Fluorescence plays an important role for biologists studying living cells and their components. Among the many fluorescence-based contrasting techniques available to microscopists, fluorescence lifetime imaging microscopy (FLIM) is rapidly gaining popularity. The recent proliferation of fluorescent probes such as green fluorescent protein and its mutants, the ever-widening array of excitation light sources, and the advent of high-speed, high-sensitivity imaging detectors has paved the way for FLIM to provide valuable insights into ionic and molecular processes in many environments, both micro and macro.
In standard fluorescence imaging, a sensitive detector, such as a cooled CCD, images emission from a fluorophore. Steady-state imaging suffers from the fact that fluorescence signal intensity is extremely dependent on the intensity of the excitation light and the concentration of fluorophore. Fluorescence lifetime is independent of fluorophore concentration, however, but dependent on local environment. FLIM thus allows researchers to obtain precise quantitative data about both fluorophore distribution and local environment.
Two principal approaches to FLIM implementation exist: one in the time domain, another in the frequency domain. For time-domain measurements, a laser or LED excites the sample with femtosecond to nanosecond pulses. A gated detector, typically an intensified CCD (ICCD) camera system, captures the exponential decay of the fluorescence. In a simple situation, the investigator can compute the lifetime of a fluorophore with single-exponential decay by acquiring only two images at two different points in time after the excitation.
To create a single-exponential lifetime map of the sample, we determine the time constant τ by
where S1 represents the intensity at time t1 and S2 represents the intensity at time t2.
Realistically, many biological probes have complex, multi-exponential decay. Assuming this scenario, the investigator must know the excitation-pulse profile and the temporal response of the detector to acquire accurate measurements. Sweeping the gate window with a light pulse that is very short compared to the gate time of the intensifier lets the investigator determine this temporal responsethat is, the rise and fall times. Deconvolving the detector response from the result yields the lifetimes of the various components in the probe.
Figure 1. With a frequency-domain configuration (left), the phase shift gives the lifetime of the fluorophore (right).
In the frequency-domain approach, a continuous laser acts as the excitation-light source. A high-power radio-frequency generator and an acousto-optical modulator modulate both source and detector, usually in the region of 10 to 300 MHz. Generally, the modulating pulse is applied either to the intensifier's photocathode (gate modulation) or microchannel plate (gain modulation). We can modulate the detector at the same frequency as the excitation for homodyne detectionor at a slightly different frequency for heterodyne detection. The process causes the detected emission to undergo a phase shift Φ compared to the excitation, and that phase shift directly relates to the lifetime of the component of the fluorophore (see figure 1).
We can derive the lifetime of the fluorophore from Φ and the modulation M as follows:
For the exponential decay of a single component, τΦ and τM are equal. When working in the frequency domain, we must collect reference data, which typically consists of specular reflections from the laser source or a sample with a known lifetime, to account for the detector's response. FLIM for FRET
Due to the quantitative nature of its data, researchers are adopting FLIM for an increasingly broad variety of applications. One significant use is as an alternative method for measuring fluorescence resonance energy transfer (FRET).
FRET is a dual-label phenomenon that indicates the proximity of two molecules to one another. These can be two molecules of DNA, two proteins, a DNA molecule and a protein molecule, a protein and a chemical, and so forth. Most often, both are proteins. FRET is of great value to much of biology, as protein-protein and protein-DNA interactions regulate the majority of cellular activities. FRET commonly serves as a binary indicator that molecules are within 10 nm of each other and hence are interacting.
In the FRET process, one molecule labeled with a donor fluorophore transfers its excitation energy to a neighboring molecule labeled with an acceptor fluorophore, quenching the donor while stimulating the acceptor to fluoresce. For an observer to acquire both donor and acceptor fluorescence for this dual-image ratiometric measurement, the fluorescent probes must exhibit spectral overlap.
Because donor fluorophore fluorescence lifetime decreases during the FRET process, measuring the lifetime gives the exact location and time of the donor-acceptor interaction. Therefore, in situations in which filter-based FRET imaging techniques cannot discriminate between multiple fluorophores, FLIM is able to do so by measuring the respective fluorescence lifetimes.
As noted earlier, fluorescence lifetime is independent of probe concentration, and thus typically immune to common photodestructive phenomena such as the bleaching of samples over time. Since the lifetime of the probe directly relates to microenvironment, the method can obtain precise information about the surrounding molecular environment; for example, the lifetime of NADH, the active coenzyme form of vitamin B3 that helps cells produce energy, changes from 0.5 ns to 1 ns when bound to protein. FLIM-based techniques can therefore image and distinguish between free and protein-bound NADH, providing researchers with functional maps of cellular metabolism.
Recent reports have suggested that FLIM can work as an alternative to measurements of pH or ratiometric ions such as Ca2+, Na+, and Mg2+. It may also provide a noninvasive method for accurately imaging and mapping intracellular oxygen concentrations. Time or Frequency
Time-domain and frequency-domain FLIM measurements each present a number of unique benefits and drawbacks. Time-domain FLIM, which is more intuitive and simpler to implement than its frequency-domain counterpart, is typically preferred for measuring relatively long-lived components.
Figure 2. A second-generation image intensifier can be fiber-optically connected to a scientific-grade CCD for time-domain or frequency-domain FLIM.
As previously mentioned, time-domain FLIM requires only two images to construct a single-exponential-lifetime map, but it requires special care to measure accurate time delays and the temporal response of the detector and pulse width of the light source. High-performance ICCD cameras provide the combination of excellent low-light sensitivity and temporal resolution this method requires. An image intensifier, which is often bonded fiber-optically to the photosensitive CCD of the system, amplifies incident photons and allows detection of low-light signals (see figure 2). Temporal resolution is possible because the voltage-controlled intensifier can be turned on and offgatedin very short intervals.
Scientific ICCD camera systems are engineered with any one of a variety of second- or third-generation intensifiers optimized for good quantum efficiency to image most popular fluorophores, including green fluorescent protein and Cy3/Cy5. Until recently, such intensifiers could only be gated quickly enough to measure lifetimes equal to or greater than 1 ns via time-domain FLIM. New developments in gating technology now allow second-generation intensifiers to perform time-domain measurements of fluorescence lifetimes as brief as 500 ps, without sacrificing quantum efficiency.
Although time-domain FLIM can measure lifetimes approaching 100 ps, doing so requires special image intensifiers designed with conductive mesh. Unfortunately, these modified intensifiers suffer from significantly lower quantum efficiency. Because they use lower gating voltages, these ultrafast ICCD camera systems also have reduced on/off ratios, which limit their performance. An on/off ratio is a direct measure of intensifier-gating quality. A high on/off ratio is required to eliminate background light and faithfully reproduce transient phenomena of interest.
Frequency-domain FLIM is generally favored for measuring fluorescence lifetimes shorter than 500 ps. Once researchers know the modulation and the phase shift, they can easily identify and map the lifetimes of individual components within multicomponent decays. Although this imaging technique does not require gating, the gain of the detection systems with higher quantum efficiency, described above, can be modulated, making these ICCD cameras well-suited for FLIM applications in both the frequency and time domains. oe
T. Gadella Jr., Fluorescent and Luminescent Probes for Biological Activity, Second Edition, W. Mason, editor, Academic Press, San Diego, CA (1999).
Mark Christenson, Scott Sternberg
Mark Christenson is life sciences business manager and Scott Sternberg is business development manager at Roper Scientific Inc., Tucson, AZ 85706.