I am looking at a protein of unknown function that is localised in the thylakoid membrane of the chloroplast with a punctate expression pattern. Previous CO-IP and Yeast-2-hybrid analysis has determined three putative interactive proteins. I am trying to confirm their interaction with my protein of interest using FRET techniques.
As currently I do not have stable transformant lines and I am using tobacco for transient expression I have been advised to carry out FRET-FLIM instead of just FRET. However, the results from my FRET-FLIM experiment generated very noisy data which is potentially due to the punctate expression of my protein (one very distinct point per chloroplast). I was hoping someone might have some suggestions regarding this issue. Alternatively, would it be possible to carry out FRET on transient expression tobacco instead of stable transformant lines?
Any help is appreciated!:)
Hi Mercedesz,
Very few details are provided in the problem description so its hard to tell what caused the noisy data. If the punctate distribution is physiological, it shouldn't by definition lead to noisy data in FRET FLIM.
In my opinion, you could perform normal, intensity-based FRET measurements even in transiently expressing cells, but you need to determine overspill correction parameters. A couple of questions that would help me help you:
1. What exactly do you mean by noisy data?
2. What are your fluorescence labels, i.e. what is your donor and acceptor?
3. How did you do the analysis?
Best,
Peter
Hi Peter Nagy ,
Thank you for the answer! Sorry for the few details I am still familiarising myself with the technique.
The punctate distribution is physiological and located in the chloroplast.
And the answers to your question:
1. By noisy data I mean the fluorescence decay curve is not as uniform as expected. It has a sawtooth pattern (I have attached it below). I have been told that this might influence the quality of my data and can be due to the punctate expression of my gene of interest.
2. The fluorescence labels are GFP as a donor and RFP as an acceptor.
3. To perform the analysis I have used the SymPhoTime software, which measured the amplitude averaged fluorescence lifetimes (τamp) and lifetime decays were fit to multi-exponential models. Chi square was also considered.
Thank you for your time!
Best wishes,
Mercedesz
Hi Mercedesz,
I don't think your curves are extremely noisy. It is perfectly natural for such a curve to have higher noise at longer times. Your fitted lifetime is 1.7 ns, at around 10 ns you have hardly any excited and emitting GFP left. This can be read from your plot as well, which has a log Y axis, but it is even more obvious if the fluorescence decay is plotted on a linear scale (see attached, the graph was generated assuming a lifetime of 1.7 ns). You can see that practically no fluorescence remains by 10 ns, and the noise, i.e. SD/mean, is inversely proportional to the square root of the mean due to Poissonian statistics. Assuming an initial count of 100 I generated such a simulated plot, which exhibits increased noise at long acquisition times.
If you look at an application note from PicoQuant (https://www.picoquant.com/images/uploads/page/files/7350/appnote_flim_overview.pdf), you can also see this feature.
Does the lifetime of 1.7 ns correspond to GFP, or GFP in the presence of the acceptor?
Best regards,
Peter
- Badges
- Science topic