Problems

The aim of this section is to stimulate discussion and experimentation on two aspects of chaperonin function inside cells. These aspects are the quantitative importance of the chaperonins in assisting protein folding in vivo and the effects of the highly concentrated intracellular milieu on their mechanism of action.

A. Number of Proteins Using GroEL to Fold in Vivo

There appears to be little or no specificity for the type of aminoacyl sequence that can bind to GroEL in vitro and be released productively, so it is possible in principle that the folding of the majority of proteins is assisted by GroEL in E. coli. Moreover, the genes encoding both GroEL and GroES are essential for viability and the proteins are required for growth at all temperatures. One genetic study employed a temperature-sensitive mutation to shut off the production of functional GroEL in vivo and concluded that a minimum of about 30% of the soluble proteins of E. coli requires GroEL to fold correctly (Horwich et al., 1993). However, it is difficult to rule out that only a small number of proteins actually interact with GroEL, the folding of the remainder being disturbed by indirect pleiotropic effects. For example, if one of the proteins folded by GroEL is DnaK, the hsp70 homolog of E. coli, any decline in DnaK after the temperature shift could affect the folding of other proteins; this possibility arises since it is known that DnaK binds to many nascent chains and prevents their aggregation (see Chapter 8). How important this effect is will depend on the rate at which DnaK turns over in the dying cells. So the gene for GroEL could be essential because GroEL is required to fold a small number of essential proteins.

Another way to approach this problem is to calculate from available data on the cellular content of GroEL and rates of protein synthesis the fraction of proteins that could interact with GroEL under defined growth conditions. From the information provided by Neidhardt (1978) and by Bremer and Dennis (1987) for the growth of E. coli strain B/r on a glucose minimal medium at 37°C, and by Herendeen et al. (1979) for the content of GroEL, the following values emerge: (1) doubling time is 40 min; (2) total number of polypeptide chains per cell (average size 40 kDa) is 2.35 X 106; (3) cytoplasmic volume of cell is 0.6 /u,m3; and (4) GroEL content per cell is 1.65% of total soluble protein or 2.75 mg/ ml or 3.4 pM or 1224 oligomers per cell. The percentage varies less than 20% over doubling time changes of fivefold (Pedersen et al., 1978) so that the concentration of GroEL remains approximately constant as the cells change size with doubling time.

Thus each cell under the above growth conditions makes about 10,000 polypeptide chains every 10 s and contains about 1000 molecules of GroEL. If all these proteins were folded by GroEL this would require each GroEL to fold one chain every second. However, the actual requirement will be less than this because many proteins do not fold in the cytosol. There are four intracellular compartments known where proteins fold, but that do not appear to contain any chaperonin, i.e., the endoplasmic reticulum, intermembrane mitochondrial space, and intrathylakoid chloroplast lumen of eukaryotic cells, and the periplasmic space of bacteria. The exported proteins of E. coli identified in the study of Horwich et al. (1993) do not appear to require GroEL function. The proteins of E. coli that constitute the periplasmic space and the outer and inner membranes are estimated to be about one-third of total protein (Good-sell, 1991), so if we assume these fold independently of GroEL, the folding requirement reduces to two-thirds of 10,000 or 6666 chains folded per 10 s.

These values need to be compared with those for the activity of GroEL at refolding polypeptides in vitro. An immediate problem is that these studies are almost always conducted at around 25°C rather than at 37°C, so they are likely to be too low by a factor of two- to threefold. If we assume that the average polypeptide chain requires three rounds of ATP hydrolysis to be correctly folded by one GroEL oligomer, this takes about 60 s at 25°C (F-U. Hartl, personal communication, 1995). If we assume that it takes 20 s for one GroEL to fold one chain at 37°C, then the 1000 molecules of GroEL could fold 500/6666 or 7.5% of the total chains folding in the cytoplasm.

Despite the uncertainties in this calculation the results suggest that GroEL folds only a minority of proteins in E. coli—probably in the range 7.5-30%. This does not mean that any proteins fold spontaneously, however, since the cytoplasm contains types of molecular chaperone other than GroEL that bind to nascent chains, and in higher concentrations. For example, the DnaK content of E. coli is reported to be about 5000 molecules per cell (Neidhardt and VanBogelen, 1987). Another possibility is that GroEL functions significantly faster in vivo than it does in the in vitro conditions that have so far been tried. There is one striking aspect of the in vivo situation whose significance for models of chaperonin action is only just starting to be appreciated—the phenomenon of macromolecular crowding.

B. Macromolecular Crowding and Models for Chaperonin Action

The basic assumption of biochemistry as a discipline is that molecular properties determined using either cell-free extracts or pure components derived from these reflect to a significant degree properties relevant to the function of these molecules inside the living cell. Proteins are flexible and sensitive molecules, so the properties of proteins vary with the environment within which they are studied. It is relatively easy to use in vitro protein concentrations, temperatures, pH values, and concentrations of components such as metal ions, nucleotides, and other known effectors that are similar to those in which the proteins under study function in vivo. In the case of the chaperonins, the concentrations of GroEL oligomers commonly employed in protein refolding studies in vitro are in the range 1-2 ¡jlM. This range is close to the reported concentration of GroEL inside the cytoplasms of E. coli cells growing under defined conditions (3.4 ¡xM, see Section V,A). Whether this correspondence is due to luck or good judgment is not clear, but the fact that most measurements of GroEL activity are performed at 25°C suggests that the former is correct! However, there is one aspect of the in vivo situation that is rarely mimicked in biochemical studies, but that has a large effect on certain properties relevant to understanding how the chaperonins function in vivo—this is the macromolecular crowding (or excluded volume) effect.

Macromolecular crowding is the term used to describe the fact that the concentration of total macromolecules inside living cells is so high that a significant proportion of the cellular volume is physically occupied by molecules, and is thus unavailable to other molecules. Thus the total RNA and protein inside E. coli occupy a substantial fraction of the total volume (340 g/liter). In the words of Zimmerman and Minton (1993) "it is not generally appreciated that, from a quantitative point of view, biochemical rates and equilibria in a living organism bear scant resemblance to those measured in a bath of solvent." Both theory and experiment have shown that the thermodynamic activity of each macromolecular species in a crowded environment exceeds the activity of that species at an identical concentration in an uncrowded solution. For example, it has been estimated that the ratio of thermodynamic activity to concentration (i.e., the activity coefficient) for a spherical protein of 30 nm radius inside E. coli is in the range 100-1000 (Zimmerman and Trach, 1991). By contrast, the effect of high volume occupancy on the activity coefficient of small ions and other solutes is small. Figure 2 attempts to illustrate the difference between the uncrowded environment in which GroEL functions in vitro and the crowded environment in which it functions in vivo.

There are two main effects of macromolecular crowding. First, the more crowded the environment the larger the association constants of interacting macromolecules; macromolecular association constants for proteins in E. coli are predicted to exceed those in dilute solution by several orders of magnitude (Zimmerman and Minton, 1993). Second, crowding will reduce the diffusion rates of macromolecules; the diffusion coefficient decreases about 10-fold as the total protein concentration increases to 300 g/liter. Why are these considerations relevant to models for chaperonin action?

In vitro

Fig. 2. Macromolecular crowding or the airport terminal effect (after Goodsell, 1991). Each square represents the face of a cube with an edge 100 nm in length. The left-hand square (in vitro) illustrates the 1.5 fiM concentration of GroEL, similar to that used in many protein refolding experiments; there is one GroEL oligomer per cube (labeled G and approximately to scale). The right-hand square (in vivo) illustrates the 3 ¡xM concentration of GroEL in the cytoplasm of E. coli; there are about two oligomers per cube, and the whole cytoplasm consists of about 600 such cubes. The other structures in the right-hand square represent the sizes and concentrations of the macromolecular components in the cytoplasm that create crowding by occupying so much space; thus each cube contains in addition to two GroEL molecules (G) about 30 ribosomes, 340 tRNA molecules, and 500 other protein molecules. The right-hand square is reprinted with permission from Goodsell (1991).

A study of Chapters 7-9 will reveal that there are two main classes of model proposed for the ability of GroEL/ES to increase the yield of correctly refolded protein after dilution from dénaturant. Both models propose that binding to GroEL reduces the probability of aggregation by segregating individual partially folded protein chains inside the central cavity of GroEL, and may also alter some noncovalent interactions that hold the chain in a kinetically trapped conformation, but they differ in their view of subsequent events. According to one view, ATP hydrolysis releases the bound chain into the free solution where it can fold spontaneously; if it fails to bury all its hydrophobic residues before it encounters another GroEL molecule, it will bind again, and the process will be repeated until folding is complete. The problem with this model is that it does not offer any protection against aggregation when the chain is

Fig. 2. Macromolecular crowding or the airport terminal effect (after Goodsell, 1991). Each square represents the face of a cube with an edge 100 nm in length. The left-hand square (in vitro) illustrates the 1.5 fiM concentration of GroEL, similar to that used in many protein refolding experiments; there is one GroEL oligomer per cube (labeled G and approximately to scale). The right-hand square (in vivo) illustrates the 3 ¡xM concentration of GroEL in the cytoplasm of E. coli; there are about two oligomers per cube, and the whole cytoplasm consists of about 600 such cubes. The other structures in the right-hand square represent the sizes and concentrations of the macromolecular components in the cytoplasm that create crowding by occupying so much space; thus each cube contains in addition to two GroEL molecules (G) about 30 ribosomes, 340 tRNA molecules, and 500 other protein molecules. The right-hand square is reprinted with permission from Goodsell (1991).

In vivo jumping between GroEL oligomers, but we know that aggregation is a real problem in vivo (Horwich et al, 1993). The alternative model proposes that ATP hydrolysis releases the chain into the internal cavity of GroEL where it can internalize its hydrophobic groups in a sequestered environment and so avoid aggregation; this type of model (Martin et al., 1993) is called the Anfinsen cage model by this author to emphasize the point that folding as observed in an Anfinsen refolding experiment is postulated to initiate inside the GroEL cavity (Ellis, 1994). The problem with this model is that once the chain unbinds from GroEL it is free to diffuse away into the free solution unless it is constrained in some way, such as being located in the cavity under the GroES component. The diffusion rate in vitro is fast compared to the rate of folding, but in vivo diffusion is reduced by macromolecular crowding, and folding need not proceed to completion, but only to the point where the danger of aggregation has passed. So one basic unresolved question is whether the sole action of GroEL is to bind sticky folding intermediates and so reduce self-aggregation, or whether, in addition, GroEL increases the time for internalizing the sticky sites by a sequestration action. It can be seen that time is of the essence in distinguishing between these possibilities.

It can be calculated that a compact partially folded polypeptide chain 50 kDa in size takes 0.5 ms to diffuse 100 nm, which is the average distance between GroEL oligomers in a typical protein refolding experiment (see Fig. 2). This is a much shorter time than it takes the average protein to refold, and even allowing for the probability that binding is not diffusion-limited and that several transfers could take place, these values suggest that proteins will not have time to fold appreciably while jumping between GroEL oligomers, as one study suggests (Weismann et al., 1994). In the crowded environment found in vivo, diffusion will be slowed by 10-fold, but a more important difference is that the association of both GroES and the partially folded chain with GroEL will be greatly favored. Both these factors suggest that in vivo the folding chain will complete its folding in close association with the same GroEL/ES complex to which it first binds.

It is possible to create a highly crowded environment in vitro by adding large amounts of certain high-molecular-weight synthetic polymers (Zimmermann and Minton, 1993). Use of such polymers with GroEL shows that the crowding restricts the jumping of folding intermediates between GroEL oligomers observed in their absence (F-U. Hartl, personal communication, 1995). Future research will increasingly use such methods to try to elucidate the extraordinary ability of the chaperonins to assist protein folding in the intact cell.

1. Chaperonins: Introductory Perspective ACKNOWLEDGMENT

I express my thanks to F-U. Hartl for stimulating discussion.

Was this article helpful?

0 0

Post a comment