Hi!

I don't know how to reduce the autofluorescence, except maybe by culturing the cells in as happy conditions as possible.

However, you can optimize your imaging at least slightly by using FluoroBrite DMEM or similar non-phenol red containing medium (will reduce background) and Prolong Live antifade reagents to make your fluorescent proteins withstand bleaching better.

And of course, for fixed cells, you can amplify the signal very well by using anti-GFP staining.

However, if despite these efforts your endogenously tagged protein signal is very weak, I'd suggest using a split-GFP strategy, as shown here Article Versatile protein tagging in cells with split fluorescent protein

and recently used by us for examining endogenous seipin (low expressed ER protein) in live cell imaing in this paper:

https://www.ncbi.nlm.nih.gov/pubmed/31178403

A futher strategy could be to use SNAP-tagging together with far red dyes instead, as those autofluorescent structures might be less problematic in the 640 (far red) emission range. However, there you have to do labeling of the cells prior to experiments, and the label itself may also induce some background (ie excess label will show signal that is not your protein of interest).

kindly,

Veijo

Hey Veijo,

thanks for your answer. Already using Fluorobrite-Medium (highly recommendable), and SNAP/Halo tagging is something I try at the moment. I like the approach of the tandam GFP tagging with multiple copies, would really love to try this.

Which vectors did you use as template for this, or did you build your own one (I am still awaiting your cell press article from our livrary, so sorry if I ask for something you described in your manuscript).

Best

Christian

Hi!

It's described in detail in the paper, we used codon-optimised version of Addgene 70219 for the GFP1-10 part, we've deposited it to Addgene (so should be there soon), and cut the GFP11x7 from here https://www.addgene.org/70224/ for the HDR template. It worked well for us, but of course some proteins might be sensitive to such a big tag. Good luck!

Usually autofluorescence is broad enough in wavelength that you can use an "empty" wavelength to subtract. Thus you might image the specific signal in the green, but leave the RFP channel open (no dye used) to visualize the autofluorescence and take an image there. Then, using imaging software, you will "subtract" the red image from the green image, thus subtracting out the autofluorescence. It's a little tricky, though, since you have to make a judgement about how much to subtract without over-subtracting. Also, it's important to have a no-dye control (in any wavelength) to determine the extent of the autofluorescence and to make sure that what you are seeing isn't due to bleedthrough from the green channel.

Another way you could do it is to use a spectral deconvolution imaging system, such as a Zeiss 710 Meta confocal system, where you use a no-dye control to take a spectral reading of the autofluorescence, then you can digitally subtract the intensity of the autofluorescence from the specific signal as long as your specific dye has a peak emission that is at least 10 nm different from your specific dye.. But not many labs have access to this sort of instrument.

And, as mentioned by Christian, you will want to use a no-autofluorescent imaging buffer, like Fluorobrite (which is a DMEM alternative) or Live Cell Imaging Solution (which is a HEPES derivative), both from Thermo Fisher.

Hey Jason,

thanks fopr your response. Unfortunately, many of these 'spots' are votile, probably vesicles, which dislocate within seconds. So substraction will not work in this case.

Btw, using Halo-tag with far red dyes like SiR works quite promising, as autofluorescence is rather low in this spectral area.