Three-dimensional Fluorescence Lifetime Imaging ... - OSA Publishing

Three-dimensional Fluorescence Lifetime Imaging with a Single Plane Illumination Microscope provides an improved Signal to Noise Ratio Klaus Greger,1,5 Manuel J. Neetz,2,5 Emmanuel G. Reynaud,1,3,* Ernst H.K. Stelzer4 1

Cell Biology and Biophysics Unit (CBBU), EMBL, Meyerhofstraße 1, D-69117 Heidelberg, Germany 2 MPI-CBG, Pfotenhauerstr. 108, D-01307, Dresden, Germany 3 School of Biology & Environmental Science, UCD Science Centre, Belfield, Dublin 4, Ireland 4 Goethe University Frankfurt, Max-von-Laue-Str. 960438, Frankfurt am Main, Germany 5 Contributed equally *[email protected]

Abstract: We designed a widefield frequency domain Fluorescence Lifetime Imaging Microscopy (FLIM)setup, which is based on a Single Plane Illumination Microscope (SPIM). A SPIM provides an inherent optical sectioning capability and reduces photobleaching compared to conventional widefield and confocal fluorescence microscopes. The lifetime precision of the FLIM was characterized with Rhodamine 6G solutions of different quencher concentrations [KI]. We demonstrate the high spatial resolution of the SPIM-FLIM combination in the intensity domain as well as in the lifetime domain with latex bead samples and multiple recordings of three-dimensional live Madine-Darby Canine Kidney (MDCK) cysts. We estimate that the bleaching rate after 600 images have been recorded is below 5%. ©2011 Optical Society of America OCIS codes: (170.2520) Fluorescence microscopy; (170.3650) Lifetime-based sensing; (180.6900) Three-dimensional microscopy.

References and links 1.

2. 3. 4.

5. 6.

7.

8.

9.

A. Periasamy, P. Wodnicki, X. F. Wang, S. Kwon, G. W. Gordon, and B. Herman, “Time-resolved fluorescence lifetime imaging microscopy using a picosecond pulsed tunable dye laser system,” Rev. Sci. Instrum. 67(10), 3722–3731 (1996). T. W. J. Gadella, Jr., T. M. Jovin, and R. M. Clegg, “Fluorescence lifetime imaging microscopy (FLIM): Spatial resolution of microstructures on the nanosecond time scale,” Biophys. Chem. 48(2), 221–239 (1993). H. Wallrabe and A. Periasamy, “Imaging protein molecules using FRET and FLIM microscopy,” Curr. Opin. Biotechnol. 16(1), 19–27 (2005). S. E. D. Webb, Y. Gu, S. Lévêque-Fort, J. Siegel, M. J. Cole, K. Dowling, R. Jones, P. M. W. French, M. A. A. Neil, R. Juškaitis, L. O. D. Sucharov, T. Wilson, and M. J. Lever, “A wide-field time-domain fluorescence lifetime imaging microscope with optical sectioning,” Rev. Sci. Instrum. 73(4), 1898–1907 (2002). M. J. Booth and T. Wilson, “Low-cost, frequency-domain, fluorescence lifetime confocal microscopy,” J. Microsc. 214(1), 36–42 (2004). R. R. Duncan, A. Bergmann, M. A. Cousin, D. K. Apps, and M. J. Shipston, “Multi-dimensional time-correlated single photon counting (TCSPC) fluorescence lifetime imaging microscopy (FLIM) to detect FRET in cells,” J. Microsc. 215(1), 1–12 (2004). D. M. Grant, J. McGinty, E. J. McGhee, T. D. Bunney, D. M. Owen, C. B. Talbot, W. Zhang, S. Kumar, I. Munro, P. M. Lanigan, G. T. Kennedy, C. Dunsby, A. I. Magee, P. Courtney, M. Katan, M. A. Neil, and P. M. French, “High speed optically sectioned fluorescence lifetime imaging permits study of live cell signaling events,” Opt. Express 15(24), 15656–15673 (2007). A. Deniset-Besseau, S. Lévêque-Fort, M. P. Fontaine-Aupart, G. Roger, and P. Georges, “Three-dimensional time-resolved fluorescence imaging by multifocal multiphoton microscopy for a photosensitizer study in living cells,” Appl. Opt. 46(33), 8045–8051 (2007). J. Huisken, J. Swoger, F. Del Bene, J. Wittbrodt, and E. H. K. Stelzer, “Optical sectioning deep inside live embryos by selective plane illumination microscopy,” Science 305(5686), 1007–1009 (2004).

#149898 - $15.00 USD

(C) 2011 OSA

Received 6 Jul 2011; revised 2 Sep 2011; accepted 5 Sep 2011; published 4 Oct 2011

10 October 2011 / Vol. 19, No. 21 / OPTICS EXPRESS 20743

10. K. Greger, J. Swoger, and E. H. K. Stelzer, “Basic building units and properties of a fluorescence single plane illumination microscope,” Rev. Sci. Instrum. 78(2), 023705 (2007). 11. A. Elder, S. Schlachter, and C. F. Kaminski, “Theoretical investigation of the photon efficiency in frequencydomain fluorescence lifetime imaging microscopy,” J. Opt. Soc. Am. A 25(2), 452–462 (2008). 12. A. D. Elder, J. H. Frank, J. Swartling, X. Dai, and C. F. Kaminski, “Calibration of a wide-field frequencydomain fluorescence lifetime microscopy system using light emitting diodes as light sources,” J. Microsc. 224(2), 166–180 (2006). 13. Q. S. Hanley, V. Subramaniam, D. J. Arndt-Jovin, and T. M. Jovin, “Fluorescence lifetime imaging: multi-point calibration, minimum resolvable differences, and artifact suppression,” Cytometry 43(4), 248–260 (2001). 14. P. J. Keller, F. Pampaloni, and E. H. K. Stelzer, “Three-dimensional preparation and imaging reveal intrinsic microtubule properties,” Nat. Methods 4(10), 843–846 (2007). 15. E. G. Reynaud, U. Krzic, K. Greger, and E. H. Stelzer, “Light sheet-based fluorescence microscopy: more dimensions, more photons, and less photodamage,” HFSP J 2(5), 266–275 (2008).

Fluorescence lifetime spectroscopy provides information about the local chemical environment of a fluorophore, e.g. pH, ionic strength or neighboring charge distributions. It is also able to distinguish spectrally similar fluorophores. The lifetimes of the commonly used fluorophores ranges from hundreds of ps to several ns. Fluorescence Lifetime Imaging Microscopy (FLIM) extends the spatial approach by including spectroscopic discrimination. Two-dimensional FLIM has been implemented in the time domain [1] and in the frequency domain [2]. As a part of a modern microscope it localizes specific molecular interactions, e.g. via Fluorescent Resonant Energy Transfer (FRET) [3], and provides information complementary to intensity images. Measuring lifetimes of three-dimensional fluorophore density distributions is of great interest for tissue engineering and three-dimensional cell cultures. Several approaches to FLIM with optical sectioning capabilities have been realized but three-dimensional lifetime maps are still difficult to achieve. Widefield FLIM with structured illumination does not provide a sufficient signal-to-noise ratio (SNR) for biological applications [4] and all scanning approaches are still limited by long recording times and high photo bleaching rates [5, 6]. The combination of a CCD camera with a spinning Nipkow disc [7] or a multifocal multiphoton microscope [8] may improve the acquisition speed but does not provide for the necessary sensitivity. We implemented a frequency domain FLIM, which is suitable for widefield imaging, with a Single Plane Illumination Microscope (SPIM) [9, 10]. In frequency domain FLIM the intensity of the incident excitation light and hence the fluorescence emission is usually modulated sinusoidally with a frequency between 20 and 80 MHz [e.g 2,5.]. Since fluorescence excitation is a resonant effect, i.e. energy is stored, the emission intensity exhibits a phase shift φ and a demodulation m relative to the excitation intensity. The fluorescence emission detection system is modulated with the same frequency as the excitation light intensity. We recorded images with 6-12 equidistant phase shifts relative to the detection and the excitation modulations. An absolute minimum of three phase shifts is necessary. It is, however, important not to undersample in order to detect the peaks of the sinusoidal properly [11]. In the case of a sinusoidal excitation and a mono-exponential decay the phase shift φ and the demodulation m can be calculated by employing only the first harmonic components of the emission signal. The angular modulation frequency ω, the modulation depth mref and the initial phase delay φref resulting from a reference measurement are known [12]. The reference measurement serves to characterize the delay between emission and detection due to intrinsic properties of the optical system. Two different lifetimes of the investigated fluorophore can then be calculated (Eqs. (1) and (2), see also [12]). 2  1  mref  2 2 τ m =   (1 + ω τ ref ) − 1 ω  mem  

#149898 - $15.00 USD

(C) 2011 OSA

1/ 2

(1)



τϕ =

1

ω

tan ϕ ref − ϕ em + tan −1 (ωτ ref ) 

(2)

τm and τφ refer to the lifetimes corresponding to the change of the modulation depth and the phase shift, respectively. φem and mem refer to the phase shift and the modulation of the emission signal of the sample. τref refers to the lifetime of a reference sample, which is usually a scattering object (τref = 0ns). In principle, every sample with a well defined monoexponential decay can be used for calibration purposes. (See [12] for a detailed calibration procedure). The measurement of one lifetime image requires at least three recordings with different phase shifts. The theoretical sensitivity of the lifetime measurements depends strongly on the modulation frequency and the optimal results are obtained by differentiation of Eq. (1) and (2) and yield Eq. (3) and (4) (see also [12]).

ω opt , m = 2 / τ

(3)

ω opt , ϕ = 1 / τ (4) In a SPIM a thin light sheet illuminates only the focal plane of the detection optics [9,10]. The beam shape and the numerical aperture (NA) of the illumination are adapted to the field of view with a beam expander and two slits in front of a cylindrical lens, which focuses the collimated laser beam in the focal plane of the detector. The other axis remains collimated. In a SPIM only the observed, i.e. focal, plane is exposed to the illuminating laser light. This is a major and very significant difference to conventional widefield and confocal fluorescence microscopy. Thus SPIM provides true optical sectioning but without generating out-of-focus light. This advantage can be quantified by the ratio of the widths of the light sheet (WLS) and the sample (WSA), e.g. in our application to MDCK cysts, that is WLS / WSA = 1.79µm / 50 µ m = 1 / 28. SPIM provides, therefore, a much more efficient application of fluorescence microscopy, which is often limited by bleaching rates (e.g. in FLIM). We use a LDH-M-C-470 laser diode (470 nm) in the excitation path and a MDL 300 driver (PicoQuant, Germany) to generate the modulated illumination. The detection arm consists of a microscope objective lens (100x, NA 1.0 W, Carl Zeiss AG, Germany), an emission filter, a tube lens, a gated image intensifier (II 18 MDS, Lambert Instruments, Netherlands) and a CCD camera (ORCA ER, Hamamatsu Photonics K.K., Japan) (Fig. 1). The field of view is defined by the effective magnification of the detection system (here 50x) and spans in our setup 173.4x132µm2. The light sheet is set in such a way, that it is a factor of sqrt(2) wider at the edges of the field of view than in the center; here it measures 1.79µm in the center. The lateral resolution of the setup was determined by fitting Gaussian curves to the images of 200 nm beads (Yellow Green FluoSphere, Molecular Probes). The full width at half maximum (FWHM) for the lateral axes was determined to be 0.79 µm (vertically) and 0.82 µm (horizontally), respectively. The laser diode and the intensifier are controlled by two frequency generators, which are integrated into the control unit of the intensifier. They provide sine waves with a computer controllable phase shift.

#149898 - $15.00 USD

(C) 2011 OSA



Fig. 1. In the SPIM-FLIM setup a laser diode illuminates the sample with a modulation frequency in the MHz range. The beam expander (BE) allows together with the cylindrical lens (f = 50mm) to form the light sheet. The detection lens dips directly into the medium of the sample chamber and no additional glass surfaces are between the sample and the detection lens. The fluorescence emission signal passes through the infinity corrected space (ICS), is spectrally filtered and enters the gated image intensifier. The final signal is integrated by the camera. The gain of the gated image intensifier and, therefore, the sensitivity of the detection system are modulated with the same frequency as the excitation signal. The sample is mounted on precision stages, which allow rotation and translation. A standard PC is used to control the hardware, acquire the images and calculate lifetime images for each plane.

The SPIM-FLIM was calibrated by measuring the lifetime of Rhodamine 6G (R4127, Sigma-Aldrich Co., Germany) solutions with different concentrations of the quencher potassium iodide [KI] (60399, Sigma-Aldrich Co.). Rhodamine 6G is reported to have a mono-exponential [13] decay and has been used in fluorescence spectroscopy for at least four decades. The dye samples were mounted in small bags of the transparent polymer Polytetrafluorethylene (PTFE), which has a refractive index similar to that of water [14]. All samples (e.g. plastic bags) in a SPIM are mounted on a precision scanning and rotation stage [10]. We chose a 1mM Rhodamine 6G solution with seven different concentrations of the quencher [KI] (0 to 0.1 M), where the ionic strength was kept constant by adding the appropriate amount of potassium chloride [KCl] (60128 Sigma-Aldrich Co.). The relationship between the lifetime τ and the quencher concentration [Q] is given by the Stern-Volmer equation (Eq. (5)

1 / τ = kq[Q] + 1 / τ 0,

(5)

where kq is the bi-molecular rate constant and τ0 the fluorescent lifetime without quencher [13]. Fitting this linear relationship to the measured lifetime values (60MHz, 12 phase shifts, binning 4) yields a value of 6.83 x 109 / (Ms) according to the modulation depth (R2 = 0.93) and 5.70 x 109 / (Ms) according to the phase shift (R2 = 0.97) for the bi-molecular rate constant (Fig. 2.). The results of the single measurements are given in Table 1, which show a good agreement with the values obtained by Hanley et al. [13].

#149898 - $15.00 USD

(C) 2011 OSA



Fig. 2. The lifetimes for aqueous 1mM Rhodamine 6G solutions with seven different concentrations of the quencher potassium iodide [KI] are shown. The linear fit of the SternVolmer-Plot verifies the expected mono-exponential behavior of the dye solution.

Table 1. Lifetimes of aqueous Rhodamine 6G solutions as a function [KI]. KI] [M] τmean [ns]1

0 0.005 0.01 0.03 0.04 0.05 0.1 4.08 ± 3.60 ± 3.18 ± 0.02 2.26 ± 0.02 1.97 ± 0.01 1.67 ± 0.02 1.05 ± 0.02 0.02 0.01 τm [ns] 4.01 ± 3.51 ± 3.12 ± 0.20 2.31 ± 0.19 2.08 ± 0.20 1.78 ± 0.20 1.26 ± 0.27 0.24 0.21 τφ [ns] 3.78 ± 3.53 ± 3.12 ± 0.25 2.17 ± 0.16 1.87 ± 0.15 1.66 ± 0.13 1.07 ± 0.13 0.35 0.31 τmean [ns] 3.90 ± 3.52 ± 3.12 ± 0.00 2.24 ± 0.07 1.98 ± 0.11 1.72 ± 0.06 1.17 ± 0.11 0.12 0.01 + 0.02 + 0.01 + 0.05 + 0.12 ∆τmean [ns] −0.18 −0.08 −0.06 + 0.01 + 0.01 + 0.03 + 0.10 ∆τmean\τmean −0.05 −0.02 −0.02 In binning mode n the charge of nxn pixels is integrated prior to current-voltage conversion. The errors for τm and τφ are standard deviations, the errors for τmean are the absolute values of τmean- τm. ∆τmean is the difference between our own values for τmean and those from [12].

In order to show the capability of the setup for distinguishing lifetimes within single images and three-dimensional sets of images, we mounted a mixture of beads with different diameters (10 µm, Yellow Green Fluoresbrite, Polysciences, Inc., USA; 2.47µm, Yellow Green FluoSphere, Molecular Probes, USA) in a cylinder of 1% low-melt agarose. We employed the binning mode to increase the framerate (and lower photobleaching) for the acqusition of whole stacks. In the binning mode n the intensity of n x n pixel is summed up. The evaluation of 155 planes (step size 330nm, binning 4) yields lifetimes (mean ± standard deviation) of τm = (2.56 ± 0.30) ns and τφ = (2.42 ± 0.24) ns for the larger beads while the according values for the smaller beads are τm = (3.74 ± 0.67) ns ns and τφ = (3.76 ± 0.52) ns. Furthermore, sections along the lateral and the optical axes show clearly that the setup is able to distinguish lifetimes along all three dimensions (see Fig. 3).

#149898 - $15.00 USD

(C) 2011 OSA



Fig. 3. Interpolation of the lifetime values for different sized beads in the lateral plane (left) and along the z-axis (right). For simplicity only the lifetime according to the phase shift is shown. The lifetimes of the different beads can be distinguished clearly along all three axes. An intensity threshold was applied to exclude the lifetime values corresponding to the background form the area outside the beads.

The ability of the FLIM -SPIM setup to record lifetime maps of large living specimens was demonstrated by investigating Madine-Darby Canine Kidney (MDCK) cysts with EGFP fused to E-Cadherin (Fig. 4) as a lateral membrane marker. The MDCK cysts were embedded in Matrigel and mounted in small PTFE chambers [14]. While the conventional mounting of a sample between two coverslips confines the sample geometry to two dimensions, the sample mounting in the SPIM maintains the physiological environment of cells to a greater extent. Relevant conditions can be realized by adding nutrients or drugs to an aqueous solution surrounding the gel cylinder. During the experiment, the temperature of the surrounding CO2 independent medium was kept constant at 37°C. The mean lifetimes of the whole sample volume were τm = (3.06 ± 0.89) ns and τφ = (2.62 ± 0.90) ns (see Table 2). Table 2. Lifetimes of E-GFP fused to E-Cadherin in MDCK cysts Recording

τm [ns]

τφ [ns]

1 3.06 ± 0.89 2.62 ± 0.90 2 3.04 ± 0.89 2.62 ± 0.89 3 3.14 ± 0.93 2.64 ± 0.90 f = 60MHz, six phase shifts, binning 2, 100 planes per stack with 0.5µm standard deviation.

τmean [ns] 2.84 ± 0.02 2.83 ± 0.20 2.89 ± 0.25 spacing. All values are given as mean ±

The relatively large standard deviation should be due to the non-uniformity of lifetimes in large biological specimen. The thickness of the sample has no effect on the lifetime measurement since the mean lifetime per plane does not change along any of the three axes (data not shown).

#149898 - $15.00 USD

(C) 2011 OSA



Fig. 4. Lifetime images of a MDCK cyst expressing EGFP E-Cadherin (60 MHz, 6 phase shifts, binning 2, 100 planes with 500nm spacing). The lifetime images are an overlay of the intensity images (normalization per plane) with a color coded lifetime map. From top to bottom 251x251 pixels cutouts of the planes 11, 66 and 82 are shown.

The combination of widefield illumination with inherent optical sectioning results in a high number of detected photons per pixel and a low background intensity. The SNR is proportional to the number of detected photons and thus highly increased for the SPIM compared to other three-dimensional fluorescence microscopes [15]. This is in particular true for the SPIM-FLIM compared to other three-dimensional FLIM methods. The multiple recordings (for different phase shifts or modulation depths) introduce photobleaching and thus deteriorate the SNR further. This photobleaching effect is much stronger for conventional approaches than for the SPIM [15]. In our experiments multiple recordings of the whole sample volume did not cause noticeable photobleaching. We could show very clearly that high quality three-dimensional data sets are easily recorded. Different detection schemes with non-sinusoidal excitation can further increase the photo-counting efficiency [11]. The SNR can also be improved by using a generation III image intensifier with a quantum efficiency of 25% while a common generation II image intensifier, as used in our setup, provides only 10% quantum efficiency. Since all widefield frequency domain FLIM setups are limited by such technicalities, the SPIM-FLIM provides an advantage, which can be utilized in the form of lower bleaching

#149898 - $15.00 USD

(C) 2011 OSA



rates and faster acquisition times. The recording time for a lifetime image (60MHz, 6 phase shifts) of 1024 x 1344 pixels was three seconds for the E-GFP fused to E-Cadherin expressed MDCK cysts. Acquiring the images directly in binning mode allows recording times of (much) less than a second for a single lifetime image or less than a minute for a sample volume of (100µm)3. This demonstrates the potential of the FLIM-SPIM setup to study dynamic changes of lifetimes in three dimensions at a high rate.

Acknowledgments We are grateful to Lambert Instruments (Netherlands) and PicoQuant (Germany) for instrumentation support. The authors would like to thank Anthony Squire (formerly EMBL) and Malte Wachsmuth (EMBL) for discussions and the mechanical and electronics workshops at EMBL for their technical assistance.

#149898 - $15.00 USD

(C) 2011 OSA

