Folding can be a problem
As we shall see in Chapter 16, proteins are synthesized as unstructured polypeptide chains. The protein-synthesizing machinery of the cell does not itself impose structure. To become fully functional, the chain must fold into its appropriate secondary and tertiary conformations and, in some cases, associate into quaternary structures. How this happens has long intrigued molecular biologists. Surely it cannot happen by a random search through possible conformations: even if there are only two orientations at each residue, a 100-residue polypeptide would have 2100 conformations. Searching only a fraction of these, at nanoseconds per search, would take billions of years. Does each kind of protein have a template protein that shows it how to fold? Then what folds the template?
In reality, it appears that protein folding often proceeds in a stepwise, orderly manner dictated by the primary structure, the information the protein is provided with as it is made. This is referred to as the self-assembly principle. Evidence for this comes from many experiments that show unfolded proteins spontaneously attaining their functional conformation after having been denatured; that is, having had their native structure unraveled. A classic example is shown in Figure 1A, which depicts the thermal denaturation of ribonuclease, a small globular protein. If ribonuclease in solution is heated above 40°C, all evidence for defined secondary and tertiary structure, as well as enzymatic activity, is lost. Upon recooling, all of these features are quantitatively regained. Furthermore, comparison of data from various techniques indicates that this is an all-or-none transition: at any intermediate temperature, a mix of native and denatured molecules is present. The argument is as follows: different techniques for studying protein structure are sensitive to different aspects of the structure, which relate to various parts of the protein. If proteins are folded piecewise, then tracking the folding by a range of methods should lead to distinguishable curves in different experiments. Instead, the curves obtained by different methods are superimposable (Figure 1B). Thus, we know that folding, at least in this case, is a fast all-or-none process, occurring in seconds or minutes.
Fig1. Protein folding. (A) Thermal denaturation of ribonuclease A. When ribonuclease is heated above a certain temperature, it undergoes a conformational transition from a highly ordered, catalytically active structure into a catalytically inactive, random coil that experiences continuous fluctuation among a large number of extended conformations. When the enzyme is inactivated by elevated temperature, its disulfide bonds remain intact, allowing for faster renaturation upon reduction of the temperature; other denaturation treatments may lead to rupture of these bonds and hence to slower renaturation. If disulfide bonds are reduced concomitant with denaturation, reoxidation will lead to a scrambled structure with incorrect bonds. However, removal of denaturant prior to reoxidation yields the correct structure. (B) A range of physical methods has been used to study the unfolding (denaturation) and the refolding (renaturation) of proteins. (C) Hypothetical free energy map for a protein. The vertical axis represents free energy, and the x- and y-axes represent different conformations; actually, there should be about 2100 such dimensions, but that is hard to draw. The map illustrates that there is one true free energy minimum, representing the most stable conformation or native state, and many pseudominima at slightly different conformations in which the molecule could, theoretically, become temporarily trapped. [A, B, adapted from Mathews CK, van Holde KE, Appling DR et al. (2012) Biochemistry, 4th ed. With permission from Pearson Prentice Hall. C, from Dill KA & Chan HS (1997) Nat Struct Biol 4: 10–19. With permission from Macmillan Publishers, Ltd.]
Such experiments have shown that, for many proteins, the amino acid sequence, and thus the DNA sequence of the gene that dictates it, can define the conformation and functionality without external aid. The native structure represents a free energy minimum, the state of optimum stability under physiological conditions (Figure 1C). This is a very important principle, one of the cornerstones of molecular biology. However, having stated it, we must add some important qualifications.
First, after a protein is synthesized in the cell, it may be subject to chemical modification. Such post-translational modifications (PTMs) are of two general types: a part of the protein may be cut off by specific proteolysis, or side-chain residues may be chemically modified. A list of the more important modifications is given in Table 1. A powerful and extraordinarily precise tool for locating and identifying such modifications is mass spectrometry (MS) (Box 3.9). Although post-translational modifications are not directly dictated by the DNA sequence, they are certainly influenced by it; the correct amino acid residue must be at the correct position for modification to occur. Furthermore, modification always requires specific enzymes, which themselves are specified in the DNA. So DNA still has the final say.
Table1. Most important postsynthetic modifications of proteins.
Second, the folding of many proteins, especially in the crowded environment of the cell, is undoubtedly more complicated and difficult than in the test tube. Interactions with other cellular molecules, including proteins, may block correct folding. Or there may exist other states, distinct from the native state, in which the protein can be trapped. A way of looking at this is shown in Figure 1C. The deepest point in the free energy landscape is the native conformation. In the complex kinetic process of becoming folded, the protein may be temporarily trapped in an adjacent, shallow energy well.
It has recently been proposed that multiple conformations close in energy to the native conformation may actually serve biological purposes. A protein that exhibits such behavior might be promiscuous in function, serving more than one purpose or allowing easy evolvability into new functions. But such alternative available states may also lead to misfolding, which can have serious consequences.
Misfolded proteins are generally useless and may even be harmful to the cell, especially through their tendency to aggregate: they must be dealt with. There are two major ways to deal with misfolded proteins. First, they can be given a chance to avoid misfolding or they can be allowed to refold to their native, functionally active state. Macromolecular structures that can help to accomplish these activities are called chaperones. Second, misfolded proteins may be sent to specialized cellular machineries for degradation (Figure 2). Both refolding and specific degradation require energy expenditure through ATP hydrolysis, which is believed to lead to conformational transitions in the respective proteins.
Fig2. Three possible fates of proteins after synthesis. Some proteins can correctly fold by themselves. The cell has evolved two alternative pathways to take care of misfolded proteins: they are either refolded with the help of specialized folding machines, known as chaperones, or sent for degradation by a destruction machine, the proteasome.
Chaperones help or allow proteins to fold
Chaperones are huge protein complexes usually considered as protein folding machines. Chaperones enclose the protein, protect it from the cellular environment, and give it a chance to fold properly. One might say they serve as protein incubators. More than 50 chaperone families have been described to date. We must note, though, that chaperones have a number of additional cellular functions, sometimes seemingly opposite to their role in folding. For example, proteins destined to function in or be transported through membranes should not fold before being inserted into the membrane, and their premature folding is prevented by special chaperones. Chaperones also interact with nascent polypeptide chains emerging from their site of synthesis on ribosomes, so that they do not associate (aggregate) with chains synthesized nearby.
As a specific example, we consider the best-studied bacterial chaperone system, GroEL/GroES, also known as chaperonin 60/chaperonin 10. The GroEL/GroES system is indispensable for the survival and growth of the bacterial cell as numerous proteins, both small and large, depend on it for proper folding. The subunit composition and structure of this large protein complex are shown in Figure 3, and the reaction cycle is schematically presented in Figure 4.
Fig3. Structure of GroEL/ GroES. The two heptameric rings of GroEL have a characteristic double doughnut structure, as modeled from electron microscopic maps. Ribbon representations of the atomic structure of two GroEL protomers are superim posed on the model. Each GroEL subunit consists of an equatorial domain shown in light blue in the top molecule, which contains the nucleotide-binding site, an intermediate hinge domain shown in jade, and an apical domain shown in royal blue. The apical domain is located at the opening of the GroEL cylinder and contains the binding site for both GroES and the chaperoned polypeptide. The polypeptide binds at a hydrophobic groove in the apical domain facing the central channel. GroES, also termed the lid, is a dome-shaped heptamer that consists almost exclusively of β-sheets.
Fig4. Reaction cycle of the bacterial GroEL/GroES chaperonin system. The nucleotide-binding abilities of the two GroEL rings are mutually exclusive: thus, ATP or ADP can be bound to only one ring at a time. The GroES lid binds to the ring that has the nucleotide bound, creating a cage. The binding of the unfolded or partly folded substrate protein to the open GroEL ring is followed by binding of ATP and GroES. The ring cavity now closes and the substrate is released into the cage, where it can fold. Following ATP hydrolysis, the binding of ATP to the opposite ring triggers the dissociation of GroES and the dissociation of folded substrate protein from the complex. The chaperonin complex is now ready to accept another polypeptide. [Adapted from Ellis RJ (2006) Nature 442: 360–362. With permission from Macmillan Publishers, Ltd.]
Another important member of the chaperone family is Hsp90 (Heat-shock protein 90). This protein, originally identified as one of the proteins induced by stress conditions, including heat, is the most abundant heat-shock protein under normal physiological conditions. It has roles in both protein folding and protein degradation and has been implicated in multiple biochemical pathways, including signal transduction. Although we know quite a bit about Hsp90 (Figure 5), its mode of function is still not fully understood, nor do we fully understand why or how it can exhibit such diverse functions.
Fig5. Chaperone Hsp90 and its ATPase cycle. (A) Structure of chaperone Hsp90 from yeast. The active protein is a dimer, with each subunit consisting of three distinct domains. The N-terminal domain (ND), shown in red, possesses a deep ATP-binding pocket, where ATP binds in an unusual, kinked conformation. The N-terminal domain in eukaryotes is connected to the middle domain (MD), shown in green, by a long flexible linker sequence. The C-terminal domain (CD), shown in royal blue, is the dimerization domain and contains, in eukaryotes, an amino acid motif MEEVD that binds to a variety of co-chaperones whose function is to regulate the ATPase activity of the N-terminus. (B) ATPase cycle of Hsp90. (Step 1) Following ATP binding, the lid flaps over the ATP-binding pocket. (Step 2) Lid binds to ND of the other subunit, producing a strand-swapped, transiently dimerized conformation. (Step 3) ATP hydrolysis; NDs dissociate from MDs; monomers separate N-terminally; lid opens. (Step 4) ADP and inorganic phosphate are released. The exact mode of client-protein binding remains unknown, as does the actual mechanism of protein folding. [A, adapted from Wikimedia. B, adapted from Mayer MP, Prodromou C & Frydman J (2009) Nat Struct Mol Biol 16: 2–6. With permission from Macmillan Publishers, Ltd.]
ATP hydrolysis by Hsp90 is rather slow: in humans, Hsp90 hydrolyzes one ATP mole cule in about 20 min. The slow hydrolysis indicates that the complex conformational rearrangements that occur in the Hsp90 dimer during its action are coupled to ATP hydrolysis (see Figure 5). Such coupling is a recurring theme in many reactions that involve ATP hydrolysis. The current view is that ATP binding and hydrolysis are instrumental in shifting the protein from one conformation to another; these conformational transitions, in turn, determine the activity of the protein. The equilibrium between different protein conformations is usually regulated in the cell; in the case of Hsp90, there are a large number of co-chaperones that bind to the C-terminus of Hsp90 and regulate its ATPase activity.
If the conformational transitions are inhibited one way or another, the protein is inactivated. Drugs that inactivate Hsp90 by blocking the ATP-binding site, and thus preclude the ATP-mediated conformational transitions, are widely used in clinical practice as anti-tumor agents. Presumably, these drugs induce apoptosis through inhibiting growth-factor signaling pathways; Hsp90 stabilizes growth-factor receptors.
