Supercontinuum white-light lasers offer excitation of many fluorescent probes with one light source.
William G. Telford, National Cancer Institute, and Husain Imam, Koheras A/S
Flow cytometry is a central technology in the life sciences. Flow cytometry systems use lasers, a hydrodynamically focused cell delivery system, wavelength-specific optics and photomultiplier tube detectors to analyze large numbers of single cells for a wide variety of properties. For example, the interior structure, physiological state and identity of cells can be analyzed. The technique has permitted detailed study of the immune system’s complex cells and has been used to characterize tumor cells and to identify a variety of human diseases (such as HIV/AIDS). Fluorescence-activated cell sorting — an extension of flow cytometry that allows physical separation of complex cell mixtures — has allowed subsequent genomic and proteomic analysis of purified cell preparations, greatly advancing our insights into cell biology.
Critical to flow cytometry are fluorescent probes — molecular tags that can be detected with appropriate excitation/emission conditions. Fluorescent probes can be used to detect receptors on cells, to determine the health and physiological state of a cell and even to measure gene expression in individual cells. Probes suitable for biomedical analysis number in the thousands, and new ones are being developed constantly. Flow cytometry relies almost exclusively on lasers as a source of excitation for these probes.
Light source limitations
Although the coherence, stability and power level of lasers make them ideal for illuminating individual cells, their discrete wavelengths limit the range of excitation bandwidths available for fluorescent probe excitation. Most clinical and laboratory flow cytometers have only one or two single-wavelength lasers — usually an argon-ion or a solid-state 488-nm unit and a red HeNe or laser diode emitting in the 633- to 635-nm range. Although very useful, these relatively simple instruments can excite only a limited number of fluorescent probes. Even the most modern multilaser flow cytometers typically provide no more than six discrete laser wavelengths, and most provide fewer. Coverage of the ultraviolet-to-infrared spectrum is therefore never complete, leaving large gaps in excitation capabilities.
Although fluorescent probes useful for biomedical analysis span the entire visible spectrum, only a small fraction of them are accessible to flow cytometric analysis. Green fluorescent proteins (GFPs) are a well-known example of this class of probes. When the gene for this protein is inserted into cells, the fluorophore is expressed and can be detected by flow cytometry, rendering it a sensitive marker for gene expression. GFP, which can be excited at 488 nm, can be detected on most conventional flow cytometers.
However, red fluorescent proteins (RFPs), a more recent generation of the probes, require green or yellow laser excitation, which virtually is unavailable on commercial instruments. These proteins therefore have seen little use in flow cytometry, despite their excellent properties of high fluorescence yield and good biological tissue penetration. The detection of RFPs would require lasers operating at single wavelengths that often are not readily available from solid-state sources (such as the 590- to 610-nm range). Flow cytometry would benefit considerably from a more flexible excitation source that could accommodate both routine analysis and newly developed fluorescent probes with unusual excitation characteristics.
White light from a laser
One potential answer to this problem is the supercontinuum white-light laser, a recently developed fiber laser technology that emits continuously over a wide bandwidth from the near-ultraviolet to the infrared. The supercontinuum laser phenomenon was observed empirically more than 30 years ago and has since been the subject of intense study and technological development. Supercontinuum light is generated by invoking high optical nonlinearity in an optical material. Laser sources emitting in the near-infrared (1064 nm) are pulsed in the nanosecond, picosecond or femtosecond range to induce a nonlinear effect in the material. The nonlinear effect “breaks” the pulse into a supercontinuum spectrum.
Early supercontinuum lasers emitted primarily in the infrared because they were limited to Ti:sapphire femtosecond lasers and bulk nonlinear materials. More recent developments in picosecond fiber sources and in the use of microstructured optical crystal fibers have extended the supercontinuum range from violet (~400 nm) to the infrared (>2000 nm). The use of optical fiber as the nonlinear material makes them much smaller in size, very reliable and easier to integrate into existing biomedical instrumentation. Their enormous visible and infrared bandwidth means that they emit at relatively high power levels — several watts, typical — with a light mixture that appears white to the human eye. Because they rely on high-power pulsed infrared lasers as pump sources, supercontinuum lasers themselves are pulsed, but at frequencies of >50 MHz.
The broad emission range of supercontinuum lasers makes them an intriguing excitation source for flow cytometry. Because flow cytometers traditionally rely on conventional lasers as excitation sources, a supercontinuum source that is specifically filtered for several narrow-wavelength ranges theoretically could be employed instead. Such sources already have been used for laser confocal microscopy, employing acoustic optical filters to select specific wavelengths for fluorescent probe excitation. Acoustic optical filters are not yet practical as optical filtering devices for flow cytometry because of their relative inefficiency and high noise levels.
However, simply inserting a narrow bandpass filter with the desired spectral range in the white light laser path should allow the user to pick any desired excitation wavelength for flow cytometric excitation. Because bandpass filters can be built at virtually any wavelength and bandwidth, this should provide almost any visible excitation wavelength for flow cytometry. Although supercontinuum sources are pulsed, their rapid repetition rates allow them to function as quasicontinuous-wave sources, which means that they can provide adequate excitation for the typical 0.5- to 1-ms laser interrogation times used for single cells in a flow cytometer.
Figure 1. A SuperK Extreme supercontinuum white-light laser system from Koheras A/S was integrated into a BD Biosciences LSR II flow cytometer that included dielectric mirrors, an infrared filter and narrow bandpass filter optics.
To test the applicability of supercontinuum lasers for flow cytometry, we integrated a Koheras SuperK Extreme source (6-W total output) into a BD Biosciences LSR II flow cytometer (Figure 1). To remove the considerable infrared component of the laser beam, we initially reflected the white-light laser off two dielectric mirrors and passed it through an infrared filter. A filter wheel containing narrow bandpass filters ranging from 450 to 700 nm — with 10- to 30-nm transmission ranges — was inserted into the beam, allowing only the desired wavelength range to pass. The resulting filtered laser light had power levels from approximately 2 to 4 mW per nanometer with a practical total emission of 10 to 50 mW, well within the useful range for biological analysis. The resulting laser light for several bandpass filters is shown in Figure 2.
Figure 2. Filtered supercontinuum laser light is shown at 488 nm (top left), 530 nm (top right), 575 nm (bottom left) and 630 nm (bottom right).
We aligned the laser beam to the cytometer sample stream and examined cells labeled with a variety of fluorescent probes. For example, we analyzed cell lines expressing the red fluorescent proteins DsRed and dTomato (excitation maxima in the 550- to 560-nm range) using the supercontinuum laser with 530/30-nm filters (Figure 3). The measured power level was 17.6 mW.
The data from this analysis was expressed as fluorescence histograms, with the fluorescent cells shown in the shaded peak and cells with no fluorescent protein (cellular autofluorescence) shown in the unshaded peak. Expression of both fluorescent proteins was easily detectable using the filtered supercontinuum source. For comparison purposes, the cells also were analyzed using a diode-pumped solid-state laser emitting at 532 nm — roughly the same spectral range and power level (20 mW) as the filtered white-light source. Detection sensitivity was comparable between the supercontinuum and the traditional green diode-pumped solid-state source.
Figure 3. Flow cytometric data for SP2/0 cells expressing the red fluorescent proteins DsRed or dTomato are shown. Cells were excited using either supercontinuum laser light filtered with a 530/30-nm bandpass filter (top), or using a conventional diode-pumped solid-state 532-nm laser emitting at 20 mW (bottom).
The supercontinuum white-light laser source should permit excitation of a variety of biologically important fluorescent probes with a single laser source, using selective filtration to gain the desired wavelength range. An obvious advantage is the flexibility: Rather than depending on single-wavelength laser sources, one can select any wavelength. In addition, the supercontinuum source will allow fine-tuning of excitation wavelengths to particular probes, producing maximum sensitivity.
In the past, we frequently had to use an excitation source that was not optimal for our probes. For example, the RFP DsRed conventionally is excited using 488-nm lasers, which overlap the excitation spectrum at a far from optimal point. The conventional blue-green lasers are an adequate but far-from-optimal laser source for this green-excited fluorescent probe. The compromises should no longer be necessary with a supercontinuum source.
Multiple single-wavelength lasers also should be replaceable by a single supercontinuum source. For example, blue, green, yellow, orange and red laser bandwidths could be simultaneously generated and aligned to multiple interrogation points on a flow cytometric sample stream, allowing complex multilaser, multiprobe experiments to be executed using a single laser source. It is likely that future flow cytometers will use a combination of single-wavelength and supercontinuum sources to provide the historically accustomed wavelengths while also incorporating the flexibility required for the ever-expanding variety of fluorescent probes used in biomedical analysis.
Supercontinuum lasers already are used in some confocal microscopes, in spectrofluorimeters and in other instruments that require broad ranges of visible and infrared light. Implementing these sources in flow cytometers should eliminate the restrictions of limited laser excitation wavelengths on flow cytometry, enabling the use of virtually any fluorescent probe, regardless of excitation/emission requirements, and allowing us to easily modify our instrumentation to accommodate our assays. Almost any fluorescent assay should be possible with this novel laser technology.
Meet the authors
William G. Telford is a staff scientist at the Experimental Transplantation and Immunology branch of the National Institutes of Health’s National Cancer Institute in Bethesda, Md.
Husain Imam is manager of Koheras A/S in Birkerød, Denmark.
1. C. Gmachl et al (Feb. 21, 2002). Ultra-broadband semiconductor laser. Nature, pp. 883-887.
2. V. Kapoor et al (Sept. 1, 2007). New lasers for flow cytometry: filling the gaps. Nature Methods, pp. 678-679.
3. N.C. Shaner et al (Dec. 1, 2004). Improved monomeric red, orange and yellow fluorescent proteins derived from Discosoma sp. red fluorescent protein. Nature Biotechnology, pp. 1567-1572.