Autofluorescence
KatherineHolzem
Posts: 21
For anyone who has gotten to imaging yet, have you noticed a lot of autofluorescence in the green and red wavelength ranges? I am working with heart tissue, which I know is notorious for autofluorescence, but sounds like this may be true for other tissues as well. To try to get around this, I am testing blue and far-red dyes, but I would like to know other thoughts or suggestions.
Also, according to the SeeDB article posted about by Bertrand_Vernay on the FocusClear discussion (http://forum.claritytechniques.org/discussion/1/focusclear#latest), this autofluorescence issue could be due to the Maillard reaction, which means it could potentially be prevented... however I don't think the chemicals used to do this would be practical to add to the ETC buffer during the clearing process (2-ME or 1-thiogylcerol). The Maillard reaction is reversible though.
Also, according to the SeeDB article posted about by Bertrand_Vernay on the FocusClear discussion (http://forum.claritytechniques.org/discussion/1/focusclear#latest), this autofluorescence issue could be due to the Maillard reaction, which means it could potentially be prevented... however I don't think the chemicals used to do this would be practical to add to the ETC buffer during the clearing process (2-ME or 1-thiogylcerol). The Maillard reaction is reversible though.
Comments
I'm making a carry over from the ResearchGate thread, but I wanted to point out that the browning of the tissue during electrophoresis could be due to the Maillard reaction, promoted by the high pH and temperature in the chamber.
I'm wondering if adding a reducing agent in the clearing solution could help with this? In the recent seeDB paper, they add 0.5% of 1-thioglycerol. This looks doable, even in the high volume of clearing solution we use here.
It seems that the paper suggests that the Maillard reaction that produces the browning of the tissue is due to the fact that they use a fructose solution, and fructose is a major reagent in the Maillard reaction.
It makes sense to me that the Maillard reaction may be contributing to the browning and autoflourescence in our ETC tissue, but I would imagine that we would see it to a lesser degree than those soaking their brains in fructose.
I should say that my preliminary images have had very high autoflourescence - which completely prevented me from seeing my native GFP.
I'm honestly wondering if I have a step wrong somewhere at this point, especially since the original paper shows fairly clean eYFP signal - considering where the autofluorescence is showing up (for me at least) it seems like their results would also have it.
My other thought, for the heart at least, was that this could be lipofuscin (collections of oxidized protein). I think this would be a harder problem to deal with than the Maillard.
https://docs.google.com/file/d/0B6RE82086JGFZ0JBUUphckJyX3c/edit?usp=sharing
Here are two pictures which show one of my brains before and after clearing - I didn't remove enough of the surrounding gel on this run, which caused the cleared tissue to be a bit sloppy, but we definitely see some good clearing. We also see yellowing, which is frustrating, but not too bad - I have not attempted index matching.
When I got this amount of clearing, I decided to run it for one more day to see what would happen.
https://docs.google.com/file/d/0B6RE82086JGFRmdGRklrMjFRcFE/edit?usp=sharing
This was the result.
Don't run your brains too long folks - they melt.
(edit) We also see some of the black tarnish that is building up on the cathode making its way to the sample and causing those black spots. My new chamber has an extra mesh in place to prevent this. I should have some pictures of the new results in a few days.
I have a strong expression of eGFP in my brain. I can see it already in the brain
under a microscope without any clearing or hydrogel, but after embedding I started
to see some autofluorescence.
Obviously 74 hours is far too long - melted brains are simply useless.
Fabio brings up an interesting point. Would the crosslinking step change the amount of autoflourescence we see?
Have you tried to check for autofluorescence after hydrogel embedding?
Im really interested in this.
Yesterday we compared a mouse heart that had been cleared and placed in focusclear to one that had been perfused with hydrogel monomer but not yet crosslinked - both seemed to exhibit significant auto-fluorescence in the green and red channels, but minimal auto-fluorescence in the blue, and basically none in the far-red (see @KatherineHolzem's comment above). I remember reading somewhere on this forum that someone had seen less auto-fluorescence before the crosslinking, but based on what we saw yesterday, this does not seem to be the case (at least not for mouse heart tissue).
We have cleared our heart tissue at 30-35 V and keeping the water bath at 32 degrees. Our tissues have taken about 2 weeks for clearing.
http://forum.claritytechniques.org/discussion/32/loss-of-gfp-signal#latest
But to be more succinct there is an excellent guide here:
http://www.uhnresearch.ca/facilities/wcif/PDF/Autofluorescence.pdf
http://www.sciencedirect.com/science/article/pii/S0165027013002896
Thanks @haisond, have you photo-irradiated the tissue with any endogenous fluorescent proteins?
Here is a quick summary of the posts (thanks everyone!), please correct or update if I'm wrong.
@joelrosiene has reported the same issue with Thy-1 GFP mouse.
@Fabio has reported that strong GFP expression was uncovered after clearing.
@kadipietro found the same autofluorescence issue.
@KatherineHolzem found minimal autofluorescence in far-red, but still a lot on the other channel where fluorescent protein usually emits.
@KeithSiew brought up techniques to reduce noise but I'm not sure it's directly related to the fluorescent protein issue.
Despite the beautiful images in the original paper, no one has ever succeeded to show good pics?? Any ideas?
Earlier you mentioned you also see some autofluorescence after embedding, but do you mean you still have an enough signal over the noise? Could you tell us what kind of mice with EGFP do you test if you don't mind?
I don't see YFP/GFP/CFP/RFP or TdTomato even under our confocal. One of their expression should be nuclear restricted, two of them should be membrane, others cytoplasmic. I believe that their signal is much weaker than H2B-GFP though. (I could see strong signal of all in thick sections without ETC).
1. Make sure you check in the Confocal scan mode to assess the level of auto-fluorescence in your samples. We would expect to see very blurred signal under the eyepiece of a Confocal or a wide-field microscope. In fact, that's a good sign implying that your sample is very god cleared, and the light from out of focus plane is contaminating your focal plane.
2. Optimize the staining protocol (eg, buffers used: PBS or Boric Acid etc, and wash very well). If using transgenic, use as high expressing transgenic line as available - good for signal-to-noise as well as for photo-bleaching. We have tested Thy1-eYFP, Thy1-GFP, Cre-driven reporter, as well as viral injected brains. All look good.
3. Make sure your sample looks normal cleared (not much yellowish, not burned or damaged). If it does not, go down in voltage applied.
4. DO NOT leave your sample in focus clear for longer than needed. The white precipitation from focus clear produce heavy auto-fluorescence and also occlude the imaging depth.
Hope these help.
I don't see any strong or blurred epi-fluorescence in my sample, and the loss of signal seems to happen during clearing.