Fast FLIM system images live-cell signaling
Researchers have a new tool for studying live-cell signaling events, thanks to the efforts of a group from the UK. By extending and enhancing existing technology and techniques, the investigators developed a new time domain optically sectioned fluorescence lifetime imaging (FLIM) microscope. The device could prove useful in ferreting out the information needed to target disease at the molecular level as well as in other applications.
Researchers developed a time domain fluorescence lifetime imaging microscope for high-speed live-cell imaging. A pulsed supercontinuum source (lower left) provides the light for a Nipkow disc scan head (lower right). A variable-delay generator incorporated in the gated optical intensifier (GOI) enables capture of intensity images at different intervals after pulsed excitation. Images reprinted from Optics Express.
Composed of researchers from Imperial College London, from the London-based Institute of Cancer Research, and from PerkinElmer in Seer Green, UK, the group created its microscope using Nipkow spinning-disc (wide-field) confocal microscopy technology along with time-gated imaging.
Team member David M. Grant, a graduate student at Imperial College London, noted that the wide-field time-gated imaging often is seen as one of the less accurate FLIM techniques. However, the advantages of parallel pixel acquisition afforded by Nipkow disc microscopy and wide-field time-gated detection more than outweigh the penalties associated with additional intensifier noise or with sampling economy, he said.
The ability to capture optically sectioned FLIM images at high frame rates could be useful from a biology perspective, particularly for live-cell Förster resonance energy transfer (FRET) experiments.
There are several ways to perform optical sectioning and various ways to observe FRET. The scientists pared down the combinations by starting with FLIM, which generally is considered to be less artifact-prone than other FRET observation techniques. They also wanted the imaging to be fast, which favored single-photon excitation. The desire for speed also made parallel pixel acquisition an advantage. Finally, they wanted to do optical sectioning.
Grant noted that Nipkow disc systems, in which a spinning disc with a pinhole array provides parallel confocal illumination and detection, fulfill all of these criteria. The researchers used a Nipkow microscope head from Yokogawa Electric Corp. of Japan. In addition to the first disc, this unit also had a second disc containing a microlens array that focused incoming excitation light onto the pinholes in the first disc. The result was a significant increase in the percentage of light getting through the disc and reaching the target.
That is particularly important because the amount of light at the sample is key to the imaging speed of the microscope. Indeed, the illumination at the sample was the speed-limiting factor in the scientists’ previous version of the instrument.
With the new microscope, the group increased the imaging speed by two orders of magnitude. One reason they could do this was the use of a new supercontinuum source from Fianium Ltd. of Southampton, UK, which provided picosecond pulses at a 50-MHz repetition rate with a spectral power density of 2 mW/nm over wavelengths from 470 to 490 nm.
Shown is a sectioned fluorescence lifetime image stack of a COS 7 cell expressingH-Ras-mRFP and Raf-RBD-enhanced GFP. The cell displays Förster resonance energy transfer (FRET) between EGFP donor and mRFP acceptor fluorophores at the plasma membrane after stimulation by EGF. Each image was recorded in 6 s, with a 100-s total acquisition time (Scale bar = 10 μm).
The researchers also performed other optimizations, primarily by adjusting their sampling strategy. Because the imaging was time-gated, they could choose where in time to position the gates with respect to the fluorescence decay, how many gates to use and, finally, what type of gate duration to employ. Grant noted that the source change and modeling-based time gate optimization strategy played nearly equal roles in the imaging acceleration, with the first being more beneficial than the second. “On reflection, I’d say 60:40 was a good balance between the two,” he noted.
The researchers used a gated optical image intensifier from Kentech Instruments Ltd. of Wallingford, UK. They sent the output to a cooled scientific-grade Hamamatsu CCD camera for data acquisition.
In general, they used time gates as wide as is practically possible to capture the most photons. The maximum imaging rate depends on how many time gates are needed to determine the characteristics of the fluorescence decay curve, with a minimum of two sampling points required. Given the speed of the equipment, the researchers estimated that FLIM imaging at up to about 10 fps could be achieved. In practice, imaging rates are related less to the need to capture more than to the bare minimum time gates in each FLIM image. Even so, tests and theoretical calculations show that the new system is much faster than one based on single-photon-counting techniques.
To demonstrate the capabilities of their instrument, the investigators performed studies of Ras activation in a model system, GTPase H-Ras in MDCK and COS 7 cells. Ras proteins are studied in cancer research because many tumor cells carry mutations in their Ras genes. Ras activation can be monitored by detecting FRET between the bound Ras protein, which the researchers tagged with the fluorescent protein mRFP, as well as Raf-RBD, which they fluorescently labeled with enhanced GFP. With their microscope, the researchers followed the interaction through a series of rapidly acquired optically sectioned images. Their work is detailed in the Nov. 26 issue of Optics Express.
On the left are representative images of enhanced GFP-expressing cells acquired with the Nipkow disc microscope, while on the right are those from a laser scanning confocal time-correlated single-photon-counting system. Acquisition times range from 1 to 10 s (top to bottom). The white pixels are where an erroneous lifetime has been calculated. The Nipkow system can acquire images in 1 s, while the confocal one has trouble with much longer acquisition times.
They also did a demonstration using model-enhanced GFP-mRFP FRET constructs in live cells. Grant noted that they were trying to push the instrument to its limits. With windowing and the minimum number of time gates, they found that the maximum rate was about 10 fps.
They also imaged cholesterol depletion in live cells, finding that the reduction in lipid order in cell membranes happened faster than had been assumed previously. This work was done at room temperature, but the researchers are repeating it at physiological temperatures, which are higher and could accelerate the cholesterol depletion rate.
They also are looking into other applications of the technology. Some that would be of commercial interest involve screening for high content analysis, high-speed time-lapse imaging of live-cell molecular interactions and automated high-throughput FLIM microscopy of live sample arrays.
As for further technological improvements, Grant said that faster imaging would be possible with more excitation power or with a microscope head more efficient at getting light through to the sample. “Such advances offer promise for increasing the rates at which one could image, but issues of photobleaching and phototoxicity must always be considered — especially for time-lapse imaging of live cells.”
Contact: David Grant, Imperial College London; e-mail: firstname.lastname@example.org.
MORE FROM PHOTONICS MEDIA