Violet-excited fluorochromes form an important class of probes for cell biology. Cascade Blue, Alexa Fluor 405, Pacific Blue, Cascade Yellow, Alexa Fluor 430, and Pacific Orange are important fluorochromes for fluorescent immunophenotyping, supplementing more conventional probes for polychromatic analysis of cell surface markers (1, 2). Cyan Fluorescent Protein is a widely used expressible probe, and one of several fluorescent proteins that requires violet excitation (3). Violet excited probes are also available to analyze cell cycle analysis, viability assessment, and cell physiology.
Until recently, the only practical source of violet laser excitation was a krypton-ion laser. Krypton-ion lasers can produce several violet laser lines, the dominant one at 407 nm. This large, expensive water-cooled laser was only usable with large-scale cell sorters (such as the BD FACSVantage, Beckman Coulter Altra, and DakoCytomation MoFlo), limiting violet applications to these instruments. More recently, smaller violet laser diodes (VLDs) with an emission range from 395 to 415 nm have been integrated into cuvette-based flow cytometers, allowing these smaller instruments to utilize these increasingly valuable probes (4–9). Since their initial development, single mode VLDs have increased in power from 5 mW to greater than 50 mW, providing more than adequate excitation for cuvette cytometers.
In 2003, our group published data for the integration of a VLD into a stream-in-air cell sorter (7). Stream-in-air instruments have signal collection optics with significantly lower numerical apertures, making them less sensitive than cuvette instruments. At that time, the maximum VLD output commercially available was less than 25 mW. Our results at that time showed that VLDs at this power level gave adequate excitation for violet-excited phenotyping fluorochromes and Cyan Fluorescent Protein, if the probe in question was bright relative to background, or abundantly expressed. Nevertheless, since that time, we and several other groups have demonstrated that this power level is not ideal for violet excited probes with lower quantum efficiency, or under circumstanced where the signal-to-background is low (such as low expression of a cell surface marker). Cyan Fluorescent Protein (CFP)-YFP fluorescence resonance energy transfer (FRET) has also been shown to require higher levels of violet excitation when carried out on stream-in-air instrumentation (10, 11).
Since that time, commercially available VLDs appropriate for cytometry have increased in power. Single mode lasers exceeding 50 mW are now available, and multimode lasers can exceed 150 mW in total output (James Jackson, personal communication). As a postscript to our previous study, we have therefore integrated a more powerful dual module VLD into a FACSVantage cell sorter. This unit combines the beams of two 55 mW VLDs using a polarizing beamsplitter, resulting in a single mode violet laser beam exceeding 100 mW. This power level approaches that previously available from krypton-ion sources. In this follow-up study, we have installed this dual module VLD on a stream-in-air sorter and evaluated its ability to resolve low-fluorescence microsphere standards, as well as a variety of violet-excited fluorochromes, including the blue-emitting probes Cascade Blue and Alexa Fluor 405, the green emitting probe Cascade Yellow, and the newly developed orange-emitting probe Pacific Orange. High levels of VLD laser light were able to resolve the fluorescence of these systems at levels approaching or comparable to those observed with lower power lasers in cuvette cytometry systems. These studies suggest that higher laser powers can confer a small but measurable difference in the ability to resolve dim violet-excited probes, and are particularly valuable in stream-in-air systems.
RESULTS AND DISCUSSION
- Top of page
- MATERIALS AND METHODS
- RESULTS AND DISCUSSION
- LITERATURE CITED
In previous studies, we have evaluated VLDs as replacement for larger ion lasers where violet excitation is required. These laser diodes have seen extensive use in cuvette systems. While VLDs have also been successfully installed in stream-in-air systems, the relatively low power level available from earlier modules have limited their use to relatively bright fluorochromes the identification of densely expressed surface markers (6, 7). Anecdotal reports suggest that this power level may be insufficient for more photon-demanding applications such as CFP-YFP FRET, where more complete saturation of the donor fluorochrome is desirable (10, 11).
The more powerful system used in this study is illustrated in Figure 1. Two polarized 404 nm ∼55 mW VLDs with detached diode modules (connected to their cooling/power supply units via an umbilicus) were mounted at 90° angles to one another (Fig. 1a, designed and built by Power Technology). The polarization orientation of each laser was similarly at perpendicular angles. Both beams entered a polarized beamsplitter cube, and were either transmitted or reflected based on the polarization state of the beam. Both beams were thus merged into a single apparent beam with a power level of ∼100 mW (Fig. 1b). The individual and merged beams profiled by CCD camera and were found to be both approximately Gaussian in shape and indistinguishably collinear to a minimum distance of 3 m (Fig. 1c). This unit was mounted on a FACSVantage DiVa cell sorter and the beams aligned to the secondary position on the sample stream. The individual laser modules could be turned on or off to produce power levels of 100 mW, or 55 mW at different polarization angles.
For the individual modules, the polarization direction was either vertical (parallel to the sample stream) or horizontal (perpendicular to the sample stream).
Results from the individual and merged beams are shown in Figure 2. InSpeck Blue microsphere sensitivity arrays, consisting of six populations with descending fluorescence plus unlabeled (left column) were analyzed using either both beams (100 mW) or individual 55 mW beams. The same microspheres were analyzed on a cuvette-equipped BD LSRII, using the same emission filter as on the FACSVantage DiVa. When the brightest bead population was normalized to a fixed channel number, the VLDs at both 55 and 100 mW allowed all bead populations to be visualized, at a sensitivity level apparently similar to the LSR II with a 25 mW source. The apparent C.V. of the dimmest bead population did improve slightly at 100 mW excitation compared to 55 mW. Violet laser light was not optimal for Spherotech Rainbow microsphere excitation; small differences in excitation efficiency were therefore very apparent using these beads. The 100 mW power level discriminated three dim, poorly separated populations that could not be well-resolved with the 55 mW units individually. The 25 mW laser on a cuvette system surpassed all both in the resolution of these bead populations, but with only a small improvement over the 100 mW system. The 100 and 25 mW cuvette systems were also able to clearly distinguish the two dimmest bead populations, while 55 mW excitation alone could not. But interestingly, both vertically and horizontally polarized beams excited more or less equally well. The horizontally polarized beam would be expected to perform less well, due to the orientation of the fluorochrome dipole relative to the detection optics of the cytometer (Gary Durack, personal communication). These results nevertheless indicate that the merged beam (including the horizontally polarized beam) gave a small but detectable increase in the detection of low-fluorescence objects on the stream-in-air system compared to lower power levels, an observation that was not obvious using the brighter standards.
Figure 2. Microbead sensitivity standards. Left column, InSpeck Blue microsphere mixtures (Molecular Probes Invitrogen) analyzed on the indicated instrument with the indicated laser. Grey-filled histograms show microsphere peaks with the 100% population normalized to a fixed fluorescence intensity, with different voltage settings for each analysis (FACSVantage data only). Black peaks indicate the location of the 100% population for the 55 mW analyses using the same voltage settings as for the 100 mW analysis. Right column, Rainbow beads 8-population beads (Spherotech) analyzed on the indicated instruments with the indicated laser. Peak intensities were also normalized for the FACSVantage data. Insets show resolution of three dim microsphere populations for each instrument/laser configuration.
Download figure to PowerPoint
The dual module VLD was then used to analyze cells labeled with several violet-excited phenotyping fluorochromes. In Figure 3, human PBMCs were labeled with Cascade Blue or Alexa Fluor 405 antibodies against T cell markers and analyzed with either dual or single VLD excitation on the stream-in-air sorter, or on the cuvette instrument. Excitation at 100 mW gave a small but reproducible improvement in signal-to-background ratio over the single 55 mW beams; interestingly, 100 mW excitation also improved detection sensitivity somewhat over the cuvette instrument. When Cascade Yellow and Pacific Orange labeled cells were analyzed in Figure 4, the 55 mW power level gave signal-to-background ratios that were at least comparable to the 100 mW level. Both power levels performed comparably to the cuvette instrument. These green, yellow, and orange emitting probes were of especial interest, since fluorochromes in this region tend to be dim relative to blue-emitting probes. Nevertheless, Cascade Yellow, and the more recently developed Pacific Orange were both readily detectable on the stream-in-air instrument, with the 100 mW results showing relative parity with the cuvette instrument data.
Figure 3. Blue emitting fluorochromes. Left column, PBMCs labeled with FITC anti-CD3 and Alexa Fluor 405 anti-CD8 (top row), FITC anti-CD4 and Cascade Blue anti-CD3 (middle row) or FITC-anti-CD4 and Alexa Fluor 405 anti-CD3 (bottom row) followed by analysis on the FACSVantage with the 100 mW dual VLD; middle column, as above but with a single 55 mW VLD; right column, as above but with analysis on the BD LSR II with a 25 mW VLD. Median fluorescence intensities are indicated for specific and background fluorescence (on the same scattergram for Alexa Fluor 405 anti-CD8, below the bottom row for Cascade Blue and Alexa Fluor 405 anti-CD3). Ratios between specific and background fluorescence are indicated in bold type. Marker boxes and lines are equally placed from one another and are for visual reference only.
Download figure to PowerPoint
Figure 4. Green, yellow and orange emitting fluorochromes. Left column, PBMCs labeled with FITC anti-CD4 and Cascade Yellow anti-CD3 (top row), FITC anti-CD3 and Pacific Orange anti-CD8 (middle row) or FITC-anti-CD4 and Pacific Orange anti-CD3 (bottom row) followed by analysis on the FACSVantage with the 100 mW dual VLD; middle column, as above but with a single 55 mW VLD; right column, as above but with analysis on the BD LSR II with a 25 mW VLD. Median fluorescence intensities are indicated for specific and background fluorescence (on the same scattergrams for Cascade Yellow anti-CD3 and Pacific Orange anti-CD8, below the bottom row for Pacific Orange anti-CD3). Ratios between specific and background fluorescence are indicated in bold type. Marker boxes and lines are equally placed from each other and are for visual reference only.
Download figure to PowerPoint
Enhanced CFP-expressing cells were also analyzed on all of the above instruments (Fig. 5). Higher power levels reproducibly produced better CFP excitation on the stream-in-air system, although the 25 mW laser on the cuvette instrument performed the best.
Figure 5. Cyan Fluorescent Protein. Left column, Sp2/0 cells constitutively expressing ECFP analyzed on the BD LSR II with 25 mW VLD (top row), or on the FACSVantage DiVa with the 100 mW dual VLD (middle row) or 55 mW single VLD (bottom row). Right column, NIH 3T3 cells constitutively expressing ECFP, analyzed as above. Open peaks represent background fluorescence, filled peaks represent specific ECFP fluorescence for each cell type. Median fluorescence intensities are indicated for specific and background fluorescence. Ratios between specific and background fluorescence are indicated in bold type.
Download figure to PowerPoint
Collectively, these results show the utility of higher-power VLD sources for stream-in-air flow cytometry. Previous work using lower-power violet sources demonstrated their usefulness on cell sorters when detecting cell surface fluorochromes; however, concerns were raised as to whether these sources would be applicable to low-fluorescence applications such as CFP-YFP FRET. Laser diode technology has advanced to the point where higher power sources are now available; merging these sources using polarizing beamsplitters can produce an even more powerful beam, approaching the power levels of traditional krypton-ion sources. While the horizontally polarized beam would be expected to contribute less to overall excitation (based on its less-than-optimal position of the collection optics relative to the fluorochrome dipole moment), the rotational freedom of the excited fluorochrome molecules evidently allowed significant detection even with this direction of polarization. It apparently contributed to the low-fluorescence resolution of the merged beam as well. These high power sources appear to produce signal-to-background levels approaching or comparable to conventional ion sources, without the cost of maintenance associated with water-cooled lasers. Although lower-power laser sources are still very useful for cuvette instruments, a more powerful solid state source is recommended for stream-in-air systems to improve the quality of measurements of low-intensity fluorescence.