Nanoscale

PAPER

Cite this: DOI: 10.1039/c6nr01908k

Received 6th March 2016,Accepted 28th April 2016

DOI: 10.1039/c6nr01908k

www.rsc.org/nanoscale

Simultaneous cathodoluminescence and electronmicroscopy cytometry of cellular vesicles labeledwith fluorescent nanodiamonds†

Sounderya Nagarajan,a Catherine Pioche-Durieu,b Luiz H. G. Tizei,a Chia-Yi Fang,c

Jean-Rémi Bertrand,d Eric Le Cam,b Huan-Cheng Chang,c François Treussart*e andMathieu Kociak*a

Light and Transmission Electron Microscopies (LM and TEM) hold potential in bioimaging owing to the

advantages of fast imaging of multiple cells with LM and ultrastructure resolution offered by TEM. Inte-

grated or correlated LM and TEM are the current approaches to combine the advantages of both tech-

niques. Here we propose an alternative in which the electron beam of a scanning TEM (STEM) is used to

excite concomitantly the luminescence of nanoparticle labels (a process known as cathodoluminescence,

CL), and image the cell ultrastructure. This CL-STEM imaging allows obtaining luminescence spectra and

imaging ultrastructure simultaneously. We present a proof of principle experiment, showing the potential

of this technique in image cytometry of cell vesicular components. To label the vesicles we used fluo-

rescent diamond nanocrystals (nanodiamonds, NDs) of size ≈150 nm coated with different cationic poly-

mers, known to trigger different internalization pathways. Each polymer was associated with a type of ND

with a different emission spectrum. With CL-STEM, for each individual vesicle, we were able to measure (i)

their size with nanometric resolution, (ii) their content in different ND labels, and realize intracellular com-

ponent cytometry. In contrast to the recently reported organelle flow cytometry technique that requires

cell sonication, CL-STEM-based image cytometry preserves the cell integrity and provides a much higher

resolution in size. Although this novel approach is still limited by a low throughput, the automatization of

data acquisition and image analysis, combined with improved intracellular targeting, should facilitate

applications in cell biology at the subcellular level.

Cell imaging techniques have been constantly evolving. Fluo-rescence microscopy is most widely used, owing to the possi-bility of labeling cell compartments with high specificity byimmunofluorescence or by the expression of fluorescent pro-teins. In the last few decades fluorescence-based super-resolu-tion microscopy has opened new prospects in structuralimaging.1–3 This is particularly true in studying cellulardynamics and a striking example is the monitoring of endocyto-tic pit and actin filament formation imaged at a resolution of84 nm by structured illumination microscopy.4 However, there

still exist complexities in sample preparation, choice of fluoro-phores and difficulties in imaging the cellular morphology atsub-organelle resolution, despite this recent progress.

In contrast, electron microscopy (EM) easily reaches thisspatial resolution when conducting structural investigations.Techniques where fluorescence and EM imaging can be corre-lated on the same sample, usually coined Correlative Light andElectron Microscopy (CLEM), have attracted much interest inrecent years and are now becoming standard.5 Moreover, withthe recent development of genetically encoded tags providingboth EM and LM contrasts6 and complementing advanta-geously immunochemical nanogold labeling,7 the gap betweenEM and LM is about to be bridged.

These methods utilizing both electron and light micro-scopies are implemented either independently or in correlativeor integrated manners.8 In conventional CLEM, samples aretransferred between the light and the electron microscopes.This approach requires locating the same region of interestwith fiducial markers,9 which is not straightforward as thereare non-linear distortions of the images due to various scan-ning systems. Integrated light and electron microscopy tries to

†Electronic supplementary information (ESI) available. See DOI: 10.1039/c6nr01908k

aLaboratoire de Physique des Solides, CNRS, Univ. Paris-Sud, Université Paris-

Saclay, 91405 Orsay, France. E-mail: mathieu.kociak@u-psud.frbSignalisations, Noyaux et Innovations en Cancérologie, UMR 8126 CNRS, Univ.

Paris-Sud, Université Paris-Saclay, Gustave Roussy, 94805 Villejuif, FrancecInstitute of Atomic and Molecular Sciences, Academia Sinica, Taipei 106, TaiwandLaboratoire de Vectorologie et Thérapeutiques Anticancéreuses, UMR8203, CNRS,

Univ. Paris-Sud, Gustave Roussy, Université Paris-Saclay, 94805 Villejuif, FranceeLaboratoire Aimé Cotton, CNRS, Univ. Paris-Sud, ENS Cachan, Université Paris-

Saclay, 91405 Orsay, France. E-mail: francois.treussart@ens-cachan.fr

This journal is © The Royal Society of Chemistry 2016 Nanoscale

Publ

ishe

d on

24

May

201

6. D

ownl

oade

d by

Aix

Mar

seill

e U

nive

rsité

on

24/0

5/20

16 1

5:45

:29.

View Article OnlineView Journal

get around these issues by using a fluorescence detection setup fitted on an electron microscope,9 but it requires a lightsource aligned with the electron beam to excite the emitters.

An interesting alternative is to use electrons not only forstructural imaging, but also to take advantage of the lumines-cence of some materials when they are probed with anaccelerated electron beam, an effect known as cathodolumi-nescence (CL).10 This implies that the CL emission from thesample can be collected while it is imaged by using an electronmicroscope without the need to move it from the EM to LMsetup. This approach has been used via the combination of CLand secondary electron scanning EM (SEM) for simultaneousluminescence and structural imaging of fluorescent rare-earth-doped yttrium or gadolinium oxide (Y2O3 or Gd2O3) particlesafter their uptake in cells11,12 and europium–zinc-codopedY2O3 nanoparticles of size down to 30 nm.13 Although SEMcombined with CL (SEM-CL) provides a sub-cellular resolu-tion,11 it is not sufficient for determining the cell ultrastruc-ture. On the other hand, Scanning Transmission ElectronMicroscopy (STEM) offers such an ultrastructural resolution,and it is thus appealing to extend SEM-CL to STEM-CL so as toget both ultrastructure and luminescence information in paral-lel. Indeed we demonstrated recently that STEM-CL is sensitiveenough to address single nanoparticles, such as individual≈20 nm sized core/shell quantum dots14 or ≈100 nm sizednanodiamonds (NDs) with embedded defect centers.10,15 In ourCL-STEM configuration, the light collected is spectrally analyzedat each beam position, yielding one CL spectrum per pixel, andthe matrix of the CL spectrum is called a hyperspectral image.

NDs are actually remarkable STEM-CL labels owing to thelarge quantum yield and low electrobleaching of theembedded luminescent defects. Moreover, we have used cat-ionic polymer-coated NDs to deliver genes to Ewing sarcomacells in culture16 and have shown, in this context, that theinternalization pathways and therefore the corresponding ves-icular compartment labeled by the fluorescent ND (size≈50 nm), depends on the cationic molecules used:17 the poly-allylamine hydrochloride (PAH) coating triggers clathrin-mediated endocytosis, whereas both clathrin-mediated endo-cytosis and macropinocytosis are involved in the case of poly-ethylene imine (PEI) coating, with the predominance ofmacropinocytosis. Such coated NDs are thus interesting labelsof different internalization pathways.

In this Letter, we report a novel ultrastructural image cyto-metry method to investigate quantitatively the cellular uptakeof nanoparticles. Our approach relies on the combination ofSTEM-CL imaging with fluorescent NDs having different spec-tral signatures for each type of cationic coating. In this way,each cationic coating can be tracked through its spectral signa-ture while this is impossible by simple TEM imaging. Weapplied our method to some quantitative characterization ofthe compartments containing NDs taken up by human Ewingsarcoma cells A673 in culture. We studied single (ND-PEI orND-PAH alone) and double labeling (both types of cationicNDs put together) situations. Four hours after their cellularuptake, all NDs localized inside vesicles and were not found

free in the cytoplasm. From ultrastructure images, we inferredthe size of individual ND-containing vesicles, and used theultrastructure-correlated CL emission to count the number offluorescent NDs inside each vesicle. These counts are equi-valent to luminescence intensities in conventional cytometry.For each vesicle we obtain a set of data composed of the size,number and the type (PEI, PAH or mixed PEI/PAH coating) ofcationic NDs it contains. These data allowed us to build plotsoffering a synthetic view of all the information at the vesiclescale. Being based on ultrastructural imaging our novel cyto-metry is a method that could be further developed by extract-ing information, beyond the vesicle size.

Results and discussion

We have used NDs with diameters ≈150 nm (see the ESI†) con-taining nominally either Nitrogen Vacancy (NV) color centers(further referred as Type 1 NDs), having red CL emission15 ornitrogen vacancy nitrogen color centers (H3) with green CL emis-sion15 (further referred as Type 2 NDs). Type 1 NDs are madefrom synthetic precursors while the Type 2 NDs are made fromnatural diamond, which implies a better control of nitrogenimpurity in Type 1 NDs than in Type 2 NDs. We observed thatType 1 NDs were pure red emitters while a Type 2 ND samplewas composed of a mixture of either green or red emitting NDs.We measured the content of pure green emitting NDs in a Type2 sample and found it to be 57 ± 2% (ESI Fig. S1†), with the rest(43%) being composed of NDs with pure red emission. We alsoobserved that ND in both samples had crystals that did notshow any luminescence. We measured the fraction of fluo-rescent NDs to be 67 ± 5% for Type 1 and 62 ± 3% for Type 2.

Type 1 NDs were coated with PAH, (ND-PAH) and Type 2NDs with PEI (ND-PEI). We incubated these cationic polymer-coated NDs with A673 cells, either separately or in conjunction(see Methods). The polymer coated NDs were spontaneouslytaken up by cells, and we then analyzed their intracellular dis-tribution by STEM and CL-hyperspectral imaging. STEMenables imaging at the scale of a whole cell (Fig. 1A) down tothe ultrastructure (Fig. 1B), in either a Bright Field (BF) modewith contrasts similar to that of conventional TEM, or HighAngle Annular Dark Field (HAADF) with roughly speaking aninversed contrast. In both modes, subcellular compartmentslike mitochondria, lysosomes, endosomes and nuclear mem-branes, and at the same time NDs, are observed with a highcontrast. NDs appear as black or white features in either theBF or HAADF mode respectively. Fig. 1C displays the detectionsystem: the two images are obtained by scanning an electronbeam (acceleration voltage ≈ 60 kV) onto the sample, and col-lecting the BF and the HAADF signal at each scan position. Atthe same time, a CL spectrum, showing clear NV or H3 signa-tures (Fig. 1D) is collected at each scan position. The NV signa-ture presents a well-known form, consisting of a Zero PhononLine (ZPL) centered at 575 nm wavelength (corresponding tothe neutral form18 NV0 followed by a convoluted series ofpeaks corresponding to phonon replica). Various analyses can

Paper Nanoscale

Nanoscale This journal is © The Royal Society of Chemistry 2016

Publ

ishe

d on

24

May

201

6. D

ownl

oade

d by

Aix

Mar

seill

e U

nive

rsité

on

24/0

5/20

16 1

5:45

:29.

View Article Online

be performed on this set of data (later called a spectrumimage, Spim). The most straightforward method is to integratethe intensity in a given wavelength range (Fig. 1D) for everypixel of the Spim and map this intensity to each point of thescan, producing colored images. This colored image can berelated, pixel-by-pixel, to the morphological images (either BFor HAADF). As shown in Fig. 1E, it is possible to correlate theemission properties of the fluorescent NDs and the sample

ultrastructure (the vesicles) with nanometer resolution.However, while the green color unambiguously identifies aND-PEI, the red emission might be attributed to a PAH or aPEI coated ND, due to the non-purity of the Type 2 sample.Fortunately, the exact spectral shape can be used to disentangleboth cases. Fig. 2C shows that for known samples of ND-PEIand ND-PAH their CL spectra significantly differ in the ratio ofZPL to the phonon side band and in the phonon sideband

Fig. 1 Correlation between ultrastructure imaging and hyperspectral imaging. (A) Left: High Angle Annular Dark Field (HAADF) and right: Bright Field(BF) images of a whole cell (N: nucleus) with visible NDs on the cell membrane marked CM. (B) Magnified images (left: BF, right: HAADF) of the areaframed in yellow in (A). The nanodiamond (ND) as well as ultrastructure components such as vesicles (V), mitochondria (M) and nuclear membrane(NM) are clearly distinguished in the uranyl acetate stained cell samples. (C) Schematics of the detection part of the CL-STEM hyperspectral setup.PM: parabolic mirror; spec.: optical spectrometer; CCD: Charge Coupled Device array detector. (D) Examples of H3 (boxed in green) and NV (boxedin red) spectra, extracted from the two NDs shown in (E). The NV spectrum consists of a sharp line (zero-phonon line) followed by a convolutedseries of peaks (phonon lines). (E) HAADF image, and overlapped H3 (green), NV (red) and HAADF images taken on a vesicle framed in red in (A).

Fig. 2 Disentangling NV emission as a function of coating. (A) HAADF image showing vesicles containing NDs. (B) Green and red CL emission mapsextracted from a spectral image of the same area as in (A), and superimposed onto the HAADF image. (C) CL emission spectra of two NV-containingNDs corresponding to the violet and orange delimited areas in (A) and (B), and normalized to their ZPL maximum intensity. The phonon regions ofthe spectra have clearly different amplitudes, as a signature of different coatings (see the text).

Nanoscale Paper

This journal is © The Royal Society of Chemistry 2016 Nanoscale

Publ

ishe

d on

24

May

201

6. D

ownl

oade

d by

Aix

Mar

seill

e U

nive

rsité

on

24/0

5/20

16 1

5:45

:29.

View Article Online

shape too. These differences are visible only after PEI and PAHcoating of the NDs and can be used for segregating red emit-ting NDs in ND-PAH and ND-PEI. To do so, we took advantageof the large redundancy of the data in spectral-imaging toanalyze the data through multivariate analysis (see the ESI†).We obtained 60% identification efficiency for ND-PAH and100% for ND-PEI. The influence of the polycation coating onthe CL spectral shape may be related to the band bendingeffect induced by electrophilic surface functionalization, asreported for the hydrogenated diamond surface.19,20 Bandbending results in a decrease of occupancy of NV in the vicin-ity of the ND surface, depending on the electron affinity of thecoating. The band bending effect on the phonon lines ishowever not documented in the literature and requires moredetailed studies, which is out of scope of the present paper.From now on, we will thus only mention the types of coatings,as deduced from spectral analysis, without referring anymoreto the type of embedded defect centers.

We then investigated the distribution of ND-PEI andND-PAH inside the cells under different incubation conditions:when the cells are incubated (i) with ND-PAH alone, (ii) withND-PEI alone at the same nominal concentration as ND-PAH,and (iii) when the cells are co-incubated with both ND-PAHand ND-PEI simultaneously, each of them at the same concen-tration as under (i) and (ii), so that under condition (iii) thereis twice more NDs than under (i) and (ii). NDs were alwaysfound inside vesicular compartments and never free in thecytoplasm. From the bright field and dark field images, weinferred the size of the vesicles, given as an equivalent dia-meter calculated as the geometric mean of the major andminor axis of its best approximant ellipse, and from the CLspectra we categorized the vesicle labeling depending on thetype of coating of the NDs it contains. The number of NDs pervesicle varied between 1 and 11. We could then build cytometryplots displaying either counts of ND-PEI or ND-PAH per vesicleversus its size in the case of a single type of labeling (Fig. 3A), or

Fig. 3 Cytometric plots of internalization vesicles labeled with polymer coated fluorescent NDs. (A) The number of NDs per vesicle versus thevesicle size (equivalent diameter) in the case of single labeling (control sample) by either ND-PAH (red dots) or ND-PEI (green dots). (B) Double label-ing by internalization of ND-PEI and ND-PAH simultaneously. The number of ND-PEI vs. the number of ND-PAH per vesicle. This plot displays alsovesicles containing a single type of polymer coated ND (counts equal to 0 along the other coating axis). In the co-loaded case, we investigated 5cells in which we detected 296 CL-emitting NDs that were distributed among 88 vesicles. The diameter of the circular markers is proportional to thevesicle size (refer to the scale bar). Different vesicles with the same number of ND-PAH and ND-PEI but different diameters appear as concentriccircles. The occurrence (i.e. the number of vesicles encompassing the same number of NDs whatever its coating is) is given by the color. (C) HAADFand (D) BF STEM image showing, in the same field of view, two vesicles with different morphologies and ND densities: a small and round one con-taining a few NDs (on top), and a larger one, elongated with much more NDs inside.

Paper Nanoscale

Nanoscale This journal is © The Royal Society of Chemistry 2016

Publ

ishe

d on

24

May

201

6. D

ownl

oade

d by

Aix

Mar

seill

e U

nive

rsité

on

24/0

5/20

16 1

5:45

:29.

View Article Online

ND-PEI versus ND-PAH counts (Fig. 3B) in the case of the twotypes of labeling. These plots are similar to the ones of imagecytometry plots built from fluorescence microscopy data. The(discrete) number of fluorescent NDs of each kind is reminis-cent of, but more precise than, the (continuous) luminescentintensity at a given wavelength in conventional cytometryplots. In Fig. 3B the occurrence of a vesicle encompassing agiven (ND-PAH, ND-PEI) count is represented by the color ofthe marker. Fig. 3B thus gives a rapid, comprehensive and syn-thetic view of the data set.

In this particular proof of principle where we used twodifferent labels (Fig. 3B), several straightforward observationscan be drawn. First, the pure vesicles (single type of label) areover-represented compared to mixed vesicles (labeled with twotypes of cationic NDs). Indeed, over the 88 investigated vesi-cles, 31 are filled only with PEI and 23 only with PAH. The factthat most of the internalization seems to occur with only PEIor only PAH coating is consistent with the differential internal-ization pathways of ND-PAH and ND-PEI that we have shownin previous work,17 although here the NDs are larger (150 nminstead of 50 nm). Indeed, we had observed that PAH-coatedparticles are mainly internalized by clathrin mediated endo-cytosis while the PEI-coated ones are able to trigger anadditional pathway, the macropinocytosis. Second, the contentof mixed vesicles distributes evenly in the (number of PEI,number of PAH) space. Third, the vesicles containing a smallquantity of fluorescent NDs are over-represented (about 65%with a number smaller or equal to 3 fluorescent NDs). A STEMimage of a typical vesicle of this category is shown on the topof Fig. 3C and D, sparsely filled with NDs. It corresponds inthis case to an endosome. Fourth, the general tendency is ofcourse an increase of the diameter as the number ofembedded NDs increases, as expected from simple steric con-sideration. However, the dispersion in vesicle sizes is very largefor the same or similar (PEI, PAH) counts. An example of alarge vesicle containing densely packed NDs is shown at thebottom of STEM images Fig. 3C and D. This vesicle is mostlikely a macropinosome, considering its size, elongated form,and not well-defined delimiting membrane. Here, ultrastruc-ture images are invaluable to identify the vesicle filling.

Recently a new fluorescence microscopy based techniquewas introduced to quantify the macropinocytotic index ofcancer cells. It relies on labeling macropinosomes throughinternalization of fluorescently labeled high molecular weightdextran molecules.21 This approach provides a better selecti-vity than previous fluorescence microscopy image-based tech-niques. Our method could be applied to the measurement ofsuch a macropinocytotic index with the additional advantagesof combining both luminescence microscopy (with someselectivity brought by the polymer coating) and ultrastructuralmorphology identification.

Let us point out that the minimal vesicle size of ≈300 nmthat we observed is of the order of the optical diffraction limit,and can therefore be resolved by conventional fluorescencemicroscopy. Moreover, fluorescent NDs are known to be ideallabels for STED super-resolution microscopy22 because they

can sustain the high power of the stimulated emission laserwithout any degradation, with a resistance even larger than theone to the electron beam of CL-STEM. Conventional or STEDmicroscopy could be used to realize the fluorescence image-based cytometry of internalization vesicles. However, fromFig. 1–3 one can see that the ND labels usually do not fill thewhole vesicle, so that its size and morphology cannot beextracted from ND fluorescence in a reliable manner, while itcan be inferred from a STEM scan with nanometer precision.Therefore, despite a simpler implementation of image cyto-metry based solely on fluorescence microscopy, such anapproach does not offer a precise quantification of the vesiclesize, neither of its detailed morphology including the innercomposition.

Conclusion

We have demonstrated the potential of CL-STEM for combinedluminescence and cell ultrastructure imaging, and haveapplied this technique to multicolor image cytometry of vesi-cles. Thanks to the diversity of emission spectra of CL-nano-probes,23 CL-STEM should have a higher specificity than goldnanobeads of different sizes conventionally used in EMimmunocytochemistry when the probes are multiplexed.Though this is an advantage, the materials currently availablehave a few drawbacks. As in our case, a simple straightforwardbicolor imaging was complicated by the presence of red emit-ters in the batch of green emitters. This is an issue related tothe purification of samples and with efforts in this direction,pure emitters are not far away. As a proof of principle appli-cation we have considered vesicular localization of two types offluorescent nanodiamonds with distinct CL spectral signaturesafter internalization. The high resolution of STEM allowsacquiring images of the cellular ultrastructure both in BF andHAADF contrasts from which we can infer various information,and in particular the cross-sectional size of vesicles. By over-lapping pixel-by-pixel, the STEM images with the CL emissionsof internalized NDs, we could classify the vesicles dependingon the ND spectral labeling. The different cationic polymercoatings (PEI or PAH) of the NDs could be used in the futurein order to direct each type of ND in different vesicles. Indeed,as demonstrated in a previous study,17 ND-PAH are found onlyin clathrin-related endosomes, and therefore should labelthese types of endosomes in CL red color. In contrast, ND-PEIwere found in both clathrin-mediated endosomes and inmacropinosomes. In order to improve the specificity of func-tionalized NDs and direct them in macropinosomes, PEI couldbe replaced by dextran.21

More generally, using antibody-functionalized NDs24 com-bined with microinjection, it should be possible to label otherintracellular organelles like mitochondria and then proceedwith a detailed CL-STEM cytometry in relationship with theproblem of interest (e.g. study of local temperature fluctuationsmeasured using a ND-based thermometer25). Also, other typesof CL nanolabels, such as II–VI luminescent semiconductor

Nanoscale Paper

This journal is © The Royal Society of Chemistry 2016 Nanoscale

Publ

ishe

d on

24

May

201

6. D

ownl

oade

d by

Aix

Mar

seill

e U

nive

rsité

on

24/0

5/20

16 1

5:45

:29.

View Article Online

nanocrystals (quantum dots) could be used, as they have alsobeen demonstrated to be efficient cathodoluminescence emit-ters.13 Finally, with the emergence of commercial STEM-CLinstruments, image acquisition and analysis automationimproving the statistics, we foresee that the CL-STEM novelimage cytometry technique will find numerous applications insubcellular level cell biology.

MethodsFluorescent ND preparation

Both types of NDs were radiation-damaged with a high-energy(3 MeV) proton beam, as detailed by Su et al.26 Briefly, a thindiamond film (thickness <50 µm) was prepared by depositing≈5 mg of ND powder on a silicon wafer (1 × 1 cm2 size) andsubsequently subjecting it to ion irradiation at a dose of≈2 × 1016 H+ per cm2. Afterwards, the radiation-damaged NDswere annealed at 800 °C for 2 h to form fluorescent NDs. Toremove graphitic carbon atoms on the surface, the freshly pre-pared fluorescent NDs were oxidized in air at 490 °C for 2 hand microwave-cleaned in concentrated H2SO4–HNO3 (3 : 1,v/v) at 100 °C for 3 h. NDs coating by PAH or PEI are per-formed as described in ref. 16 and 17.

Cell samples preparation

The A673 human Ewing sarcoma cell line was used as themodel system. These cells were cultured in DMEM with 10%FBS and 1% PS. Once the cells reached 70% confluence theywere trypsinized and seeded on cover slips placed on a 6-wellplate. About 105 cells were plated and allowed to adhere over-night. These cells were then incubated with ND-PEI andND-PAH (100 µL each in 500 µL of medium) for 4 hours.Control cells were treated with only ND-PEI or ND-PAH for4 hours. The co-loaded cells and the control cells were rinsedin 1× Phosphate Buffered Saline (PBS) to remove nanoparticlesin excess. Cells were fixed using 2% glutaraldehyde in 0.1 Mcacodylate buffer, pH 7.4 for 1 h at room temperature and thenpost-fixed for 1 hour at room temperature with 1% osmiumtetroxide and 1% potassium ferrocyanide in cacodylate buffer.They were dehydrated and finally embedded in epoxy resin.The monolayer was sectioned to 200–300 nm thin sections onan ultra-microtome. The sections were exposed to chloroformvapors to stretch them and reduce wrinkles. They were thenloaded on carbon/collodion coated copper finder grids. Forthe experiments of Fig. 1, the whole procedure was the sameexcept for an additional final step of uranyl acetate stainingused to increase the image contrast.

Author contributions

S. N., J. R. B, C. D., E. L. C., F. T., and M. K. conceived anddesigned the experiments; S. N., L. H. G. T., J. R. B., C. D., andM. K. performed the experiments; S. N., C. D., E. L. C., F. T. and

M. K. analyzed the data; C. Y. F. and H. C. C. contributedmaterials/analysis tools; S. N., F. T. and M. K. co-wrote the paper.

Conflict of interest

The authors declare no competing financial interests.

Acknowledgements

This work is supported by a public grant overseen by theFrench National Research Agency (ANR) as part of the “Inves-tissements d’Avenir” program (reference: ANR-10-LABX-0035,Labex NanoSaclay). The research leading to these results hasreceived funding from the European Union Seventh Frame-work Programme [No. FP7/2007–2013] under Grant AgreementNo. n312483 (ESTEEM2).

References

1 S. W. Hell, Angew. Chem., Int. Ed., 2015, 54, 8054–8066.2 W. E. Moerner, Angew. Chem., Int. Ed., 2015, 54, 8067–

8093.3 E. Betzig, Angew. Chem., Int. Ed., 2015, 54, 8034–8053.4 D. Li, L. Shao, B.-C. Chen, X. Zhang, M. Zhang, B. Moses,

D. E. Milkie, J. R. Beach, J. A. Hammer, M. Pasham,T. Kirchhausen, M. A. Baird, M. W. Davidson, P. Xu andE. Betzig, Science, 2015, 349, aab3500.

5 P. de Boer, J. P. Hoogenboom and B. N. G. Giepmans, Nat.Methods, 2015, 12, 503–513.

6 X. Shu, V. Lev-Ram, T. J. Deerinck, Y. Qi, E. B. Ramko,M. W. Davidson, Y. Jin, M. H. Ellisman and R. Y. Tsien,PLoS Biol., 2011, 9, e1001041.

7 K. Cortese, A. Diaspro and C. Tacchetti, J. Histochem. Cyto-chem., 2009, 57, 1103–1112.

8 F. J. Timmermans and C. Otto, Rev. Sci. Instrum., 2015, 86,011501.

9 N. Liv, A. C. Zonnevylle, A. C. Narvaez, A. P. J. Effting,P. W. Voorneveld, M. S. Lucas, J. C. Hardwick, R.A. Wepf, P. Kruit and J. P. Hoogenboom, PLoS One, 2013,8, e55707.

10 M. Kociak, O. Stéphan, A. Gloter, L. F. Zagonel,L. H. G. Tizei, M. Tencé, K. March, J. D. Blazit, Z. Mahfoud,A. Losquin, S. Meuret and C. Colliex, C. R. Phys., 2014, 15,158–175.

11 E. Kimura, T. Sekiguchi, H. Oikawa, J. Niitsuma,Y. Nakayama, H. Suzuki, M. Kimura, K. Fujii and T. Ushiki,Arch. Histol. Cytol., 2004, 67, 263–270.

12 H. Niioka, T. Furukawa, M. Ichimiya, M. Ashida,T. Araki and M. Hashimoto, Appl. Phys. Express, 2011,4, 112402.

13 T. Furukawa, H. Niioka, M. Ichimiya, T. Nagata,M. Ashida, T. Araki and M. Hashimoto, Opt. Express,2013, 21, 25655.

Paper Nanoscale

Nanoscale This journal is © The Royal Society of Chemistry 2016

Publ

ishe

d on

24

May

201

6. D

ownl

oade

d by

Aix

Mar

seill

e U

nive

rsité

on

24/0

5/20

16 1

5:45

:29.

View Article Online

14 Z. Mahfoud, A. T. Dijksman, C. Javaux, P. Bassoul,A.-L. Baudrion, J. Plain, B. Dubertret and M. Kociak,J. Phys. Chem. Lett., 2013, 4, 4090–4094.

15 L. H. G. Tizei and M. Kociak, Nanotechnology, 2012, 23,175702.

16 A. Alhaddad, M.-P. Adam, J. Botsoa, G. Dantelle,S. Perruchas, T. Gacoin, C. Mansuy, S. Lavielle, C. Malvy,F. Treussart and J.-R. Bertrand, Small, 2011, 7, 3087–3095.

17 A. Alhaddad, C. Durieu, G. Dantelle, E. Le Cam,C. Malvy, F. Treussart and J.-R. Bertrand, PLoS One,2012, 7, e52207.

18 Y. Mita, Phys. Rev. B: Condens. Matter, 1996, 53, 11360.19 M. V. Hauf, B. Grotz, B. Naydenov, M. Dankerl, S. Pezzagna,

J. Meijer, F. Jelezko, J. Wrachtrup, M. Stutzmann, F. Reinhardand J. A. Garrido, Phys. Rev. B: Condens. Matter, 2011, 83, 1–4.

20 V. Petrakova, I. Rehor, J. Stursa, M. Ledvina, M. Nesladekand P. Cigler, Nanoscale, 2015, 7, 12307–12311.

21 C. Commisso, R. J. Flinn and D. Bar-Sagi, Nat. Protoc.,2014, 9, 182–192.

22 S. Arroyo-Camejo, M.-P. Adam, M. Besbes, J.-P. Hugonin,V. Jacques, J.-J. Greffet, J.-F. Roch, S. W. Hell andF. Treussart, ACS Nano, 2013, 7, 10912–10919.

23 H. Zhang, I. Aharonovich, D. R. Glenn, R. Schalek,A. P. Magyar, J. W. Lichtman, E. L. Hu and R. L. Walsworth,Small, 2014, 10, 1908–1913.

24 B.-M. Chang, H.-H. Lin, L.-J. Su, W.-D. Lin, R.-J. Lin,Y.-K. Tzeng, R. T. Lee, Y. C. Lee, A. L. Yu and H.-C. Chang,Adv. Funct. Mater., 2013, 23, 5737–5745.

25 G. Kucsko, P. C. Maurer, N. Y. Yao, M. Kubo, H. J. Noh,P. K. Lo, H. Park and M. D. Lukin, Nature, 2013, 500,54–58.

26 L.-J. Su, C.-Y. Fang, Y.-T. Chang, K.-M. Chen, Y.-C. Yu,J.-H. Hsu and H.-C. Chang, Nanotechnology, 2013, 24,315702.

Nanoscale Paper

This journal is © The Royal Society of Chemistry 2016 Nanoscale

Publ

ishe

d on

24

May

201

6. D

ownl

oade

d by

Aix

Mar

seill

e U

nive

rsité

on

24/0

5/20

16 1

5:45

:29.

View Article Online