Circular Dichroism Help!!

J Cort jcort at u.washington.edu
Tue Nov 8 16:29:44 EST 1994


In article <3971gp$cp1 at server.st.usm.edu>,
Mark Jason Logan <mjlogan at whale.st.usm.edu> wrote:
>
>I have a few questions about CD.
>
>1) When is an increase in molar ellipticity significant?  Is a change of
>3-5% significant?

This change in ellipticity could be significant, but you should rule out 
simple explanations for it.  Maybe the protein or peptide is 
thermally frayable and the temperature changed.  Maybe the spectra were 
recorded on different days and the concentration changed due to adhesion 
to the container, or from evaporation of water, or from drift in the 
sensitivity of the instrument.  A change of 3-5% certainly doesn't 
indicate a massive structural reorganization, but it could be telling you 
something.

>2) Also what would happen to the molar ellipticity if a helical bundle
>composed of four monomeric helices was transformed to bundles containing
>two monomeric helices?

As long as the helicity of each helix remains constant, the molar 
ellipticity shouldn't change, unless the contacts between helicies in the 
bundle themselves (assuming they have no effect on helicity) are influencing 
the rotatory characteristics of the individual residues and the 
helicies.  I would expect to see a decrease in the molar ellipticity upon 
dissociation of the four helix bundle if the bundle is actually 
stabilizing the helicies.

>3) Can any one give me their opinion of the several different methods
>available to calculate the total secondary structural content via
>CD(Greenfield & Fasman, Chen, Russe, etc).  I am using several of these
>methods and routinely get extremely high values for the total secondary
>structure content (as well as some negative values for certain
>structures). I realize that these methods are designed to tackle proteins
>containing a wide range of secondary structures, but how well do they work
>on proteins that are say 80% helical?

Results from these programs are often reported in the literature,
occasionally (references available on request) without any of the original
spectra from which the amounts of different secondary structures were
calculated.  This disturbs me quite a bit, not only because one cannot
reconstruct a CD spectrum from the output of these programs, but also
because the output is sometimes quite clearly wrong.  I know of one 
example in which a peptide which quite clearly consisted of mostly random 
coil was reported to be 50% beta sheet simply because that's what the 
program said. (I can provide the citation for this if anyone wants to 
look it up).  CD spectral deconvolution is an important goal, and I have 
no doubt that the deconvolution methods listed above can be useful, but 
one must be very careful when interpreting and believing the results.  
Most importantly, raw data or spectra should always accompany the output 
of a deconvolution procedure which is reported in the literature.

John Cort
john at peptide.chem.washington.edu



More information about the Proteins mailing list