Increasing brightness in eGFP channel with wide-pass filter set?

Hardware
1

I’m using an Olympus BX61 to image both eGFP and mCherry expressed in yeast cells - both FPs should be approx. equimolar.

The BX61 is currently fitted with -

  1. A narrow-band EGFP set with a short-pass excitation and emission filters to minimize autofluorescence (Ex ET480/20x, Em ET510/20m, BS T495lpxr)

  2. A mCherry (RFP) set (Exciter 562 nm > 93% over 40nm, Emitter 641 nm > 93% over 75nm, Dichroic 593nm)

Relative to mCherry the GFP channel is dim. I’m assuming this is due to the short-pass Ex and/or Em filters. Is this the reason? If so, would fitting a wide-pass Ex and Em eGFP filter set capture more of the GFP channel and increase the signal? Of the two filters which would have more of an effect on increasing the brightness or is increasing both Ex and Em important?

2

@ColinD

You did not tell us about your light source. If you are using a 488nm laser with a bandwidth of 5nm, then it won’t matter if you increase the bandwidth of the excitation filter. If that is the case, the only way to increase excitation will be to buy a laser with higher nominal power. If you are using a LED with a wider bandwidth, then maybe it can improve a bit on the excitation side. That said, your filter (red in the diagram below) is quite well placed on the excitation spectrum of EGFP (blue in the diagram below):

image
image322×602 15.5 KB

I’ve used FPbase Fluorescence Spectra Viewer to look at the plots and make those screenshots. To be honest, I wouldn’t change the excitation filter. If the nominal power of your laser/led is on the low end (<50mW), perhaps you should consider upgrading it.

On the other hand, an emission filter with a wider bandwidth might have a more significant effect. I’m assuming you are not acquiring the two channels simultaneously. However, there’s at least one thing to worry about. In the diagram below, I added the excitation spectrum of mCherry (in yellow):

image
image266×609 14.9 KB

When you use your EGFP excitation, there’s a bit of mCherry excitation as well. Therefore, if you choose a large bandwidth emission filter, or worst, a long-pass emission filter. You might start seeing mCherry in your EGFP channel. In the screenshot below, red is the emission filter ET510/20m and blue is the emission spectrum of mCherry:
image
image848×638 19.6 KB

If you absolutely want to avoid crosstalk perhaps you should limit yourself to 500-550nm bandwidth and you should get an increased signal. If the two channels are to be acquired simultaneously, then it is even worst, because you need to stop before the excitation filter of mCherry. In the diagram below, yellow is ET510/20m (sorry that it changed color…), red is Exciter 562 over 40nm, and blue is still the emission spectrum of mCherry:
image
image713×625 24.1 KB

In this case, I would choose something in the realm of 500-530nm. I must admit, adding those 10nm of bandwidth might not make a huge difference in increasing the signal.

Also, I am assuming both labels are equally bright. Which is rather a strong assumption but I know more about Optics and less about fluorophores.

I hope this makes sense. It is difficult to be more quantitative without knowing the light sources that you use and how powerful they are.

Take care,

Omni

3

Many thanks Omni,
Light source is Lumen Prior 200 metal halide. I’ve just ordered a 528/38 BP eGFP emission filter and will compare this to the 510/20 I have currently. The mCherry Ex filter is 562/40 so there will be a few nm of overlap with this 528/38 BP emission filter. The FPs are expressed as a single transcript but translated as discrete proteins due to the presence of a ribosomal skipping 2A peptide sequence between the two coding sequences.
Colin

4

The Lumen 200 spec sheet…

https://www.prior.com/Content/Downloads/Lumen_200_EN_Datasheet.pdf

…shows it has a sizable peak available for the 562/40 exc filter,

but no such peak for exciting GFP near its maximum.

This is an issue for metal halide lamps.

The Excelitas X-Cite light source has a similar spectral output. See spectra below from fpbase (the s528/38m from Chroma may not be the same filter that you just purchased)

You should get a huge boost using a really wideband exc filter like the 460/60 from Semrock.

You have the “xr” (extended reflection) version of the dichroic, so you’re all set there.

But then you are inviting more autofluorescence too…perhaps not an issue?

When choosing the excitation and emission filters, put them on a log scale (‘OD’ on Chroma’s or Semrock’s website) and try to get them as close as possible, but the product where they cross should be <10 e-10. And there should be no wiggles above 10 E-6 for either filter spectra in the 50-100nm on either side of the crossing point, where they are supposed to be blocking. The 460/60 filter above, you’ll notice, does not have such a clean cut-off, so you would need to choose the emission filter carefully to avoid reflection background.

It doesn’t matter if the GFP em filt spectra overlaps some of the mCherry exc filter spectra; but it’s important is that you keep that emission filter away from the mCherry emission.

image001.png924×336 47.6 KB

Cheers,

Jeff

5

@jmreece

It doesn’t matter if the GFP em filt spectra overlaps some of the mCherry exc filter spectra;

If the two channels are acquired simultaneously, I believe it can have an effect. Even in epi configuration, some of the red excitation light will scatter back into the detection and risk passing through the EGFP emission filter. The dichroic can filter some of it of course, but not to the same extent as the emission filter. Hence my concern.

Take care,

Omni

6

Sorry not totally clear what you mean by “acquired simultaneously” (not an expert in fluorescence microscopy although learning fast). At the moment I have separate filter cubes for the GFP and mCherry (RFP) channels and image each channel sequentially not at the same time.

7

@ColinD

Maybe I shouldn’t have mentioned it because it seems to create a lot of confusion. There are systems where the two channels are acquired exactly at the same time. If you take an image of EGFP and then the EGFP excitation turns off (and the filter turret turns), and finally you acquire an image of mCherry. Then, you can ignore what I say about simultaneously imaging both channels.

Is that clear?

Take care,

Omni