One of the earliest essays I wrote on the co-translational folding concept. This was for a course on protein folding that Prof Balaram takes at IISc. I am reproducing it in full, with missing punctuation marks added. At one point, I say that I have not been able to find out why it is assumed that refolding is a good substitute for in vivo protein folding. I did know even then, but I chose not to talk about it. I did so, since I cannot talk about the topic without calling everyone involved an idiot. Also, I have omitted the references as I cannot bring myself to format them properly. Rest assured that nothing I said in the essay is without literature support.
Co-translational folding
Introduction
Protein folding is one of the major unsolved problems of molecular biology. The problem is as follows : given that a gene encodes a unique structure (which performs a unique function), how does this information conversion take place ? it is assumed (barring few exceptions), that a particular sequence leads to only only a unique structure (fluctuations about this structure are possible but the gross topology remains the same). But, it is not necessary for a structure to have only one sequence, which leads to it. In that sense, the exact arrangement of amino acids does not seem to be absolutely necessary for the attainment of a structure and many related sequences can lead to the same or very similar structures. So, an analysis of folding of related sequences can lead to a deeper understanding of the folding process in vivo.
In that sense, the protein folding problem is how can we fold a given sequence of amino acids into a compact and stable structure ? What are the physical forces that dictate the formation of the native structure ? How does the cell fold proteins efficiently and correctly every time it synthesizes a polypeptide ? Or, does it make mistakes and these are digested away ? These are some of the questions for which a clear answer at present is not possible.
The initial experiments, which gave rise to the present state of confusion, were done on Ribonuclease A by Anfinsen and co-workers. Prior to that, experiments had shown that a protein upon denaturation loses its strucutre. This dentaturaion was thought to be an irreversible process (like boiling an egg). Anfinsen showed conclusively for the first time that denaturation need not be irreversible. At least for the case of Ribonuclease A, it was shown that upon removal of the denaturant, the protein regains close to 100% of its activity. Such experimnets done on mnay other proteins now suggest that the information required to refold the protein from the denatured state resides within the sequence of amino acids of the protein. The assertion that this native structure is the thermodynamically most stable one (global minimum) has attracted the attention of many scientists and is generally believed to be true but has not been rigorously proved for all proteins.
Refolding vs folding in the cell
Proteins can refold does not necessarily mean that they do refold in vivo also.So, the obvious question which one must ask is : how good an approximation is refolding of the process of protein folding in the cell ? This is the first question I would discuss in this essay. Recent and older reviews on refolding have helped in summarizing the following facts related to refolding. These are :
In that sense, the protein folding problem is how can we fold a given sequence of amino acids into a compact and stable structure ? What are the physical forces that dictate the formation of the native structure ? How does the cell fold proteins efficiently and correctly every time it synthesizes a polypeptide ? Or, does it make mistakes and these are digested away ? These are some of the questions for which a clear answer at present is not possible.
The initial experiments, which gave rise to the present state of confusion, were done on Ribonuclease A by Anfinsen and co-workers. Prior to that, experiments had shown that a protein upon denaturation loses its strucutre. This dentaturaion was thought to be an irreversible process (like boiling an egg). Anfinsen showed conclusively for the first time that denaturation need not be irreversible. At least for the case of Ribonuclease A, it was shown that upon removal of the denaturant, the protein regains close to 100% of its activity. Such experimnets done on mnay other proteins now suggest that the information required to refold the protein from the denatured state resides within the sequence of amino acids of the protein. The assertion that this native structure is the thermodynamically most stable one (global minimum) has attracted the attention of many scientists and is generally believed to be true but has not been rigorously proved for all proteins.
Refolding vs folding in the cell
Proteins can refold does not necessarily mean that they do refold in vivo also.So, the obvious question which one must ask is : how good an approximation is refolding of the process of protein folding in the cell ? This is the first question I would discuss in this essay. Recent and older reviews on refolding have helped in summarizing the following facts related to refolding. These are :
- Refolding is an equilibrium process. This means that under strong denaturing conditions, the protein should stay in a denatured state and upon removal of denaturing agent, the protein should be back to the native state and stay in that state.
- Refolding has not been shown for all proteins. In other words, all proteins fold, but not all proteins refold. Moreover, refolding rates are usually much slower than in vivo folding rates. For example, Firefly luciferase becomes active within seconds after its synthesis is completed, whereas the refolding of the same protein takes a few minutes to hours.
- The unfolded state is for some strange reason thought to be in a random coil state. This view of the denatures state of proteins has come under attack, based on recent NMR studies on the denatured state. These studies show that there is residual structure in the denatured state. Other studies on homo polymers of amno acids suggest that the residues do not have as much conformational freedom as it was assumed previously for random coils.
- Many different models for solving the protein folding problem have been proposed and no single model seems to be true for all proteins. These models are nucleation-condensation, hydrophobic collapse and framework model. Note that each of these models is only s asimplification of the folding process and the actual process may be a complex mixture of all these models.
- Even for single domain proteins lacking disulphide bonds and cis-prolines, multiple pathways of refolding exist and have been demonstrated experimentally.
- Since in vivo, proteins fold in conditions which favor aggregation, there must be mechanisms that prevent such aggregation. The chaperone proteins perform this function by binding to the hydrophobic patches on newly synthesized proteins. The chaperones and chaperonins provide an environment in which proteins can refold. Thus, the chaperone proteins play a passive role in protein folding.
Co-translational folding
Direct experimental demonstartion of co-translational folding has not been done and does not seem possible with the spatial and temporal resolution of current experimental techniques. Most of the experiments illustrate the fact that folding can and does start while the polypeptide has not been synthesized fully. In a typical experiment designed to test co-translational folding, a truncated polypeptide is tested for presence of secondary or tertiary structure. If the nascent polypeptide has teriary structure (which may or may not be the native structure), then that means the part of polypeptide synthesized has the ability to fold into a stable structure and can serve as a nucleation site for the formation of the rest of the native structure. Thus, in my opinion, the following assumptions are made while proposing progressive co-translational folding of proteins :
- Parts of the full protein can fold into native like or stable structures on the ribosome without the full sequence being available.
- Nucleation sites are created co-translationally upon which the rest of the structure is built either co-translationally or right after the synthesis of the polypeptide is complete.
- In a strictly co-translational process, after nucleation, each step of synthesis would result in a stable structure being formed including the residue added most recently.
The hypothesis of progressive formation of structure seems interesting because it would avoid formation of potential intra-molecular aggregates, which is the major cause of poor efficiency of refolding of multi-domain proteins. Thus, many studies on co-translational folding have focussed on proteins which have more than one domain (e.g. firefly Luciferase). The techniques applied to detect structure in a partially synthesized protein are the following :
- Resistance of limited proteolysis.
- Ability to bind conformation specific antibodies.
- Ability to bind specifically to a ligand (e.g Heme moiety).
- Enzymatic activity.
- CD and NMR spectra.
Each of these techniques has been applied to many nascent proteins, and it has been shown that such polypeptides have the ability to fold into a native-like conformation without the synthesis of the full polypeptide. Thus, the assumption that fragments of polypeptide can fold as it is synthesized does not seem unreasonable. But, till a direct analysis of the conformation attained by the polypeptide is not done, the question will remain unanswered.
Assuming that the polypeptide does fold co-translationally is not improbably also because of the fact that the time of formation of helix formation and turn formation (micro seconds) is faster than the time taken for one cycle of elongation on the ribosome (around 10-100 microseconds). Thus, if co-translational folding takes place, then it should be possible to misfold proteins by changing the speed of translation. Again in this case, direct experiments by changing the speed of ribosome, have not yet been done. Indirectly however, such a scenario can be simulated by substitution of synonymous rare codons in place of the common ones. It is found from such studies that proteins fold less efficiently when the ribosome moves faster through the mRNA. Such studies are hard to explain if folding is a purely post-translational process.
Problems with co-translational folding
One of the first experiments done of refolding of proteins showed that co-translational folding could not occur. These experiments were done on Ribonuclease in which five residues at the C_terminus were chopped off. It was shown that such a protein lacking the C-terminus, could not fold into the native state. Thus, the full sequence is required before refolding can even start. But, to my knowledge, it has not been shown that the peptide loses its structure completely. In another case (of prosubtilisin), it has been shown that a truncated subtilisin lacking the pro sequence (at the N-terminus) cannot fold into the native state. It however exists in a molten globule state, which is pushed towards the native state upon addition of the N-terminal fragment. Thus, the fact that chopping off C-terminal residues leads to non-native structure just means that for some proteins, the C-terminal residues are absolutely essential for the complete folded structure and does not necessarily mean that co-translational folding cannot occur for any protein. In this regard, it is worth mentioning that for a protein, the molten globule state is seen when only 80% of the sequence is synthesized suggesting that the full sequence is not required to form such molten globules.
Other observations that have been suggested for co-translational folding not being necessary are, (1) chemical synthesis of proteins (that proceeds from C-terminus to N-terminus opposite to in vivo synthesis) and (2) circular permutations that fold to the same structure. Chemical synthesis seems to have been done for proteins that are very small (the longest to my knowledge is 124 amino acids). These proteins would not fold co-translationally as has been mentioned earlier (because of shielding by the ribosome exit channel). Thus, the success of refolding does not prove that co-translational folding does not or cannot occur. It merely points to the fact that small proteins that are routinely used in refolding studies, are not a good model system for studies on in vivo folding.
The case of circular permutants that fold into the native state seems to be the case in my opinion, which is hard to explain if co-translational folding takes place. Still, there are some points that need to be looked into more detail. First of all, circular permutations of all proteins are not known to occur naturally. Only those proteins in which the N and C termini are close together in 3-D space seem to be amenable to circular permutations. Even for these, the mutations that fold into the native structure are restricted to certain points. If we link the termini of any protein together with a linker of amino acids and then show that it folds into the same native structure, then it would indeed mean that refolding might be closer to in vivo folding. But such experiments do not seem to work and perhaps thus have not been reported so far.
The second point, I would like to illustrate using an extreme example. Suppose there is a protein triangulin, which has three helices and three turns which maintain the angle between the helices rigidly at 60 degrees. There are absolutely no tertiary contacts. Such a protein would fold into the native structure no matter whwere it is cut and circular permutations made, as the structures are formed by local contacts alone. Such proteins do not exists in nature, but it illustrates the point that if local contacts dominate in a protein, then that protein is amenable to circular permutations. Such a study has been done using a quantification of local contacts known as 'contact order'. It ha sbeen shown that the stability of circular permutations was correlated with the contact order, and the lesser the contact order (more local contacts), the more stable are the circular permutants.
Thus, even though co-translational folding seems possible, it hasn't been shown conclusively that it is indeed the dominant model for in vivo folding. More work needs to be done in order to clarify the issues related to co-translational folding.
Other observations that have been suggested for co-translational folding not being necessary are, (1) chemical synthesis of proteins (that proceeds from C-terminus to N-terminus opposite to in vivo synthesis) and (2) circular permutations that fold to the same structure. Chemical synthesis seems to have been done for proteins that are very small (the longest to my knowledge is 124 amino acids). These proteins would not fold co-translationally as has been mentioned earlier (because of shielding by the ribosome exit channel). Thus, the success of refolding does not prove that co-translational folding does not or cannot occur. It merely points to the fact that small proteins that are routinely used in refolding studies, are not a good model system for studies on in vivo folding.
The case of circular permutants that fold into the native state seems to be the case in my opinion, which is hard to explain if co-translational folding takes place. Still, there are some points that need to be looked into more detail. First of all, circular permutations of all proteins are not known to occur naturally. Only those proteins in which the N and C termini are close together in 3-D space seem to be amenable to circular permutations. Even for these, the mutations that fold into the native structure are restricted to certain points. If we link the termini of any protein together with a linker of amino acids and then show that it folds into the same native structure, then it would indeed mean that refolding might be closer to in vivo folding. But such experiments do not seem to work and perhaps thus have not been reported so far.
The second point, I would like to illustrate using an extreme example. Suppose there is a protein triangulin, which has three helices and three turns which maintain the angle between the helices rigidly at 60 degrees. There are absolutely no tertiary contacts. Such a protein would fold into the native structure no matter whwere it is cut and circular permutations made, as the structures are formed by local contacts alone. Such proteins do not exists in nature, but it illustrates the point that if local contacts dominate in a protein, then that protein is amenable to circular permutations. Such a study has been done using a quantification of local contacts known as 'contact order'. It ha sbeen shown that the stability of circular permutations was correlated with the contact order, and the lesser the contact order (more local contacts), the more stable are the circular permutants.
Thus, even though co-translational folding seems possible, it hasn't been shown conclusively that it is indeed the dominant model for in vivo folding. More work needs to be done in order to clarify the issues related to co-translational folding.
Comments