Two-Photon Fluorescence Microscopy

Multi-Photon Fluorescence Microscopy is a relatively novel imaging technique in cell biology. It relies on the quasi-simultaneous absorption of two or more photons (of either the same or different energy) by a molecule. During the absorption process, an electron of the molecule is transferred to an excited-state molecular orbital. The molecule (i.e. the fluorophore) in the excited state has a high probability (> 10 %) to emit a photon during relaxation to the ground state. Due to radiationless relaxation in vibrational levels, the energy of the emitted photon is lower compared to the sum of the energy of the absorbed photons.

	Figure 1: Principle of fluorescence induced by one-photon absorption (left) and two-photon absorption (right). While the resolution in two-photon fluorescence mciroscopy (2PFM) is less good, photodamage is lower and penetration depth is higher compared to single-photon (confocal) fluorescence microscopy (1PFM).

The multi-photon absorption process needs a very high density of photons (0.1 - 10 MW/cm²) from a ps-to-fs-pulsed light source. This is because the virtual absorption of a photon of non-resonant energy lasts only for a very short period (10^-15 - 10^-18s). During this time a second photon must be absorbed to reach an excited state. Multi-photon absorption was predicted in 1930, and the proof-of-principle was performed in the 1960s using continuous-wave laser sources. Multi-photon fluorescence microscopy in cell biology uses pulsed lasers (pulse width in the range of 20 fs to 5 ps, typically ~ 100 fs) with high repetition rates (10 -100 MHz).

	Figure 2: Maria Göppert-Mayer (with daughter Marianne) and the title of the scientific article predicting the two-photon absorption process which was based on her PhD thesis.

Two-photon fluorescence microscopy (2PFM) is the most common multi-photon fluorescence application in cell biology for two reasons: First, the best performing, commercially available laser (a solid state mode-locked Ti:Sapphire laser) covers the spectral region from 700-1200 nm (most often operated in the range of 720 nm to 920 nm) with pulse widths of 100 - 150 fs and repetition frequencies of about 75 MHz. Second, the performance of fluorescence microscopy relies very much on the use of highly specific dye molecules (e.g. pH- or other ion indicators). The development of these dyes took place during the last three decades and was based on applications in the fields of classical microscopy using near UV and visible light sources. Therefore, most of the dyes absorb in the near-UV and the visible spectral region. These well characterised dyes are also used in 2PFM. Fluorescence microscopy ased on three- or more-photon absorption using the Ti:Sapphire laser for excitation would excite far-UV absorbing fluorophores that are not very well characterised and rarely used.

The number of absorbed photons (n_a) in 2PFM is described by the following equation:

n_a = (σ_2P P₀² π² NA⁴) / (τ_p f_p² h² c² λ²)

σ_2P is the two-photon absorption cross-section, P₀ is the average power, NA is the numerical aperture of the objective, τ_p is the pulse width, f_p is the repetition frequency, c is the speed of light, and λ is the excitation wavelength. The most important differences compared to one-photon absorption (1PA) are the quadratic dependence on the average power (linear in 1PA) and the power-of-four dependence on the NA of the microscope objective (quadratic in 1PA).

2PFM has several advantages over commonly used fluorescence microscopy techniques. If the average excitation power is moderate, the two-photon absorption process takes place only in the focus of the laser beam and thereby provides a three-dimensional resolution. Radial and axial resolutions of well below 0.5 μm and 1 μm, respectively, are achieved for a typical excitation wavelength (800 nm) by using microscope objectives with a high numerical aperture. The axial resolution of 2PFM is a great advantage compared to the classical one-photon fluorescence microscopy (1PFM). A similar axial resolution can be achieved in 1PFM using a confocal set-up. The three-dimensional spatial resolution in 1PFM is compromised by lower signal intensity compared to 2PFM. Therefore, confocal 1PFM requires higher excitation intensities. This reduces the observation time of a photo-vulnerable object - like cells or biological tissues.

	Figure 3: Z-scan (2 μm step width, the scale bar is 10 μm) of 2PFM images of two dividing HEK293 cells (blue = low intensity; red = medium intensity; yellow = high intensity). The fluorescence originates from a Calmodulin(CaM)-EGFP fusion protein, which binds to Myosin Light Chain Kinase (at the cell membrane) and to CaM-Kinase II (in the meiotic spindle poles) during cell division.

The use of near-infrared radiation in 2PFM enhances the penetration depth and at the same time reduces image deterioration due to scattering when passing through biological tissue. Since the elastic scattering of light is proportional to the inverse power of the wavelength (with an exponent between 2.2 and 4 depending on the material), the scattering - a major contributor to image detioration - is an order of magnitude less pronounced in 2PFM, and allows imaging in 10 times deeper regions compared to 1PFM.

Imaging in cell biology has an important 4th dimension (besides the three spatial dimensions) - the time. The observation of a cell or a tissue on a time scale comparable to the physiological events is a prerequisite to follow biological processes. Two of the major limiting factors in the use of fluorescence microscopy are photo-bleaching and photo-damage. Both are limited to the focal region in 2PFM, whereas in 1PFM the upper and lower regions of the excitation light cone are affected. Although 2PFM has a slightly lower resolution and requires complicated excitation sources, it can be still advantageous due to the much longer observation time.

An additional advantage of the fs-to-ps pulsed excitation with MHz repetition rates is the capability of fluorescence lifetime imaging (FLIM). The fluorescence lifetime is obtained by repetitive measurement of the time laps between the excitation pulse and the fluorescence photon (TCSPC, Time-Correlated Single Photon Counting). Each pixel contains information on both the fluorescence lifetime and the "classical" fluorescence. Fluorescence lifetime imaging is valuable especially for intracellular measurements, where the absolute number of fluorophores, either membrane permeable organic fluorophores or autofluorescent proteins (e.g. GFPs), can not be determined in a quantitative way.