Hello,
I am working with a small, single subunit protein with relatively slow refolding rate (~20 min). I have noticed that after thermal unfolding protein refolds without any precipitation. I have set up an experiment using Circular dichroism spectrometer to measure the refolding rate, but I was also interested how the CD spectrum (240 - 190 nm) changes over time. During 1 h I have measured ~30 CD spectra which showed very nice signal change upon folding of the protein. I have ploted a signal at different wavelengths versus time and observed that for different wavelengths not the same refolding rate constants were calculated.
I know that signal change on CD depends on the secondary structure of the protein. The question is why the refolding rates differs even by 2.5-fold at different wavelengths? Is it possible that one part of the protein is folding faster - which is for example detected at one wavelength and other part is folding slower? Rate constant got from different technique, which shows only if the protein is folded or unfolded, is very similar to one of the rate constants I've got from CD measurement. Could you recommend me some books or publications which could help me explain this phenomenon?
Best,
Dawid
Can you please supply the series of CD spectra so that we can inspect them and maybe suggest a solution? What type of protein are you working with? Does it have a pdb entry?
Which wavelengths between 240 and 190 are you using and which one results in the 2.5x faster/slower refold rate? It's difficult to say without knowing the spectra.
Protein is mostly beta-sheet protein, structure is not known yet.
Difference of the t1/2 of refolding was detected between, for example, 200 nm and 230 nm. Difference between 200 nm and 195 is 1.5 x (values are not in the error range).
Could you please post the raw data, as a plain-text "Comma Separated Value" (CSV) file? It might be interesting to take a crack at this with DynaFit.
http://www.biokin.com/dynafit/
Looks like a coil-to-betasheet transition. Looks like your protein is all beta,is it?
Regarding the speed of refolding, what is the buffer used, what's the pH, is it the right pH for your protein? Does the buffer include any viscogen, osmolyte or crowders and if so, what type and at what concentration/vol. percentage?
How do you invoke refolding? Is this by solvent exchange? temperature change?
Another thing that may occur is amyloid-like aggregation and the beta sheet character + the slowness of "refolding" may be a hallmark of it. In amyloid-like aggregation, proteins fibrilize into inter-protein cross-beta sheet structure.
Have you performed some sort of a molecular-size-based kinetics like gels, size exclusion chromatography, dynamic light scattering (DLS) or even fluorescence correlation spectroscopy (FCS)?
In this case, DLS or FCS would be the most sensitive to follow such possible fibrilization kinetics.
Does the kinetics of refolding depend on the protein concentraion (size-based kinetics in a titration of protein concentration).
How pure is your protein and by which criteria? What temperatures have you used and what protein concentrations? You indicated that there is no precipitate at the end of the renaturation but is it possible that there's no hysteresis and there's soluble oligomers
Dear Dawid,
thank you for sending the data in a private message. I have now analyzed the data set and it seems that (as the cliche goes...) "the mystery is solved". As I see it, there is only on rate constant which applies uniformly to the entire global data set you sent. Please see the attached image. This is a EPS file that was just generated by the DynaFit software. I am also attaching the "script" file that was used to preform this analysis. To retrace my steps, please do this:
http://www.biokin.com/dynafit/
It is available free of charge to all academic researchers.
http://www.biokin.com/dynafit/license/academic.html
./users/EDU/CH/ETH/Zyla_D/160315 and
./users/EDU/CH/ETH/Zyla_D/160315/data
As you can see, the residual plot (extremely important) is distributed perfectly randomly, which indicates very good global fit (see ref. [1] below) to the combined CD data. Importantly, the fitting model I used is a system of differential equations (auto-generated by the software) which corresponds to the simplest possible folding model. Here is an excerpt from the input file:
[mechanism]
A ---> B : k
[constants]
k = 0.001 ?
Most importantly, this model (or rather the mathematical representation that was auto-generated by the software) applies to all time-vs-CD traces included in your data set. The best-fit value of the rate constant is k = (0.0012 ± 0.0003) s-1. Caveat: These time points are not statistically independent so the formal standard error should be ignored.
Hope this helps... Please do not hesitate to contact me directly if you have any questions.
REFERENCES
1. Beechem, J. M., Global analysis of biochemical and biophysical data, Meth. Enzymol. 210 (1992) 37-54.
P.S.: I would be extremely interested to know which particular data-analysis method you utilized before, to arrive at the (misleading, as we can see now) conclusion that there should be "different rate constants" at different wavelengths. Could you please contact me directly through http://www.biokin.com/contact.html ? Thanks...
- Badges
- Science topic
- Similar topics
- Biological Science
- Molecular Biology
- Biomolecules
- Proteins