Correlative Light-Electron Microscopy of Lipid-Encapsulated

Dec 19, 2017 - Many of the FND particles (and thus the CD44 antigens) are found to be attached to the filopodia, which are as thin as 0.1 μm and prot...
0 downloads 10 Views 6MB Size
Download PDF
Technical Note Cite This: Anal. Chem. XXXX, XXX, XXX−XXX

pubs.acs.org/ac

Correlative Light-Electron Microscopy of Lipid-Encapsulated Fluorescent Nanodiamonds for Nanometric Localization of Cell Surface Antigens Feng-Jen Hsieh,†,‡,§,∥ Yen-Wei Chen,†,∥ Yao-Kuan Huang,⊥ Hsien-Ming Lee,# Chun-Hung Lin,‡,§,□ and Huan-Cheng Chang*,†,○ †

Institute of Atomic and Molecular Sciences, Academia Sinica, Taipei 106, Taiwan Taiwan International Graduate Program − Chemical Biology and Molecular Biophysics, Academia Sinica, Taipei 115, Taiwan § Department of Biochemical Sciences, National Taiwan University, Taipei 106, Taiwan ⊥ Institute of Cellular and Organismic Biology, Academia Sinica, Taipei 115, Taiwan # Institute of Chemistry, Academia Sinica, Taipei 115, Taiwan □ Institute of Biological Chemistry, Academia Sinica, Taipei 115, Taiwan ○ Department of Chemical Engineering, National Taiwan University of Science and Technology, Taipei 106, Taiwan ‡

S Supporting Information *

ABSTRACT: Containing an ensemble of nitrogen-vacancy centers in crystal matrices, fluorescent nanodiamonds (FNDs) are a new type of photostable markers that have found wide applications in light microscopy. The nanomaterial also has a dense carbon core, making it visible to electron microscopy. Here, we show that FNDs encapsulated in biotinylated lipids (bLs) are useful for subdiffraction imaging of antigens on cell surface with correlative light-electron microscopy (CLEM). The lipid encapsulation enables not only good dispersion of the particles in biological buffers but also high specific labeling of live cells. By employing the bL-encapsulated FNDs to target CD44 on HeLa cell surface through biotin-mediated immunostaining, we obtained the spatial distribution of these antigens by CLEM with a localization accuracy of ∼50 nm in routine operations. A comparative study with dual-color imaging, in which CD44 was labeled with FND and MICA/MICB was labeled with Alexa Fluor 488, demonstrated the superior performance of FNDs as fluorescent fiducial markers for CLEM of cell surface antigens.

T

is required for nanoscale analysis. A number of nanomaterials have been developed as fluorescent fiducial markers for CLEM, including dye-labeled nanogolds, dye-doped polystyrene beads (FluoSpheres), and quantum dots.6 However, these nanomaterials are not sufficiently stable to allow routine CLEM analysis because, in order to avoid damaging the markers prior to LM imaging, the specimens must be fixed by high-pressure freezing, which is a highly specialized technique and a time-consuming process. Fluorescent nanodiamond (FND) is a carbon-based nanoparticle containing a high-density ensemble of negatively charged nitrogen-vacancy (NV−) centers as light emitters.7 The nanomaterial possesses several unique optical properties. First, the NV− centers are exceptionally photostable and can be detected individually by super-resolution fluorescence micros-

he surface of a cell is covered with various types of antigens. These antigens serve as molecular markers for identification and classification of cell types as well as targets for diagnosis and therapy. CD44, for example, is a well-known cell surface receptor highly expressed in tumor cells and heavily involved in cancer cell metastases.1 Some of the antigens are attached to filopodia, which are thin actin-rich bundles protruding from cell plasma membranes. They are responsible for cell migration and invasion during cancer progression.2 Despite the importance, locating the exact positions of these and other antigens on cell surface with nanometer resolution remains a challenge in analytical chemistry. Correlative light-electron microscopy (CLEM) is a technique that combines the multicolor versatility of light microscopy (LM) and the high resolution power of electron microscopy (EM) for molecular and cellular biology research.3,4 Different from cathodoluminescence microscopy,5 CLEM is performed with two vastly different instruments and therefore highprecision colocalization of same objects in the respective images © XXXX American Chemical Society

Received: November 3, 2017 Accepted: December 19, 2017 Published: December 19, 2017 A

DOI: 10.1021/acs.analchem.7b04549 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry

(1 mL) at 60 °C and shaken for 30 s. After forming a homogeneous lipid-FND suspension, tetrahydrofuran was removed by evaporation using a rotary evaporator to obtain lipid-encapsulated FNDs. To eliminate excess lipids, the LFNDs were cleaned twice with DDW by centrifugal separation. Cell Culture and Labeling. HeLa cells were incubated in Dulbecco’s modified Eagle’s medium containing 10% fetal calf serum and 1% penicillin at 37 °C with 5% CO2. To conduct CD44 labeling, cells were seeded in 35 mm dishes with a density of 1 × 105 cells/well (70% confluence) 1 day before experiments. They were then washed with PBS three times and fixed with 4% paraformaldehyde for 15 min. After removal of paraformaldehyde by PBS wash, the fixed cells were blocked with 3% bovine serum albumin (BSA) in PBS for 30 min, followed by the addition of 1.5 μg/mL biotinylated anti-CD44 antibody in 3% BSA-PBS (1 mL) to target CD44 antigens on the surface of the fixed cells. After 30 min incubation, the antiCD44-labeled cells were washed 3 times with 3% BSA-PBS to remove free antibody and then stained with 1.5 μg/mL DyLight488-conjugated neutravidin in 3% BSA-PBS (1 mL). Following the same 30 min incubation and 3% BSA-PBS wash, the cells were stained with bL-FND (e.g., 100 μg/mL) in 3% BSA-PBS for 30 min and finally washed with 3% BSA-PBS to remove nonspecifically bound bL-FNDs. The same procedures were applied to label MICA/MICB antigens, except that the reagents used were 1.5 μg/mL antihuman MICA/MICB antibody and 1.5 μg/mL Alexa Fluor 488 antimouse IgG antibody. Flow Cytometry. Cells attached to dishes were trypsinized, fixed with 4% paraformaldehyde, and divided into 15 mL tubes (TPP Techno Plastic Products AG) with a total number of 2 × 105 cells per tube. After labeling with bL-FNDs, the cells (∼100000) from each sample were analyzed by using a flow cytometer (FACSArray Bioanalyser, BD Bioscience) equipped with a 532 nm laser for the excitation and the resulting fluorescence was detected at wavelengths >590 nm. Confocal Fluorescence Microscopy. Fluorescence images were acquired with a laser-scanning confocal fluorescence microscope (SP-8, Leica) equipped with a supercontinuum white-light laser operating at 488 and 561 nm for the excitation of DyLight488 and FND, respectively. The corresponding fluorescence was collected through an oil-immersion objective (63×, NA 1.4) and detected with either a photomultiplier (PMT) for DyLight488 or a hybrid detector (HyD) for FND at emission wavelengths of 500−550 and 650−800 nm, respectively. Scanning Electron Microscopy (SEM). Single layer HeLa cells (∼1 × 105) were seeded on a 35 mm dish with a gridded glass bottom (MetTek) and labeled with biotinylated antiCD44 antibody, neutravidin, and bL-FND as described earlier. They were then fixed with 2.5% glutaraldehyde and 1% osmium tetroxide at 4 °C in the dark for 1 h and dehydrated through an ascending concentration series of ethanol (30−100%) for 15 min in each step to reduce the water content. After being fully dehydrated in a critical point dryer (Samdri PVT-3B, Tousimis), the cells were coated with platinum before being imaged with a field emission scanning electron microscope (S4800, Hitachi) operating at 10 kV.

copy.8 Second, having a dense carbon core, the particles are visible by EM in cells and tissue samples.9−12 Third, the fluorescence imaging of FND is perfectly compatible with its EM imaging because the NV− centers are embedded deeply inside the chemically inert diamond matrix and thus are protected from damages caused by heavy metal staining (e.g., uranyl acetate and lead citrate staining). Our previous experiments have shown that the optical properties of FNDs (diameter ∼ 100 nm) are nearly unaffected by strong acid and base treatments in aqueous solution at room temperature,13 suggesting that these nanoparticles are useful as a robust and versatile new tool for nanometric localization of subcellular components by CLEM. In applying FND as a CLEM marker for nanometric localization of cell surface antigens, two hurdles must be overcome: (1) particle agglomeration in cell medium and (2) nonspecific binding with undesired molecules on cell surface, both of which can lead to false positive results. Several endeavors have been made to address these issues.14−18 Notable examples include covalent conjugation of carboxylated FNDs with polymers such as hyperbranched polyglycerol14,18 and poly[N-(2-hydroxypropyl)methacrylamide]17 to form highly biocompatible protein-resistant coating. Although the validity of these approaches has been experimentally demonstrated, a more general and effective method to facilitate the coating of FNDs for biolabeling applications is needed. Here, we present a simple protocol to encapsulate FNDs in lipid layers for bioimaging with CLEM. Instead of using the thin-film hydration technique,19 we took advantage of the Ouzo effect,20 which is a spontaneous emulsification phenomenon involving the addition of a mixture of hydrophobic solute (e.g., lipids) and water-miscible solvent (e.g., tetrahydrofuran) into water to form stable microdroplets that act as carriers for the compounds of interest (e.g., FNDs). Subsequent evaporation of the solvent in vacuum allows the cargo to be encapsulated in the hydrophobic layers. The method enables not only robust coating but also the synthesis of lipid-coated FNDs (L-FNDs) grafted with various functional groups such as biotin. The LFND particles exhibit exceptionally high dispersibility in biological buffers, well suited for specific labeling and targeting applications.

■

EXPERIMENTAL SECTION FND Production. Diamond powders were radiationdamaged by 40-keV He+ ions to create vacancies, followed by annealing in vacuum at 800 °C to form NV− centers.21 The asproduced FND particles were then oxidized in air at 450 °C for 2 h to remove graphitic carbon atoms on surface. Particles of 50 and 30 nm in diameter were obtained by high-pressure crushing of 100 nm FNDs, separated by differential centrifugation, and characterized by dynamic light scattering in either distilled deionized water (DDW) or phosphate-buffered saline (PBS) with a particle size analyzer (Delsa Nano C, BeckmanCoulter).22 Preparation of bL-FNDs. Surface-oxidized FNDs were dispersed in DDW (1 mg/mL) by sonication. L-α-Phosphatidylcholine from chicken eggs (Egg PC, 13 mg), cholesterol (8 mg), 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N[methoxy(polyethylene glycol)-2000] (DSPE-PEG2000, 10 mg), and 1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N[biotinyl(polyethylene glycol)-2000] (DSPE-PEG2000-biotin, 1.2 mg) were mixed together in tetrahydrofuran (1 mL). The lipid mixture was then slowly dropped into the FND solution

■

RESULTS AND DISCUSSION Lipid Encapsulation. The lipid mixture used to synthesize L-FNDs consisted of Egg PC and cholesterol (a membrane stabilizer). A key step in this synthesis is the surface B

DOI: 10.1021/acs.analchem.7b04549 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry

In a demonstration for the usefulness of bL-FNDs as fluorescent fiducial markers, we applied the particles to label CD44 antigens on the outer membrane of HeLa cells by threestep sandwich immunostaining (Figure 1b). To conduct the FND labeling, HeLa cells were first fixed with 4% paraformaldehyde and then sequentially stained with biotinylated anti-CD44 antibody, neutravidin, and bL-FND. To determine the optimal content of biotin on the FND surface, we produced bL-FNDs with different molar ratios (0%, 0.5%, or 1%) of DSPE-PEG2000-biotin and 9% DSPE-PEG2000 in the lipid layer. Flow cytometric analysis of cells labeled with these particles in the far-red channel indicated that the bLFNDs containing a larger amount of biotin have a higher targeting efficiency, but the effect reaches a plateau at ∼1% (Figure S3). We thus used bL-FNDs synthesized with 1% DSPE-PEG2000-biotin and 9% DSPE-PEG2000 in subsequent experiments. Nonspecific labeling is a major concern in this study. To address the concern, we performed dose-dependent measurements by flow cytometry for bL-FND-labeled HeLa cells. As displayed in Figure 1c, the amount of FNDs on the cell surface increased steadily with the increasing particle concentration up to 700 μg/mL. Although the higher concentration of bL-FND boosted the fluorescence intensity, it also substantially leveled up the background signals. At 700 μg/mL, the control experiment without using neutravidin in the immunostaining showed a contribution of ∼6% to the total signals by the nonspecific labeling. To minimize the interference from nonspecific labeling in subsequent analysis, bL-FNDs with a concentration of ≤300 μg/mL were applied for nanometric identification and localization of surface antigens by CLEM. In addition to flow cytometry, fluorescence microscopy also provides a means to confirm the absence of nonspecific labeling of bL-FNDs. In this experiment, we first labeled the cells with biotin-anti-CD44 antibody, DyLight488-conjugated neutravidin, and bL-FND at a particle concentration of 100 μg/mL for 30 min. Confocal fluorescence microscopy revealed good colocalization of FND and DyLight488 on the cell membrane (Figure 2a−c), proving the high specific labeling ability of the lipid-coated fluorescent nanoparticles. An enlarged view of the image of the cells showed that many of the red fluorescence spots have diffraction-limited sizes, likely derived from single FNDs (Figure 2d). Notably, these particles did not photobleach, in stark contrast to DyLight488 whose fluorescence emission could be barely detected in the second scan of the excitation laser through the zoom-in region (Figure S4). There is essentially no nonspecific labeling of the cells as confirmed by control experiments (Figure S5). Correlative Light-Electron Microscopy. The absence of nonspecific labeling of bL-FND at concentrations less than 300 μg/mL led us to apply CLEM to find the positions of CD44 antigens on the HeLa cell surface with scanning electron microscopy (SEM). An intrinsic problem of SEM is that the samples must be dehydrated by critical point drying prior to imaging. This inevitably causes irregular morphological changes of the samples even after fixation with 2.5% glutaraldehyde and 1% osmium tetroxide.24 As a result, the LM images taken before sample dehydration cannot be directly superimposed on the EM images after sample dehydration with high precision. With the availability of bL-FNDs as biolabels, the aforementioned difficulties can be easily surmounted because one can observe the fluorescence emission from these carbon nanoparticles after EM imaging without problem. Figure 3a

modification of FNDs. We found that air oxidation of the particles at 450 °C for 2 h effectively terminated the diamond surface with oxygen atoms. Roughly 30% of the surface carbon atoms are bound with oxygen and the rest of them are in sp2and sp3-C forms.23 The presence of these hydrophobic groups significantly promotes the interaction of the hydrocarbon tails of Egg PC with the FND surface. Figure 1a displays a plausible

Figure 1. (a) Proposed structure of a biotinylated lipid-encapsulated FND (bL-FND). (b) Sandwich immunostaining of HeLa cells with bL-FNDs. (c) Flow cytometric analysis of HeLa cells labeled with 100 nm bL-FNDs at the concentration of 0−700 μg/mL. No neutravidin was added for the labeling in control experiments.

structure of the lipid-encapsulated FND. Any desired functional groups (e.g., biotin, carboxyl, and amino groups, etc.) can be added onto the mixed lipid layer through minor changes of the compositions. Biotinylated PEGylated DSPE is one of such compounds to be incorporated in the lipid mixture. To optimize the lipid/FND ratio for sample preparation, we measured the size distribution of L-FNDs to examine their dispersibility in DDW and PBS. For air-oxidized FNDs of ∼100 nm in diameter, the optimal lipid/FND ratio to ensure good dispersion of the particles in both solutions was found to be about 30:1 in weight. The L-FNDs dispersed well in PBS even after 3−6 washes to remove free micelles in the buffer. The high dispersibility could also be maintained if the lipid layer contained 1% DSPE-PEG2000-biotin and 9% DSPE-PEG2000 in molar ratio (Table S1). The number-averaged hydrodynamic diameters of these biotinylated L-FNDs (denoted as bL-FNDs) in DDW and PBS are 144 ± 38 and 149 ± 41 nm, respectively. In contrast, the mean size of the FNDs without lipid coating is more than 500 nm in PBS (Figure S1a). Similar results were obtained for 50 and 30 nm FNDs before and after coating with bL (Figure S1b,c). Stability tests of the bL-FNDs in PBS revealed that their particle sizes are essentially unchanged after storage in PBS for 6 h at room temperature (Figure S2a). Moreover, the bL-FNDs prepared 6 months ago and stored at 4 °C have good dispersibility in PBS (Figure S2b). They are ready for use without sonication. C

DOI: 10.1021/acs.analchem.7b04549 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry

Figure 2. (a−c) Bright-field and Z-stacked confocal microscopy images of HeLa cells labeled with 100 nm bL-FND by sandwich immunostaining. The cells were first labeled with biotin-anti-CD44 antibody, DyLight488-neutravidin (green), and then bL-FND (red). Yellow spots in (c) indicate the colocalization of DyLight488neutravidin in (a) and bL-FND in (b). (d) Zoom-in image of the region enclosed in the white box in (b). Scale bars: 10 μm (c) and 1 μm (d). Figure 3. (a) SEM image of a HeLa cell labeled with anti-CD44 antibody, neutravidin, and 100 nm bL-FND. (b, c) Zoom-in EM, (d, e) bright-field LM, (f, g) fluorescence, and (h, i) CLEM images of the white boxes in (a). The tilt angles to achieve complete overlaps of the FNDs are 1° and 39° for (b, f, h) and (c, g, i), respectively. (j) Enlarged view of the white box in (h). (k) Intensity profiles of the white dashed line drawn in (j). Scale bars: 10 μm (a) and 2 μm (b, c, f, g).

shows a typical SEM image of HeLa cells labeled with biotinanti-CD44 antibody, neutravidin, and bL-FND. Clearly, the EM imaging (Figure 3b,c) provides considerably more structural information than the corresponding LM imaging (Figure 3d,e). Although some bL-FNDs can be identified under high magnification with SEM, unambiguous discrimination of them from underlying complex nanostructures on the cell surface is not possible with the EM alone. Taking advantage of the exceptionally high chemical and photophysical stability of FNDs, we were able to detect their fluorescence even after strong irradiation of the Pt-coated cell samples with 10 kV electrons (Figure 3f,g). Superimposition of the LM and EM images with one of them tilted by an angle, deduced by using the eC-CLEM software,25 allowed us to determine the locations of the CD44 antigens on the HeLa cell membrane with high accuracy (Figure 3h,i). There is no need to correct the distortion caused by dehydration of the cells in the analysis. In Figure 3j, we show an enlarged view of some representative fluorescence spots and their corresponding entities in the SEM image. Many of the FND particles (and thus the CD44 antigens) are found to be attached to the filopodia, which are as thin as 0.1 μm and protrude from the cell body. With the use of bL-FNDs in combination with CLEM, the positions of CD44 on the filopodia can be measured with an accuracy of ∼50 nm (Figure 3k), which is about 1/5 that of the diffraction limit of light. The accuracy is expected to be further improved to be better than 20 nm if smaller FND particles (such as the 30 nm ones26) are used in the measurements. To further demonstrate the superior performance of FNDs as fluorescent fiducial markers, we performed a comparative study with dual-color CLEM, in which CD44 was labeled with

FND and MICA/MICB was labeled with Alexa Fluor 488. MICA and MICB are transmembrane glycoproteins related to the major histocompatibility complex (MHC) class I molecules, functioning as stress-induced antigens.27 Prior to the SEM imaging, confocal fluorescence microscopy revealed the presence of both CD44 (red) and antigens MICA/MICB (green) on the HeLa cell surface, and they showed markedly different distributions on the membrane (Figure 4a−c). However, following all chemical treatments (including dehydration and metal staining) necessary for EM, no green emission could be observed any more for the cell, whereas the red emission from the FNDs persisted even after the SEM imaging (Figure 4d−f). Again, the high registration of LM and EM images allowed us to identify the positions of the individual FNDs (or CD44 antigens) with nanometric resolution (Figure 4g−i). Moreover, the positions of the MICA/MICB antigens could be measured with an accuracy of ∼100 nm, despite that substantial morphological change of the cells after the SEM imaging had occurred. Such a measurement is made possible by nonlinear transformation of the LM images pixel by pixel using the multidimensional registration software,25 together with the use of more than 40 FND particles on the cell membrane as fiducial markers (Figures 4j−l and S6). D

DOI: 10.1021/acs.analchem.7b04549 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry

Figure 4. (a−f) Bright-field and Z-stacked confocal microscopy images of a HeLa cell first labeled for MICA/MICB antigens with the Alexa Fluor 488 dye and then for CD44 antigens with 100 nm bL-FND by immunostaining before (a−c) and after (d−f) SEM imaging. (c, f) Zoom-in fluorescence images of the regions indicated by the white boxes in (a, d) before and after SEM imaging. Note the significant morphological change after SEM imaging and the disappearance of green emission in (f). (g) SEM image of the corresponding MICA/MICB- and CD44-labeled HeLa cell. (h, i) Zoom-in SEM and CLEM images of the region indicated by the white box in (g). The CLEM image was constructed by superimposing the images in (f) and (h), and the tilt angle to achieve complete overlap of the FNDs is 23°. (j−l) Nonlinear transformation of the LM image in (c) with the eC-CLEM software. The deformed grid used to perform the nonlinear transformation is given in (j) and the resulting LM and CLEM images are shown in (k) and (l), respectively. Scale bars: 20 μm (a, b, d, e, g) and 5 μm (c, f, h, i).

commercially available and they all serve well the purpose after minor modification of the protocols presented in this report.

Apart from SEM, the presently developed method is also compatible with transmission electron microscopy (TEM).28 Figure S7a−c display TEM images of suspended HeLa cells either labeled with biotin-anti-CD44 by the sandwich immunostaining or without labeling. In order to reduce the background signals arising from illumination of the resins, the fluorescence signals were collected at the time longer than 3 ns after pulsed laser excitation.29 Using this time-gating technique, we were able to identify the individual 100 nm FND particles on the surface of the cells in the thin-sectioned resins without difficulty. Moreover, the EM (Figure S7d) and LM (Figure S7e) images can be readily superimposed on each other with FNDs as the fiducial markers (Figure S7f). The localization accuracy of the CD44 antigens on the outer membrane of the cells is ∼50 nm, again limited by the size of the FND particles. In conclusion, we have demonstrated that FNDs surfaceoxidized in air can be facilely encapsulated in lipids by utilizing the Ouzo effect, and these lipid-encapsulated FNDs are useful as specific cell targeting agents after proper conjugation of the lipid layers with bioactive molecules such as biotin. We have applied the particles for nanometric localization of CD44 antigens on HeLa cell membrane with CLEM to prove the principle. More importantly, the correlative analysis can be routinely and repeatedly carried out, thanks to the exceptional chemical inertness of the electron-dense crystal matrix and the excellent photostability of the hosted color centers. To the best of our knowledge, FND is the only carbon nanoparticle capable of acting both as a targeting agent and as a fiducial marker for CLEM so far. The method is general and applicable to other biomolecules as well, since a variety of lipid derivatives are now

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.analchem.7b04549. Experimental details, supporting figures, and supporting table (PDF).

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Huan-Cheng Chang: 0000-0002-3515-4128 Author Contributions ∥

These two authors contributed equally to this work.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by Academia Sinica and the Ministry of Science and Technology, Taiwan, with Grant No. 104-2811M-001-149. We thank Su-Jen Ji, Ching-Yen Lin, Ya-Yun Yang, and the staffs of TC5 Bio-Image Tools, Technology Commons, College of Life Science, National Taiwan University for the assistance with SEM & TEM sample preparation and measurements. E

DOI: 10.1021/acs.analchem.7b04549 Anal. Chem. XXXX, XXX, XXX−XXX

Technical Note

Analytical Chemistry

■

REFERENCES

(1) Senbanjo, L. T.; Chellaiah, M. A. Front. Cell Dev. Biol. 2017, 5, 18. (2) Pusch, A.; Boeckenhoff, A.; Glaser, T.; Kaminski, T.; Kirfel, G.; Hans, M.; Steinfarz, B.; Swandulla, D.; Kubitscheck, U.; Gieselmann, V.; Brustle, O.; Kappler, J. Biochim. Biophys. Acta, Mol. Cell Res. 2010, 1803, 261−274. (3) Muller-Reichert, T.; Verkade, P. Correlative Light and Electron Microscopy; Elsevier Science, 2012. (4) de Boer, P.; Hoogenboom, J. P.; Giepmans, B. N. G. Nat. Methods 2015, 12, 503−513. (5) Kociak, M.; Zagonel, L. F. Ultramicroscopy 2017, 176, 112−131. (6) Kukulski, W.; Schorb, M.; Welsch, S.; Picco, A.; Kaksonen, M.; Briggs, J. A. G. Methods Cell Biol. 2012, 111, 235−257. (7) Hsiao, W. W.-W.; Hui, Y. Y.; Tsai, P.-C.; Chang, H.-C. Acc. Chem. Res. 2016, 49, 400−407. (8) Arroyo-Camejo, S.; Adam, M.-P.; Besbes, M.; Hugonin, J.-P.; Jacques, V.; Greffet, J.-J.; Roch, J.-F.; Hell, S. W.; Treussart, F. ACS Nano 2013, 7, 10912−10919. (9) Vaijayanthimala, V.; Cheng, P.-Y.; Yeh, S.-H.; Liu, K.-K.; Hsiao, C.-H.; Chao, J.-I.; Chang, H.-C. Biomaterials 2012, 33, 7794−7802. (10) Chu, Z. Q.; Zhang, S. L.; Zhang, B. K.; Zhang, C. Y.; Fang, C.Y.; Rehor, I.; Cigler, P.; Chang, H.-C.; Lin, G.; Liu, R. B.; Li, Q. Sci. Rep. 2015, 4, 4495. (11) Nawa, Y.; Inami, W.; Lin, S.; Kawata, Y.; Terakawa, S.; Fang, C.Y.; Chang, H.-C. ChemPhysChem 2014, 15, 721−726. (12) Nagarajan, S.; Pioche-Durieu, C.; Tizei, L. H. G.; Fang, C.-Y.; Bertrand, J.-R.; Le Cam, E.; Chang, H.-C.; Treussart, F.; Kociak, M. Nanoscale 2016, 8, 11588−11594. (13) Kuo, Y.; Hsu, T.-Y.; Wu, Y.-C.; Hsu, J.-H.; Chang, H.-C. Proc. SPIE 2013, 863503−8. (14) Zhao, L.; Takimoto, T.; Ito, M.; Kitagawa, N.; Kimura, T.; Komatsu, N. Angew. Chem., Int. Ed. 2011, 50, 1388−1392. (15) Lee, J. W.; Lee, S.; Jang, S.; Han, K. Y.; Kim, Y.; Hyun, J.; Kim, S. K.; Lee, Y. Mol. BioSyst. 2013, 9, 1004−1011. (16) Chang, B.-M.; Lin, H.-H.; Su, L.-J.; Lin, W.-D.; Lin, R.-J.; Tzeng, Y.-K.; Lee, R. T.; Lee, Y. C.; Yu, A. L.; Chang, H.-C. Adv. Funct. Mater. 2013, 23, 5737−5745. (17) Slegerova, J.; Hajek, M.; Rehor, I.; Sedlak, F.; Stursa, J.; Hruby, M.; Cigler, P. Nanoscale 2015, 7, 415−420. (18) Shingo, S.; Ryuji, I.; Jun, I.; Yuta, K.; Hidehito, T.; Yoshie, H.; Masahiro, S. Chem. Lett. 2015, 44, 354−356. (19) Moore, L.; Chow, E. K.; Osawa, E.; Bishop, J. M.; Ho, D. Adv. Mater. 2013, 25, 3532−3541. (20) Vitale, S. A.; Katz, J. L. Langmuir 2003, 19, 4105−4110. (21) Chang, Y.-R.; Lee, H.-Y.; Chen, K.; Chang, C.-C.; Tsai, D.-S.; Fu, C.-C.; Lim, T.-S.; Tzeng, Y.-K.; Fang, C.-Y.; Han, C.-C.; Chang, H.-C.; Fann, W. Nat. Nanotechnol. 2008, 3, 284−288. (22) Su, L.-J.; Fang, C.-Y.; Chang, Y.-T.; Chen, K.-M.; Yu, Y.-C.; Hsu, J.-H.; Chang, H.-C. Nanotechnology 2013, 24, 315702. (23) Sotoma, S.; Akagi, K.; Hosokawa, S.; Igarashi, R.; Tochio, H.; Harada, Y.; Shirakawa, M. RSC Adv. 2015, 5, 13818−13827. (24) Zhang, Y.; Huang, T.; Jorgens, D. M.; Nickerson, A.; Lin, L.-J.; Pelz, J.; Gray, J. W.; López, C. S.; Nan, X. PLoS One 2017, 12, e0176839. (25) Paul-Gilloteaux, P.; Heiligenstein, X.; Belle, M.; Domart, M.-C.; Larijani, B.; Collinson, L.; Raposo, G.; Salamero, J. Nat. Methods 2017, 14, 102−103. (26) Tzeng, Y.-K.; Faklaris, O.; Chang, B.-M.; Kuo, Y.; Hsu, J.-H.; Chang, H.-C. Angew. Chem., Int. Ed. 2011, 50, 2262−2265. (27) Beckman, E. M.; Brenner, M. B. Immunol. Today 1995, 16, 349−352. (28) Prabhakar, N.; Peurla, M.; Koho, S.; Deguchi, T.; Näreoja, T.; Chang, H.-C.; Rosenholm, J. M.; Hänninen, P. E. Small 2017, 1701807. (29) Wu, T.-J.; Tzeng, Y.-K.; Chang, W.-W.; Cheng, C.-A.; Kuo, Y.; Chien, C.-H.; Chang, H.-C.; Yu, J. Nat. Nanotechnol. 2013, 8, 682− 689.

F

DOI: 10.1021/acs.analchem.7b04549 Anal. Chem. XXXX, XXX, XXX−XXX