another big message on GFP

ichida at BIOMAIL.UCSD.EDU ichida at BIOMAIL.UCSD.EDU
Fri Oct 31 17:54:13 EST 1997


Dear GFPers,

During the preparation of our meeting summary Eugene Kandel posed some
interesting questions. He has given me approval to send his questions along
with Beat Ludin's responses to fluorpro. Hopefully people who think about
these questions every day will be generous enough to comment to the news
group. We all hope to keep fluorpro a friendly and easy going environment,
so don't be shy. Nothing ventured, nothing gained.

EUGENE SAYS:
1. Is GFP toxic? Since  transgenics are normal and people reported
successful establishment of GFP-positive cells from the same cell lines in
which others have reported toxicity, I would conclude that GFP toxicity is a
myth. This statement does not cover cases when the protein is expressed to
extremely high levels, but  in these conditions every protein becomes a
burden. It also does not covers the cases when cells are constantly
irradiated with the light absorbed by GFP.

2. How reliable are the observations  made with GFP fused proteins? I would
have several questions in this regard:
 a. Is fusion protein functional? Several groups reported that some of the
fusions clearly behave odd. On the bright side: several groups showed that
their chimeras can complement genetic defects, other showed that the protein
is active in biochemical assays. However, I  would not trust a study that is
not backed with  at least one of these arguments.
 b. Does the ability of the GFP tag to dimerize produce any artifacts? This
is in particular important when the tagged protein is naturally regulated by
dimerization. For example, extra dimerization domain (= GFP tag) can make
bcl2 dimerize preferentially with another bcl2 or bax molecule that contains
such a tag, at the expense of interaction with other natural counterparts
(e.g. Bad or BclX).
 c. Since GFP tags absorb huge amount of energy (GFP extinction coefficient
is only ~60%!), this energy is likely to be turned into heat. Is it the heat
that results in such a beautiful mobility of cellular structures, some of
which were thought before to be very rigid? Does it contribute to high
diffusion rates shown in the movies? Is it toxic to the cells and are the
cells continuously viewed for hours simply displaying the picture of a dying
cell? I would be a bit apprehensive to accept a study that does not rule out
these artifacts.
 d. When comparison of Ab-stained and GFP-tagged samples reveals the
difference is GFP always right? Some , but not all groups elegantly
demonstrated that Ab-staining produces artifacts that can be reproduced by
subjecting GFP-tagged cells to specific fixation procedures. I believe this
has to be done for every case in which there is a discrepancy between GFP-
and Ab-based methods. Someone suggested injection of labeled Ab. I would
think that in this case labeled Fab fragment should be less prone to artifacts.
 e. When GFP-tagged protein produces no signal in some of the cell
compartments, does it mean that this protein does not go there or it is not
fluorescent there (e.g. due to low pH)? This is very important when people
try to drive quantitative conclusions from distribution of fluorescence in
different areas of the cell. Also, we may keep in mind that pH in various
cell compartments can be changing in response to various stimuli.

3. FRET applications: 
  a. How does dimerization, uneven bleaching rates and different
pH-dependance of FRET components contribute to the interpretation of results? 
  b. Is a non-dimerizing GFP possible? 
  c. Since unfavorable dipole orientation can bring FRET to 0 even if the
molecules are very close, is it possible to interpret negative results of FRET? 
  d. In the case when FRET components are already in the same complex, but
the distance is thought to be modulated by the binding of another molecule
(e.g. Ca): does this molecule bring together the parts of the complex, or
does it improve the orientation of dipoles? 

4. Some groups observed high background fluorescence in various tissues and
organs. Can this problem be solved (e.g. by using different GFP variants or
different detection techniques)?

5. Since it is known that  prolonged culture of mammalian cells, especially
starting from low number of original clones, often results in spontaneous
changes in transformation status (increase more common than decrease), is it
informative to study tumors that first had to be passed through weeks, if
not months, of culture in order to obtain stable GFP-expressors?

6. Several groups (with industry ties) claimed that they got the brightest
GFPs. Which one of them says the truth? Has anyone compared all these GFP
forms side by side to make such claims? (transient transfection without
control for transfection efficiency does not count)

7. When people talk about gene therapy and show constructs containing only
neo and GFP, which one of these genes is considered a marker, which one -
therapeutic agent? (just kidding)

8. When a company claims that they have made a particular product, but
insist that the information about this product will not be released, is it
advertisement or scientific communication? 


Below follows Beat Ludin's response to some of Eugene's points. Please be
sure to refer to the whole context of Eugene's questions, so Beat's
comments make sense. I have done some modest editing (not changing any
content, I hope.).

BEAT SAYS:
in regard to point 1.I wouldn't say that GFP cytotoxicity is a myth.
However, current evidence 
seems to point mostly against it, IMO.


in regard to point 2. However, I  would not trust a study that is not
backed with  at least one of these arguments.
I wouldn't go as far as that. Every method can produce artifacts, and one 
has to keep oneself aware of it.


EUGENE said:b. Does the ability of the GFP tag to dimerize produce any
artifacts? This
is in particular important when the tagged protein is naturally regulated by
dimerization. For example, extra dimerization domain (= GFP tag) can make
bcl2 dimerize preferentially with another bcl2 or bax molecule that contains
such a tag, at the expense of interaction with other natural counterparts
(e.g. Bad or BclX).
BEAT:
Good point. That's definitely something to investigate more closely.

EUGENE said: c. Since GFP tags absorb huge amount of energy (GFP extinction
coefficient
is only ~60%!), this energy is likely to be turned into heat. Is it the heat
that results in such a beautiful mobility of cellular structures, some of
which were thought before to be very rigid? Does it contribute to high
diffusion rates shown in the movies? ...
BEAT:
Interesting point (except that it's the quantum efficiency and not the 
absorption coeff that is 60%). I'll do an estimation of a thermal energy 
distribution during excitation as soon as i get round to it. Of course, 
most chemical reactions (e.g. hydrolysis of ATP) are going to produce 
local temperature spikes too, so it will be interesting to calculate (I 
don't know if I can do it) whether the heat output from GFP illumination 
is significant.

EUGENE said:
d. When comparison of Ab-stained and GFP-tagged samples reveals the
difference is GFP always right? Some , but not all groups elegantly
demonstrated that Ab-staining produces artifacts that can be reproduced by
subjecting GFP-tagged cells to specific fixation procedures. I believe this
has to be done for every case in which there is a discrepancy between GFP-
and Ab-based methods. Someone suggested injection of labeled Ab. I would
think that in this case labeled Fab fragment should be less prone to 
artifacts.
BEAT:
Here I don't agree. Since there is no precedence of mis-targeting by 
GFP, I would usually trust results were a GFP-tagged proteins is clearly 
localized to any kind of structure. The two things that GFP-tagging can 
cause are a) abortive targeting so the fusion protein becomes 
cytoplasmic and b) aggregation. Immunostaining in vivo or in fixed 
specimens is more prone to errors because it depends on the accessibility 
of the epitope. However, it's clear again that none of the techniques can 
deliver perfect proof.

EUGENE:
3. FRET applications:
b. Is a non-dimerizing GFP possible?
BEAT:
I'm actually amazed that nobody seems to have tried that yet. I don't 
think there is any direct evidence that it can't be done, although W.W. 
Ward guesses that dimerization is crucial for fluorophore maturation 
('greening').

EUGENE:
 c. Since unfavorable dipole orientation can bring FRET to 0 even if the
molecules are very close, is it possible to interpret negative results of
FRET?
BEAT:
Negative results with FRET can only be meaningful if you have a built-in 
positive control with the same molecules, i.e. a regulatable interaction. 
As a stand-alone result, the absence of FRET doesn't prove anything at 
all.

EUGENE:
d. In the case when FRET components are already in the same complex, but
the distance is thought to be modulated by the binding of another molecule
(e.g. Ca): does this molecule bring together the parts of the complex, or
does it improve the orientation of dipoles?
BEAT:
W/o additional information, the only statement you can make is that there 
has been some sort of conformational change.

EUGENE:
4. Some groups observed high background fluorescence in various tissues and
organs. Can this problem be solved (e.g. by using different GFP variants or
different detection techniques)?
BEAT:
The problem can be alleviated by reducing the oxidative pressure which 
results in a lower production of fluorescent falvin co-enzymes and 
apparently decreases lipofuscin formation. However, changing the redox 
balance of the cell can also affect its differentiation.

EUGENE:
5. Since it is known that  prolonged culture of mammalian cells, especially
starting from low number of original clones, often results in spontaneous
changes in transformation status (increase more common than decrease), is it
informative to study tumors that first had to be passed through weeks, if
not months, of culture in order to obtain stable GFP-expressors?
BEAT:
This IS problematic (but then again, all model systems are prone to 
produce artifacts). However, tumor cells are 'passaged' many times in 
the body to, albeit under slightly different conditions. In the end, 
every single tumor, every single cell even, is an individual with its 
unique properties anyway. But of course, shortening the time to produce 
stables (e.g. using viruses or viral sequences) would certainly increase 
the confidence..

EUGENE:
6. Several groups (with industry ties) claimed that they got the brightest
GFPs. Which one of them says the truth? Has anyone compared all these GFP
forms side by side to make such claims? (transient transfection without
control for transfection efficiency does not count)
BEAT:
The experiments will have to be done in vitro from constructs cloned into 
identical vectors expressed in the same cells/bacteria. That's the only 
fair comparison I can think of, but it's still got its shortcomings. 

EUGENE:
8. When a company claims that they have made a particular product, but
insist that the information about this product will not be released, is it
advertisement or scientific communication?
BEAT:
It should be called a 'commercial announcement' as long as they don't try 
directly to make you buy something.


THAT'S ALL for now folks!!




More information about the Fluorpro mailing list