Brightness of dyes
roederer at nih.gov
Sun Mar 24 08:05:47 EST 2002
The discussion on the brightness of dyes on the listserv has been
useful to introduce the complexities involved in trying determine
which dye is "better".
One cannot rely solely on data such as what Dick Haugland presented
-- if you read his "fine print", the graph is the product of the
measured quantum yield (relative to that of a reference dye)
multiplied by the dye:protein labeling ratio. Since I have no idea
what the "reference dye" is, and I have no idea in what way the
dye:protein ratio scales with the number of photons coming out of the
conjugate, I am taking this graph with a grain of salt.
Indeed, for flow cytometry, "brightness" is a very different issue
than for microscopy. Several people noted that Alexa488 is superior
to Fluorescein in microscopy--primarily because it bleaches less
quickly (therefore, it's not that it's necessarily "brighter" to
begin with--it just stays brighter over the integrated time of the
experiment). However, bleaching is much less of an issue for flow
cytometry, where the dye stays in the laser for a few microseconds.
Thus, whereas in microscopy, Alexa488 may be several times brighter
than fluorescein, on a flow cytometer, this ratio may only be 1 or 2.
If nothing else, this shows the limited utility of the graph that
Others have reported variable results in comparing Alexa488 to
fluorescein by flow cytometric experiments. This might be explained
by the different conditions of the flow cytometers, the two most
important being stream velocity (and thus how long the dye is in the
laser), and the laser power.
One of the important brightness-related issues in flow cytometry,
where a dye is in the excitation beam for a very short time, is the
fluorescence lifetime of the dye (i.e., how long after it is excited
by the laser does it emit a photon). The shorter the lifetime, the
faster the dye is available for re-excitation, and, thereby, the
"brighter" it will be during the time it is in the laser. Note that
while quantum yield depends on fluorescence lifetime, it also depends
on other factors. I think (but am not positive), that Dick's graph
of brightness really should be normalized by the fluorescence
lifetime to get a somewhat more accurate picture of brightness.
In addition, the absorption coefficient (epsilon) of a dye can be
important to brightness--but much more so on low laser power systems
(like benchtop instruments) than high laser power systems (like
sorters). This is because on high power systems, there are more than
enough photons present to excite the dye; it will always be in the
excited state; on the low power systems, the exciting light is
subsaturating and hence the absorption coefficient of the dye becomes
important for "brightness". This might explain why different groups
see different relative brightnesses of Alexa488 and fluorescein: the
relative brightness may well be dependent on the laser power.
But there are many other considerations that lead to "relative"
brightness. For example, one reason we switched from using TexasRed
to Alexa595 is that the former dye generates conjugates that are much
more sticky than Alexa595. Thus, while there may not have been any
difference in their "brightness", practically the Alexa595 was far
superior because the background was lower. This increased the
"relative brightness" (signal of the positive cells vs. the negative
cells) of Alexa595 vs. TexasRed.
These issues simply illustrates that "brightness" on the flow
cytometer is very complex, and really must be determined empirically
for any particular conditions. One need not conjugate dozens of
antibodies; one need only conjugate a single antibody. However, to
properly do the comparison, one must do a titration of dye:protein
for both conjugates, optimize both conjugates, and then compare
them--on stained cells, on the flow cytometer, and taking into
account background binding (i.e., staining cells that do not express
My conclusion is that the "brightness" of any dye, conjugate, or
staining system must be evaluated empirically for the type of
experiment that you do, utilizing your particular hardware. What may
be better for some people could be worse for your system because of
hardware differences. Rather than trying to model all of these
differences to come to some prediction of brightness, just do the
experiment once and determine the answer for your system.
PS. Howard Shapiro ("The Howard") has long criticized me for
referring to the relative signal intensity of positive and negative
cells as brightness--he is quite correct that brightness is not
really the right word here. That's why I've tried to put the word in
quotation marks as much as possible. By "brightness" in the above
discussion, I try to convey what we think of as "useful brightness"
or "biological brightness"--a value that is empirically determined
and is useful from a practical standpoint. What chemists and
physicists think of as brightness is only partly useful to flow
Mario Roederer, Ph.D.
Chief, ImmunoTechnology Section and Flow Cytometry Core
Vaccine Research Center, NIAID, NIH
40 Convent Dr., Room 5509
Bethesda, MD 20892-3015
Phone: 301 594-8491
FAX: 301 480-2651
More information about the Cytometry