OSA's Digital Library

Optics Express

Optics Express

  • Editor: Andrew M. Weiner
  • Vol. 22, Iss. 6 — Mar. 24, 2014
  • pp: 7028–7039
« Show journal navigation

Simultaneous dual-color 3D STED microscopy

Christian Osseforth, Jeffrey R. Moffitt, Lothar Schermelleh, and Jens Michaelis  »View Author Affiliations


Optics Express, Vol. 22, Issue 6, pp. 7028-7039 (2014)
http://dx.doi.org/10.1364/OE.22.007028


View Full Text Article

Acrobat PDF (3868 KB)





Browse Journals / Lookup Meetings

Browse by Journal and Year


   


Lookup Conference Papers

Close Browse Journals / Lookup Meetings

Article Tools

Share
Citations

Abstract

We describe the design and implementation of a stimulated emission depletion (STED) microscope which allows simultaneous three-dimensional super-resolution imaging in two colors. A super-continuum laser source is used to provide all spectral bands necessary for excitation and efficient depletion to achieve a lateral and axial resolution of ~35 nm and ~90 nm respectively. We characterize the systems' performance by imaging colloidal particles and single fluorescent molecules. Its biological applicability is demonstrated by dual-color imaging of nuclear pore complexes and of DNA replication sites in mammalian cells.

© 2014 Optical Society of America

1. Introduction

Considerable improvements in the field of super-resolution fluorescence microscopy have been achieved over the last years [1

1. L. Schermelleh, R. Heintzmann, and H. Leonhardt, “A guide to super-resolution fluorescence microscopy,” J. Cell Biol. 190(2), 165–175 (2010). [CrossRef] [PubMed]

,2

2. C. Coltharp and J. Xiao, “Superresolution microscopy for microbiology,” Cell. Microbiol. 14(12), 1808–1818 (2012). [CrossRef] [PubMed]

]. Arguably the most important area for the application of the various approaches to overcome the diffraction limit is found in the field of life sciences. However the application of super-resolution fluorescence imaging to biological systems poses a number of experimental challenges since structures can be altered by the light used for the imaging and by the labeling procedure itself. Moreover, not only the high-resolution structure of a certain entity is of importance but also the relative spatial arrangements of structures formed by different entities are generally of interest. Therefore, multi-color super-resolution imaging is necessary. Furthermore revealing three-dimensional structures of biological complexes also requires super-resolution microscopy in all three dimensions.

Stimulated emission depletion (STED) microscopy was the first super-resolution technique to overcome the diffraction limit in the far-field [3

3. S. W. Hell and J. Wichmann, “Breaking the diffraction resolution limit by stimulated emission: stimulated-emission-depletion fluorescence microscopy,” Opt. Lett. 19(11), 780–782 (1994). [CrossRef] [PubMed]

]. In its most common implementation a confocal laser scanning microscope is equipped with an additional depletion beam. This beam depletes the excited state of fluorophores via stimulated emission before spontaneous emission takes place. Due to the toroidal, “donut”-like intensity profile of the depletion beam only fluorophores in the central area of the excitation beam, where the intensity of the depletion beam is negligible are left to fluoresce spontaneously. The depletion efficiency is dependent on the intensity of the depleting beam and therefore increasing its intensity will yield decreasingly smaller effective point spread functions (PSF). The full-width-at-half-maximum (FWHM) of the effective PSF can then be approximated by: λ/(2NA1+I/Is) where I is the intensity of the depletion beam and Is is the saturation intensity at which 50% of the fluorophores are emitting via stimulated emission [4

4. D. Wildanger, E. Rittweger, L. Kastrup, and S. W. Hell, “STED microscopy with a supercontinuum laser source,” Opt. Express 16(13), 9614–9621 (2008). [CrossRef] [PubMed]

]. While the use of a donut beam increases the lateral resolution, the axial extend of the effective PSF is not affected, thus leading to a considerable disparity of lateral resolution (~20-40 nm) and axial resolution (~900 nm). This might not present a disadvantage when imaging a correctly oriented, inherently two-dimensional sample. However, when imaging three-dimensional biological samples that extend along the optical axis, the poor axial resolution will drastically compromise image quality. Dual-color 3D STED has been implemented previously using a 4pi/STED approach (isoSTED [5

5. R. Schmidt, C. A. Wurm, S. Jakobs, J. Engelhardt, A. Egner, and S. W. Hell, “Spherical nanosized focal spot unravels the interior of cells,” Nat. Methods 5(6), 539–544 (2008). [CrossRef] [PubMed]

],). This design offers the highest 3D resolution (40 - 45 nm in 3D); however, it is very elaborate to implement due to the required coherent superposition of all beams. An alternative approach has been introduced for improving axial resolution using circular phase masks [6

6. D. Wildanger, R. Medda, L. Kastrup, and S. W. Hell, “A compact STED microscope providing 3D nanoscale resolution,” J. Microsc. 236(1), 35–43 (2009). [CrossRef] [PubMed]

]. This method is far simpler than isoSTED and is readily adopted with existing STED microscopes. Here we use the circular phase mask approach to create a dual-color, table-top 3D STED microscope that is compact, easy to align and maintain, and straightforward to construct.

When expanding a one-color STED microscope to a dual-color setup an important peculiarity has to be taken into account. In a conventional dual-color confocal fluorescence microscope one would spectrally separate the two detection channels as much as possible in order to minimize spectral crosstalk. In the case of a STED microscope this causes additional problems. Special care has to be taken to prevent overlap of the depletion band of the short wavelength channel (channel 1) with the absorption spectrum of the dye used in the second, higher wavelength channel (channel 2). If those spectra do coincide, one has to subsequently scan first channel 2 and then channel 1, to avoid excitation and photobleaching of the fluorophores in channel 2 by the depletion beam of channel 1 [7

7. L. Meyer, D. Wildanger, R. Medda, A. Punge, S. O. Rizzoli, G. Donnert, and S. W. Hell, “Dual-color STED microscopy at 30-nm focal-plane resolution,” Small 4(8), 1095–1100 (2008). [CrossRef] [PubMed]

]. This early dual-color STED implementation essentially prevents multiple optical sections to be taken as necessary for e.g. a z-image stack in two colors as during a single scan the fluorophores of channel 2 are bleached by the depletion beam of channel 1. Somewhat counterintuitively the performance of a dual-channel STED microscope can be significantly improved by bringing the spectral bands of the two fluorophores as close together as possible while still preventing excessive crosstalk. This approach moves the depletion band out of the absorption spectrum thus allowing simultaneous imaging of two different fluorophores over multiple scans [8

8. D. Neumann, J. Bückers, L. Kastrup, S. W. Hell, and S. Jakobs, “Two-color STED microscopy reveals different degrees of colocalization between hexokinase-I and the three human VDAC isoforms,” PMC Biophys 3(1), 4 (2010). [CrossRef] [PubMed]

]. Spectral crosstalk of the fluorescence emission can be successfully reduced by introducing a temporal delay between the pulses of the two channels together with a synchronized detection.

In the following we present the custom design of a 3D dual-color STED microscope that fulfills all the aforementioned requirements for multi-wavelength, three-dimensional imaging at the nanoscale. We present a thorough characterization of the setup by imaging point-like fluorescent objects such as single-molecules and fluorescent beads. We further demonstrate the biological applicability of the instrument by dual-color 3D nanoscopy of various components of the nuclear pore complexes, and of DNA replication structures in mammalian cell nuclei.

2. Dual-color 3D STED setup

The dual-color 3D-STED microscope uses a single, randomly-polarized super-continuum laser source (SC-450-PP-HE, Fianium Ltd., Southampton, UK) to deliver all necessary spectral bands for the various beam paths [Fig. 1
Fig. 1 Schematic of experimental setup. The left part shows the laser source and the optical components necessary to prepare a total of six different beams of desired spectral bands for excitation and depletion. The six beams are coupled into six polarization maintaining fibers. The part on the right depicts the actual microscope starting with the output of the six fibers, the components for spatial co-alignment of the various beams, the phase plates for the purpose of shaping the PSF of the depletion beam, the objective lens and the detection beam path for the two fluorescence channels.
, Table 1

Table 1. Fiber couplers, fibers and collimators used in the 3D dual-channel STED microscope. All parts from Schäfter & Kirchhoff GmbH, Germany unless stated otherwise

table-icon
View This Table
| View All Tables
, Table 2

Table 2. Key components used for the custom-built 3D STED microscope

table-icon
View This Table
| View All Tables
]. The laser has a repetition rate of 1 MHz. Due to the 1 µs interval between pulses excitable metastable dark states can relax efficiently ensuring minimal photobleaching [4

4. D. Wildanger, E. Rittweger, L. Kastrup, and S. W. Hell, “STED microscopy with a supercontinuum laser source,” Opt. Express 16(13), 9614–9621 (2008). [CrossRef] [PubMed]

]. The pulse width is ~100 ps which allows for efficient depletion without the need for time-gated detection (T-STED/G-STED [9

9. J. R. Moffitt, C. Osseforth, and J. Michaelis, “Time-gating improves the spatial resolution of STED microscopy,” Opt. Express 19(5), 4242–4254 (2011). [CrossRef] [PubMed]

,10

10. G. Vicidomini, G. Moneron, K. Y. Han, V. Westphal, H. Ta, M. Reuss, J. Engelhardt, C. Eggeling, and S. W. Hell, “Sharper low-power STED nanoscopy by time gating,” Nat. Methods 8(7), 571–573 (2011). [CrossRef] [PubMed]

],) to maintain the best resolution at a given power. The laser output is divided into vertical and horizontal polarization by the use of a polarizing beam splitter (PBS-1). Each path is further split by a first dichroic mirror, which is centered at 730 nm (D-1) transmitting the spectrum necessary for the longer wavelength depletion beams (XY and Z STED-2 beams). The reflected light of D-1 is subsequently separated by another dichroic mirror (D-2) into the spectral range necessary for the depletion beams of the shorter wavelength STED beams (channel 1) in transmission and the two excitation beams in reflection. The respective spectral ranges for the depletion beams are further fine-tuned to the correct spectral band for the corresponding fluorophore by means of prism-based monochromators (PMC [4

4. D. Wildanger, E. Rittweger, L. Kastrup, and S. W. Hell, “STED microscopy with a supercontinuum laser source,” Opt. Express 16(13), 9614–9621 (2008). [CrossRef] [PubMed]

],). The spectral range typically used for the depletion beams is ~715 nm ± 10 nm and ~750 nm ± 10 nm for channel 1 and 2, respectively. The excitation wavelengths of 568 nm and 633 nm are selected by a pair of bandpass filters (BP-1 and BP-2). All six beams are coupled into polarization-maintaining single-mode fibers using wavelength selected fibers and lenses (SMF-x, Table 1) for the purpose of spatial cleaning and mechanic decoupling from the rest of the microscope. This design allows for independent control of the intensities of all beams by adjusting the respective fiber coupling efficiencies. Typically we used an average power of ~1.5 mW for each depletion beam and ~1 µW for each excitation beam (measured before the objective). The length difference (~20 m) between the fibers of channel 1 and channel 2 introduces a temporal delay of the pulses between the two channels of ~90 ns.

In order to achieve maximal depletion efficiency the timing of the excitation and depletion pulse is critical. The excitation pulse has to arrive first, with the STED pulse being delayed by a few picoseconds [3

3. S. W. Hell and J. Wichmann, “Breaking the diffraction resolution limit by stimulated emission: stimulated-emission-depletion fluorescence microscopy,” Opt. Lett. 19(11), 780–782 (1994). [CrossRef] [PubMed]

]. To ensure correct relative timing of excitation and depletion pulses for each channel, the fiber couplers are mounted on manual linear stages and adjusted accordingly. Correct synchronization between the respective excitation and STED pulses is based solely on beam path length thus not requiring any drift-sensitive electronic timing device.

The spatial co-alignment of the six beams is done using polarizing beam splitters and dichroic mirrors [Fig. 1]. The beam paths of the depletion beams are described here in detail exemplary for channel 1 (but also apply to channel 2). The depletion beam STED-XY-1 is sent through a Vortex Phase plate (STED-XY-PP-1) to imprint the phase pattern necessary to yield the Laguerre-Gaussian LG01 mode in the focal plane. This light pattern is used to increase the lateral resolution of the microscope in x and y. To ensure a perfect circular polarization of this beam it passes through a combination of a λ/2 retarder (λ/2-WP) and a custom-selected achromatic λ/4 wave-plate (λ/4-WP) before being directed into a high NA, 100x oil immersion objective (HCX PL APO 100x/1.40-0.70 oil CS, Leica, Germany). To obtain optimal circular polarization for the STED-XY beams a homebuilt polarization analyzer is inserted between the λ/4 retarder and the objective lens. It consists of a rotating polarizer and a photodiode which detects the modulated light intensity that is transmitted through the polarizer. The signal is analyzed by a custom-written software (LabVIEW, National Instruments, USA). The observed sine modulation is reduced to a minimum by adjusting the angle of the two retarders one at a time in an iterative fashion. The fine-tuning of the circular polarization is crucial to the performance of the STED microscope. In order to deplete the fluorescence also along the axial direction a different phase mask is introduced into the STED-Z-1 beam (STED-Z-PP-1). The mask consists of a sputtered MgF2 cylinder in the center [6

6. D. Wildanger, R. Medda, L. Kastrup, and S. W. Hell, “A compact STED microscope providing 3D nanoscale resolution,” J. Microsc. 236(1), 35–43 (2009). [CrossRef] [PubMed]

] with a thickness which results in a phase shift of π. The size of the cylinder is chosen such that half of the intensity of the beam is phase shifted. When focused the beam will show destructive interference in the focal point and two prominent axial side lobes above and below the focal plane. Similar to the effect of the STED-XY depletion beam on the resolution in x and y, the STED-Z beam will lead to an enhanced resolution along the optical axis, z. Note that the interference pattern also shows a ring of light resembling the donut pattern of the STED-XY beams. This also increases resolution in x and y but to a much lesser extend as the dedicated STED-XY beam. The STED-Z-1 beam is superimposed onto the STED-XY-1 beam by the use of a polarizing beam splitter (PBS-2). In contrast to the STED-XY beams the polarization state is not important for the STED-Z beams. All STED beams, i.e. STED-XYZ-1 and STED-XYZ-2, are subsequently combined by a dichroic mirror (D-1). Moreover, the two excitation beams are coupled into the optical path by a pair of dichroic mirrors, respectively (D-2 and D-4) and all beams are combined using dichroic mirror D-3 (D-3). In order to minimize wave front distortions D-1 - D-4 are all on 5 mm thick substrates.

The fluorescence light is collected by the same microscope objective and after passing D-3 and D-4, it is split into the two fluorescence channels by another dichroic (D-5) and subsequently filtered by a pair of emission filters (EF-1 and EF-2). The respective fluorescence light is focused onto a multimode fiber which serves as a confocal pinhole to reject out-of-focus light (MMF, M31L01, Thorlabs, USA). For detection a pair of fiber-coupled avalanche photodiodes is used (APD, SPCM-AQRH-13/14-FC, Perkin-Elmer, USA).

While a careful choice of emission filters and dichroic mirrors allows minimizing spectral crosstalk, it is still not negligible. Moreover, excitation of the fluorophores in detection channel 2 by the STED beams of channel 1 is an even more severe problem. Therefore, it is necessary to introduce a time-delay between the two channels to minimize these effects, thus effectively disregarding counts from the inactive channel. To this end an electronic gating device was designed and implemented. It uses the pulse trigger output of the laser source together with adjustable electronic delays to synchronize a user-defined detection time-gate for the arrival time of the photons of the respective channel. By keeping the time-gate short (~10’s of ns) dark counts from both detectors are efficiently ignored resulting in low background images. As a result, for all experiments conducted no crosstalk had to be removed after imaging.

The sample was scanned with a fast piezo scanning stage (Stage 733.2DD; Controller E-710 with dynamic digital linearization feature for better tracking accuracy, Physik Instrumente, Germany). Detector events were counted with a National Instruments computer card (PCIe-6259, National Instruments, USA). Instrument control and image acquisition was performed by custom-written software (LabVIEW, National Instruments, USA).

3. Results

The PSFs of the depletion beams were experimentally measured by scanning across immobilized individual 80 nm gold colloids dispersed on a coverslip at a dilution of 1:1000 (British Biocell, UK) and recording the scattering using a channel photomultiplier module (MD-963, Perkin-Elmer, USA). The recorded intensity patterns show deep minima with little residual intensity [Fig. 2
Fig. 2 Experimentally measured point-spread functions (upper panels; green for excitation and magenta for depletion patterns; shown representatively for channel 2) and corresponding fluorescence images (lower panels). Point-spread functions were measured by recording the scattering from an immobilized 80 nm gold bead scanned through the focus. For simplicity, only data from x-z scans are presented showing both the lateral as well as the axial PSF. The fluorescence images were recorded using 100 nm fluorescent beads dispersed on a cover slip (fluorescence images are shown representatively for channel 1). A) and E) Confocal mode using only the excitation beam and resulting in a diffraction limited PSF and fluorescence image which is considerably spread along the z-axis. B) and F) The XY-STED beam improves only lateral resolution; axial resolution is not changed. C) and G) The Z-STED beam acts mainly in the axial dimension. Additionally, resolution is also slightly improved along the lateral dimension. D) and H) By spatially co-aligning the XY and Z STED beams resolution is dramatically enhanced in all three dimensions.
]. The quality of the depletion beam pattern is crucial for STED microscopy in order to produce high-contrast images. More so, for dim objects such as single molecules a true central zero is essential to detect any signal above background at all. The effect of the two individual depletion beam types and their summed intensity profile on the effective PSF were measured by imaging 100 nm fluorescent nanospheres dispersed on a coverslip at a dilution of 1:1000 [Fig. 2]. By selectively activating the depletion beams by means of mechanical shutters the setup can be operated as a standard confocal diffraction limited microscope [all depletion beams deactivated, Fig. 2(a) and Fig. 2(e)], a conventional 2D STED [only STED-XY depletion beams active, Fig. 2(b) and Fig. 2(f)], a Z-STED with only the z depletion beam [Fig. 2(c) and Fig. 2(g)] or a 3D STED microscope [all depletion beams activated, Fig. 2(d) and Fig. 2(h)]. It is important to note that this 3D mode works with a single objective lens simply by non-coherent co-alignment of the two fundamental depletion patterns. This greatly simplifies the operation and stability of the setup compared to other techniques such as 4Pi-microscopy [11

11. M. Nagorni and S. W. Hell, “4Pi-confocal microscopy provides three-dimensional images of the microtubule network with 100- to 150-nm resolution,” J. Struct. Biol. 123(3), 236–247 (1998). [CrossRef] [PubMed]

].

One common criticism of the STED technique is that it requires rather high light intensities for the depletion laser to achieve high spatial resolution, which could be detrimental to the sample. Therefore, we used single dye molecules attached to short pieces of DNA to test both resolution and sensitivity of our instrument in a single experiment [Fig. 3
Fig. 3 Dual-color imaging of a mixture of immobilized DNA molecules labeled with single Atto590 or Atto647N fluorophores. A) Single molecules imaged in STED-2D mode. Green LUT shows fluorescence intensity of Atto590, magenta shows that of Atto647N molecules. Arrows illustrate the position of the cross-sections shown in B) (Atto590) and C) (Atto647N). D) Cross-section along the z-axis of a single Atto647N molecule shows a FWHM of ~83 nm. Image A) was post processed for printing purposes but not for data analysis, fits in B), C) and D) show raw data. Pixel size is 10 nm2 (panels B and C) or 20 nm2 (panel D) at a pixel dwell time of 400 µs. Fitting and plotting was performed using the Python distribution Python(x,y) [12].
]. To this end a 1 nM 1:1 mixture of 60 bp long DNA molecules labeled with either Atto590 or Atto647N and biotin were immobilized to cover slips using NeutrAvidin and imaged in a dual-color experiment in 2D mode. Since single molecules are individual quantum systems, photobleaching occurs in a one-step process. Therefore, if molecules bleach faster than their illumination time, the fluorescence images of the molecules exhibit sudden drops in intensity. In contrast, most of the fluorescence spots observed in our STED experiments showed rather round fluorescence patterns [Fig. 3(a)], indicating that photobleaching was not limiting in these experiments. In order to determine the resolution of the microscope the recorded 2D intensity profile from several molecules were fitted using rotation-symmetric 2D-Gaussian functions. The calculated average lateral FWHM (mean ± standard deviation) for channel 1 was 40.2 nm ± 2.6 nm (n = 32) and for channel 2 we found 34.2 nm ± 2.5 nm (n = 51). Extracting intensity values along a line across a representative single molecule yielded slightly better results, namely ~34 nm for Atto590 and ~31 nm for Atto647N [Fig. 3(b) and Fig. 3(c)]. One possible reason for the observed improved resolution when only single line scans are analyzed could be the fact that blinking of the molecules adds noise of the system and thus reduces the quality of the 2D Gaussian fits. Independent of the analysis procedure it is clear that the microscope has single molecule sensitivity and that a lateral resolution of ~35 nm can be obtained. Axial resolution for channel 2 was measured in the same experiment. Fitting a Gaussian function to axial cross-sections of single Atto590 and Atto647N molecules imaged with the STED-Z beam active shows an axial FWHM of 89.4 nm ± 14.5 nm and 88.4 nm ± 27.4 nm [mean ± standard deviation, n = 10 and n = 5, respectively, Fig. 3(d)]. While combining the STED-XY and STED-Z intensity profiles leads to an increased intensity in the x-y plane and should thus theoretically lead to a slight improvement in lateral resolution (~5%), we observed a similar lateral resolution in 3D STED mode as compared to STED-XY mode due to limited experimental accuracy.

Having determined the resolution and sensitivity of the microscope we wanted to demonstrate its capabilities on biological samples. Due to their size of roughly 100 nm and precise localization in the nuclear envelope, nuclear pore complexes (NPCs) have been proven to be an excellent test system for super-resolution optical microscopy [13

13. A. Löschberger, S. van de Linde, M. C. Dabauvalle, B. Rieger, M. Heilemann, G. Krohne, and M. Sauer, “Super-resolution imaging visualizes the eightfold symmetry of gp210 proteins around the nuclear pore complex and resolves the central channel with nanometer resolution,” J. Cell Sci. 125(3), 570–575 (2012). [CrossRef] [PubMed]

15

15. F. Göttfert, C. A. Wurm, V. Mueller, S. Berning, V. C. Cordes, A. Honigmann, and S. W. Hell, “Coaligned Dual-Channel STED Nanoscopy and Molecular Diffusion Analysis at 20 nm Resolution,” Biophys. J. 105(1), L01–L03 (2013). [CrossRef] [PubMed]

]. We therefore investigated fixed human bone osteosarcoma U2OS cells with nuclear pore complexes immunostained for dual-color imaging in 2D mode [Fig. 4
Fig. 4 Dual-color STED microscopy in 2D mode of fixed U2OS cells using antibodies targeting different parts of NPCs. Green: Primary antibody against FG-repeat containing nucleoporins detected via secondary antibody conjugated to Alexa594. Magenta: Primary antibody against inner nuclear pore basket protein Tpr detected via secondary antibody conjugated to Atto647N. Images were background corrected and contrast was enhanced for printing purposes. Pixel dwell times were 180 µs. Depending on the z-height of the imaging plane different views of the cell nucleus are obtained. A) Apical slice shows a co-localization of the two secondary antibodies with a FWHM of single nuclear pore complexes of ~80 nm. B) Mid-section slice through the nucleus shows a clear separation between the two secondary antibodies. In contrast to Tpr, the FG-repeat antibody typically displays two spots, in the central part and the cytoplasmic side of NPCs (white arrows), demonstrating the ability to separate different epitopes within this large ~125 MDa sized complex. C) 3D isosurface reconstruction of 34 slices (19 × 19 µm) of one nucleus atop a maximum intensity projection of the same data using Avizo 8.0.1 (FEI, USA) and Fiji [17]. Inset shows zoomed in volume rendering of individual NPCs.
]. Primary antibodies were targeted against the FG-repeat region common to several nucleoporins in the central or the cytoplasmic region of the NPC (αFG-repeat, channel 1, secondary antibodies conjugated to Alexa594). We also used primary antibodies to histone H3.X/Y which showed a cross-reaction with nuclear pore complexes [16

16. S. M. Wiedemann, S. N. Mildner, C. Bönisch, L. Israel, A. Maiser, S. Matheisl, T. Straub, R. Merkl, H. Leonhardt, E. Kremmer, L. Schermelleh, and S. B. Hake, “Identification and characterization of two novel primate-specific histone H3 variants, H3.X and H3.Y,” J. Cell Biol. 190(5), 777–791 (2010). [CrossRef] [PubMed]

]. Sequence alignment indicates that it is most likely targeted to the inner basket protein Tpr (αTpr, channel 2, secondary antibodies conjugated to Atto647N). Cells were imaged using a voxel size of 20 × 20 × 250 nm (xyz).

In an apical slice viewing top-down onto the nuclear membrane co-localization of the two colors is observed [Fig. 4(a)]. The mean FWHM ± standard deviation of single dot-like structures is 81 nm ± 6.6 nm and 80 nm ± 5.9 nm in channel 1 and 2, respectively.

In contrast, when looking at the mid-section slice of the nucleus a distinct staining pattern can be observed in which antibodies directed against NPC are oriented towards the cytoplasm relative to the signal from the Tpr antibody [Fig. 4(b)]. Strikingly two distinct regions targeted by the αFG-repeat antibody with a separation of about 92.2 nm ± 11.4 nm (n = 8) can be observed for a single nuclear pore complex. In a similar experiment using 3D-structured illumination microscopy (3D-SIM) different epitopes recognized by the very same primary antibody could not be resolved [18

18. L. Schermelleh, P. M. Carlton, S. Haase, L. Shao, L. Winoto, P. Kner, B. Burke, M. C. Cardoso, D. A. Agard, M. G. Gustafsson, H. Leonhardt, and J. W. Sedat, “Subdiffraction multicolor imaging of the nuclear periphery with 3D structured illumination microscopy,” Science 320(5881), 1332–1336 (2008). [CrossRef] [PubMed]

]. It should be pointed out that the commercially bought antibody does in fact react with Nup62, Nup214, Nup358 and Nup153 [18

18. L. Schermelleh, P. M. Carlton, S. Haase, L. Shao, L. Winoto, P. Kner, B. Burke, M. C. Cardoso, D. A. Agard, M. G. Gustafsson, H. Leonhardt, and J. W. Sedat, “Subdiffraction multicolor imaging of the nuclear periphery with 3D structured illumination microscopy,” Science 320(5881), 1332–1336 (2008). [CrossRef] [PubMed]

]. While the first three proteins are located towards the cytoplasmic side on the nuclear pore complex, Nup153 is found more centrally on the same level as the nuclear lamina [18

18. L. Schermelleh, P. M. Carlton, S. Haase, L. Shao, L. Winoto, P. Kner, B. Burke, M. C. Cardoso, D. A. Agard, M. G. Gustafsson, H. Leonhardt, and J. W. Sedat, “Subdiffraction multicolor imaging of the nuclear periphery with 3D structured illumination microscopy,” Science 320(5881), 1332–1336 (2008). [CrossRef] [PubMed]

]. Since in the 2D mode the lateral resolution of the herein presented setup is higher than that of 3D-SIM the observation of two distinct label positions strongly suggests that we can in fact discriminate between different epitopes on a single nuclear pore complex using the same primary antibody. The antibody directed against Tpr that constitutes the inner basket locates even further at the inner nucleoplasmic side of the NPC [16

16. S. M. Wiedemann, S. N. Mildner, C. Bönisch, L. Israel, A. Maiser, S. Matheisl, T. Straub, R. Merkl, H. Leonhardt, E. Kremmer, L. Schermelleh, and S. B. Hake, “Identification and characterization of two novel primate-specific histone H3 variants, H3.X and H3.Y,” J. Cell Biol. 190(5), 777–791 (2010). [CrossRef] [PubMed]

].

The mid-section slice can reveal many details as the object of interest is correctly oriented in the focal plane and shows only a spot like pattern of small size. This is due to the fact that for this orientation the structure of interest, i.e. location of the respective epitopes, is oriented in the direction of optimal resolution. This fortunate combination of circumstances is seldom found when imaging more complex structures in cells with strongly asymmetric effective PSFs. Additionally to infer trustworthy results representative of the whole cell nucleus from only a single slice is rather difficult and oftentimes unjustified. The downside of low axial resolution becomes apparent when calculating a 3D reconstruction of 34 slices of one nucleus [Fig. 4(c)]. In all but the mid-section slice information gained from imaging is compromised due to the poor axial resolution of the 2D STED. Therefore, three-dimensional super-resolution microscopy is necessary in order to quantitatively interpret microscopy data from three-dimensional biological structures that extend along the optical axis. For the given example of NPC the correct sectioning can lead to two orthogonal 2D projections of the 3D structures which yields already a wealth of information, however, for more complex three-dimensional structures without pre-determined spatial orientation 3D super-resolution data is required.

To test the performance of the 3D mode in a dual-color experiment we imaged replication complexes in mouse C2C12 myoblast cells in early to late S-phase using a pulse labeling experiment. 5-ethynyl-2’-deoxyuridine (EdU) was incorporated into cells for 15 min after which cells were formaldehyde fixed. EdU labeled postreplicative DNA was detected with Alexa594 (channel 1) via click chemistry and the replication machinery was immunostained with primary antibody against proliferating cell nuclear antigen (αPCNA) and secondary antibodies conjugated to Atto647N (channel 2). A series of z-slices was acquired which allowed a 3D reconstruction of the replication foci within the nuclear volume [Fig. 5(A)
Fig. 5 3D reconstruction of replication foci in a C2C12 cell nucleus during late S-phase labeled during a 15 min EdU pulse. EdU incorporation was visualized by covalent coupling of Alexa594 dye via click chemistry (first channel; green). The second channel (magenta) shows immunostaining of PCNA using Atto647N-conjugated secondary antibodies. A) 3D isosurface reconstruction of 38 dual-color image planes taken in 3D mode with a voxel size of 50 nm3 at a pixel dwell time of 400 µs using Avizo 8.0.1 and Fiji [17]. The bottom plane (20 × 20 µm) shows the maximum intensity projection of the data set. B) X-y slice of a different nucleus in early S-phase imaged in 3D mode with a pixel size of 20 nm2 and pixel dwell time of 200 µs. C) Upper panel: Representative x-z-slice assembled from the first channel (EdU) of the image stack used for subfigure A. Lower panel: Gaussian function fitted to raw data of upper panel, position indicated by arrow. D) Upper panel: Representative x-z-slice constructed from the second channel (αPCNA) of the image stack used for subfigure A. Lower part: Gaussian function fitted to raw data of upper panel at position indicated by arrow. Image data was post processed to remove background and improve contrast for printing purposes. Figure C and D uses a multicolor look-up table to better visualize the dynamic range of the images. Fitting and plotting was performed with raw data using the Python distribution Python(x,y) [12].
].

4. Discussion

We have shown the design and implementation of a compact, table-top 3D two-color STED microscope. The system can simultaneously record fluorescence signals from two differently labeled species and obtain resolution on the order of 35 nm laterally and 90 nm axially. These values are similar to published values for lateral resolution of a STED-XY microscope [4

4. D. Wildanger, E. Rittweger, L. Kastrup, and S. W. Hell, “STED microscopy with a supercontinuum laser source,” Opt. Express 16(13), 9614–9621 (2008). [CrossRef] [PubMed]

], given similar laser powers, or 3D resolution of a single color 3D STED microscope [6

6. D. Wildanger, R. Medda, L. Kastrup, and S. W. Hell, “A compact STED microscope providing 3D nanoscale resolution,” J. Microsc. 236(1), 35–43 (2009). [CrossRef] [PubMed]

]. The sensitivity of the microscope allows even the signal from single dye molecules to be detected and to be used for the characterization of the instruments resolution. Moreover, we used the microscope for the investigation of different biological structures. Our results exemplify the need for three-dimensional super-resolution microscopy when moving away from in vitro experiments on the cover slip towards more complex biological systems, e.g. to visualize internal cellular structures remote from the coverslip. When increasing resolution in the lateral dimension only a concurrent increase in axial resolution will allow faithful reconstruction of objects heterogeneously distributed in space.

We present a straightforward design of a 3D dual-channel super-resolution STED microscope based on non-coherent spatial co-alignment of two differentially shaped depletion beams per fluorescence channel. There is no theoretical limit to which resolution can be improved in stimulated emission depletion microscopy. Practical restrictions are mainly set by the laser power available, the photostability of the employed dyes, and optical aberrations causing the central minimum of the depletion beams to deviate from zero. In our setup we make use of a super-continuum laser, which provides all spectral bands necessary. This permits easy and - more important - stable temporal alignment without having to synchronize multiple excitation and depletion sources. A present downside is the relatively low power available which presently sets the bottleneck in terms of resolution. Very recently the use of a single, high pulse power laser operating at 775 nm at 20 MHz was demonstrated in a dual-channel setup running in 2D mode [15

15. F. Göttfert, C. A. Wurm, V. Mueller, S. Berning, V. C. Cordes, A. Honigmann, and S. W. Hell, “Coaligned Dual-Channel STED Nanoscopy and Molecular Diffusion Analysis at 20 nm Resolution,” Biophys. J. 105(1), L01–L03 (2013). [CrossRef] [PubMed]

]. This source could easily be used in a microscope design similar to the one presented here. Such a setup would not be limited by available depletion beam power and would represent a step forward towards higher resolution imaging.

In this study we optimized the instrument for the fluorophores with spectral properties similar to Atto590 (channel 1) and similar to Atto647N (channel 2). Since the utilized white light laser covers a spectral range from ~450 nm to ~2000 nm a combination of other dyes should be straightforward to implement by the use of different dichroic filters as long as the given design rules for a dual-color STED instrument are observed, making the application of the design very general. Moreover, other schemes have been published where the same depletion laser is used for two super-resolved channels making use of long Stokes shift dyes. Furthermore in other schemes the lifetime of fluorophores have also been used for image multiplexing [5

5. R. Schmidt, C. A. Wurm, S. Jakobs, J. Engelhardt, A. Egner, and S. W. Hell, “Spherical nanosized focal spot unravels the interior of cells,” Nat. Methods 5(6), 539–544 (2008). [CrossRef] [PubMed]

,21

21. J. Bückers, D. Wildanger, G. Vicidomini, L. Kastrup, and S. W. Hell, “Simultaneous multi-lifetime multi-color STED imaging for colocalization analyses,” Opt. Express 19(4), 3130–3143 (2011). [CrossRef] [PubMed]

]. These schemes could certainly be combined with the presented experimental design thus opening the possibility for 3- or 4-color STED schemes.

Furthermore the systems design and the technique itself does not preclude the imaging of living cells [22

22. V. Westphal, S. O. Rizzoli, M. A. Lauterbach, D. Kamin, R. Jahn, and S. W. Hell, “Video-rate far-field optical nanoscopy dissects synaptic vesicle movement,” Science 320(5873), 246–249 (2008). [CrossRef] [PubMed]

]. The image acquisition speed in this study is limited by the relatively slow repetition rate of the laser used which allows the imaging of a 1 µm × 1 µm area at a pixel size of 20 nm2 in approximately ~0.5 seconds. Recently faster laser systems with two depletion outputs have become available which operate at 20 MHz. The use of these modern sources will allow for video rate imaging speed ideally suited for live-cell imaging.

In summary the instrument described provides a versatile tool for 3D two-color super-resolution microscopy and should, due to its simple, modular and tunable design, find many applications in life science.

Acknowledgments

We thank Sandra Hake for providing us with the αTpr antibodies, Andreas Maiser for help with preparation of labeled cells, Lars Kastrup for providing us with the custom made waveplates for the Z-STED beams. This work was funded by the European Union through the ERC starting grant Remodeling.

References and links

1.

L. Schermelleh, R. Heintzmann, and H. Leonhardt, “A guide to super-resolution fluorescence microscopy,” J. Cell Biol. 190(2), 165–175 (2010). [CrossRef] [PubMed]

2.

C. Coltharp and J. Xiao, “Superresolution microscopy for microbiology,” Cell. Microbiol. 14(12), 1808–1818 (2012). [CrossRef] [PubMed]

3.

S. W. Hell and J. Wichmann, “Breaking the diffraction resolution limit by stimulated emission: stimulated-emission-depletion fluorescence microscopy,” Opt. Lett. 19(11), 780–782 (1994). [CrossRef] [PubMed]

4.

D. Wildanger, E. Rittweger, L. Kastrup, and S. W. Hell, “STED microscopy with a supercontinuum laser source,” Opt. Express 16(13), 9614–9621 (2008). [CrossRef] [PubMed]

5.

R. Schmidt, C. A. Wurm, S. Jakobs, J. Engelhardt, A. Egner, and S. W. Hell, “Spherical nanosized focal spot unravels the interior of cells,” Nat. Methods 5(6), 539–544 (2008). [CrossRef] [PubMed]

6.

D. Wildanger, R. Medda, L. Kastrup, and S. W. Hell, “A compact STED microscope providing 3D nanoscale resolution,” J. Microsc. 236(1), 35–43 (2009). [CrossRef] [PubMed]

7.

L. Meyer, D. Wildanger, R. Medda, A. Punge, S. O. Rizzoli, G. Donnert, and S. W. Hell, “Dual-color STED microscopy at 30-nm focal-plane resolution,” Small 4(8), 1095–1100 (2008). [CrossRef] [PubMed]

8.

D. Neumann, J. Bückers, L. Kastrup, S. W. Hell, and S. Jakobs, “Two-color STED microscopy reveals different degrees of colocalization between hexokinase-I and the three human VDAC isoforms,” PMC Biophys 3(1), 4 (2010). [CrossRef] [PubMed]

9.

J. R. Moffitt, C. Osseforth, and J. Michaelis, “Time-gating improves the spatial resolution of STED microscopy,” Opt. Express 19(5), 4242–4254 (2011). [CrossRef] [PubMed]

10.

G. Vicidomini, G. Moneron, K. Y. Han, V. Westphal, H. Ta, M. Reuss, J. Engelhardt, C. Eggeling, and S. W. Hell, “Sharper low-power STED nanoscopy by time gating,” Nat. Methods 8(7), 571–573 (2011). [CrossRef] [PubMed]

11.

M. Nagorni and S. W. Hell, “4Pi-confocal microscopy provides three-dimensional images of the microtubule network with 100- to 150-nm resolution,” J. Struct. Biol. 123(3), 236–247 (1998). [CrossRef] [PubMed]

12.

Python(x,y) 2.7.5.1,” http://code.google.com/p/pythonxy/.

13.

A. Löschberger, S. van de Linde, M. C. Dabauvalle, B. Rieger, M. Heilemann, G. Krohne, and M. Sauer, “Super-resolution imaging visualizes the eightfold symmetry of gp210 proteins around the nuclear pore complex and resolves the central channel with nanometer resolution,” J. Cell Sci. 125(3), 570–575 (2012). [CrossRef] [PubMed]

14.

A. Szymborska, A. de Marco, N. Daigle, V. C. Cordes, J. A. Briggs, and J. Ellenberg, “Nuclear pore scaffold structure analyzed by super-resolution microscopy and particle averaging,” Science 341(6146), 655–658 (2013). [CrossRef] [PubMed]

15.

F. Göttfert, C. A. Wurm, V. Mueller, S. Berning, V. C. Cordes, A. Honigmann, and S. W. Hell, “Coaligned Dual-Channel STED Nanoscopy and Molecular Diffusion Analysis at 20 nm Resolution,” Biophys. J. 105(1), L01–L03 (2013). [CrossRef] [PubMed]

16.

S. M. Wiedemann, S. N. Mildner, C. Bönisch, L. Israel, A. Maiser, S. Matheisl, T. Straub, R. Merkl, H. Leonhardt, E. Kremmer, L. Schermelleh, and S. B. Hake, “Identification and characterization of two novel primate-specific histone H3 variants, H3.X and H3.Y,” J. Cell Biol. 190(5), 777–791 (2010). [CrossRef] [PubMed]

17.

J. Schindelin, I. Arganda-Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, S. Preibisch, C. Rueden, S. Saalfeld, B. Schmid, J.-Y. Tinevez, D. J. White, V. Hartenstein, K. Eliceiri, P. Tomancak, and A. Cardona, “Fiji: an open-source platform for biological-image analysis,” Nat. Methods 9(7), 676–682 (2012). [CrossRef] [PubMed]

18.

L. Schermelleh, P. M. Carlton, S. Haase, L. Shao, L. Winoto, P. Kner, B. Burke, M. C. Cardoso, D. A. Agard, M. G. Gustafsson, H. Leonhardt, and J. W. Sedat, “Subdiffraction multicolor imaging of the nuclear periphery with 3D structured illumination microscopy,” Science 320(5881), 1332–1336 (2008). [CrossRef] [PubMed]

19.

D. Baddeley, V. O. Chagin, L. Schermelleh, S. Martin, A. Pombo, P. M. Carlton, A. Gahl, P. Domaing, U. Birk, H. Leonhardt, C. Cremer, and M. C. Cardoso, “Measurement of replication structures at the nanometer scale using super-resolution light microscopy,” Nucleic Acids Res. 38(2), e8 (2010). [CrossRef] [PubMed]

20.

Z. Cseresnyes, U. Schwarz, and C. M. Green, “Analysis of replication factories in human cells by super-resolution light microscopy,” BMC Cell Biol. 10(1), 88 (2009). [CrossRef] [PubMed]

21.

J. Bückers, D. Wildanger, G. Vicidomini, L. Kastrup, and S. W. Hell, “Simultaneous multi-lifetime multi-color STED imaging for colocalization analyses,” Opt. Express 19(4), 3130–3143 (2011). [CrossRef] [PubMed]

22.

V. Westphal, S. O. Rizzoli, M. A. Lauterbach, D. Kamin, R. Jahn, and S. W. Hell, “Video-rate far-field optical nanoscopy dissects synaptic vesicle movement,” Science 320(5873), 246–249 (2008). [CrossRef] [PubMed]

OCIS Codes
(000.1430) General : Biology and medicine
(100.6640) Image processing : Superresolution
(180.2520) Microscopy : Fluorescence microscopy
(180.6900) Microscopy : Three-dimensional microscopy
(350.5730) Other areas of optics : Resolution

ToC Category:
Microscopy

History
Original Manuscript: January 9, 2014
Revised Manuscript: March 12, 2014
Manuscript Accepted: March 12, 2014
Published: March 18, 2014

Virtual Issues
Vol. 9, Iss. 5 Virtual Journal for Biomedical Optics

Citation
Christian Osseforth, Jeffrey R. Moffitt, Lothar Schermelleh, and Jens Michaelis, "Simultaneous dual-color 3D STED microscopy," Opt. Express 22, 7028-7039 (2014)
http://www.opticsinfobase.org/oe/abstract.cfm?URI=oe-22-6-7028


Sort:  Author  |  Year  |  Journal  |  Reset  

References

  1. L. Schermelleh, R. Heintzmann, H. Leonhardt, “A guide to super-resolution fluorescence microscopy,” J. Cell Biol. 190(2), 165–175 (2010). [CrossRef] [PubMed]
  2. C. Coltharp, J. Xiao, “Superresolution microscopy for microbiology,” Cell. Microbiol. 14(12), 1808–1818 (2012). [CrossRef] [PubMed]
  3. S. W. Hell, J. Wichmann, “Breaking the diffraction resolution limit by stimulated emission: stimulated-emission-depletion fluorescence microscopy,” Opt. Lett. 19(11), 780–782 (1994). [CrossRef] [PubMed]
  4. D. Wildanger, E. Rittweger, L. Kastrup, S. W. Hell, “STED microscopy with a supercontinuum laser source,” Opt. Express 16(13), 9614–9621 (2008). [CrossRef] [PubMed]
  5. R. Schmidt, C. A. Wurm, S. Jakobs, J. Engelhardt, A. Egner, S. W. Hell, “Spherical nanosized focal spot unravels the interior of cells,” Nat. Methods 5(6), 539–544 (2008). [CrossRef] [PubMed]
  6. D. Wildanger, R. Medda, L. Kastrup, S. W. Hell, “A compact STED microscope providing 3D nanoscale resolution,” J. Microsc. 236(1), 35–43 (2009). [CrossRef] [PubMed]
  7. L. Meyer, D. Wildanger, R. Medda, A. Punge, S. O. Rizzoli, G. Donnert, S. W. Hell, “Dual-color STED microscopy at 30-nm focal-plane resolution,” Small 4(8), 1095–1100 (2008). [CrossRef] [PubMed]
  8. D. Neumann, J. Bückers, L. Kastrup, S. W. Hell, S. Jakobs, “Two-color STED microscopy reveals different degrees of colocalization between hexokinase-I and the three human VDAC isoforms,” PMC Biophys 3(1), 4 (2010). [CrossRef] [PubMed]
  9. J. R. Moffitt, C. Osseforth, J. Michaelis, “Time-gating improves the spatial resolution of STED microscopy,” Opt. Express 19(5), 4242–4254 (2011). [CrossRef] [PubMed]
  10. G. Vicidomini, G. Moneron, K. Y. Han, V. Westphal, H. Ta, M. Reuss, J. Engelhardt, C. Eggeling, S. W. Hell, “Sharper low-power STED nanoscopy by time gating,” Nat. Methods 8(7), 571–573 (2011). [CrossRef] [PubMed]
  11. M. Nagorni, S. W. Hell, “4Pi-confocal microscopy provides three-dimensional images of the microtubule network with 100- to 150-nm resolution,” J. Struct. Biol. 123(3), 236–247 (1998). [CrossRef] [PubMed]
  12. Python(x,y) 2.7.5.1,” http://code.google.com/p/pythonxy/ .
  13. A. Löschberger, S. van de Linde, M. C. Dabauvalle, B. Rieger, M. Heilemann, G. Krohne, M. Sauer, “Super-resolution imaging visualizes the eightfold symmetry of gp210 proteins around the nuclear pore complex and resolves the central channel with nanometer resolution,” J. Cell Sci. 125(3), 570–575 (2012). [CrossRef] [PubMed]
  14. A. Szymborska, A. de Marco, N. Daigle, V. C. Cordes, J. A. Briggs, J. Ellenberg, “Nuclear pore scaffold structure analyzed by super-resolution microscopy and particle averaging,” Science 341(6146), 655–658 (2013). [CrossRef] [PubMed]
  15. F. Göttfert, C. A. Wurm, V. Mueller, S. Berning, V. C. Cordes, A. Honigmann, S. W. Hell, “Coaligned Dual-Channel STED Nanoscopy and Molecular Diffusion Analysis at 20 nm Resolution,” Biophys. J. 105(1), L01–L03 (2013). [CrossRef] [PubMed]
  16. S. M. Wiedemann, S. N. Mildner, C. Bönisch, L. Israel, A. Maiser, S. Matheisl, T. Straub, R. Merkl, H. Leonhardt, E. Kremmer, L. Schermelleh, S. B. Hake, “Identification and characterization of two novel primate-specific histone H3 variants, H3.X and H3.Y,” J. Cell Biol. 190(5), 777–791 (2010). [CrossRef] [PubMed]
  17. J. Schindelin, I. Arganda-Carreras, E. Frise, V. Kaynig, M. Longair, T. Pietzsch, S. Preibisch, C. Rueden, S. Saalfeld, B. Schmid, J.-Y. Tinevez, D. J. White, V. Hartenstein, K. Eliceiri, P. Tomancak, A. Cardona, “Fiji: an open-source platform for biological-image analysis,” Nat. Methods 9(7), 676–682 (2012). [CrossRef] [PubMed]
  18. L. Schermelleh, P. M. Carlton, S. Haase, L. Shao, L. Winoto, P. Kner, B. Burke, M. C. Cardoso, D. A. Agard, M. G. Gustafsson, H. Leonhardt, J. W. Sedat, “Subdiffraction multicolor imaging of the nuclear periphery with 3D structured illumination microscopy,” Science 320(5881), 1332–1336 (2008). [CrossRef] [PubMed]
  19. D. Baddeley, V. O. Chagin, L. Schermelleh, S. Martin, A. Pombo, P. M. Carlton, A. Gahl, P. Domaing, U. Birk, H. Leonhardt, C. Cremer, M. C. Cardoso, “Measurement of replication structures at the nanometer scale using super-resolution light microscopy,” Nucleic Acids Res. 38(2), e8 (2010). [CrossRef] [PubMed]
  20. Z. Cseresnyes, U. Schwarz, C. M. Green, “Analysis of replication factories in human cells by super-resolution light microscopy,” BMC Cell Biol. 10(1), 88 (2009). [CrossRef] [PubMed]
  21. J. Bückers, D. Wildanger, G. Vicidomini, L. Kastrup, S. W. Hell, “Simultaneous multi-lifetime multi-color STED imaging for colocalization analyses,” Opt. Express 19(4), 3130–3143 (2011). [CrossRef] [PubMed]
  22. V. Westphal, S. O. Rizzoli, M. A. Lauterbach, D. Kamin, R. Jahn, S. W. Hell, “Video-rate far-field optical nanoscopy dissects synaptic vesicle movement,” Science 320(5873), 246–249 (2008). [CrossRef] [PubMed]

Cited By

Alert me when this paper is cited

OSA is able to provide readers links to articles that cite this paper by participating in CrossRef's Cited-By Linking service. CrossRef includes content from more than 3000 publishers and societies. In addition to listing OSA journal articles that cite this paper, citing articles from other participating publishers will also be listed.

Figures

Fig. 1 Fig. 2 Fig. 3
 
Fig. 4 Fig. 5
 

« Previous Article  |  Next Article »

OSA is a member of CrossRef.

CrossCheck Deposited