Green fluorescent protein (GFP) from Aqueorea Victoria  and its multicolored variations on the theme are among the most routinely fluorescent tracers used for biological visualization . Interest has grown in more precise localization studies of protein activity and movement within a cell and we could say that a new revolution started with the advent of photoactivatable fluorescent proteins [3, 4]. Fluorescence of proteins effectively brought a "new light" in molecular and cellular biology studies [5, 6], the "fluorescence toolbox" is growing  and steps towards macromolecular-scale resolution, using optical microscopes, are becoming reality . Within this pivotal scenario we focused on pa-GFP as photoactivatable fluorescent protein , and on the indispensable tools offered by confocal and two-photon microscopy [9–11]. Particle tracking inside the cell largely benefits from the ability to spatially and temporally mark specific structures to follow their "signalling" over a "dark" background as made possible since the advent of pa-GFP. Pa-GFP results from a site specific mutagenesis substituting threonine 203 with histidine in wild-type GFP. This leads an excellent photo-convertible molecule producing up to a 100 fold increase in 488 nm excited fluorescence after irradiation with high energy flux at 405 nm . Selective photoactivation by means of confocal microscopy immediately demonstrated that pa-GFP can be considered an optimal tool to study spatial and temporal dynamics of proteins in vivo, as tracking of lysosome and mitochondria [3, 4]. At the same time, when one is performing experiments using pa-GFP, one the first key aspects is related related to the ability to perform spectral fingerprinting . This is particularly useful when considering the "sea of autofluorescent molecules" present within living cells and tissues .
Now, in terms of spatial confinement of the photo activation process, the use of two-photon or even multiphoton excitation [14, 15] provides several favorable aspects compared to single photon confocal microscopy in photomarking biological structures to be tracked [16–18]. The highly confined excitation volumes, of the order of magnitude of subfemtoliter, due to the non-linear requirements provide a unique control of the excitation and consequently photoactivation in the 3D space.
Even though single photon confocal laser scanning microscopy can efficiently modulate excitation power in a planar sub-micron region, it fails to elicit the same control along the optical axis, being the excitation volume extended to the entire illumination cone of the objective . The ability to mark specific cells in living embryos by photoactivating biomolecular markers can provide a unique tool in developmental biology studies for understanding cell fate and mechanisms of differentiation .
Materials and methods
Confocal laser scanning microscopy
Single photon photoactivation measurements wereb performed on a Leica TCS SP2 confocal microscope equipped with a 63×/1.40 (OIL CS HC×PL APO) objective lens (Leica), employing the 405 nm line of a 20 mWatt laser diode. Imaging of pa-GFP pre and post-activation was obtained by the 488 nm laser line. The spectral window used to collect fluorescence spanned from 500 to 600 nm. The main channel used was channel 2.
Ti:sapphire laser source was directly coupled into the scanning head of a Leica TCS SP2 AOBS confocal microscope through the infrared port. Measurements were collected using an average power of <P>min = 2.5 mW up to <P>max = 12.5 mW at the focal plane for two photon induced photoactivation. Imaging of the activated protein was obtained using the 488 nm line of a 20 mW Argon ion laser. Images were collected using a 100x/1.4 (OIL HCX PL APO) objective. The two-photon activation process followed a previously established procedure : it was first primed by focusing a pulsed infrared laser beam on a region of the sample with λ = 750 nm; subsequently, the unzoomed area was excited with λ = 488 nm and <P> = 0.04 mW (as measured at the objective back plane) to visualize the activated proteins. The spectral window used for collecting fluorescence was 500–600 nm, channel 2. AOBS and spectral unit were used for proper checking of the protein emission spectra before and after the photoactivation process, Figure 1.
Phoenix and HeLa cells were grown in standard culture conditions at 37 °C, 5 % CO2 in DMEM medium supplemented with 10 % North American Fetal Bovine Serum (Gibco Europe, Paisley, UK). pa-GFP was a generous gift from Dr. George Patterson. HeLa transient transfection was performed using FuGene (Boehringer-Ingelheim Italia S.p.A., Milan, Italy) reagent according to manufacturer instructions. Cells were harvested and fixed after 48 hours. Before mounting, cover slips were stained with the DNA dye TOPRO 3 (Molecular Probes Europe, Leiden, Netherlands) to facilitate cell identification. Phoenix cells were transfected with pa-GFP encoding DNA using Calcium Phosphate. After 24 hours, cells were lysed in cold 1X PLC buffer (50 mM HEPES, 1.25 % glycerol, 150 mM NaCl, 166 mM MgCl2, 1 % TRITONX-100, 0.001 % EGTA, 10 mM Napyruvate, 0.1 mM NaOrtovanadate, 0.01 mM PMSF, Aprotinin, pepstatin and leupeptin) on ice for 1 hour. Cell lysates were incubated with anti GFP polyclonal antibody and subsequently with Protein A sepharose beads, used as phantom samples. For imaging, beads were resuspended in a 90 % glycerol solution containing diazabicyclo-(2.2.2) octane antifade (Sigma-Aldrich S.r.l., Milan, Italy). The modified spheres were sandwiched between ethanol/acetone cleaned cover slip and glass slide. To avoid drying, the sample was sealed with ordinary nail polish .
Results and discussion
Two-photon activation, similarly to two-photon excitation, is expected to be a three-dimensional confined process. This can be proved by imaging selected volume of interest within artficial samples containing the pa-GFP or living cells able to express the protein. We checked the photoactivation properties as a function of activation wavelength and power  using phantom samples made by pa-GFP immobilized on beads. Green fluorescence was initially excited at 488 nm to visualize the pre-activation intensity. The activation process was then primed by focusing a pulsed infrared laser beam on a 22 μm2 region (512 × 512 pixel). For the activation process the dwell time per pixel was varied between 4.88 μs and several milliseconds (ms). Subsequently, the unzoomed area was imaged with the acquisition parameters used before the activation process.
Figure 2 shows different areas photoactivated under different conditions on the surface of the pa-GFPsphere used as phantom. The images before and after activation were analyzed for determination of mean intensity values. The ratio (C) of mean fluorescence intensities of the activated areas was taken as a measure for the efficiency of the photoconversion process as function of the photo-activation wavelength, Figure 3.
Due to the locally restricted non-linear excitation probability, activated volumes of finite thickness were expected in dependence on the provided laser light intensity [14, 15], whereas single photon induced activation should extend along the whole beam path of the light in the sample. A comparison between the one-photon λ = 405 nm) and two-photon activation process is shown in Figure 4. The thickness of the two-photon activated volume is limited to a narrow region inside the cell, while one-photon activation results in a fully activated cell in axial direction.
The intensity profile I(x, y, z) of a photoactivated volume within the nucleus is shown in Figure 5. The intensity profile is also a function of the power being used. The change of the activated volume is in accordance with the expectation based on the change in the intensity distribution I(z) that follows an inverse fourth power law as a function of the distance, z, from the geometrical focus of the lens. Using parameters like the excitation wavelength, the numerical aperture, the refractive index of the sample, it is possible to predict an arbitrary intensity value in dependence on the distance from the focal plane .
The intensity of the fluorescence emitted after photoactivation represents the edge between activated and not-activated volume and can therefore be considered as a threshold indicator for determining the activation process parameters. The thickness of the activated volume increased as a function of the laser light energy. The values for different volumes of interest can be summarized using the following data triplets reporting scan time per pixel [μs], activation power [mW] and FWHMI(z) [μm], namely: (19.6, 5, 2.35); (4.9, 17, 3); (9.8, 10, 3.5).
To quantify the error that can be made due to the location on the sphere, of different areas, covering the whole sphere, with the same settings were activated. The difference in the fluorescence intensity after irradiation with 488 nm was found to be 10 %. This uncertainty of the measured intensity values is intrinsic and will not be displayed additionally to the statistical errors. To identify the best working efficiency a full activation spectrum would be useful. With our set-up, a range of wavelength between 720 nm and 920 nm is accessible which covers very well the known absorption cross sections of wild-type GFP and at least the absorption peak of EGFP. For the acquisition of the spectrum the different probability of absorbing two photons in dependence on the wavelength was taken into consideration (power correction for higher wavelengths).
To measure the success or efficiency of the activation process the ratio of the fluorescence intensities (irradiation with l = 488 nm) after two-photon activation and before was considered. It was found that the conversion efficiency drops dramatically for wavelengths above 830 nm (for high powers and long times it is nevertheless possible) . Further, the efficiency shows a high level for wavelengths between 720 nm and 750 nm. Up to 800 nm another efficiency level applies. Unfortunately, the following values show strong fluctuations, no neither a clear assignment to the former level, nor an assignment to the low level efficiency shall be made at this point without further investigations.
Fig. 6: Projection of optically sectioned cells, obtained with Leica TCS SP2 AOBS confocal microscope: in violet 405 nm low intensity excitation (not activated pa-GFP), in green 488 nm excitation (activated pa-GFP) of wt-pa-GFP transfected HeLa cells. First row – not all activated; second row – after photoactivation of a portion that results in a black rectangle in the non-activated column (violet) and bright green in the photoactivated; third row – post photoactivation x-z section.
In the present work, the photophysical properties of pa-GFP under multiphoton excitation have been outlined showing the possibility to employ two-photon microscopy in combination with photoactivatable markers for protein dynamic studies. Pa-GFP exhibited the same photoconversion properties when excited in far red wavelength range using two-photon excitation in comparison with photoactivation at 405 nm: the absorption of photons can induce an increase of absorption cross section at 488 nm. This effect can be efficiently employed to mark submicron regions in the whole cell selecting, case by case, the appropriate volume of interest. Examination of the activated spatial volume as a function of the excitation energy showed that intensity modulation can be efficiently used to induce spatially controlled protein photoconversion along the optical axis providing a unique possibility to dynamically identify single 3D structures and considering final 4D (x-y-z-t) processes.
Two-photon activation can be efficiently used to track the fate of proteins and other biological macromolecules in living cells following cell cycle steps, and applications in terms of 3D memories could also be pursued. It is worth noting that coupling the selective expression of a protein within biological samples with the high localization level of non-linear processes produces a formidable tool for scientific reserch. Moreover, the exploitation of non linear processes involved in the interaction between light and proteins can lead to macromolecular resolution levels while keeping the advantages of using optical microscopy [8, 19, 20].
Authors are indebted to George Patterson and Jennifer Lippincott-Schwartz for pa-GFP availability.
- Shimomura O: The discovery of aequorin and green fluorescent protein. J Microscopy 217 (2005) 3.
- Tsien RY: The Green Fluorescent Protein. Annual Review of Biochemistry 6 (1998) 509.
- Patterson GH, Lippincott-Schwarz J: A Photoactivatable GFP for Selective Photolabeling of Proteins and Cells. Science 297 (2002) 1873.
- Patterson GH, Lippincott-Schwartz J: Selective photolabeling of proteins using photo activatable GFP. Methods 32 (2004) 445.
- Miyawaki A.: Fluorescent proteins in a new light. Nat Biotechnol 22 (2004) 1374.
- Diaspro A: Shine on ... proteins. Microsc Res Tech 69 (2006) 149.
- Giepmans BN, Adams SR, Ellisman MH, Tsien RY: The fluorescent toolbox for assessing protein location and function. Science 312 (2006) 217.
- Donnert G, Keller J, Medda R, Andrei MA, Rizzoli SO, Luhrmann R, Jahn R, Eggeling V, Hell SW: Macromolecular scale resolution in biological fluorescence microscopy. Proc Natl Acad Sci USA 103 (2006) 11440.
- Pawley J (ed): Handbook of Biological Confocal Microscopy. Plenum (2006).
- Diaspro A (ed): Confocal and Two-Photon Microscopy: Foundations, Applications, and Advances. Wiley-Liss (2001).
- Yuste R: Fluorescence microscopy today. Nature Methods 2 (2005) 902.
- Patterson G, Day RN, Piston D: Fluorescent protein spectra. J Cell Sci 114 (2001) 837.
- Piston DW, Patterson GH, Knobel SM: Quantitative imaging of the green fluorescent protein (GFP). Methods Cell Biol 58 (1999) 31.
- Diaspro A, Chirico G, Collini M: Two-photon fluorescence excitation and related techniques in biological microscopy. Quarterly Reviews of Biophysics 38 (2005) 97.
- Diaspro A, Bianchini P, Vicidomini G, Faretta M, Ramoino P, Usai C: Multi-photon excitation microscopy. Biomed Eng Online 5 (2006) 36.
- Schneider M, Barozzi S, Testa I, Faretta M, Diaspro A: Two-photon activation and excitation properties of PA-GFP in the 720–920-nm region. Biophys J 89 (2005) 1346.
- Post JN, Lidke KA, Rieger B, Arndt-Jovin D-J: Photoactivation of PAGFP in live Drosophila embryos: two-photon activation and phototoxicity. FEBS Letters 579 (2005) 325.
- Chen Y, MacDonald PJ, Skinner JP, Patterson GH, Muller JD: Probing nucleocytoplasmic transport by two-photon activation of PA-GFP. Microsc Res Tech 69 (2006) 220.
- Diaspro A: New World Microscopy. IEEE Engineering In Medicine And Biology Magazine 15 (1996) 29.
- Willig KI, Kellner RR, Medda R, Hein B, Jakobs S, Hell SW: Nanoscale resolution in GFP-based microscopy. Nature Methods 3 (2006) 721.