Quantitative assessments of the distinct contributions of polypeptide backbone amides versus side chain groups to chain expansion via chemical denaturation
J. Am. Chem. Soc. 137, 2984 - 2995
Holehouse, A.S., Garai, K., Lyle, N., Vitalis, A., and Pappu, R.V.
In the last sixty years or so, there has been extensive interest in the "protein folding" problem. Protein folding is highly relevant for a range of diseases, and to better understand how it occurs, we (as a field) need more detailed insight into the specific mechanisms by which folding and unfolding occur. In this paper, we used a combination of computer simulations, pen-and-paper theory, and experiments to tease apart some of the processes involved in protein unfolding and developed new ideas about how this may occur. Of course, before we can begin to ask about protein folding, we need to step back and ask; what are proteins?!?
Proteins don't just make you big and strong
Proteins are basically tiny molecular machines, which are themselves made up of a well defined and specific sequence of amino acids. There are twenty different amino acids, and so each protein is some very specific combination of those amino acids stuck together in a linear chain (see below)
In the image above, each circle is an amino acid, and you can see here our 'protein' is made of 14 amino acids, although just six unique ones. Importantly, the order of the amino acids in the chain matter.
We have tens of thousands of individual proteins in our body, and each protein has one or more functions: they help digest food, they help make bones, they make up the lens in your eye. Basically everything we do is, at some level, directly or indirectly carried out by proteins (often with the help of other molecules).
Proteins can fold
Protein folding is the process through which a specific sequence of amino acids in a protein causes that protein to fold up into some well defined and stable 3D structure (conformation). Not all proteins undergo folding but a large fraction of them do. As an analogy, imagine a lego set which spontaneously assembles itself into the fully formed model just because the pieces were close to one another. In a very real sense, this is what protein folding is, except the "pieces" (amino acids) are actually attached to one another in a linear chain. This chain connectivity provides a lot of constraints regarding how the protein can fold. The folding process also demonstrates why the order of the amino acids is so important - by having them as a linear sequence the specific order means certain amino acids are close to one another and some are far away. If you disrupt that order you may totally destroy the protein's ability to fold.
In the schematic above, our protein undergoes folding into a nice regular rectangle. The amazing thing about protein folding is that it's spontaneous and rapid. This means if you can unfold your protein in some way you can refold it and watch it refold, and this can be done many times to try and decipher how it's refolding. A huge amount of effort has gone into this approach and using it to define folding pathways, specific routes that a protein may take to get from an unfolded state to a folded state.
Going back to our lego analogy, the instruction book which comes with lego gives you a specific "pathway" (i.e. a step-by-step set of instructions) to go from the individual bricks to the assembled model. However, it is possible (though, from a lego perspective not recommended) to construct the individual bits of your lego model in a different order. In both cases the end result is the same (your lego model) but you have taken different routes to get to that final result. This is analogous to protein folding - it may be possible to take a variety of pathways to go from an unfolded protein to a folded one. That said, there may be specific points where all the pathways have to perform the same series of steps - i.e. there may be places where multiple different pathways converge (true for proteins and for lego!).
To watch proteins fold, we have to unfold them first
As mentioned, the typical way you study protein folding is to unfold the protein and then watch it refold (where there are various tools and approaches you can use to 'watch'). There are a few ways you can do this unfolding, but the most common approach for protein folding studies is to use chemical denaturants. Chemical denaturants are basically just really nasty chemicals which cause proteins to unfold. The two most common denaturants used are urea and guanidinium chloride (GdmCl), both of which have been used for over seventy years to study protein biochemistry and biophysics.
Despite their extensive use in biochemistry and biophysics, the explicit molecular mechanism of how urea and GdmCl can convert a folded protein into its unfolded form has remained a mystery. In this paper, we address this question and provide a solution which helps unify several apparently opposing viewpoints. Specifically, we believe the way these denaturants interact with proteins and peptides is synergistic with how water interacts with them, meaning their mode of action is extremely related to the protein's intrinsic solution behavior.
Sidechains and backbone
The specific insight from this paper relates to the interplay between solvent and solute and the impact of that interplay on peptide and protein conformation.
Broadly speaking, there have been several proposed mechanisms for denaturant action on polypeptides: 1) denaturant-backbone interaction, 2) denaturant-sidechain interaction, or 3) denaturant-sidechain and denaturant-backbone interaction (I'm not including the solvent-disruption hypothesis - a fairly old hypothesis which has failed to garner any support in recent years). Proponents of all three hypotheses have published compelling work in the last ten years, especially for option three, which argues that denaturants can interact with both the sidechains and the backbone.
In this paper, we published the outcome of close to four years of work (not all by me, I should add - a large chunk of this work was done by Nick Lyle, and although our interpretation and marriage with the experimental data came after Nick had left, he really trouble shot these simulations and got them running in an impressively rigorous manner). The order I'm going to describe these results is not the order in which they were obtained, but I think this presents a much more coherent train of thought.
We found that polyglycine - which we used as a proxy for the protein backbone - forms a highly collapsed, disordered globule in water, and that this collapsed globule only undergoes very minimal expansion in 8 M Urea and 7 M GdmCl. We're using chain expansion here as a proxy for unfolding. The naive interpretation from this could be that given polyglycine lacks any sidechains and we fail to see any impact of denaturants, denaturation must occur exclusively through interaction with the sidechains. However, simulation results (which involved extensive MD simulations of polyglycine and two sidechain-containing peptides) suggested something quite different. Instead, we showed that the sidechain-water interaction primes the backbone and the sidechains for interaction with the denaturant, and sets an initial degree of expansion from which denaturants can further drive chain expansion and denaturation.
In the case of polyglycine, we find that water is such a poor solvent for the peptide backbone that even 8 M denaturant doesn't functionally change the fact that you're at about 55 M water, so polyglycine remains in a highly collapsed state. However, in sidechain-containing polypeptides, the sidechain functional groups do have an intrinsic interaction energy associated with the solvent, with hydrophilic sidechains preferring a watery interface than a peptide one. This sets an initial level of expansion for the peptide chain, and primes the backbone, meaning it is much more exposed than it would be if the sidechains were absent.
As the concentration of denaturant is increased, the urea or Gdm- molecules interacts with the protein through the sidechains directly, facilitating some degree of unfolding. However, in addition to this sidechain-denaturant interaction, there is also a degree of denaturant-backbone interaction which was not possible without the effect of sidechain priming.
Essentially what this suggests is that sidechains are effectively providing a kind of initial solvation shell around the backbone. How that solvation shell (which is made up of sidechains) interacts with the bulk solvent has a major impact on the protein's extent of collapse/expansion, but it can also influence the extent to which the backbone is accessible for solute interaction too.
As a specific example, if we had a poly-lysine peptide (at pH 7.0) then the positively charged sidechains would have an extremely favorable free energy of solvation and would repel one another, which in turn would drive the chain into a highly extended conformation. Upon the the addition of urea to such a system, one might observe might a minor compaction of the chain (or have no impact, I don't really know). On the other hand, if the protein in question was polyglutamine (which has an amide-containing sidechain) then in water polyglutamine would lie in a collapsed conformation, and the addition of urea would have a modest effect (at best) because the glutamine-water priming is very weak (despite the fact that glutamine-urea interaction should be strong).
Of course, all this ignores the role of secondary structure, which is going to have an impact on the effectiveness of denaturants as it contributes to the stability of the folded state. However, it won't fundamentally change any of the conclusions.
An attractive outcome from this work is that both sidechains and backbone are relevant and operate in a kind of synergy. We propose that the fact that most folded proteins show similar denaturant stability behavior stems from the fact that the sidechain priming effect will define the degree of expansion for the unfolded state under folding conditions. As a result, there is likely to be a protein folding-sweet spot where proteins are semi expanded to facilitate rapid folding (i.e. minimizing internal friction), but not so expanded that folding becomes too entropically unfavorable. This sweet-spot is likely to make the overall sidechain priming effect fairly similar for all proteins (in a chemical potential sense) such that folded proteins will have relatively similar responses to similar concentrations of denaturants.