Much of the information in this review, along with the figures and diagrams are from Bukau, B. and Horwich, A.L. 1998. The Hsp70 and Hsp60 Chaperone Machines. Cell 92, 351-366.
Ribosomes synthesize proteins as a linear chain of amino acids whereby the sequence of amino acids is dictated by the sequence of the nucleotides in the mRNA. This amino acid sequence, in turn, is responsible for the three-dimensional conformation (shape) of the protein. Here, the particular amino acid sequence is a particular sequence of amino acid side chains of differing sizes with differing chemical properties (hydrophilic, hydrophobic, basic, and acidic). Thus, by positioning these chemically different side-chains in a particular manner, a particular set of interactions is ultimately favored and these interactions give rise to a particular conformation that is most thermodynamically stable.
Scientists have long recognized that the primary structure (the amino acid sequence) of a polypeptide chain is sufficient cause for protein folding. Simple experiments have shown this. In these experiments, proteins are isolated and subject to denaturing conditions (detergents, salts, etc.). These conditions disrupt side-chain interactions so that proteins unfold. And because of these denaturants, the proteins exist as linear chains of amino acids. However, if the denaturants are removed (for example, with dialysis), the amino acid chains can again fold up into their original three-dimensional conformation. This ability to refold clearly demonstrates only the amino acid sequence is needed to fold the protein (however, there are some exceptions, where proteins fail to refold properly after removal of the denaturant).
These experiments have long led scientists to believe that protein folding in the cell was a rather simple process that did not need input from other systems. That is, once the protein was synthesized by the ribosome, it would simply fold into its proper shape as dictated by the amino acid sequence of the protein. However, recent studies are beginning to show that protein folding events are more complicated inside the cell. Hindsight should have told us that this should have been the case, as there are two very good reasons for suspecting protein folding to be a complex event in a cell (and not as simple as that seen in the test tube).
The first reason for this is that the intracellular conditions are not like the test tube conditions of the lab. In the cell, protein folding occurs in an environment where the concentration of proteins (synthesized and being synthesized) is quite high. This is important, because newly-made proteins which have yet to fold are exposing hydrophobic residues (recall that globular proteins bury most of their hydrophobic residues in the core of the protein). These residues are like micro-oil patches that can stick to other hydrophobic residues on other unfolded proteins. When proteins begin to stick, they can form a 'seed' that serves to form protein aggregates (this is analogous to crystal formation). These aggregates can grow and eventually disrupt cellular processes. In fact, there is growing evidence that prion diseases (like Mad Cow Disease) and maybe even Alzheimer's Disease is causes by the abnormal formation of excessive protein aggregates. What this means is that an unfolded (or partially folded) protein can proceed in two different directions - towards a folded state or towards an aggregate. And given the very high concentration of proteins inside a cell, the equilibrium between these two states is predicted to strongly favor the aggregate state.
The second reason for expecting protein folding to be a complicated affair inside a cell is that unfolded and partially folded proteins have another potential fate, namely, degradation. Proteins that are improperly folded tend to be recognized by machinery that degrades the proteins (we will discuss this in the next lecture). That is, a partially folded protein may display degradation signals that are normally buried inside the folded core. Inside a cell, there is probably a kinetic competition between the degradation of partially folded proteins and their folding into the proper globular conformation.
In a test tube, the concentration of proteins is usually much lower than that found within the cell (and even then, proteins tend to stick to each other). Also, there is no degradation machinery. Thus, in the test tube, the folding pathway does not have to compete against the aggregation pathway or the degradation pathway. Yet in the cell, these alternative fates not only exist, but would seem to be favored. These considerations alone would seem to allow us that there is more to protein folding inside a cell than there is inside the test tube.
So how does the cell prevent the unwanted aggregation and degradation of unfolded proteins? There are two possible solutions.
1. Protect the unfolded proteins. When a protein exists in an unfolded state, block the hydrophobic patches. This would prevent aggregation and hide degradation signals.
2. Quarantine the unfolded proteins. Create a separate environment that favors protein folding. This can be done by sequestering unfolded proteins so they cannot interact with other and form aggregates. Ideally, we would want to capture each distinct protein and allow it to fold in isolation from the rest of the cell.
Both strategies would thus serve to increase the rate of on-pathway folding to out-compete aggregation and degradation. Well, what do you think the chaperones do?
Chaperones are defined as proteins and protein assemblies that help other proteins fold into their proper conformation. They were originally identified as heat-shock proteins (two classes we will consider are hsp60 and hsp70). They were called heat-shock proteins because when microorganisms were briefly exposed to high temperatures, these proteins were synthesized. It was thus proposed that heat-shock proteins were inducible repair proteins that worked to reverse the protein denaturation caused by the heat exposure.
Since this time, it has been discovered that chaperones not only are overexpressed in response to heat, but they are constantly being made and are involved in many cellular processes. In fact, chaperones are associated with protein synthesis, as the growing polypeptide that emerges from the ribosome interacts with chaperones. Chaperones also interact with proteins and facilitate their transport across membranes (we will
see that proteins often have to be unfolded to get across a membrane and then refolded once they get across the membrane). Furthermore, chaperones may function in an assembly-line fashion, where they pass proteins (in differently folded states) to each other (see Science 274, pp. 1613-1614) as part of the overall folding process. The bottom-line is that chaperones are everywhere in the cell and seem to be involved with many cellular processes. With this in mind, let's look at two different classes of chaperones. Hsp70 is a simple chaperone that is found in all living organisms. It functions to protect unfolded proteins. Hsp60 is a molecular machine that functions to isolate unfolded proteins and provide the optimal environment for on-pathway folding.
Hsp70 is a single, monomeric protein that is found throughout the cell. Very closely related proteins are DnaK (studied in bacteria) and BiP (associated with the endoplasmic reticulum). Hsp70 plays many roles in the cell. Not only does it assist in the process of folding, but it also guides proteins across membranes (a process known as translocation). Hsp 70 can also assist in the disassembly of large protein complexes. It has also been implicated in the control of regulatory factors (such as transcription factors).
To imagine the rough shape of hsp70, hold out your hand in front of you so that your thumb and index finger faces you (imagine you are going to use your hand as a shadow-puppet). The space between your thumb and index finger is where hsp70 will clamp onto an unfolded protein. To see what hsp70 really looks like, click here (and come back). Hsp70 has two domains. The N-terminal domain constitutes 2/3 of the protein's mass and is the bottom portion as shown in the figures (of your thumb if you are using your hand). This domain contains an ATPase activity (meaning it breaks down ATP into ADP and P). The C-terminal region (the top portion or your index finger) is the substrate-binding domain. There is a hinge region in this domain which allows the C-terminal portion to bend open and allow for substrate binding. The opening and closing of this hinge is coupled to the ATP hydrolysis in the N-terminal domain.
Hsp70 can only bind unfolded proteins by recognizing an extended region of the polypeptide chain that is rich in hydrophobic residues. Thus, a floppy chain containing many hydrophobic amino acids will be bound by hsp70, but not the compact globular proteins which have hydrophilic amino acids on their surface. By clamping onto these regions of unfolded proteins, aggregation is prevented. However, not only is aggregation prevented, but so too is further folding. Thus, hsp70 seems to provide a temporary blocking solution. Let's now consider the mechanism of hsp70 binding.
The key to understanding hsp70 function is to remember that ATP hydrolysis in the N-terminal domain drives conformational changes (focused on the hinge) in the C-terminal domain. As such, it follows that hsp70 can exist in two conformations. These are the ATP-bound state (before hydrolysis) and the ADP-bound state (after hydrolysis). Let's consider some facts that are known about these two states.
The ATP-bound state has a low affinity for the substrate. This is because the hinge is flexed so that the substrate binding pocket is open. Thus, there are fast exchange rates meaning an unfolded protein that enters the pocket is likely to quickly leave the pocket.
The ADP-bound state has a high affinity for the substrate, because in this state, the hinge is flexed so that the binding pocket closes and clamps down on the substrate protein. Thus, there is a slow exchange rate as the unfolded protein is held in the pocket by a series of amino acid interactions with the C-terminal domain (functioning like teeth on a clamp).
Essentially, hsp70 exists in an open state (bound to ATP) where unfolded proteins can enter the binding pocket (between the thumb and index finger) and once inside, ATP is hydrolyzed and the pocket snaps shut. The unfolded protein can then be clamped by several hsp70 molecules to prevent aggregation. Yet this simple mechanism comes with its own set of problems.
First, what triggers the pocket to open up again? Ideally, we would like the N-terminal domain to release the ADP and bind to a new ATP molecule, as this would trigger a conformational change back to the open state. Yet when hsp70 is clamped onto a protein, ADP is locked in place and ATP can not access its binding site (not to mention its binding site is not opened-up correctly).
Secondly, what triggers the pocket to clamp shut in the first place? Recall that in the open state, substrate protein does not remain in the pocket long as it has only a weak interaction with hsp70. One could imagine that there was a type of signal on the substrate protein that interacted with hsp70 to trigger ATP hydrolysis, but this would entail putting these signals throughout the sequences of all proteins in the cell. Thus, this hypothetical signal mechanism would not be very efficient.
Third, if there is no trigger, hsp70 could simply open and shut on a continual basis. This could be used by the cell, but remember that such cycling uses ATP. This would then be an example of a futile ATPase cycle, where energy is simply wasted.
The cell solves these problems by recruiting other proteins to solve the first two problems (and thus prevent the futile ATPase cycling). Put simply, one protein brings the substrate to the binding pocket of hsp70 and passes it on so that hsp70 is triggered to clamp, while another protein can interact with closed-hsp70 to get it to open up. Let's consider one model of such activity as shown below.
The cycle begins when a substrate protein interacts with a co-chaperone known as DnaJ. DnaJ protein then interacts with the ATP-bound form of hsp70 (if the figure is confusing you, recall hsp70 is known as DnaK in bacteria). This interaction involves two coupled steps: 1. the unfolded substrate protein is transferred to the binding pocket of hsp70 and 2. ATP hydrolysis is triggered in hsp70. As a consequence of ATP hydrolysis, ADP is generated and hsp70 closes its binding pocket. This also causes DnaJ protein to dissociate. In this state, the substrate is now firmly bound by hsp70. When it is time to release the substrate, another protein known as GrpE binds to hsp70. GrpE binds to the N-terminal domain of hsp70, causing it to shift its conformation so that the ADP is released (click here to see what the two proteins look like in complex). The ATP-binding site is now exposed so ATP can bind. When ATP binds, the binding pocket of hsp70 is opened to allow the substrate to be released. Hsp70 now exists in the state that started the cycle.
Once the substrate is released, it has four possible fates.
1. It can complete folding into the proper conformation.
2. It can be passed on to other chaperones, like hsp60.
3. It can form aggregates.
4. It can rebind to hsp70 for another cycle.
Future research needs to be done to better understand not only what determines which outcome will occur, but also to determine the manner in which both DnaJ and GrpE function is regulated.
Hsp60 - The Chaperonins
Remember that hsp70 mainly serves to block aggregation. The hsp60 class of chaperones do something entirely different - they provide an "isolation chamber" in which individual unfolded proteins can fold unimpeded. Whereas hsp70 prevents improper folding and aggregation, hsp60 promotes proper folding.
Hsp60 is a huge protein complex composed of 14 proteins. These machines are organized into two 7-protein rings that are stacked on each other like two tires (each protein that is a member of a ring is the same protein and is called a subunit). The space within this hollow tube will serve as the site of protein folding. Hsp60 is best understood in E. coli where it is called GroEL. The GroEL chamber also is capped by a smaller protein lid called GroES. GroES is composed of a simple ring of 7 smaller proteins. Click here to see the structure of GroEL with and without its cap. The subunits also bind and hydrolyze ATP near the center of the complex.
The isolation chamber mechanism
The simple mechanism behind GroEL/hsp60 is this: a single unfolded protein gains access to the interior of the chamber and it is capped. Once capped, the unfolded protein can begin to fold in isolation from all other unfolded proteins. This has led some scientists to propose this as infinite dilution, where a molecule can fold in complete isolation. But the unfolded protein is not really alone, as the walls of the chamber are themselves made of protein. So what keeps the unfolded protein from gunking up the chambers?
The mechanism by which GroEL promotes folding is astounding. Like hsp70, GroEL can exist in two states that correspond to the state of bound-ATP.
1. The binding active state. In this state, ATP is bound to the inner ring and the end of the tube is open. Hydrophobic residues line the walls of the tube and serve to capture unfolded proteins.
2. The folding-active state. In this state, a conformation change occurs in all the subunits of the rings. This conformation change is induced by the hydrolysis of ATP and the binding of the GroES lid. Let's consider the conformation change in more detail.Click here to see GroEL in these two states (binding state on upper left and folding state on upper right) and use this picture to understand the following description.
When GroEL interacts with a substrate protein, it changes from the binding active state to the folding active state. The conformation change causes the volume inside the chamber to increase two-fold and also causes the end of the tube to narrow and close-shut with binding to the GroES lid. This change comes about as all the ring members change their conformation in unison. When this shape change occurs, the hydrophobic walls of the chamber are thought to pull apart the substrate protein (that interacts with the walls through its hydrophobic residues). This may help to unfold misfolded portions of the substrate (such misfolding may have occurred earlier). Then, as the shape continues to change, all the subunits twist away and concurrently present a new wall of hydrophilic amino acids. Now, interactions between the unfolded protein and the wall of the chamber are broken. What's more, the hydrophilic environment favors the burial of hydrophobic amino acids in the substrate protein and exposure of hydrophilics on its surface. Thus, proper folding is promoted inside the chamber. The chamber remains closed for approximately 15 sec, which is sufficient time for most proteins to fold. After this time, the lid is released and the folded protein diffuses into the cytoplasm.
To understand the conformation of the chamber, consider the conformation change of the individual subunits (remembering that chamber shape changes come about as all the subunits change their shapes in unison). Again, click here to consider the individual subunits . Note the subunit has three domains:
1. The apical domain shown in red.
2. The intermediate domain shown in green.
3. The equatorial domain shown in blue.
The conformation change involves the intermediate domain moving downwards at an 25 degree angle. This locks ATP in place as ATP binds in the crevice between the equatorialand intermediate domain. At this point, ATP will be committed to hydrolysis as the proper amino acids are brought into place to promote this reaction. Yet once ATP is hydrolyzed, ADP remains in bound. At the same time this is occurring, the apical domain is undergoing radical conformation changes. Not only does it rise at a 60 degree angle, it turns sideways at a 90 degree angle. The 60 degree rise opens up the chamber volume, closes the ends, and makes interactions with the GroES lid possible. The 90 degree turn shifts moves the hydrophobic wall and replaces it with a hydrophilic wall.
The GroEL/hsp60 cycle
Recall that hsp60/GroEL actually exists as two stacked rings. But the folding chamber itself is composed only of one ring. Hsp60 is thus actually a two-chamber complex. But these two chambers exist in an asymmetric fashion with respect to substrate binding. That is, only one ring at a time holds and folds a substrate protein. Why is this?
One ring by itself can capture an unfolded protein, close off, and hydrolyze ATP. But it is then unable to open back up to release the folded product. Thus, let us see why two rings are needed. Refer to the text below to understand the following diagram of the hsp60 cycle.
Let's begin with a substrate the binds to the open chamber formed by one of the rings (the substrate was probably brought to hsp60 by another chaperone, such as hsp70). This is followed by ATP binding inside the chamber and GroES binding in concert with the conformation change (the folding active state). As the substrate folding is occuring inside this isolation chamber, ATP hydrolysis occurs (this is a slow step and is partly responsible for the 15 sec "clock" on the closed-chamber). When hydrolysis occurs, the chamber is primed for opening as the high affinity between the GroES lid and the edges of the chamber are loosened by slight shape changes. At this point, ATP can bind to the second open chamber. This binding in the second chamber sends an allosteric signal which sends a wave of conformation changes so that the GroES lid is jettisoned, and the folded protein diffuses away. The ADP is also released, triggering a conformation change in the ring that folded the protein so it resumes its original shape (the binding active state).
And thus ends our summary of two ways in which chaperones help proteins to fold inside a cell.