Hi Flowers,
first of all: Thanks for plenty of replys and advice on my former questions
about viability measurements with AAD and PI. I did some more experiments
and ended up with 5 microgram AAD /ml. 1 or 2 microgram work as well, but
there is a better separation with 5 microgram. There are differences
between PI and AAD, which also depend on how you set the regions
(especially with PI, which renders quite variable results) - but they are
not too bad when I use lower concentrations of PI than I did before, like
many of you recommended. I did it like Howard Shapiro recommended: I
decided what is positive, and I felt best with the set-up that I have now
chosen. I still incubate 20min on ice, again because separation is clearer,
5min would do as well. I cut the peak in the middle. If anybody is
interested in more data on this, I can supply conrete numbers. And, of
course, if anyone has any more comments, please let me know!
A lot of you asked me to post a summary of replies. Sorry it took so long,
but here they are:
The Question was:
I am stuck with a very basic problem. I am trying to separate dead from
viable cells. People in my lab have always done this with 10µgPI/ml plus
RNase for 20min room temperature. I have tried two other protocols: PI
2µg/ml 5min on ice and AAD 1µg/ml 30min on ice. Every protocol yields
different counts of dead cells - from 16 +/-3% (AAD), 22+/-7% (PI) to
27+/-5% (PI combined with RNase).
Is this a probem of concentration - are DNA-Dyes titreable like an antibody?
Is it a question of temperature or time?
We do not understand this, because we always thought that a dead cell is
dead, has permeabilised membranes and therfore takes up the dye - not
depending on concentration. We always thought that higher concentration
only gives you a better separation of peaks.
How do I know which concentration is the right one for my AAD?
And there ist still another problem with PI: If I stay with PI - is it more
accurate to compensate PI out of FL2 (which in fact means 99,9%
compensation) or is it better to leave the PI signal where it is, since the
dead cells are gated out anyway before the PE stained cells in FL2 are
analyzed?
And these were the answers:
1. Howard Shapiro wrote:
Dead cells with permeabilized membranes should take up
otherwise impermeant nucleic acid dyes such as 7-AAD and PI. Propidium (PI)
and ethidium (EB) should not be used interchangeably; ethidium has only a
single positive charge, and can get into intact cells, especially at
relatively high pH, but is pumped out. Propidium has two positive charges
and generally does not get in. TO-PRO-1 and TO-PRO-3 also have two
positive charges, and, in my limited experience (unpublished) using PI (488
nm excitation) and TO-PRO-3 (633 nm excitation) together, cells that take
up one dye take up the other.
However, my colleagues and I have also observed (Novo et al, Antimicrob
Agents Chemother 2000, 44:827-834 [should be on PubMed]; Shapiro, Cytometry
2001, 43:223-226 and Shapiro, J Microbiol Methods 2000, 42:3-16 [this is in
a special issue on Microbial Analysis at the Single Cell Level, and it and
all other articles are available free at
http://www.elsevier.com/locate/jmicmeth]) that, under some conditions,
bacteria that maintain their membrane potentials, and which therefore have
intact membranes, take up PI or TO-PRO-3.
The "permeabilization" that permits PI uptake is usually conceived as a big
hole in the membrane, which should also allow equilibration of ion
concentrations across the membrane, reducing membrane potential to
zero. This is clearly not the situation just described; the PI or TO-PRO-3
would have to get in via some more specific mechanism, e.g., a transport
protein, and I have heard that such transport proteins have been
described. Suffice to say that the "membrane permeability" which is taken
as evidence of nonviability when dye exclusion tests are used may be
substantially more complex than was originally thought.
At a simpler level, there is the possibility that the different
experimental conditions you use could affect the development of membrane
permeability in dying (e.g., apoptotic) cells; while early apoptotic cells
exclude PI and can be identified only by other criteria (Annexin V
binding,mitochondrial membrane potential changes), these cells eventually
end up permeable to PI. Along the way, you may see weak staining, so you
have to decide what is "positive".
>How do I know which concentration is the right one for my AAD?
That would depend on whether you can resolve your initial problem or not.
Under normal circumstances, you shouldn't need to titrate nucleic acid dyes
for dye exclusion tests.
>And there ist still another problem with PI: If I stay with PI - is it more
>accurate to compensate PI out of FL2 (which in fact means 99,9%
>compensation) or is it better to leave the PI signal where it is, since
>the dead cells are gated out anyway before the PE stained cells in FL2 are
>analyzed? I'd leave the signal as is, since it puts the dead (or
permeabilized) cells
off scale and lets you gate them out.
2. Mario Roederer wrote:
You're using way too high a PI concentration. Go down to 1 ug/ml. I
suspect you're getting some background uptake by live cells at that
concentration. Don't compensate the PI signal--just gate the + cells out.
3. Mark KuKuruga wrote:
Dead cells are stained with PI almost immediately. There's really no
reason to wait several minutes. Also, RNAse is added when we stain for DNA
content analysis . . . it improves resolution . . . and is not necessary in
a "live/dead" application. I haven't actually looked at it, but I also
believe 7-AAD staining of DNA in dead cells will be "immediate."
I'm not surprised there's a concentration dependant difference, although I
think you'll see this more with 7-AAD than with PI (since PI is so much
brighter). I've gone as low as 0.5 ug/ml PI and still seen good dead cell
detection, but also know that others will recommend higher dye
concentrations . . .
Regarding your "compensate/no compensate" question . . . I think this
depends somewhat on what you use the PI for. If you're interested in
detecting live vs dead cells, I think it's important to correctly
compensate so that you can determine live/dead cutoffs amongst all
parameters. This is easier to achieve at lower dye concentrations, since
the amount of crossover will increase with signal intensity. In a dead
cell exclusion scenario, compensation is irrelevant since the PI positive
cells are excluded by gating. We usually compensate these anyway . . .
again because it looks better . . . but it's really not necessary.
4. Craig Shadbolt wrote:
I have been using DiBAC (from molecular probes) and PI as injury/dead
stains on E. coli and Listeria innocua. I use simulated food processing
conditions in broth to measure viability at different time/temperature
combinations.
The sad thing I find is that flow cytometry is only effective down to
0.1-1% of starting numbers (about 2-3 log cfu/ml). Considering I regularly
examine viability over 8log cfu/ml this is not very helpful. Flow is good
for looking at the distribution of healthy/injured/dead cells over that
first 99% or so (which is interesting) but beyond that I often find my
viable counts start to increase in direct contrast to my plate counts.
Other people have published information on this (reference to follow) in
which they speculate that DNA leaks out from permeabilised membranes,
forming "ghost cells" which appear as viable on a flow output because they
do not stain with PI. I am working on a method to see this ghost cell
formation using video imaging. Just waiting on some equipment to arrive!
The reference which describes these ghost cells is:
Brul, S., Rommens, A.J.M., Verrips, C.T., 2000. Mechanistic studies on the
inactivation of Saccharomyces cerevisiae by high pressure. Innov. Food Sci.
Emerg. Technol. 1, 99-108.
I hope this may be of some help. Perhaps you are working with eukaryotic
cells which, obviously, is a different kettle of fish.
4. Horst von Bernuth wrote (roughly translated from german):
Concentration and incubation time are important with PI. After a longer
incubation with high concentration even live cells are stained without
previous perforation/tratement with detergent. There are different staining
protocols for the nucleus that take advantage of that.
A good prtocol for PI: 1 microgram/ ml PI, 1 min before start of measurement
5. Jeff Carrell wrote:
You can definitely reduce your PI concentratation. We use 0.5 ug/ml; the
compensation is never more than about 30%.
6. Louis King wrote:
As for PI-- First do a trypan blue viability count so you know
what your cells are by trypan blue exclusion. Next, remember you are not
staining DNA for cell cycle! If you are going to use PI, throw out RNAse
and detergents immediately and keep the cells in some protein (2-4% heat
inactivated and filtered fetal bovine serum, BSA, your favorite protein),
use PI at 1 ug/ml. Now setup 10 samples and do a PI timed stain course.
I've chosen 3 minutes immediately prior to running because that
gives me the % dead nearest the value obtained for trypan blue. Finally,
that PI exclusion is not a absolute thing and it will eventually get into
any cells that have undergone manipulation. I can not tell you whether
that means all manipulated cells have slightly damaged membranes and thus
PI permeable or PI will eventually permeate cells!
Repeat the above paragraph for AAD at 1 ug and add that staining
with this dye appears to "quench" the flr in the PE channal. To my joy
DAPI at 1 ug staining exactly like you do for cell cycle
(without detergents of course) gives vary closely trypan blue counts for up
to 45 minutes post DAPI addition! The only problem here is you have to
have a uv laser to do viability. The good news is the dye does not affect
data on the primary laser.
7. Gerhard Nebe-von-Caron wrote:
To clarify a few points about live and dead. Permeabilized does not
necessarily mean dead as there is plenty of literature telling you about
transient permeabilisation. However, if you can exclude that it's the
closest you get to dead cell detection. The use of solvents and detergents
can also alter dye
exclusion properties or add to toxicity.
Concentration will change your staining kinetics and Howard has already
mentioned the problems of active dye transport in either way that will add
to that effect. At low dye concentration you will also experience more
signal variation based on the amount of binding sites available to the dye.
Thus the
amount of free double stranded RNA will reduce the amount of free PI and
your histograms will change with your cell numbers.
The way you set your regions is another factor contributing to variation in
% positive cells, particular if you include weak positive cells. Perhaps
you can put some example data on your web site for further discussion
Kerstin Buettner
Institute of Biophysics and Radiobiology
University of Hamburg
Martinistr. 52
20246 Hamburg
Phone: 0049-40-42803-2938
Fax: 0049-40-42803-5139
E-Mail: k.buettner@uke.uni-hamburg.de
This archive was generated by hypermail 2b29 : Wed Apr 03 2002 - 11:57:36 EST