Confocal of membrane proteins.

[A. Hamdoun]

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

I am attempting to capture the distribution of membrane proteins on microvilli using a
Zeiss LSM 510 Meta. I find that although I have what seems to be fantastic expression of
mCherry or eGFP tagged protein (as determined by epi-fluor) the signal on the confocal
seems excessively photon-starved.

I am wondering what else could be causing this beyond the usual limitation of confocal
pinhole. For instance could this is related to the suboptimal wavelength of the HeNe laser
(543 vs. the excitation of 587 for mCherry)? The limit of the detector signal:noise? Or  
just some problem in how I am using the confocal?

I find that I have to image the membrane at 40x, zoomed in and using channel mode
rather than lambda (i.e long pass emission collection across many of the spectral
channels of the meta PMT), and that I have to open up the pinhole to 2 Airy disk units.

This also leads to second question which is how realistic is spectral detection if it further
reduces the signal that can be captured from any given FP. Has anyone here had any
experience with spectral detection on this scope using multiple FPs and membrane
proteins? Is there a physical limitation for imaging imposed  by how many FPs are in the
membrane (as compared to cytosolic proteins)?

thanks in advance for any advice.

[Stephen Cody]

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

The Green HeNe laser by comparison tends to be very weak irrespective of which brand confocal you use. If the excitation max is a fair distance away from your laser, then these two problems will compound. Opening the pinhole as you have done seems like a great idea. I would also look at the objective you are using. To maximise signal you must use the highest NA available, and match the refractive index of your specimen to the lens. Are you using a water, oil or glycerine lens? You should also look at using a lens with an adjustable collar, I like to adjust these by trial and error looking at the sharpness of widefield fluorescence by eye. You need to adjust the collar then the fine focus, and repeat over until the sharpest image is reached. When you do this (spherical aberration correction) much more light will go through the pinhole.
 
Also it is no secret that there will be some signal loss when using spectral detectors.
Stephen H. Cody
Microscopy Manager
Central Resource for Advanced Microscopy
Ludwig Institute for Cancer Research
PO Box 2008 Royal Melbourne Hospital
Parkville, Victoria,      3050
Australia
Tel: [phone]    Fax: [phone]
email: [hidden email]
www.ludwig.edu.au/labs/confocal.html
www.ludwig.edu.au/confocal

Tip: Learn how to receive reminders about you microscope booking:
http://www.ludwig.edu.au/confocal/Local/Booking_Hint.htm Type your signature here

________________________________

From: Confocal Microscopy List on behalf of A. Hamdoun
Sent: Fri 08/08/2008 9:53 AM
To: [hidden email]
Subject: Confocal of membrane proteins.



Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

I am attempting to capture the distribution of membrane proteins on microvilli using a
Zeiss LSM 510 Meta. I find that although I have what seems to be fantastic expression of
mCherry or eGFP tagged protein (as determined by epi-fluor) the signal on the confocal
seems excessively photon-starved.

I am wondering what else could be causing this beyond the usual limitation of confocal
pinhole. For instance could this is related to the suboptimal wavelength of the HeNe laser
(543 vs. the excitation of 587 for mCherry)? The limit of the detector signal:noise? Or
just some problem in how I am using the confocal?

I find that I have to image the membrane at 40x, zoomed in and using channel mode
rather than lambda (i.e long pass emission collection across many of the spectral
channels of the meta PMT), and that I have to open up the pinhole to 2 Airy disk units.

This also leads to second question which is how realistic is spectral detection if it further
reduces the signal that can be captured from any given FP. Has anyone here had any
experience with spectral detection on this scope using multiple FPs and membrane
proteins? Is there a physical limitation for imaging imposed  by how many FPs are in the
membrane (as compared to cytosolic proteins)?

thanks in advance for any advice.

[Julio Vazquez]

Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

-

It could be a simple thing such as a bad laser, or a misaligned pinhole. A few controls might help diagnose the problem. Try imaging your sample of the regular detectors (not META), and see what laser powers and PMT gains you need to get an image. Compare to the settings needed on META. Meta will be probably less sensitive, maybe 50-70% of regular PMT (ball park, but all other things being equal as far as you can tell, such as having a detection band on your META similar to the Emission filter on your normal PMT channel).   If the difference is significantly greater, then something is wrong. Maybe do a similar comparison with another slide stained with FITC or Rhodamine, or such. If the red channel looks bad in both cases, but green channel looks good, maybe the 543 laser is dying. You can check its actual power with a power meter at the lens, and compare that to the expected value. 

Another test you can do is to take different images on the META with different pinhole apertures. If you see a sudden increase in image intensity, at values greater than 1 Airy unit, it probably means your pinhole is misaligned. If the pinhole is well aligned, you should see image intensity increase progressively. 

Exciting at 543 nm versus 587 should reduce your excitation efficiency roughly by half, so you lose a little, but it shouldn't be that dramatic. 

If you can exclude a laser problem or a pinhole problem, then there is probably some more complex issue and you may need to have a qualified person have a look at the instrument. 

--

Julio Vazquez

Fred Hutchinson Cancer Research Center

Seattle, WA 98109-1024

http://www.fhcrc.org/

-

On Aug 7, 2008, at 4:53 PM, A. Hamdoun wrote:

Search the CONFOCAL archive at

http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

I am attempting to capture the distribution of membrane proteins on microvilli using a 

Zeiss LSM 510 Meta. I find that although I have what seems to be fantastic expression of 

mCherry or eGFP tagged protein (as determined by epi-fluor) the signal on the confocal 

seems excessively photon-starved. 

I am wondering what else could be causing this beyond the usual limitation of confocal 

pinhole. For instance could this is related to the suboptimal wavelength of the HeNe laser 

(543 vs. the excitation of 587 for mCherry)? The limit of the detector signal:noise? Or  

just some problem in how I am using the confocal?

I find that I have to image the membrane at 40x, zoomed in and using channel mode 

rather than lambda (i.e long pass emission collection across many of the spectral 

channels of the meta PMT), and that I have to open up the pinhole to 2 Airy disk units. 

This also leads to second question which is how realistic is spectral detection if it further 

reduces the signal that can be captured from any given FP. Has anyone here had any 

experience with spectral detection on this scope using multiple FPs and membrane 

proteins? Is there a physical limitation for imaging imposed  by how many FPs are in the 

membrane (as compared to cytosolic proteins)?

thanks in advance for any advice.

[G. Esteban Fernandez]

Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

I second Julio's suggestion that pinhole misalignment is a likely culprit.  Use Maintain > Pinhole in the LSM software and play with the X and Y positions for Pinhole 1 (META detector pinhole) @ 1-2 Airy units while FastXY-ing and see if the signal gets better (maybe Z position too).  Before you start moving the pinhole, write down the original position values.

 

-Esteban

 

On Thu, Aug 7, 2008 at 7:58 PM, Julio Vazquez <[hidden email]> wrote:

Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

-

It could be a simple thing such as a bad laser, or a misaligned pinhole. A few controls might help diagnose the problem. Try imaging your sample of the regular detectors (not META), and see what laser powers and PMT gains you need to get an image. Compare to the settings needed on META. Meta will be probably less sensitive, maybe 50-70% of regular PMT (ball park, but all other things being equal as far as you can tell, such as having a detection band on your META similar to the Emission filter on your normal PMT channel).   If the difference is significantly greater, then something is wrong. Maybe do a similar comparison with another slide stained with FITC or Rhodamine, or such. If the red channel looks bad in both cases, but green channel looks good, maybe the 543 laser is dying. You can check its actual power with a power meter at the lens, and compare that to the expected value. 

Another test you can do is to take different images on the META with different pinhole apertures. If you see a sudden increase in image intensity, at values greater than 1 Airy unit, it probably means your pinhole is misaligned. If the pinhole is well aligned, you should see image intensity increase progressively. 

Exciting at 543 nm versus 587 should reduce your excitation efficiency roughly by half, so you lose a little, but it shouldn't be that dramatic. 

If you can exclude a laser problem or a pinhole problem, then there is probably some more complex issue and you may need to have a qualified person have a look at the instrument. 

--

Julio Vazquez

Fred Hutchinson Cancer Research Center

Seattle, WA 98109-1024

http://www.fhcrc.org/

-

On Aug 7, 2008, at 4:53 PM, A. Hamdoun wrote:

Search the CONFOCAL archive at

http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

I am attempting to capture the distribution of membrane proteins on microvilli using a 

Zeiss LSM 510 Meta. I find that although I have what seems to be fantastic expression of 

mCherry or eGFP tagged protein (as determined by epi-fluor) the signal on the confocal 

seems excessively photon-starved. 

I am wondering what else could be causing this beyond the usual limitation of confocal 

pinhole. For instance could this is related to the suboptimal wavelength of the HeNe laser 

(543 vs. the excitation of 587 for mCherry)? The limit of the detector signal:noise? Or  

just some problem in how I am using the confocal?

I find that I have to image the membrane at 40x, zoomed in and using channel mode 

rather than lambda (i.e long pass emission collection across many of the spectral 

channels of the meta PMT), and that I have to open up the pinhole to 2 Airy disk units. 

This also leads to second question which is how realistic is spectral detection if it further 

reduces the signal that can be captured from any given FP. Has anyone here had any 

experience with spectral detection on this scope using multiple FPs and membrane 

proteins? Is there a physical limitation for imaging imposed  by how many FPs are in the 

membrane (as compared to cytosolic proteins)?

thanks in advance for any advice.

--
G. Esteban Fernandez, Ph.D.
Associate Director
Molecular Cytology Core Facility
University of Missouri
120 Bond Life Sciences Center
Columbia, MO 65211

http://www.biotech.missouri.edu/mcc/

573-882-4895
573-884-9395 fax

[Julio Vazquez]

Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal -

...but I also wonder: eGFP and mCherry are pretty well separated spectrally. Why don't just use standard confocal channels, with for instance Ex=488, Em = 500-550 for EGFP and Ex= 543 and Em around 585-660 nm (or your best matched filter) for mCherry?. This should give you low or negligible bleed-through and significantly better images than with the META.  The META detector has limitations in terms of sensitivity and noise, and you probably should use it mainly when your GFPs are too close to be separated with your regular filters,  or to remove background autofluorescence. 

--

Julio Vazquez

Fred Hutchinson Cancer Research Center

Seattle, WA 98109-1024

http://www.fhcrc.org/

----

On Aug 7, 2008, at 4:53 PM, A. Hamdoun wrote:

Search the CONFOCAL archive at

http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

I am attempting to capture the distribution of membrane proteins on microvilli using a 

Zeiss LSM 510 Meta. I find that although I have what seems to be fantastic expression of 

mCherry or eGFP tagged protein (as determined by epi-fluor) the signal on the confocal 

seems excessively photon-starved. 

I am wondering what else could be causing this beyond the usual limitation of confocal 

pinhole. For instance could this is related to the suboptimal wavelength of the HeNe laser 

(543 vs. the excitation of 587 for mCherry)? The limit of the detector signal:noise? Or  

just some problem in how I am using the confocal?

I find that I have to image the membrane at 40x, zoomed in and using channel mode 

rather than lambda (i.e long pass emission collection across many of the spectral 

channels of the meta PMT), and that I have to open up the pinhole to 2 Airy disk units. 

This also leads to second question which is how realistic is spectral detection if it further 

reduces the signal that can be captured from any given FP. Has anyone here had any 

experience with spectral detection on this scope using multiple FPs and membrane 

proteins? Is there a physical limitation for imaging imposed  by how many FPs are in the 

membrane (as compared to cytosolic proteins)?

thanks in advance for any advice.

[leoncio vergara]

Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Dear Confocalists

 

One of the users of our facility has prepared cells in which he stained a surface protein by using an external epitope tag and a red labeled secondary, he also stained the same protein using an internal epitope and a green labeled secondary in order to asses the total level of expression per cell.

 

He wants to show the level of surface expression relative to the total cell expression of that protein. First he tried to approach this just by showing representative images of the different treatment conditions, of course the reviewers of his manuscript asked him for quantification in the form a ratio of surface to total protein expression.

 

I told him that he must be aware that although it is easy to calculate a ratio of the red intensity to the green intensity, the resulting value would be just proportional to the absolute ratio of surface expression and could only be used to compare the relative changes among the different treatments, but he would have a lot of variability due to uncontrolled variation in illumination sources, etc etc….

 

Of course trying to segment the green image in peripheral (near and at the surface) and total fluorescence would not be a solution either, due to the obvious limitations in the optical resolution.

 

I would appreciate any comments in that sense and if a relative ratio of red to green fluorescence would be in anyway acceptable to answer the reviewer.

 

Thanks in advance

 

Leoncio A. Vergara MD

Assistant Professor
Laboratory of Protein Misfolding Diseases (lab-PMD),
George and Cynthia Mitchell Center for Neurodegenerative Diseases Research.

Director of the Optical Imaging Lab. (OIL),

Dept. of Neuroscience and Cell Biology

University of Texas Medical Branch (UTMB)

301 University Blvd

Galveston, Texas 77555-0641

OIL phone: [phone] 

Lab-PMD phone: [phone]

fax: [phone]

 

 

[Julio Vazquez]

Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal -

Hi Leoncio, 

If I understood the experiment correctly, and the red label is membrane protein and the green label is total protein, then calculating the ratio between red and green (after background subtraction and/or other possibly needed corrections) should give you a quantitative measure matching the visual representation provided by the images (that is, you are comparing the relative amounts of red and green), so this is probably what the reviewer wants, and I don't see why they wouldn't be satisfied with it.

Now, the actual biology may be a different thing. Suppose all the potential membrane sites are occupied by the protein provided there is sufficient protein made. If now you make 10 times more protein, the excess will all be cytoplasmic. therefore, the ratio of red to green will decrease, but that has no relevance to how much there is in the membrane. So the question really is, which numbers matter... is it the absolute amount of protein in the membrane, or the ratio of membrane to free protein... this is something for your scientist to decide...

--

Julio Vazquez

Fred Hutchinson Cancer Research Center

Seattle, WA 98109-1024

http://www.fhcrc.org/science/shared_resources/imaging/

===

On Aug 8, 2008, at 9:37 AM, Vergara, Leoncio A. wrote:

Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Dear Confocalists

 

One of the users of our facility has prepared cells in which he stained a surface protein by using an external epitope tag and a red labeled secondary, he also stained the same protein using an internal epitope and a green labeled secondary in order to asses the total level of expression per cell.

 

He wants to show the level of surface expression relative to the total cell expression of that protein. First he tried to approach this just by showing representative images of the different treatment conditions, of course the reviewers of his manuscript asked him for quantification in the form a ratio of surface to total protein expression.

 

I told him that he must be aware that although it is easy to calculate a ratio of the red intensity to the green intensity, the resulting value would be just proportional to the absolute ratio of surface expression and could only be used to compare the relative changes among the different treatments, but he would have a lot of variability due to uncontrolled variation in illumination sources, etc etc….

 

Of course trying to segment the green image in peripheral (near and at the surface) and total fluorescence would not be a solution either, due to the obvious limitations in the optical resolution.

 

I would appreciate any comments in that sense and if a relative ratio of red to green fluorescence would be in anyway acceptable to answer the reviewer.

 

Thanks in advance

 

Leoncio A. Vergara MD

Assistant Professor
Laboratory of Protein Misfolding Diseases (lab-PMD),
George and Cynthia Mitchell Center for Neurodegenerative Diseases Research.

Director of the Optical Imaging Lab. (OIL),

Dept. of Neuroscience and Cell Biology

University of Texas Medical Branch (UTMB)

301 University Blvd

Galveston, Texas 77555-0641

OIL phone: [phone] 

Lab-PMD phone: [phone]

fax: [phone]

 

 

[Gabriel Lapointe]

Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal Maybe not a answer to your question. But for that kind of quantification, I think a biochemical approach will be more reliable. You could for example:
A)
Label the surface protein with your antibody and a cross-linker (which of course doesn't pass the cell membrane. Then , lyse the cells and do a western blot. Proteins with the external epitope will have a higher apparent molecular weight.

B) Treat the cell with a suitable protease (one that doesn't cleave the eitope nor cut too many times). collect the supernatant and precipitate with TCA. Wash the cell extensively and lyse them. Do a western and compare the signal. The external protein will have been cut away (so smaller) from the cell and will be present in the precipitate of the media, the internal protein will be protected by the cell membrane and therefore only full size in the cell membrane. Only use this if the external part of the protein is big enough.


IMHO, both approach will be more reliable in estimating the amount of proteins than trying to quantify fluorescence.

Gabriel

Vergara, Leoncio A. wrote:

Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal Search the CONFOCAL archive at http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Dear Confocalists

 

One of the users of our facility has prepared cells in which he stained a surface protein by using an external epitope tag and a red labeled secondary, he also stained the same protein using an internal epitope and a green labeled secondary in order to asses the total level of expression per cell.

 

He wants to show the level of surface expression relative to the total cell expression of that protein. First he tried to approach this just by showing representative images of the different treatment conditions, of course the reviewers of his manuscript asked him for quantification in the form a ratio of surface to total protein expression.

 

I told him that he must be aware that although it is easy to calculate a ratio of the red intensity to the green intensity, the resulting value would be just proportional to the absolute ratio of surface expression and could only be used to compare the relative changes among the different treatments, but he would have a lot of variability due to uncontrolled variation in illumination sources, etc etc….

 

Of course trying to segment the green image in peripheral (near and at the surface) and total fluorescence would not be a solution either, due to the obvious limitations in the optical resolution.

 

I would appreciate any comments in that sense and if a relative ratio of red to green fluorescence would be in anyway acceptable to answer the reviewer.

 

Thanks in advance

 

Leoncio A. Vergara MD

Assistant Professor
Laboratory of Protein Misfolding Diseases (lab-PMD),
George and Cynthia Mitchell Center for Neurodegenerative Diseases Research.

Director of the Optical Imaging Lab. (OIL),

Dept. of Neuroscience and Cell Biology

University of Texas Medical Branch (UTMB)

301 University Blvd

Galveston, Texas 77555-0641

OIL phone: [phone] 

Lab-PMD phone: [phone]

fax: [phone]

 

 

[A. Hamdoun]

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Thanks ll for useful advice. Interestingly I found that when imaging the mCherry the
spectral detection channel gave a slightly brighter signal than the standard PMT and filters. I
attributed this to the wide spectrum of emitted wavelengths collected. But it could also be
that the filters are not optimal (TRITC, I think) or that there is something else like pinhole
alignment at work.
It does seem like the standard PMT should be much brighter than the PMT.

[regan m]

Search the CONFOCAL archive at
http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal

Well try to check for the alignment of fiber optical cable in the
scanner head.... Let an experienced person do it

You will realsie that the alignment of pin hole along with knowing the
lazer power output on to the sample and the efficency of the Fiber
optic cable.. all play a critical role

And if u want to capture teh PMT trans Mode images beter to put the
filter in DIC mode ...


On Sat, Aug 9, 2008 at 7:22 PM, Amro Hamdoun <[hidden email]> wrote:

> Search the CONFOCAL archive at
> http://listserv.acsu.buffalo.edu/cgi-bin/wa?S1=confocal
>
> Thanks ll for useful advice. Interestingly I found that when imaging the mCherry the
> spectral detection channel gave a slightly brighter signal than the standard PMT and filters. I
> attributed this to the wide spectrum of emitted wavelengths collected. But it could also be
> that the filters are not optimal (TRITC, I think) or that there is something else like pinhole
> alignment at work.
> It does seem like the standard PMT should be much brighter than the PMT.
>