This paper compared the crystal structures of 2 globular proteins in the absence and presence of sub-denaturing concentration of GuHCl. The main effect was a decrease in B-factor in the presence of GuHCl. This was interpreted as a loss of mobility caused by "non-covalent crosslinking" by GuHCl. Perhaps something similar is happening with your protein.

http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2143764/pdf/9260285.pdf

Hi， Garg：

your observation can be fully explained by protein conformation concept, protein thermodynamics, and partition fucntion of protein conformaitonal state. see my paper: Partition Function of the Protein Conformational State. 

briefly speaking, at low concentration of (GdnCl), the protein thermodynamic state (or internal motion) is changed and protein volume is increased. at high concentration of denaturant, the protein sreucture is destroyed. 

the protein may be sensitive to urea at high temperature. the effect of 3.26M  (GdnCl) equals to 3.25x6=19.5M urea for protein unfolding. 

hope this comment is helpful. 

I do not understand how you can say that the protein gains structure. To gain structure or become more compact the tryptophan(s) should be more buried because of the  increase of average hydrophobicity and therefore it increases (not decrease) its quantum yield and therefore its fluorescence intensity. The same applies to the ANS which is little fluorescent when it is in contact with the solvent. Something is not clear also because to far UV UV changes little.  What about the maxima of emission? Another thing not clear is if your protein is formed from a single chain with two structural domains, or is formed by several subunits, each with two structural domains. To denature a protein with more subunits or different domains, you must have right spectroscopic reporter groups properly located in the structure (i.e. of which you know the exact structural location), otherwise you will always see the sum of the various spectroscopic effects. Techniques to separate the different spectroscopic contributions often require the use of different reporter groups or techniques. Furthermore, if the protein is structurally complex, it's likely to find equilibrium intermediates.

What I see is that the dimer dissociates to monomers. You see it from your size-exclusion chromatography. No compaction (if it means that the protein becomes more compact and denser) occurs. This is evident from the fluorescence data. The fluorescence intensity of both ANS and Trp would increase if the protein became more compact. Furthermore, in such a case, you would observe not only a fluorescence intensity increase for both fluorophores but also the blue shift of the spectrum (to shorter wavelengths. Unfortunately, you said nothing about the peak position. So, I may suggest that dissociation of the dimer leads to that Trps in monomers become more accessible to water, which results in the fluorescence intensity decrease. No compaction occurs at all. So, it is important to check if there is any shift in fluorescence spectra. It is also important to know the number of Trps in the monomer and experimental conditions (in particular, the excitation wavelength).

I think comments above are quite intelligent. (I also do not understand the fluorescence trends you describe, by the way.)  Unlike them, though, I'll propose a dumb experiment:

Hyperthermophilic proteins tend to be rich in surface-exposed salt bridges.  Do you get similar observations when you simply increase the ionic strength of your solutions?  You don't say anything regarding whether you see the same trends in urea denaturation or not, but if the fluorescence trends are not parallel between the two, maybe the surface charges are doing something.  

I assume that you don't have access to a CD instrument, so maybe doing denaturations in a series of different solvents (dioxane+water mixtures could be another) would tell you something about the process, assuming you are interested in the said process, of course. :)

I fully agree with Ismet´s suggestion. GdmHCl at low concentration may simply act as a salt, so you could be seen its screening effect on (surface) electrostatics. Experiments with similar concentrations of different salts (three or four, with different ions found in the Hofmeister series) could be helpful to interpret you results at low GdmHCl.

Important, try to reproduce this effect using urea as well.

Good luck

I would try to do one more experiment. I would do the same measurements but with respect to the monomer alone (under conditions inhibiting association of monomers) to see what kind of differences, if any, occur in your system.

Fully agree with Alexander. If you have access to a method of determining the actual molar mass with SEC (like SEC-MALS) then this would prove that you have dimers or monomers or a mixture at certain states - I did like the fact that you knew SEC is a measure of hydrodynamic volume and not molar mass!

If you could combine this with DLS online it would show any conformation changes (or Rg from MALS if it is big enough) rather than rely on teh elution time as the elution is strongly affected by other factors (like column interactions)