>
> There is a real problem emerging here that people on the list may like to
> comment on... As fluorophores, optical systems and detectors all become
> more efficient, getting complete discrimination between different
> fluorophores in multiple-labeled preparations is becoming more and more
> difficult with the current lot of labels. We recently got some very narrow

...text deleted...

> It may also be worth pointing out that the 488 nm blue line is actually
> quite good at exciting Cy3, especially in well labelled preparations, so
> you can get cross-excitation as well! (Texas Red is not excited
> significantly by this line).
>
488nm can excite TR also, and whether it is significant depends on the
degree of staining (relative quantity of dye on the sample). We have
only an argon ion laser, which makes it even harder to separate
multiple stains. Here are some options:

1. If you have it, use the 568nm line or 543nm line to excite
rhodamines without exciting fluorescein. However, the cleanest and
easiest way to do multiple dye excitation is with one excitation line,
since then you only do one scan, and at least avoid any change in
excitation volume location caused by the change in excitation
wavelength.

2. I have a description of how I do FITC/CY3/TR triple stained samples
on my web page. I usually use 514nm excitation, and detect FITC with a
20nm wide filter centered at 535nm (HQ535/20). The FITC is quite
strong in the CY3 channel, but not in the TR channel. I set up the
mixers on the Biorad MRC1024 to subtract off the FITC signal from the
CY3 channel, and I subtract the CY3 (or TRITC) signal from the TR
channel.

> So in the end, we recommend that everyone using multiple-labelled
> preparations must make a series of single-labelled controls and explicitly
> test for bleed through and cross excitation on a run-by-run basis...
>
absolutely. However, you can save some work here, because the FITC
channel is pretty clean, and you can usually find features that are
isolated, so that you can use them as a reference to adjust the mixing
parameters. In one case, where TRITC was being used to stain tight
junctions, it was easily identified and subtracted off of the TR
channel.

I describe the filter set and link to an image in URL
http://optics.jct.ac.il/~aryeh/Confocal/filter_set.html

For qualitative results (localization and such) you can do well with
careful balancing of the channels and staining. But if you have a
stain that is 1000 times stronger than the others, you will have
trouble. Also, for quantitative measurements (eg, relative strength of
different stains), you have to do very careful controls.

I should add that if I had 12 bits dynamic range, it would be *much*
easier to correct the channels in postprocessing. This is a pet peeve
of mine, because in accumulation mode there is no reason not to
provide this dynamic range.

--aryeh
Aryeh Weiss                          | email: [log in to unmask]
Department of Electronics            | URL:   http://optics.jct.ac.il/~aryeh
Jerusalem College of Technology      | phone: 972-2-6751146
POB 16031                            | FAX:   972-2-6422075
Jerusalem, Israel                    | ham radio: 4X1PB/KA1PB