This document is a summary of message
from the Purdue University Cytometry Laboratories Cytometry Discussion list.
http://www.cyto.purdue.edu/hmarchiv/index.htm
------------------------------------------------------------------------------------------
Re: CFP and GFP... The Summary Posted February 11, 2003
Summary by: Tara McDonald t.mcdonald@centenary.usyd.edu.au
Below are the responses from my recent enquiry regarding the use of CFP and GFP and the issue of stray 488nm light finding its way into the detector used for CFP (filter at 485/22). During the last few days I have actually changed our primary laser, as this is easy for us to do, from 488 to 514 to excite the GFP (need to go up in the GFP filter to 555/30) and kept the 457 laser as is. There is now no problem with stray SSC light finding its way to any detector. I am using this for now but over the course of time will investigate the other options suggested to find the best for us and the lab doing the research. Happy reading.
Cheers, Tara McDonald
I'd be interested in hearing the responses you get from the list. We, too, are doing the same sort of thing and have had a heck of a time blocking the 488 from the CFP detector (we have the same setup minus the DiVa option). I spoke with a guy at Omega Optical, and he recommended the following filters: for CFP, 475AF15; for GFP, 545AF75. These should do a good job for the application -- I just haven't had a chance to try it yet. In looking at the transmission plots, however, these filters look like they will work well...
--Greg Faulkner
*******
I did run lots of CFP and YFP lots in our lab. Generally there are two different setting regarding laser usage.
1. The dual-laser system, YFP was excited by 488 nm whileas CFP was excited by 407 nm in Krypton laser.
2. Single laser system, 458 nm line from argon-ion laser can excite both CFP and YFP with right filters as you have. This system also can used for FRET measurement from cfp to yfp if they are physically close below 10nm. Our paper describing it will be coming soon in "Cytometry".
Hopefully it will help you!
Liusheng He, Ph.D, MD
*******
That is a very difficult one to do without violet excitation. If you must use the 457nm line (which also excites the wild type GFPs), you should get a custom emission filter with VERY sharp cutoffs to exclude 457nm on the low end and everything above about 482nm on the upper end – which leaves a spec. near 470/16 for CFP. Suboptimal. but will do a better job. GFP is better with a "wide" FITC filter (525/50) or a GFP (510/20) band pass filter, since the 530/30 is a bit long.
Best, Joe
*******
Advice I have received in the past for dual staining for microscopy is combine GFP with BFP or alternatively use CFP with YFP. These combinations produce more easily separated emission spectra with specific fluorescence cubes (see www.chroma.com for filter specs). The latter combination is excited with longer wavelengths and doesn't fade as fast. I have NO experience with these in flow but I thought this info might help.
Esther
*******
try tuning your laser to 407 instead of 457. also, use a filter with a more narrow and lower detection - something like a 460/20 for cfp. also, if you get bleeding into yfp, try a 546/10. these are what we sue and they come out with no problem.
good luck,
do
*******
Your results are to be expected. We use YFP instead of GFP for our 2 fluorescent protein work and use the 514nm line to excite YFP. Your only hope to use GFP and CFP together is to use a 488nm laser blocking line. Filters which block a large range of wavelengths are cheaper to buy and the ones that block only a few wavelengths on each side of 488nm are more expensive.
Good luck.
Glenn MIT
*******
That is a tough one to deal with, since the spatial separation on every sorter I have seen does not prevent stray laser light from entering the optical channels of other lasers. Since, in your case, you will need a bandpass filter with a low end at least 10-12 nm above 457 nm and a high end 10-12 nm below 488 nm, that doesn't leave much room to measure CFP. I have not used, nor have I heard of anyone else using, a notch filter centered on 488 nm in conjunction with your 485/22 to block the scattered light from the 488 nm line. But it may be possible.
Here are several other suggestions.
1) Substitute YFP for GFP. Then you can use the 514 nm line to excite it and have plenty of room to detect CFP as well.
2) Substitute YFP for CFP and excite both with 488 nm. Although there is significant spectral overlap, you can use compensation to separate the signals. This has worked well for us.
2) Trade in your second argon for a violet-enhanced krypton laser and use it a 407 nm to excite the CFP. This works quite well in conjunction with GFP, and even better with YFP. This will be costly.
3) Get one of the new benchtop analyzers from BD (LSR-II) or Cytomation (CyAn) with a violet diode laser. This will even be more costly. My impression is that the violet diodes are too low in power to be useful retrofits for the DiVa, but Cytomation did sell them for use on the MoFlo, so someone may have some experience here. If so, these lasers are much less costly than the krypton.
4) Find someone local who has an appropriate setup to run your current samples (should not be too costly).
Good luck!
Marty
*******
If your primary laser is a tunable argon, try pushing it up to 496 or 502 nm. This will decrease the CFP excitation, while still maintaining good GFP excitation. Use a 470 to 485 nm filter (10 or 15 nm wide) for CFP off of the 457 line; then pick up GFP off of the 502 nm line. The other alternative is to just use a 470/10 filter off of the 457 (I have a 470/15 which passes too much 488 light), and try to gate out the CFP from
the 457 plot. 457 in my hands excites CFP and GFP really well.
FYI, I got pretty good separation of GFP and YFP transients, using a 510/20 filter and a 550/25 filter, separated by a 555 DCLP. I wouldn't publish it, but it serves to illustrate that careful filter selection can really maximize separation of two populations. Maybe making the two beams (488 and 457) colinear, then using a 470/10 filter with a 510/10 or 510/20 filter, might give you enough separation to write home about.
Please let me know what you discover, as I do a fair amount of FP work too.
Thanks!
Andrew Beernink
*******
I have something like this, too. If you try a 510 or even a 530 instead of the 485, you may see better signal. CFP fusion proteins tend to be greener than cyan, and the spectrum does have that long tail into the green. I've sorted CFP-only cells using only the 457 line as a single laser and found better signal with the 530!
*******
We did this on a moflo using the 457 from the primary laser. We had to replace the filter in front of the
forward scatter diode and the side scatter pmt with 450/20's then the cfp gfp and yfp were easily detected with 485/22, 510/21, and 550/30 respectively.
best of luck,
sb
*******
Just use the 457 to excite the GFP as well and turn off the 488. It works well.
Larry
*******
Even though there is spatial separation between the laser lines... you still need to be sure that your emitters fully block your laser lines... there will always be some scatter in the system and the scattered laser light will
be very bright as compared to your fluorescence emission... especially CFP... If you have the 514 nm line available, I would use that for the GFP and then shift the GFP emission to 540/30... the other option is to use a very narrow 488 notch filter in the path but these are very expensive.
Thanks
Chris Baumann Ph.D.
Even with spatial separation, you're still getting plenty (relatively speaking) of 488nm laser into the CFP channel. With log amplification, it's worse. Try a bandpass filter that doesn't include the 488nm region.
Also, you can include (stack) a laser rejection filter in front of the CFP detector -- 488nm band rejection filter, or maybe just a 480 shortpass -- in combination with the CFP bandpass. Finally, you can reduce the laser crossover effect by increasing the laser interval. That can be tricky . . . you want the longest delay possible, while still obtaining reasonable signal intensities.
MAK.
*******
When analyzing CFP in the presence of 488 nm laser light, we use a 488 nm notch filter in front of the CFP detector - this is particularly important when you are using a more powerful 488 nm laser, and a less-than-optimal excitation source for CFP (like the 457 nm line). Reducing your 488 nm laser power might also help reduce the contribution of scattered laser light to the CFP signal, but you will still need the notch filter.
An even better option - can you get your primary laser to emit 457 nm, or place your 457 nm laser in the primary position? You can do CFP, GFP and scatter off the 457 nm line, and do away with the 488 nm line altogether. We have a tunable Coherent I-90 in the primary position, and have done this with 450/20 bandpass filters in front of the forward scatter diode and the side scatter PMT, in place of the usual 488/10s.
The "best" laser lines for CFP (that are readily obtainable, anyway), are the violet laser lines available from krypton or diode sources. Krypton-ion lasers are expensive and a little fussy maintenance-wise, but a high-power (15-30 mW) violet laser diode with mounting hardware can be had for a good bit less than $10,000. We've mounted both the Power Technologies and the Coherent violet laser diodes on our Vantage and have gotten surprisingly good results for CFP excitation. You would then do GFP off the 488 nm primary laser. Some data and instructions for mounting a violet diode on a Vantage can be found on our website at http://home.ncifcrf.gov/ccr/flowcore/index.htm (go to the Projects page, and look for the violet laser diode section).
Hope this helps.
Bill Telford
NCI-NIH
*******
I had the same problem on my Vantage SE DiVa with the laser setup that you are using. This is what works for me: Tune your primary laser to 457nM, put 457/15 filters in front of FSC and SSC. For CFP, I use a 505SP dichroic for a beamsplitter and a 485/22 filter in front of the PMT. I detect CFP in the "traditional
FL1" PMT. Make sure your iris is closed all the way.
Tune the second laser (I have a Coherent Spectrum) to 488nM. Detect GFP in the appropriate PMT for that laser using the 530/30 filter. An even better combination is CFP / YFP; you can tune your second
laser to 514nM and use a 550/30 filter for YFP. With that combination on my instrument, I do not need compensation.
Barbara J Taylor I've realized that investigators often don't want to change FP's, as they've already done the molecular bio with the ones they have. If they had consulted with a flow jock/jockette first, they would have been able to figure out a better combination than CFP/GFP.
As a follow-up to the low power diode discussion, I tested a couple of different diode lasers for use with CFP. As it turns out, 10 mW at 430 or 442 gave excellent excitation of CFP (though poor CV's) on a MoFlo. I believe the price from Company X was $15K plus or minus, so they're not cheap. I know that there are other options at or near those wavelengths, probably less than $10K. The good news is, they're relatively small, run on 110V, and don't require forced air cooling. In one case, I triggered off of UV SSC, gated on a gate of UV SSC vs UV-excited PI, and collected CFP fluorescence at 485 nm.
In all, I've tested 5 different diode/SS lasers, so feel free to contact me directly -- I'm happy to share the data and the manufacturers' info.
Andrew
-------------------------------end of summary – this summary posted February 11, 2003 ---------------------------
