Nanoscale Resolution in Far-field Fluorescence Microscopy Part I Basics and Axial Resolution Improvement Stefan W. Hell Max Planck Institute for Biophysical Chemistry Department of NanoBiophotonics Göttingen, Germany
Fluorescence microscopy importance in the life sciences: Electron microscopy, etc. 20 % Ratio 80 % Fluorescence Microscopy
Fluorescence excitation and emission of labels: S 1 1-5 ns λ exc λem S 0 Synth. organic fuorophore Fluoresc. Proteins (GFP)
500 nm Δx = λ 2n sin α α 200 nm λ Wavelength Lens λ S 1 λ fl S 0
Sheppard, Kompfner, Appl Opt. (1978) Denk, Strickler, Webb, Science (1990) 500 nm Δx = λ 2n sin α α 200 nm λ Wavelength Lens ~I(r/2) ~I(r) 2 2 ~I(r) Multiphoton processes S 1 2λ λ don t work for resolution improvement! S 0 S 0 2λ S 1 λ fl
Z- resolution improvement
4Pi- Microscopy: 70-140 nm r r r (,, ϕ ) = (,, ϕ) + (,, ϕ) 4Pi E r z E1 r z E2 r z Coherent illumination and/or fluorescence detection Hell, Europ Patent OS 0491289 (1990) Hell, et al JOSA A (1992) Hell, et al, Appl. Phys. Lett., 64, 11 (1994)
1997 4Pi type A Detector 2 photon 750 nm Piezo Detector r r E1 + E2
~I(r/2)2 ~I(r) 2-Photon Exc. 1-Photon Exc. Z Z Excitation 800 nm Excitation 488 nm X X Detection 510 nm Detection 510 nm Effective PSF Effective PSF
Z Confocal 4Pi X 1μm Green and red beads, 100 nm diameter
Z Microtubules, mouse fibroblast Immunofluor, Oregon Green 2 µm 2 µm X confocal 4Pi
How does the linear lobe filter really work?
1-Photon PSF 2-Photon Z Z Excitation 488 nm Excitation 800 nm X. X. Detection 510 nm Detection 510 nm = = Consequence of outer part suppresion is Effective PSF PSF separable in (XY) and (Z) Effective PSF h4 Pi ( x, y, z) c( x, y) h( z) Z Z
Separability: h4 Pi ( x, y, z) c( x, y) h( z) -> treat the lobe problem separately, in the z-direction Z X = Z =
Real space 1.25 algebraic inversion 1 1 = 0 z -0.42 z 0 z Frequency domain. = 1 kz kz 0 k z
Real space Filtering the PSF = z Frequency domain. = k z
Z 4Pi confocal = X z 4Pi 1D-FFT 1D-FFT -1. = k z
Non-linear image restoration: - Know the full PSF (not just height and position of lobes) - Positivity of image (and PSF) assumed Examples: Richardson-Lucy Algorithm (RL) Maximum-Likelihood Estimation Algorithm
Z Confocal 4Pi X Confocal + Restoration (RL) 4Pi + Restoration (RL) 1μm
Z Microtubules / Mouse fibroblast Immunofluor. Oregon Green 100 nm 100 nm X Confocal 4Pi + restor. (RL) Confocal 4Pi + restor. (RL)
Higher Speed: Multifocal 4Pi - microscopy - 4Pi-foci in parallel: 16-64 - 1 Second / xy-image - Water immersion Microlens Array Egner,Jakobs & Hell, PNAS, 99, 3370 (2002)
Application: Mitochondrial network in S. cerevisiae 3D-reconstruction Electron Microscopy B. J. Stevens & J. G. White, 1975
GFP 4Pi microscopy Mitochondrial compartment in live yeast @ 100 nm 3D-resolution Egner,Jakobs,Hell, PNAS, (2002)
Number of nodes increases 4-fold on Glucose on Glycerol 5.8 ± 2.1 23.0 ± 9.3 Quantitative live cell 4Pi microscopy
Diameter increases by 20 nm on average on Glucose on Glycerol 339 ±4 nm 359 ±5 nm Quantitative live cell 4Pi microscopy Surface 2.8 x larger
4Pi Image Golgi in live Vero Cell UDP-Galactosyltransferase- EGFP z x 2.5 mm Egner & Hell, TiCB (2004)
4Pi Image Golgi in live Vero Cell UDP- Galactosyl- transferase- EGFP z 2.5 mm x 3D-resolution: ~100 nm
Practical issues
Practical issues 1) Alignment of lenses Dark exit of Sagnac interferometer Piezo closed loop alignment (< 50 nm precision)
Checking out phase changes and aberrations...induced by live yeast cell Δφ n ( r, z) dz Yeast Cell Cover Slip Water immersion Fluorescent layer
Checking out object induced phase changes and aberrations... Water Glass Shift in absolute position < 100 nm Yeast Cell 2 μm.
Practical issues 1) Alignment of lenses 2) Need establish PSF (phase, lobes): point or planar objects in sample 3) Refractive index changes: index matching (water lenses) Z erbb1-egfp In Chinese Hamster Ovary (CHO) Cells 18 x 33 x 16 µm (x,y,z)
4Pi type C (coherent detection + illumination)
1997 4Pi type C Detector r r E1 + E2 2 photon 750 nm Piezo Detector r r E1 + E2
Z-Response of a layer (Experiment) Confocal 4 Pi (A) Illumination x z 1.0 0.8 0.6 0.4 0.2 500 nm 0.0 1.0 0.8 0.6 0.4 0.2 Exp. 2-Photon / 840 nm NA = 1.35 Oil Theory 525nm -1.0-0.5 0.0 0.5 z[µm] 1.0 135nm 0.0-1.0-0.5 0.0 0.5 z[µm] 1.0 4 Pi (B) Detection 1.0 0.8 0.6 0.4 127nm 0.2 0.0-1.0-0.5 0.0 0.5 z[µm] 1.0 4 Pi (C) Illumination + Detection 1.0 0.8 0.6 0.4 96nm 0.2 0.0-1.0-0.5 0.0 0.5 z[µm] 1.0
Z-Response of a layer (Theory) Confocal z 1.0 0.8 0.6 0.4 2-Photon / 840 nm NA = 1.35 Oil 520nm x 500 nm 0.2 0.0-1.0-0.5 0.0 0.5 z[µm] 1.0 4 Pi (A) Illumination 1.0 0.8 0.6 0.4 132nm 0.2 0.0-1.0-0.5 0.0 0.5 z[µm] 1.0 4 Pi (B) Detection 1.0 0.8 0.6 0.4 0.2 122nm 0.0-1.0-0.5 0.0 0.5 z[µm] 1.0 4 Pi (C) Illumination + Detection 1.0 0.8 0.6 0.4 95nm 0.2 0.0-1.0-0.5 0.0 0.5 z[µm] 1.0
Compact 4Pi-microscope Z- resol < 100 nm (live cells /aqueous cond.) 1 & 2 photon
How essential is the use of focused (spherical) wavefronts? 4Pi Isn t it just about making a focal interference standing wave pattern?
Standing Wave microscopy Non-Confocal! Bailey et al, Nature (1993) Excitation: flat standing wave Detection: epifluorescence No 4Pi element Does not work for z-resolution improvement! PSF x z exc. det. eff. z-profile k z OTF k x k z -profile
I 5 M Non-Confocal! Gustafsson et al, 1997 Conceptually fine, but not very robust Excitation: flat standing wave Detection: 4Pi, spherical waves PSF x z k z exc. det. eff. z-response k z -profile OTF k x
Comparison of the effective PSFs z conventional x confocal LUT 0 100% 2ph 4Pi (A) standing wave 1ph 4Pi (C) I 5 M 2ph 4Pi (C)
Comparison of the effective OTFs k z conventional k x confocal 2ph 4Pi (A) LUT 0 100% standing wave 1ph 4Pi (C) I 5 M 2ph 4Pi (C)
PSFs: Overview Widefield z Confocal SWM I 5 M 4Pi C 2 hν 4Pi A 2 hν 4Pi C r Excitation PSF Detection PSF PSF LUT 0 50% 100% 4% 2% z z z z z z z z z z z z z z
Widefield Confocal SWM I 5 M 4Pi C 2 hν 4Pi A 2 hν 4Pi C k z OTFs: Overview k r Excitation OTF Detection OTF OTF k z k z k z k z k z k z k z
No. Essential physical element: increase in solid angle ( 4π ) α 68 75 Isn t it just about making a focal interference standing wave pattern?
Two opposing lenses give 3-7 fold increased z-resolution 60-150 nm..provided that they increase the aperture angle of the microscope
Nanoscale Resolution in Far-field Fluorescence Microscopy Part II Breaking the diffraction barrier Stefan W. Hell Max Planck Institute for Biophysical Chemistry Department of NanoBiophotonics Göttingen, Germany
500 nm Δx = λ 2n sin α α 200 nm λ Wavelength Lens
Jena, Germany
"Breaking the diffraction resolution limit by stimulated emission" Hell & Wichmann, Opt. Lett., 19, 11, (1994) axial > 500 nm Δx = λ 2n sin α Z a Airy-Disk X lateral > 200 nm
Stimulated Emission Depletion (STED) S1 4 3 τ fl 1ns Fluorescence Excitation Stimulated Emission S0 2 1 τ vib p 1ps Fluoresc. Spectrum Exc Stim. Emission ~ 200 ps ~50 ps 2 Pulses 500 600 λ [nm] Stim. Em. Excit. t
STED microscope y x z focal spot 1 µm Detector 50 ps 200 ps Excitation Sync Depletion (STED) Fluorescence 1.0 0.5 0 0.5 1.0 1.5 time [s] S 1 Absorption S 0 Fluorescence τ 1ns fl Stimulated Emission τ vib p 1ps Fluorescence 1.0 0.5 0 ON OFF 2 4 6 I STED [GW/cm 2 ]
STED microscope y x z x 200 nm Detector PhaseMod 2π 0 y 50 ps 200 ps Excitation Depletion (STED) Fluorescence 1.0 0.5 0 2 4 6 I STED [GW/cm 2 ]
STED microscope y x z x 200 nm Detector PhaseMod 2π 0 y 50 ps 200 ps Excitation Depletion (STED) Fluorescence 1.0 0.5 0 2 4 6 I STED [GW/cm 2 ]
Fluorescence 1.0 0.5 0 I I s 20 2 4 6 I STED [GW/cm 2 ] Focal spot... probed with 1 molecule STED 48 nm 200nm Confocal 254 nm -250-150 -50 50 150 250 x [nm] λ STED = 770 nm Westphal, Hell, PRL (2005)
40 nm beads (Molecular Probes) Confocal STED... just physics! 10 counts/0,3ms 204 5 counts/0,3ms 89 1µm X Y
Biological imaging. (Immunofluorescence)
SNAP- 25 / plasma membrane...just physics! Confocal: STED: K. Willig et al, NJP (2006)
Syntaxin 1A/ plasma membrane...just physics! c d Confocal STED 9 x 9 µm e f g Sample: T. Lang, J. Sieber, R. Jahn
Neuronal communication Presynaptic Synaptic bouton Synaptotagmin I antibody labeled Postsynaptic Confocal STED New biology: Synaptic vesicle protein synaptotagmin I is clustered after exocytosis Y X Willig, Rizzoli, Jahn, Hell, Nature (2006)
Nonbiological imaging.
65 nm Colloids: Silica beads with fluorescent core 130 nm Confocal STED X Y 0 counts/0,2ms 210 500nm 5 counts/0,2ms 80 Sample by A van Blaaderen Willig et al, New J Phys (2006)
65 nm Colloids: Silica beads with fluorescent core 130 nm Confocal + Lin Deconv. STED + Lin Deconv.... + maths! X Y 500nm Willig et al, New J Phys (2006)
Confocal + Lin. Deconv. STED + Lin. Deconv.... + maths! Field: (6x6) µm; Pitch 120 nm; Line width: 60 nm E-beam lithography on dyed PMMA Westphal et al, J Phys B (2005)
STED microscope y x z x 200 nm Detector y PhaseMod 2π 0 Sync 1.0 Excitation Depletion (STED) Fluorescence 0.5 0 2 4 6 I STED [GW/cm 2 ]
STED microscope y x z x 200 nm Detector y PhaseMod 2π 0 Sync 1.0 Excitation Depletion (STED) Fluorescence 0.5 0 2 4 6 I STED [GW/cm 2 ]
STED microscope y x z x 200 nm Detector y PhaseMod 2π 0 Sync 1.0 Excitation Depletion (STED) Fluorescence 0.5 0 2 4 6 I STED [GW/cm 2 ]
STED microscope y x z x 200 nm Detector y PhaseMod 2π 0 Sync 1.0 Excitation Depletion (STED) Fluorescence 0.5 0 2 4 6 I STED [GW/cm 2 ]
STED microscope y x z x 200 nm Detector y PhaseMod 2π 0 Sync 1.0 Excitation Depletion (STED) Fluorescence 0.5 0 2 4 6 I STED [GW/cm 2 ]
In far-field field fluorescence microscopy: Diffraction resolution barrier does no longer exist! Molecular scale resolution is possible with visible light and regular lenses. With a rectangular depletion curve, the resolution could be increased to infinity. Hell & Wichmann, Opt. Lett. 19 (1994)
STED Microscopy y x z x 200 nm Detector y λ /2 Phaseplate Grating SHG Delay OPO Ti:Sa Depletion (STED) Excitation Westphal & Hell, PRL 94 (2005)
1.0 0.5 I I I I s 20 90 sat... 14 x below Abbe Fluorescence 0.0 0 3 6 9 I STED λ STED = 770 nm STED 16 nm λ STED / 46 STED 200nm Confocal -250-150 -50 50 150 250 x [nm] Westphal & Hell, PRL 94 (2005)
Triplet Relaxation STED microscopy..provides <20 nm resolution (x,y) Donnert et al, PNAS (2006)
Heavy subunit of neurofilaments in neuroblastoma Donnert et al, PNAS (2006)
Axial resolution improvement by STED - 4Pi microscopy
S.W. Hell, in Topics Fluoresc Spectr V, Plenum Press (1997)
Microtubules / HEK cell Immunofluorescence XY-Overview Confocal STED-4Pi... 15 x improved 53 nm Monolayer M. Dyba &SWH, Nature Biotech 21, (2003)
How does resolution scale with intensity?
Resolution gain with intensity I/ I S 0 50 100 150 200 221nm Resolution gain ~ I I s 10 FWHM r / [nm] 118nm 95nm 60nm 44nm 34nm 27nm 22nm Resolution gain 0 0 0.5 1.0 1.5 Intensity (STED) / [GW/cm²] 0 I S ~30 MW /cm 2
The basic principle...
BRIGHT fluorescent absorbing trans A k AB = σ I( r) k BA = optical, thermal... B DARK Non-fluorescent Non-absorbing cis 1.0 1.0 N B / N A B 0.5 0.5 0.0 0.0 0.5 1.0 r / λ 0.0 0 I( r) A I s = k BA σ
BRIGHT fluorescent absorbing trans A k AB = σ I( r) k BA = optical, thermal... B DARK Non-fluorescent Non-absorbing cis 1.0 1.0 N B / N A B 0.5 0.5 0.0 0.0 0.5 1.0 r / λ 0.0 0 I( r) A I s = k BA σ
BRIGHT fluorescent absorbing trans A k AB = σ I( r) k BA = optical, thermal... B DARK Non-fluorescent Non-absorbing cis 1.0 1.0 N B / N A B 0.5 0.5 0.0 0.0 0.5 1.0 r / λ 0.0 0 I( r) A I s = k BA σ
BRIGHT fluorescent absorbing trans A k AB = σ I( r) k BA = optical, thermal... B DARK Non-fluorescent Non-absorbing cis 1.0 1.0 N B / N A B 0.5 0.5 0.0 0.0 0.5 1.0 r / λ 0.0 0 I( r) A I s = k BA σ
BRIGHT fluorescent absorbing trans A k AB = σ I( r) k BA = optical, thermal... B DARK Non-fluorescent Non-absorbing cis 1.0 1.0 N B / N A B 0.5 0.5 0.0 0.0 0.5 1.0 r / λ 0.0 0 I( r) A I s = k BA σ
BRIGHT fluorescent absorbing trans A k AB = σ I( r) k BA = optical, thermal... B DARK Non-fluorescent Non-absorbing cis 1.0 1.0 N B / N A B 0.5 0.5 0.0 0.0 0.5 1.0 r / λ 0.0 0 I( r) A I s = k BA σ
BRIGHT fluorescent absorbing trans A k AB = σ I( r) k BA = optical, thermal... B DARK Non-fluorescent Non-absorbing cis 1.0 1.0 N B / N A B 0.5 0.5 0.0 0.0 0.5 1.0 r / λ 0.0 0 I( r) A Δ x π n λ I I s I k BA s = = σ 1 στ BA
BRIGHT fluorescent absorbing trans A k AB = σ I( r) k BA = optical, thermal... B DARK Non-fluorescent Non-absorbing cis 1.0 1.0 N B / N A B 0.5 0.5 0.0 0.0 0.5 1.0 r / λ 0.0 0 I( r) A Δ x π n λ Iστ BA I k BA s = = σ 1 στ BA
BRIGHT fluorescent absorbing trans A k AB = σ I( r) k BA = optical, thermal... B DARK Non-fluorescent Non-absorbing cis STED microscopy S 1 A Absorption Fluorescence τ ~ 1 ns k AB = σ I( r) S 0 B 1 16 2 Is = 10 cm 10 στ 9 s I s 30 MW/cm 2
BRIGHT fluorescent absorbing trans A k AB = σ I( r) k BA = optical, thermal... B DARK Non-fluorescent Non-absorbing cis There must also be Other Molecular States & transitions to break the diffraction barrier
BRIGHT fluorescent absorbing trans A k AB = σ I( r) k BA = optical, thermal... B DARK Non-fluorescent Non-absorbing cis Ground State Depletion Microscopy Confocal 40 nm beads Ground State Depletion Micr. t~20ns t ~ 1 ns t~1µs A Singlet B Triplet 500nm S. Bretschneider, C. Eggeling, SWH Phys Rev Lett (2007) Hell & Kroug, Appl. Phys. B 60 (1995) I s ~ kw/cm 2
BRIGHT fluorescent absorbing trans A k AB = σ I( r) k BA = optical, thermal... B DARK Non-fluorescent Non-absorbing cis Optical Bistability (Synthetic compounds) A fluorescent dark B Δ x π n λ Iστ BA trans - O 3 S cis - - O 3 S SO 3 N O O N O O N N + O N O O O O O N O O + N - SO 3 O O N O O I s ~ W/cm 2 Photoisomerisation Hell, in Topics Fluoresc Spectr V, Plenum Press (1997) Dyba & Hell, Phys. Rev. Lett. 88 (2002) Hell, Jakobs, Kastrup, Appl. Phys. B 77 (2003)
BRIGHT fluorescent absorbing trans A k AB = σ I( r) k BA = optical, thermal... B DARK Non-fluorescent Non-absorbing cis Optical Bistability (Fluorescent proteins) A fluor dark B Δ x π n λ Iστ BA I s ~ W/cm 2 Photoswitchable fluorescent proteins Hell, Jakobs, Kastrup, Appl. Phys. B 77 (2003) Hofmann, et al PNAS (2005)
Fluor. On Fluor. Off Object 0.6 λfl nsinα Read CCD-camera Out Image
Fluor. On Fluor. Off Object 0.6 λfl nsinα S 1 A Absorption Fluorescence S 0 B Read CCD-camera Camera detection possible. Zeros & scanning are required! Out Image MGL Gustafsson, PNAS (2005)
Sequential read-out with at least one Intensity Zero + Reversible Saturable/Switchable Optical Linear (Fluorescence) Transition = IZ RESOLFT
RESOLFT Switching molecule ensembles (confinement by zeros) STORM / PALM Switching molecules individually! Object 0.6 λfl nsinα A B 0.6 λfl nsinα SW Hell Nature Biotech (2003) Betzig, Hess et al Science Rust, et al Nature Meth Hess et al Biophys J (2006) Read CCD-camera Read CCD-camera Out Image Out Center Gravity
RESOLFT Switching molecule ensembles (confinement by zeros) STORM / PALM Switching molecules individually! Object 0.6 λfl nsinα A B 0.6 λfl nsinα SW Hell Nature Biotech (2003) Δx λ 2nsinα N CCD-camera CCD-camera Read Out Image Read Out
What is the common enabling physical element behind all these approaches?
Establishing a state A contrasted with B in a sub-λ/2 region moved through the object and detected sequentially A A B B B B B B B B B B Δr<<λ/2 BB B B B B B A B B B B
Energy state: S 0, S 1, T 1,.. S 1 Fluorescence T 1 (Photophysics) Absorption A B Conformational states: cis, trans, binding S 0 (Photochemistry) A need not be fluorescent, but just detectable (e.g. heat). Long lifetimes of states τ are advantageous Key is the states, not the density of photons!
m-photon absorption does not really work for resolution! S 1 2- photons 2 λ 2 λ Fluorescence α ~ I 2 S 0 STED S 1 A 1.0 Absorption Fluorescence Stimulated emission 0.5 ~ exp (-Iτσ) S 0 B 0 2 4 6 I STED [GW/cm 2 ]
... Abbe s equation? 1.0 0Signal 2 Intensity 4 6 Concepts using zeros...expanded. SWH, Nature Biotech 21, 1347 (2003) SWH, Phys. Lett. A 326, 140 (2004) Westphal, SWH, Phys. Rev. Lett. (2005)
Acknowledgements References Pictures/Movies www.nanoscopy.de SW Hell (Review article) Science, May 25, (2007) Physics: G. Donnert et al PNAS, July 24 (2006) K Willig, J Keller, M Bossi, SWH New J Phys, (2006) V Westphal, SWH Phys Rev Lett, 94 (2005) L Kastrup, H Blom, C Eggeling, SWH Phys Rev Lett, 94 (2005) Applications: K Willig, S Rizzoli, R Jahn, SWH Nature, April 13, (2006) R Kittel, et al Science, May 19, (2006)