[Protein-analysis] Re: how to make protein unfold curve

Frank Küster frank at kuesterei.ch
Mon Aug 29 08:50:38 EST 2005


"subberry" <zhangpy at gmail.com> wrote:

> Thank you, Frank
>
> What do you mean with "are not ordered"?
>
> i mean the spectrum of buffer containing guanidine is not smooth, just
> like something with sllipicity, so is guanidine interferes with cd
> measurment?and just subtracting the rugged background result the real
> spectrum?

In which wavelength range did you measure the spectrum, was the
pathlength, and did you record the high voltage values of the
spectrometer?  Guanidine has quite a high absorbance in the far-UV
region; if hardly any light gets through due to the guanidine
absorbance, you can't measure the ellipticity (i.e. the difference in
absorption between the two circular polarization directions). 

Therefore, with guanidine in the solution, it only makes sense to look
at wavelengths where there is still enough light getting through.  With
the instruments I've used so far (different Aviv and Jasco
spectrometers), the borderline was around 600V for the high voltage of
the detector.  With 6M guanidinium, you'll need pathlengths below 0.1mm
to get below 200 or even 210 nm, approximately.

> "take the values at these wavelengths lambda_max from all the
> spectra you have measured"
>
> If there are several wavelenghs lambda_max, may i just take one of them
> to get value? 

Assuming that it's not just noise because you measured too low, yes
there can be multiple maxima/minima.  Often the spectra intersect
somewhere, and then you'll get at least one maximum and one minimum; but
there may be more intersections.  

On the one hand it's best to take the wavelength with the highest
(absolute) difference between native and unfolded.  But if two
wavelengths give a well measurable difference, it doesn't hurt to do the
analysis at both: You'll get a first indication whether it is two-state,
or not, or whether some artifacts obscure the real transition.

> And is the spectrum the difference spectrum or the
> original spectrum?

One usually takes just the values from the original spectra at each
GdnHCl concentration.  Mathematically it makes no difference, though.
Absolute values are just easier to compare with transitions you've
already seen from different proteins.

> "measure the spectrum to make sure
> that it is actually denatured"
>
> What kind of spectrum to make sure it is denatured?

(Note that this is only necessary if you've got an oligomeric protein.
For a monomeric protein, there's no reason to assume that it unfolds
slower, or not at all, if the protein concentration is increased):

Any spectrum that will show you the difference between native and
unfolded will do.  Assuming that you measured far-UV-CD, you can just
take the denatured, concentrated sample and measure the spectrum.  In
case the absorbance is too high, try a lower pathlength.

If that isn't feasible, an alternative approach would be to dilute the
native protein into 6M GdnHCl, put it into the CD-spectrometer as fast
as possible and measure the denaturation kinetics at one wavelength were
you know the signal changes.  If that is fast (or so fast that you only
see the resulting constant, unfolded signal), try with higher protein
concentrations (3 or 4 steps up to the highest concentration you can
still measure).  If you notice that the reaction gets slower, try to
extrapolate this to the concentration you will use to produce the
denatured stock solution, and decide how long you'll have to denature
it.  Mind that the extrapolation need not be linear...  

> And can i use light scattering to make sure above which guanidine
> concentration the protein doesnot aggregate?

Yes, light scattering is a good method for that.  However, if you
observe aggregation, most probably you'll have a hard time in finding
conditions where the unfolding process is completely reversible.
Aggregation usually means that there is an intermediate I which is
formed with or even without guanidine, and the intermediate molecules
then lump together to form the aggregates.  This means your minimal
reaction scheme is

F <=> I <=> U
      |
      |
      v
  aggregates

getting F (folded) and U (unfolded) in equilbrium, while keeping the I
--> aggregates reaction slow enough that it doesn't matter is really
hard, and in many cases it simply doesn't work.  But if F and U are not
yet in equilibrium, you cannot measure the thermodynamic stability of F

Regards, Frank
-- 
In der Zeit, in der man einen defekten Riegel derartig seziert hat, dass man
verlässlich bestimmte Speicherbereich als "heile" garantieren(!) kann, kann 
man auch Pfandflaschen im Stadtpark sammeln, dass reicht dann für einen 
doppelt so großen neuen Riegel. ;)       [Jörg Rossdeutscher in d-u-g at l.d.o]


More information about the Proteins mailing list