Quantum Dot Triexciton Imaging with Three-Dimensional

Nano Lett. , 0, (),. DOI: 10.1021/nl9012387@proofing. Copyright American Chemical Society. * Corresponding Authors: M.H., phone +49-521-106-5442, fax ...
9 downloads 3 Views 2MB Size
NANO LETTERS

Quantum Dot Triexciton Imaging with Three-Dimensional Subdiffraction Resolution

2009 Vol. 9, No. 6 2466-2470

Simon Hennig, Sebastian van de Linde, Mike Heilemann,* and Markus Sauer* Applied Laser Physics and Laser Spectroscopy and Bielefeld Institute for Biophysics and Nanoscience, Bielefeld UniVersity, UniVersita¨tsstrasse 25, 33615 Bielefeld, Germany Received April 18, 2009; Revised Manuscript Received May 8, 2009

ABSTRACT We describe a simple method that improves optical resolution in fluorescence microscopy approximately 1.7-fold in all three dimensions and can be implemented on any basic confocal scanning microscope. This approach is based on three-photon absorption of commercially available quantum dots generating a triple exciton (triexciton) and subsequent blue-shifted fluorescence emission following recombination of the triexciton. As a pure physical approach, the resolution enhancement is independent from the nanoenvironment and demonstrated to work in living cells.

Light microscopy enables the noninvasive investigation of optically transparent material such as cells under a variety of conditions. Through the development of efficient fluorescent labels that can be covalently attached to antibodies and other biomarkers and the associated possibility to specifically label cellular structures, the expansion of fluorescence microscopy in laboratory applications and research has been substantially accelerated. However, the ability to spatially resolve a structure has a physical limit which is caused by the wave nature of light.1 Therefore, any lensbased microscope cannot resolve objects that are closer than the order of half the wavelength of light, i.e., for visible light in the range of ∼200 nm. A variety of methods to realize high-resolution fluorescence microscopy has been introduced in the recent past.2-10 These approaches can be discriminated by the underlying strategy that is used to achieve subdiffraction resolution in fluorescence imaging. Prominent examples are stimulated emission depletion (STED),2 structured illumination methods,7 or methods that use reversible transitions of fluorophores between a fluorescent and a nonfluorescent (“dark”) state, such as molecular photoswitches.11,6,12,13 However, subdiffraction imaging with standard microscopes and fluorescent probes remains challenging. Most methods require tailored laser systems and optical components that are not easily assembled especially by nonexperts. These problems seriously limit the routine use of novel subdiffraction imaging methods in research areas that could profit enormously. This undermines the quest for new and simpler methods that are * Corresponding Authors: M.H., phone +49-521-106-5442, fax +49-521106-2958, e-mail [email protected]; M.S., phone +49-521106-5451, fax +49-521-106-2958, e-mail [email protected]. 10.1021/nl9012387 CCC: $40.75 Published on Web 05/19/2009

 2009 American Chemical Society

compatible with very low need in instrumentation and are available in practically all biological or medical laboratories working with fluorescence microscopy. Here, we introduce quantum dot triexciton imaging (QDTI) a comparably simple and novel 3D subdiffraction imaging method that uses three-photon absorption in single QDs to generate triple excitonic states (triexciton states) whose emission can be selectively detected at shorter wavelengths.14,15 QDs are crystals of a few nanometers in diameter with an energy band gap directly related to their size. Due to their spectral tunability, high fluorescence brightness, and photostability they attracted attention as substitutes for organic fluorophores.16,15 Especially, the unique spectral properties of QDs suggests them as ideal probes for spectral multiplexing applications.17 Specific surface functionalization with biomolecules such as antibodies or proteins paved the way for QD applications in biological imaging, such as extracellular targeting and tracking.14,18 Even internalization of QDs into living cells has been successfully achieved.19,20 The underlying photophysical processes of QD fluorescence are described as photon absorption, subsequent formation of an electron-hole pair (exciton), and fluorescence emission as the result of exciton recombination.16 It is well-known that multiexcitonic states can be generated in QDs,21,22 i.e., states that are characterized by the existence of more than one electron-hole pair in one QD. As a result of various recombination pathways each of the multiexcitonic states exhibits different fluorescence emission properties. The generation of multiple excitons in a single QD can be realized by multiphoton absorption. Combined with fluorescence microscopy, multiexciton fluorescence imaging of QDs is thus comparable to other methods that are based

Figure 1. (A) Triexciton fluorescence emission is centered around 590 nm and thus well separated from the monoexciton emission centered at 655 nm. (B) Configuration of a confocal microscopic system as experimental setup for QDTI with subdiffraction resolution. QDs are excited at 445 nm with a pulsed laser diode (5 MHz repetition rate) and exciton emission is detected on two spectrally separated APDs. (C) Fluorescence images of individual QDot655 adsorbed on dry glass surface are recorded subsequently (scale bar 1 µm). The mono- and biexciton image is generated upon excitation with a relatively low intensity of 20 W/cm2 to avoid saturation effects. The triexciton image is measured by applying an excitation intensity of 200 W/cm2. The surface was scanned from top to bottom and from left to right with a resolution of 50 nm/pixel (1 ms integration time per pixel). Mono- and biexciton emission is recorded between 650 and 700 nm, whereas triexciton emission is selectively detected between 550 and 600 nm. (D) Fluorescence intensity profiles of single scan lines demonstrate a reduced fwhm of the PSFs of single QDs measured by QDTI. 50 nm, 1 ms/pixel.

on the absorption of two23 or three24 photons. The main advantage of multiphoton microscopy is reduced out-of focus bleaching, together with an increase in penetration depth. However, one has to be aware of the fact that the gain in lateral resolution is canceled out by the long wavelength of the infrared laser source used for efficient multiphoton excitation.25 In contrast, the generation of multiple excitons in QDs can be realized with visible light, such that, in addition to axial resolution, an increase in lateral resolution can be anticipated. Triexcitonic states have been characterized to be spectrally distinct from energetically closer monoand biexcitonic states in CdSe QDs and are commonly observed at higher energies.26,21,22,27,28 The blue-shifted triexciton emission can be easily spectrally separated from mono- and biexciton emission22 (Figure 1A). Therefore, QDTI can be used advantageously in combination with standard laser scanning microscopes for 3D imaging with subdiffraction optical resolution. Previous theoretical studies enlightened the various recombination pathways of multiexcitons in CdSe QDs and described that emission of a photon following triexciton recombination depends on the size of the nanocrystal.28 Following these findings, we studied commercially available CdSe QDs of different diameter and selected QDot655 as efficient triexciton source for QDTI. The fluorescence emission of the triexciton recombination is expected around 590 nm from theoretical calculations28 and was confirmed in experiments.22 As the generation of a triexcitonic state requires the successive absorption of three photons, we expect a point spread function (PSF) that is reduced by a factor of 3 in lateral and axial directions.24 In order to generate multiple excitonic states in single QDot655, we used an excitation wavelength of 445 nm provided by a laser diode operated with a repetition rate of 5 MHz and a custom-built confocal microscope that spectrally separates the emission signal on two detectors (Figure 1B),29 Nano Lett., Vol. 9, No. 6, 2009

recording the mono- and biexciton emission of single QDot655 between 650 and 700 nm (the signal is dominated by monoexciton emission) and the triexciton emission selectively between 550 and 600 nm. To prevent saturation effects mono- and biexciton imaging was performed at relatively low excitation intensities in the range of 10-100 W/cm2. On the other hand, triexciton imaging required the application of higher excitation intensities and was performed at 0.1-1 kW/cm2. Scan images of QDot655 on a dry glass surface (Figure 1C) clearly demonstrate that the PSF on the triexciton channel is smaller, which is corroborated by the cross sectional profile of a scan line shown in Figure 1D. Statistical analysis of the full width half-maximum (fwhm) of the PSF obtained from mono- and triexciton fluorescence emission of single QDot655 was made by fitting the emission profiles to Gaussian functions. As a result, the lateral resolution improved 1.6-fold for QDTI (see Figure 1 in Supporting Information). Three-dimensional confocal scanning of single QDot655 embedded in a poly(vinyl alcohol) (PVA) matrix was performed to obtain the 3D profile of the PSF. Here, an axial resolution improvement of 1.7-fold was determined (see Figure 1 in Supporting Information). These findings demonstrate that unlike in multiphoton microscopy, QDTI exhibits a substantial gain in axial and lateral direction and thus opens novel and facile possibilities for 3D subdiffraction fluorescence imaging. To prove the applicability of our approach for biological imaging, we applied QDTI for cellular imaging. COS-7 cells were fixed and stained with primary antibodies directed against β-tubulin and secondary antibodies labeled with QDot655. Confocal scan images from monoexciton (Figure 2A) and triexciton emission (Figure 2B) measured under different excitation conditions reveal a higher resolution of the triexciton image. A direct comparison of magnified image areas (Figures 2C,D) shows that the fluorescence image 2467

Figure 2. Immunofluorescence images of the microtubulin network of COS-7 cells stained with QDot655 labeled secondary antibodies, recorded with the confocal microscope depicted in Figure 1. with a resolution of 50 nm/pixel. (A) Mono- and biexciton image recorded with 20 W/cm2 excitation intensity and (B) triexciton image measured on the short-wavelength detector between 550 and 600 nm applying an excitation intensity of 200 W/cm2 (scale bar 5 µm; 1 ms integration time per pixel). Two boxed regions are shown in a magnified view in (C) and (D), together with line profiles for the monoexciton (red) and triexciton (green) emission channel.

recorded at the triexciton channel resolves structures that are blurred in the regular monoexciton channel. An overlay of both line profiles measured in the mono- and triexciton emission channel clearly demonstrates the potential of QDTI for cellular subdiffraction imaging. An important advantage of QDTI results from the long lifetime of mono- and biexciton states of quantum dots. This fact is in contrast to multiphoton microscopy, where the excited state is populated through consecutive absorption of two or more photons at very short time scales, such that expensive and complex short-pulsed laser sources are necessary. The available time to generate triexciton states in QDs is much longer, and as a consequence, any continuous-wave laser source providing the appropriate wavelength can be used for the required three-photon absorption. To demonstrate the simplicity of QDTI, we used a commercial confocal line scanning microscope (LSM 710), equipped with a 2468

Figure 3. Three-dimensional immunofluorescence QDTI performed on a standard confocal laser scanning microscope (LSM 710; Zeiss) with continuous wave excitation at 488 nm. The microtubulin network of a fixed COS-7 cell was stained with QDot655 labeled antibodies. (A) Monoexciton channel and (B) triexciton channel recorded using a higher detector gain (size of image box, 67 µm × 67 µm × 13 µm; the color code of the images represents the z-coordinate). The two-dimensional cross sections (scale bar 10 µm) shown in (C) for monoexciton and (D) triexciton emission demonstrate the improved axial resolution in the triexciton channel.

continuous wave argon ion laser with multiple emission lines and photomultiplier tubes to image the microtubular network in COS-7 cells (Figure 3). The spectral range of the two detector channels was defined identically as in the previous experiments (650-700 nm for the mono- and biexciton channel whereas 550-600 nm for the triexciton channel), and a whole-cell three-dimensional image of the microtubulin network of a fixed COS7 cell was recorded upon excitation at 488 nm using a reduced detector gain on the monoexciton channel (Figure 3, panels A and B). The cross sectional images (Figure 3, panels C and D) show substantially improved axial resolution and demonstrate the potential of QDTI for optical sectioning. Nano Lett., Vol. 9, No. 6, 2009

Figure 4. In vivo QD imaging of QDot655 internalized into live COS-7 cells. As previously already observed30,20 most QDs accumulate at the nuclear membrane. Two-dimensional images of the (A) monoexciton and (B) triexciton channels (scale bar 5 µm) and three-dimensional images of monoexciton (C, E) and triexciton emissions (D, F) (box size 42 µm × 42 µm × 14 µm). The magnified 3D images (E, F) demonstrate QD clustering near the nucleus and clarify the superior 3D resolution of QDTI.

Here it has to be pointed out that QDTI is a pure physical way to reduce the fwhm of the PSF through three-photon absorption, and as such, similar to other multiphoton fluorescence imaging methods, QDTI does not depend on the nanoenvironment of the fluorescent probe. This further broadens the spectrum of possible applications, such as live cell imaging, which is demonstrated by internalizing streptavidin-coated QDot655 into living COS-7 cells (Figure 4). Monoexciton and triexciton images were recorded on a standard LSM (Figure 4, panels A and B) and show the localization of QDs in the cytoplasm and around the nuclear membrane. Comparing the two emission channels, a similar reduction of the fwhm of the PSFs in the triexciton channel (Figure 4B) is observed as for QDot655 on a dry glass surface (Figure 1C). The increased axial resolution is clearly visible by the reduced out of focus signals observed in the triexciton channel (Figure 4B). The 3D images of QDot655 internalized into a COS-7 cell demonstrate the appearance of QD clusters inside the living cell centered around the Nano Lett., Vol. 9, No. 6, 2009

nucleus (Figure 4, panels C and D), a phenomenon that has been described previously.30,20 The higher resolution of the triexciton image (Figure 4D) is clearly visible and further clarified by the expanded images of the monexciton (Figure 4E) and triexciton channels (Figure 4F). Our results demonstrate that QDTI can almost double the resolution of conventional microscopy in three dimensions. The method is simple and can be realized on any confocal microscope system equipped with a continuous-wave laser light source providing appropriate excitation wavelengths, with low excitation intensities. Three-photon absorption by single QDot655 quantum dots does not depend on the environment of the nanocrystals, and no special buffer conditions are required. As a consequence, the method is widely applicable, in particular for biological imaging in living cells. Importantly, QDTI is not limited to higher image resolution in both lateral and axial directions. Single-particle tracking with a localization accuracy that is proportional to the standard deviation of the PSF and the number of emitted 2469

photons31 is also enhanced by the theoretical factor of 3 in all three dimensions. Although the resolution is less than that afforded by other techniques such as various localization microscopy methods11,6 and STED,2,9 QDTI is currently the only sub-diffraction-resolution imaging technique that can provide images of whole fixed and living cells with enhanced resolution in both axial and lateral directions using standard fluorescence microscopes available in almost any lab. Acknowledgment. This work was supported by the Biophotonics and the Systems Biology Initiative (FORSYS) of the German Ministry of Research and Education (BMBF, Grants 13N9234 and 0315262). Supporting Information Available: A detailed description of the microscopic setup, experimental techniques, and sample preparation. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11)

2470

Abbe, E. Arch. Mikrosk. Anat. 1873, 9, 413–420. Hell, S. W.; Wichmann, J. Opt. Lett. 1994, 19, 780–783. Hell, S. W. Science 2007, 316 (5828), 1153–8. Bossi, M.; Folling, J.; Belov, V. N.; Boyarskiy, V. P.; Medda, R.; Egner, A.; Eggeling, C.; Schonle, A.; Hell, S. W. Nano Lett. 2008, 8 (8), 2463–8. Gustafsson, M. G. Nat. Methods 2008, 5 (5), 385–7. Huang, B.; Wang, W.; Bates, M.; Zhuang, X. Science 2008, 319 (5864), 810–3. Schermelleh, L.; Carlton, P. M.; Haase, S.; Shao, L.; Winoto, L.; Kner, P.; Burke, B.; Cardoso, M. C.; Agard, D. A.; Gustafsson, M. G.; Leonhardt, H.; Sedat, J. W. Science 2008, 320 (5881), 1332–6. Shroff, H.; Galbraith, C. G.; Galbraith, J. A.; Betzig, E. Nat. Methods 2008, 5 (5), 417–23. Westphal, V.; Rizzoli, S. O.; Lauterbach, M. A.; Kamin, D.; Jahn, R.; Hell, S. W. Science 2008, 320 (5873), 246–9. Hell, S. W. Nat. Methods 2009, 6 (1), 24–32. Heilemann, M.; van de Linde, S.; Schuttpelz, M.; Kasper, R.; Seefeldt, B.; Mukherjee, A.; Tinnefeld, P.; Sauer, M. Angew. Chem., Int. Ed. 2008, 47 (33), 6172–6.

(12) Irvine, S. E.; Staudt, T.; Rittweger, E.; Engelhardt, J.; Hell, S. W. Angew. Chem., Int. Ed. 2008, 47 (14), 2685–2688. (13) Heilemann, M.; Dedecker, P.; Hofkens, J.; Sauer, M. Laser Photonics ReV. 2009, 3, 180–202. (14) Medintz, I. L.; Uyeda, H. T.; Goldman, E. R.; Mattoussi, H. Nat. Mater. 2005, 4 (6), 435–446. (15) Michalet, X.; Pinaud, F. F.; Bentolila, L. A.; Tsay, J. M.; Doose, S.; Li, J. J.; Sundaresan, G.; Wu, A. M.; Gambhir, S. S.; Weiss, S. Science 2005, 307 (5709), 538–44. (16) Alivisatos, A. P. Science 1996, 271 (5251), 933–937. (17) Goldman, E. R.; Clapp, A. R.; Anderson, G. P.; Uyeda, H. T.; Mauro, J. M.; Medintz, I. L.; Mattoussi, H. Anal. Chem. 2004, 76 (3), 684– 688. (18) Xing, Y.; Chaudry, Q.; Shen, C.; Kong, K. Y.; Zhau, H. E.; WChung, L.; Petros, J. A.; O’Regan, R. M.; Yezhelyev, M. V.; Simons, J. W.; Wang, M. D.; Nie, S. Nat. Protocols 2007, 2 (5), 1152–1165. (19) Chen, F.; Gerion, D. Nano Lett. 2004, 4 (10), 1827–1832. (20) Parak, W. J.; Pellegrino, T.; Plank, C. Nanotechnology 2005, 16 (2), R9-R25. (21) Fisher, B.; Caruge, J. M.; Chan, Y. T.; Halpert, J.; Bawendi, M. G. Chem. Phys. 2005, 318 (1-2), 71–81. (22) Fisher, B.; Caruge, J. M.; Zehnder, D.; Bawendi, M. Phys. ReV. Lett. 2005, 94 (8), 087403. (23) Denk, W.; Strickler, J. H.; Webb, W. W. Science 1990, 248 (4951), 73–6. (24) Sheppard, C. J. R. Bioimaging 1996, 4, 124–128. (25) Zipfel, W. R.; Williams, R. M.; Webb, W. W. Nat. Biotechnol. 2003, 21 (11), 1368–1376. (26) Bonati, C.; Mohamed, M. B.; Tonti, D.; Zgrablic, G.; Haacke, S.; van Mourik, F.; Chergui, M. Phys. ReV. B 2005, 71 (20), 205317. (27) Oron, D.; Kazes, M.; Shweky, I.; Banin, U. Phys. ReV. B 2006, 74 (11), 115333. (28) Franceschetti, A.; Troparevsky, M. C. J. Phys. Chem. C 2007, 111 (17), 6154–6157. (29) Heilemann, M.; Tinnefeld, P.; Sanchez Mosteiro, G.; Garcia Parajo, M.; Van Hulst, N. F.; Sauer, M. J. Am. Chem. Soc. 2004, 126 (21), 6514–5. (30) Parak, W. J.; Boudreau, R.; Le Gros, M.; Gerion, D.; Zanchet, D.; Micheel, C. M.; Williams, S. C.; Alivisatos, A. P.; Larabell, C. AdV. Mater. 2002, 14 (12), 882–885. (31) Thompson, R. E.; Larson, D. R.; Webb, W. W. Biophys. J. 2002, 82 (5), 2775–83.

NL9012387

Nano Lett., Vol. 9, No. 6, 2009