Dear all,
When I perform FLIM-TCSPC measurements of a certain probe in solution or in a porous medium (polysaccharide substrate) I can fit the TCSPC data with only one term that corresponds to the expected lifetime of this probe (in my case, Atto-655 conjugated to a protein, 1.8 ns). When the measurements are performed in the same experimental conditions with a solid/dense substrate, the TCSPC data must be fitted with at least two terms. In most cases, one term is a short lifetime varying from 0.35 to 0.75 ns. Since in a tutorial from PicoQuant a lifetime of approximately 0.75 ns can be observed, besides the lifetimes of the employed probes, should I assume that these short lifetimes are some sort of artifacts? Light-scattering?
I have also observed that in some cases I have to employ a third term with a longer lifetime (> 3 ns) in order to fit the data. Should I consider this behavior some kind of artifact (such as light-scattering) or a real alteration of the probe lifetime (quenching)?
Here are some information about my detection configuration and data analysis procedure:
"Several FLIM measurements were performed for the multiple samples available, immediately after adding the labeled enzyme. The focus height was gradually varied, so that images from multiple layers could be acquired up to some tens of micrometers, depending on each object analyzed."
"The measurements were performed in a time-resolved fluorescence confocal microcope with single-molecule sensitivity MicroTime 200 (PicoQuant, Berlin, Germany) using the FLIM/TCSPC mode. Most of the acquisitions were images of 80 µm x 80 µm, 256 pixels x 256 pixels (density of 3.2 pixel/µm), with a scan speed of 0.60 ms/pixel in a bidirectional pattern."
"Glass coverslips containing immobilized carbohydrates and the arabinoxylan film were mounted upon an UPlanSApo 60x NA = 1.2 water immersion objective (OLYMPUS, Shinjuku, Tokyo, Japan). The objective was centered on a solid fragment of carbohydrate. Thirty microliters of a 1 µM solution of Enzyme-AttoOxa11 were applied upon the coverslip. By using the wide field illumination of the Olympus IX71 inverted microscope (OLYMPUS, Shinjuku, Tokyo, Japan) and observing the image through the binocular, the objective was centered in a corner between the object and the solution. The sample was excited with a 640 nm pulsed laser beam, which was generated by a LDH640 diode laser head (PicoQuant, Berlin, Germany). The laser focus height in the Z-axis was adjusted to the glass-sample interface based on the backscattered excitation light observed in a camera. The dichroic beam splitter z638rdc (Chroma, Bellows Falls, VT, USA) and the emission filter HQ 690/70 (Chroma, Bellows Falls, VT, USA) were employed to bar the exciting photons from the emission line, so that just emission photons would reach the detector. The optical power of the exciting laser was of 40–200 nW after it passed through the dichroic. The laser pulse repetition rate was set to 25 MHz (one pulse at each 40 ns). A 50 µm pinhole was employed."
"The FLIM images were treated and corrected by using the tools in the software SymPhoTime 32. Firstly, the mean lifetime of a whole raw image was adjusted by one, two or three exponentials, according to each sample. Then, the resulting lifetimes from each exponential were fixed for all pixels in the image. Finally, the FLIM image was recalculated, so that the amplitudes for each lifetime in each individual pixel were obtained. The images were colored according to the mean fluorescence lifetime in each pixel, and the colors intensity were determined based on the intensity image."
Thank you very much for your time.
Best regards,
Gustavo
Edit 1: Added procedure details;
Edit 2: Corrected the emission filter name.
I agree with the previous answer, 0.35-0.75 ns component might be a result of your fluorophore slowly rotating, which makes sense given that your substrate is 'solid and dense' as you say. In order to check if this is not scaterring you can also try collecting a decay with a high number of counts (10 000 at the peak, for instance) and see if you get a better fit when you restrict one lifetime to 0 instead of 0.35-0.75 ns. If your component is due to scattering, the fit should get better as scaterring is essentially a fluorescence decay with 0 lifetime. You also need to measure your instrument response function (IRF) in order to do this check. Some softwares calculate it from the decays, which would not be ideal. Anyway, playing with filters is probably a better way, as the previous answer suggests.
3 ns component should not be a result of scaterring unless you have a big problem with it i.e. your substrate is not fully transparent. Try measuring a fluorescence decay in your substrate without the dye. A flat background is what you want to see. If it is not the case, the substrate might have some fluorescent impurities.
Also, check if the lifetime Atto-655 is not known to be sensitive to polarity, viscosity or any other parameters of its environment.
Thank you very much for your answers.
So, if these short components are more likely to be due to the slow rotation of my conjugate, could I assume that it is a consequence of the protein-substrate interaction, or of the crowded microenvironment of the substrate?
I have found an article (follows attached) in which the Atto-655 lifelime increases about 75% from water to ethanol (which should result in a lifetime of 3.15 ns). Therefore, it seems possible that the longer lifetime I experienced is due to polarity quenching.
Here are some information about my detection configuration and data analysis procedure:
"Several FLIM measurements were performed for the multiple samples available, immediately after adding the labeled enzyme. The focus height was gradually varied, so that images from multiple layers could be acquired up to some tens of micrometers, depending on each object analyzed."
"The measurements were performed in a time-resolved fluorescence confocal microcope with single-molecule sensitivity MicroTime 200 (PicoQuant, Berlin, Germany) using the FLIM/TCSPC mode. Most of the acquisitions were images of 80 µm x 80 µm, 256 pixels x 256 pixels (density of 3.2 pixel/µm), with a scan speed of 0.60 ms/pixel in a bidirectional pattern."
"Glass coverslips containing immobilized carbohydrates and the arabinoxylan film were mounted upon an UPlanSApo 60x NA = 1.2 water immersion objective (OLYMPUS, Shinjuku, Tokyo, Japan). The objective was centered on a solid fragment of carbohydrate. Thirty microliters of a 1 µM solution of Enzyme-AttoOxa11 were applied upon the coverslip. By using the wide field illumination of the Olympus IX71 inverted microscope (OLYMPUS, Shinjuku, Tokyo, Japan) and observing the image through the binocular, the objective was centered in a corner between the object and the solution. The sample was excited with a 640 nm pulsed laser beam, which was generated by a LDH640 diode laser head (PicoQuant, Berlin, Germany). The laser focus height in the Z-axis was adjusted to the glass-sample interface based on the backscattered excitation light observed in a camera. The dichroic beam splitter z638rdc (Chroma, Bellows Falls, VT, USA) and the emission filter z636/10x (Chroma, Bellows Falls, VT, USA) were employed to bar the exciting photons from the emission line, so that just emission photons would reach the detector. The optical power of the exciting laser was of 40–200 nW after it passed through the dichroic. The laser pulse repetition rate was set to 25 MHz (one pulse at each 40 ns). A 50 µm pinhole was employed."
"The FLIM images were treated and corrected by using the tools in the software SymPhoTime 32. Firstly, the mean lifetime of a whole raw image was adjusted by one, two or three exponentials, according to each sample. Then, the resulting lifetimes from each exponential were fixed for all pixels in the image. Finally, the FLIM image was recalculated, so that the amplitudes for each lifetime in each individual pixel were obtained. The images were colored according to the mean fluorescence lifetime in each pixel, and the colors intensity were determined based on the intensity image."
Thank you very much for your time.
Best regards,
Gustavo
http://nathan.instras.com/MyDocsDB/doc-653.pdf
Dear Peter,
Concerning the filter, I will have to verify it with more experienced colleagues who work in the lab where I performed these acquisitions and analysis. The filter I used was in the detection filter wheel in the main optical unit and has never been changed as far as I know. It has always been used by other colleagues for data acquisition with the same laser beam, so I believe (and hope) that I used the proper filter. I will let you know as soon as I get the answer.
About the IRF, sorry for my limited knowledge. I do not remember of having obtained it.
About the fitting, do you mean that maybe I should use another algorithm, like that available in SymPhoTime 64?
Best regards,
Gustavo
Dear Peter,
I believe I understand what you mean. So, an easier but not proper way to verify if these short components are due to IRF or not is to include the IRF fitting during the analysis, getting the IRF from the own decays? The most correct way would be to get the IRF pattern by acquiring the data using buffer, without the emission filter, and then load this pattern during the analysis procedure, as you explained to me about one year ago in the link below? Should I get the IRF from a single point or from an area?
Thank you very much for your time.
Best regards,
Gustavo
https://forum.picoquant.com/viewtopic.php?f=19&t=1000&p=1857&hilit=gustavo#p1857
Hi there,
have you resolved the emission filter issue? What is the emission filter then?
If you focus to the glass-sample interface, chances are that reflection may contribute to the measurement. Have you tried a control sample without the fluorophore (all else being equal)? This would not only tell you the intensity of scattered light, but also any possible sources of unwanted fluorescence (although not very likely in far red spectral region).
Experimental IRF is important, it is often quite different from what the software can guess (at least in the case of 2-photon femtosecond excitation and becker & hickl TSCPS equipment).
Short lifetime may also be due to fluorescence quenching, e.g. FRET. It can also happen if the dye concentration is very high (e.g. due to clustering), this is poorly understood, as homo-FRET alone cannot explain lifetime shortening...
Long lifetime may also result from improper fitting of ambient light level, but that is more of an issue with higher rep lasers (80 MHz), where the decay is usually incomplete.
Just some thoughts.
Good luck, zdenek
- Badges
- Science method