods 2010, 7 (3), 168−9. (29) Wombacher, R.; et al. Live-cell super-resolution imaging with trimethoprim conjugates. Nat. Methods 2010, 7 (9), 717−9. (30) van de Linde, S.; et al. Photoinduced formation of reversible dye radicals and their impact on super-resolution imaging. Photochem. Photobiol. Sci. 2011, 10 (4), 499−506. (31) Roth, B. L.; et al. Bacterial viability and antibiotic susceptibility testing with SYTOX green nucleic acid stain. Appl. Environ. Microbiol. 1997, 63 (6), 2421−31. (32) Schnell, U.; et al. Immunolabeling artifacts and the need for livecell imaging. Nat. Methods 2012, 9 (2), 152−8. (33) Weber, K.; Rathke, P. C.; Osborn, M. Cytoplasmic microtubular images in glutaraldehyde-fixed tissue culture cells by electron microscopy and by immunofluorescence microscopy. Proc. Natl. Acad. Sci. U.S.A. 1978, 75 (4), 1820−4. (34) Stadler, C.; et al. A single fixation protocol for proteome-wide immunofluorescence localization studies. J. Proteomics 2010, 73 (6), 1067−78. (35) Small, J.; et al. Visualising the actin cytoskeleton. Microsc. Res. Tech. 1999, 47 (1), 3−17.
1381
DOI: 10.1021/nl504660t Nano Lett. 2015, 15, 1374−1381
