Dear colleagues,

In a spectrofluorometer the problem of instability of light source is solved by measuring the ratio of fluorescence intensity over incident light intensity. May be the same method can be used in confocal microscope. Part of the incident light can be reflected into reference chanel and fluorescence signal can be corrected simultaneously. 

Bets regards,

Oleg Andreev, Ph. D.
Confocal Microscopy and Fluorescence Spectroscopy Lab.
Dept. Mol. Biol. & Imm.
University of North Texas Health Science Center
3500 Camp Bowie Blvd.
Fort Worth, Texas 76107
Phone: (817) 735-2127
E-mail: [log in to unmask]

>>> Kees Jalink <[log in to unmask]> 11/05/99 09:46AM >>>
Dear confocalists,

    We want to bring up a point that is potentially of huge concern to
all of you that do "quantitative" fluorescence (or transmission,
reflection) during prolonged times. It is especially important to
physiologists and to users that compare intensities from preparation to
preparation. We observed that the light intensity delivered to the lens
in our fibre-coupled, AOTF-equipped instruments fluctuated enormously
with time. Fluctuations were up to 40%. Often, the fluctuations are
somewhat sinusoidal, with periods from 2 minutes to 30 minutes, and are
temperature dependent (but not induced by temperature changes!).
Especially the fast oscillations could easily be mistaken for
oscillations in physiological parameters (e.g. cytosolic Ca2+, in a
non-ratiometric experiment)!! Sudden drops in intensity by as much as
25% do also occur.
    Both in mixed gas lasers and in multi-laser systems, the intensities
vary independently (say,  488 oscillates by 30% with a period of 8
minutes, and 567 varies with 25% over 13 minute periods). If one were
doing a excitation ratiometric experiment, total error could thus get
much more than 50%.
    We have brought this to the attention of the manufacturer, Leica,
and they have been very cooperative.  According to them, the problem is
likely to be present in every fibre-coupled system, because it stems
from polarization issues in the fibre-couplings. Using different
alignment criteria, they were able to reduce it to 5-10%. Systems that
use direct laser coupling should be free of this artefact.

    It may be a good idea that all of you that use fibre-coupled systems
perform some test on your own setup. We have tested fluorescent mode and
transmission mode with essentially the same results. The fact that
simultaneously assayed laser lines oscillate totally independent rules
out detector fluctuations or the like as source of the problem. We have
found fluctuations on our own 3 Leica systems (extensive tests) and on
Biorad and Zeiss systems (very limited tests). The tests basically
involve imaging in transmission mode (no preparation) a few times per
minute over 2 hour periods, and plotting total image intensity versus
time.


    The big question of course is: how to work around it. Theoretically
one could add stably fluorescent objects to the preparation (e.g.
molecular probes tetraspeck beads), make sure that at least one such
bead is present along with the cells in the field of view, and determine
fluo levels of the beads (at all used lines!!) simultaneously; then
divide each wavelength picture by that level. Disadvantage is that you
have to add the beads in your experiments, that they have to be roughly
equally intense as the dyes you are using (to prevent saturation, at all
wavelengths), they have to have roughly the same spectra as the dyes you
use, and they have to be small (not to take up most of your picture) and
big (not to drift out of focus with a micron Z-axis drift) at the same
time. Clearly not very practical.
    To our opinion, a better solution would involve measuring laser line
intensities independently (using a photodiode) immediately before taking
each image (since on second time scale, the fluctuations are small), and
then ratioing the fluctuations out in software. When an AOTF is already
fitted, this should involve little hardware (photodiode) and moderate
software changes by the manufactures. In the mean time, we would be very
grateful for all suggestions.

    If there is interest, we can attach our test results as an excel
graph. We would be very interested to hear test results from your
systems!


--

Kees Jalink Ph.D.                                Lauran Oomen
Ph.D.                      Lenny Brocks Ph.D.
Dept. of Cell Biology H1                     Digital Microscopy
facility               Digital Microscopy facility

The Netherlands Cancer Institute
Plesmanlaan 121, 1066CX Amsterdam, the Netherlands