The attempt to eliminate aberrant proteins in cells of patients with inherited genetic disorders is only one important function of a more general PQC that works at all biological levels, from the whole organism to the integrity of the genetic material. Here we are concerned with the mechanisms surveying the structure and function of individual proteins, and we refer to the preceding detailed discussion in the previous section about the chaperone and protease components of these systems.
In the context of misfolding or conformational diseases, the interesting mechanisms are those that guarantee the conformational quality of newly synthesized proteins and supervise the structural integrity of old proteins as well eliminate misfolded and damaged proteins. Thus, the PQC systems first have the function to detect proteins that cannot attain functional folds; second, they should try to refold unfolded and damaged proteins; and third, if it is not possible to attain or recover the functional conformation (due, for instance, to sequence variations or oxidative modification) the systems should be able to eliminate the aberrant protein (Fig. 13.1-8).
Principle of an efficient cytosolic protein quality control system. Newly synthesized polypeptides are protected from aggregation by chaperones, mainly Hsp70 and Hsp90, before embarking on folding--either without help or assisted by Hsp70 and, in some cases, additionally by the chaperonin TRiC (the cytosolic Hsp60 analogue). If the polypeptides carry genetic variations, the smooth folding may be compromised, and all or variable amounts of the protein are captured by the proteolytic systems, especially the proteasome. The degradation pathway is also activated in case of modification or damage to the protein (indicated by the black star in the mature protein). TRiC, TCP-1 ring complex.
Quality Control of Newly Synthesized Proteins
The trafficking in the cytosol and the membrane translocation of newly synthesized proteins is in all cases assisted by chaperones. Most large, multidomain proteins need further assistance from the chaperones and--in the case of mitochondrial or cytosolic proteins--also need the chaperonins to fold properly. If processing of the nascent polypeptide chain takes place in the ER, additional assistance from lectin-chaperones may be needed. The binding and release of the chaperones/chaperonins are highly ordered and have energy-demanding processes that require ATP binding and hydrolysis according to a defined time frame. The last step in the processing may be assembling of folded monomers in homo- or heteromeric complexes, which in many cases is also assisted by chaperones, many of which are specific to the particular complex.
As discussed in the previous section, each protein molecule may go through several rounds of binding to and release from one or more chaperones before it can achieve the proper functional structure. Cytosolic and mitochondrial proteins fold in their respective compartments, while nuclear and peroxisomal proteins fold in the cytosol before transport to their respective compartments. Many membrane receptors and secretory proteins as well as lysosomal proteins are processed though the ER, where they become transport-competent. Transport competence is the imperative quality criterion before processing can proceed (Fig. 13.1-9).
Chaperone-mediated trafficking and folding (quality control) of cellular proteins. All nascent polypeptides are protected by cytosolic chaperones, notably Hsp70 and/or Hsp90. Cytosol-residing proteins use the cytosolic chaperone/protease systems, as depicted in Fig. 8. Nuclear encoded mitochondria matrix proteins containing mitochondria targeting signal peptides are translocated into the lumen, assisted by mitochondrial Hsp70 (mHsp70) and chaperoned in the folding by the chaperonin Hsp60. The turnover and elimination of misfolded proteins, in this case medium-chain acyl-CoA dehydrogenase (MCAD), is accomplished by proteases, notably ClpP and Lon. Polypeptides that reside in the endoplasmic reticulum (ER) or are processed for secretion, translocation to the lysozymes [as aspartylglutaminidase (AGA)], or membrane insertion (as the LDL receptor), are cotranslationally translocated to the ER lumen, assisted by the ER Hsp70 homologue GRP78. Depending on the amount and site of glucosylation, the chaperones that further assist in the folding are GRP78 or one or more of the lectin-chaperones, calnexin and calreticulin. Turnover and premature degradation of ER-residing or -processed proteins is accomplished--after retrotranslocation to the cytosol--by cytosolic proteases, particularly the proteasome. Both peroxisomal and nuclear-residing proteins seem to be folded in the cytosol and transported to their respective organelles by special transport proteins. Some of the peroxin (PEX) types of peroxisomal proteins, together with Hsp70, help the translocation--in this case--of acyl-CoA oxidase (ACOX). For nuclear proteins, a number of helper proteins, such as Hsp90 and importin, are also involved, respectively, in the holding and translocation of p53, for example.
If the protein is not able to attain the functional structure or transport competence within a reasonable period of time, it is--as a rule--degraded by one of the intracellular proteolytic mechanisms. The failure to degrade an aberrant protein may give rise to toxic conformations, which may inactivate wild-type partners (negative dominance) and/or form aggregates, which will be discussed in detail below.
The PQC systems of newly synthesized proteins may thus be understood as consisting of two parts: one that supervises folding and protects folding intermediates from aggregation, and one that eliminates aberrant folding intermediates and unstable proteins that would occupy (overload) the chaperone systems and damage the cell by their tendency to aggregate. The critical process is thus to sort out the proteins with a decreased ability to fold or an increased tendency toward misfolding and aggregation. The longer the folding process, the higher the probability that the protein will be captured by the proteolytic systems. The PQC systems seem to function stochastically, and the competition between the kinetics of folding and binding to proteases determines the extent to which the protein is folded or degraded.
Little is known about the balance between folding and degradation for newly synthesized proteins in general. Although this balance, as a measure of the folding efficiency, has--to our knowledge--been determined only for the CFTR protein and the δ-opioid receptor (Gelman, Kopito, 91; Petaja-Repo et al, 187; Ward, Kopito, 258), it may be quite different for other native structures. For the CFTR and δ-opioid receptor it has been estimated that 40% or more of newly synthesized protein fails to attain a transport competent conformation, and is degraded rapidly.
Despite the fact that the folding efficiency has not been determined for the fatty acid oxidation enzymes MCAD and SCAD, it is noteworthy that the duration of contact time between the chaperonin Hsp60 and the two proteins is rather different. The MCAD monomer folds easily and is transferred to the tetrameric structure within a few minutes in vitro in isolated eukaryotic mitochondria, whereas the SCAD protein is complexed with Hsp60 for longer times and slowly attains the tetrameric structure (Pedersen et al, 183).
Based on these few examples, it is suggested that the effect of missense sequence variations varies greatly within the same protein and that the effect on various proteins may be quite different. Without going into detail, this conclusion is compatible with the knowledge about perturbation of protein folds (Denton et al, 58) and with the concept that folding is guided by important interactions between specific amino acids (Chiti et al, 42; Dobson, 65; Dobson, 66; Lindorff-Larsen et al, 145; Vendruscolo et al, 250). Furthermore, it has been estimated that the free energy difference between the unstructured nascent polypeptide and the stable structure is only about 10 kcal/mol, and that the destabilization by a single amino acid alteration may be 1-2 kcal/mol, which may elevate the global minimum and increase the probability of kinetic trapping in nonnative folds, which may be prone to form aggregates (Perutz et al, 186). Unproductive folding of missense variant proteins can thus be considered a rule rather than an exception. This notion is strengthened by the estimation that only about 1000 stable protein folds exist in nature (Denton et al, 58); all other putative protein folds are unstable, meaning that polypeptides with stretches of amino acids that cannot find an energy minimum and that represents a stable fold within a reasonable time will be sorted to degradation. However, the folding chaperones of the PQC systems may counteract this, simply by catching the unstable conformations and giving them another try. Depending on the severity of the sequence alteration and the conformational space available, the efficiency of the PQC systems is decisive for the sorting of proteins between degradation, aggregation, or folding to a nonfunctional or functional variant protein, which under certain stress conditions may come to clinical attention (see below). It is no wonder that it is impossible with the current knowledge to predict the effect of a given sequence variation.
In this context it is interesting to bring into the discussion a series of investigations concerning errors in newly synthesized polypeptides and the cell-cleaning property of the PQC systems in vertebrates. In 1996 it was proposed that the ribosomal translation was error prone and that the cell produced so-called defective ribosomal products (DRiPs), which should be the precursors of self-proteins presented on MHC molecules of the immune system (Yewdell et al, 276). In 2000 the same research group (Schubert et al, 214) performed another study showing that DRiP production may account for 30% of newly synthesized proteins. The results indicated further that they were rapidly degraded by the proteasome. The authors did not go into detail with regard to the type of protein products. They showed that a significant amount of accumulated proteins following a 60 minute radioactive pulse had molecular masses of 46,000-250,000 and that a substantial fraction of these polypeptides were ubiquitinated. Furthermore, they showed that the amounts of accumulated protein, including ubiquitinated proteins, were increased in cells treated with proteasome inhibitors. Notably, the largest amount of proteins was recovered in an insoluble fraction, which was obtained by freeze-thawing followed by centrifugation, indicating that within the time window of the 60-minute chase, the polypeptides--or at least some of them--aggregated before they were degraded.
Although the implications in vivo of the formation of DRiPs are not exactly known, continuous production of truncated and missense variant proteins, which cannot attain a stable fold, must interact with the PQC systems that try to degrade them. The proteasome is involved in this process (Fuertes et al, 87), producing small peptides, which are suggested to be substrates for the MHC (Reits et al, 200; Schild, Rammensee, 211). Whether a small fraction of the aberrant proteins are able to obtain stable conformations, possibly with assistance from chaperones, is not known, but--if so--it may contribute to the somatic mutational load, which has been proposed to be important for the development of age-dependent diseases (Andreassi, 6).
Such studies demonstrate the importance and power of the PQC systems as cell-cleaning systems, and they also substantiate the contention/fact that most proteins carrying sequence alterations, including those resulting in amino acid changes, are--as a rule--rapidly degraded by the PQC systems. The exceptions are inherited sequence variations creating stable truncated missense protein variants as well as stable proteins carrying amino acid substitutions and small amino acid deletions or insertions. These aberrant proteins slip through the PQC systems and, as a consequence, may give rise to toxic conformers (Fig. 13.1-10).
Schematic events leading to different types of conformational diseases. Normally most proteins will smoothly attain the stable active conformation, or--if they miss the stable conformation--will be rapidly be degraded. Certain proteins--especially proteins carrying variations or modifications--are prone to aggregation, and the protein may be trapped in semistable conformations. Depending on the protein and the aberration, the protein may form toxic conformations, stable cell-toxic aggregates, or amorphous aggregates, which possess varying degrees of degradability.
In cases where the aberrant proteins exert negative dominance, the conformation is often only slightly different from that of the wild-type, but it is sufficiently similar to participate in a multimeric complex with wild-type molecules, and impair function.
The mechanisms that create "aggregation diseases" may be different. In these cases the protein often possesses an intrinsic propensity to stabilize in predominantly β-sheet conformations (see below), which are enhanced by inherited (or acquired) sequence variations. The β-sheet conformations are formed rapidly before the PQC systems can eliminate the precursors and--when formed--are resistant to proteolytic degradation. The properties of such proteins, which are the pathogenetic substances in early-onset amyloidosis and many neurodegenerative diseases, will be discussed in more detail below.
The situation is more complex in the many recessive diseases, where rapid degradation has been indicated to be the pathogenetic mechanism. As explained above, many aberrant proteins will probably be degraded rapidly, perhaps following a period of prolonged interaction with chaperones and possibly proteases, as has been seen or implicated for some variant MCAD and SCAD proteins (Pedersen et al, 183; Saijo et al, 208) as well as for PAH mutant proteins (Gamez et al, 88; Pey et al, 189; Waters, 262) and variant LDL receptors (Jorgensen et al, 123; Jørgensen et al, 122).
The Δ-F508 CFTR variant protein may, under certain nonstressful cell conditions, show the same behavior (Gelman, Kopito, 91). However, under cell stress--such as proteasome inhibition--and when overexpressed, the variant (as well as wild-type) CFTR protein may form aggregates; the protein assembles in so-called aggresomes (see below). The aggregation is preceded by prolonged interaction with chaperones, and the aggregates (aggresomes) contain chaperones, proteases (proteasomes), and other vesicular components (Garcia-Mata et al, 89; Johnston et al, 121; Yang et al, 274). α1-antitrypsin deficiency, in the form of the z-variant, is another case of an autosomal recessive disorder where protein aggregates accumulate temporarily and where in certain special genetic cases, the aggregate persists long enough to exert cell damage (Crowther et al, 48; Perlmutter, 185).
The extent to which cells from patients with other autosomal recessive disorders accumulate misfolded proteins, as either benign temporary or cell-toxic aggregates, is not presently known, but results from studies of a number of SCAD variant proteins could indicate that accumulation is the case more often than was previously thought (Pedersen et al, 183).
All these examples show that aberrant proteins sometimes slip through the PQC system, and they demonstrate that the efficiency of the systems is not total. As will be discussed below, the efficiency depends on the amounts of chaperone and protease components available in the PQC systems as well as on several other genetic and cellular factors. The inefficiency may indicate an unfortunate evolutionary development or--more probably--may propose a flexibility in the systems, with evolutionary significance. This is suggested by studies that indicate that the PQC systems also possess evolutionary roles by adapting to changing environments (Feder, Hofmann, 79) and internal stress, such as inbreeding (Kristensen et al, 135; Sorensen et al, 226), and by the fact that the systems--or rather some of the chaperone components--possess the ability to accumulate and buffer genetic variation, which under stress and changing environmental conditions, may contribute to the driving force of evolution (Queitsch et al, 197; Rutherford, 205; Rutherford, Lindquist, 206).
In conclusion, the PQC systems supervise the proper folding of nascent polypeptides, but the systems are not absolutely efficient, leaving some proteins with an aberrant protein fold to exert evolutionary fortunate effects or unfortunate disease-causing effects. It has also been hinted that the efficiencies of the systems are not invariable factors in the cell’s ability to produce correctly folded active proteins--i.e., that the efficiency varies with the properties of the particular protein and with cell conditions.
Maintenance Quality Control of Mature and Modified Proteins
In addition to the functions related to newly synthesized polypeptides, the PQC systems exert important functions in maintenance of the structure and function of mature proteins. Although it is not possible to distinguish functionally between the roles of PQC systems in the processing of nascent proteins and the function of the systems during the lifetime of proteins, it may be of conceptual value. In spite of the fact that factors such as modification of proteins during cellular stress and ageing are of some significance in relation to folding and stability of newly synthesized proteins, such factors--together with intrinsic properties of the proteins--become of decisive importance in the pathogenesis of some of the conformational diseases, such as amyloidosis, and the neurodegenerative Alzheimer disease, Parkinson disease, and polyglutamine diseases. We will therefore discuss intrinsic propensity to spontaneous conformational changes of some proteins and stress-dependent protein modifications in some detail below.
After trafficking and processing, most cellular proteins exist in a conformation, which is a balanced mixture of stable folds of α-helices, β-sheets, and unstructured turns. It is unknown to what degree such structures are flexible, but studies of molecular softness, e.g., with neutron scattering (Zaccai, 279) and deuterium exchange methods (Luque et al, 150), have corroborated the conventional wisdom that conformational flexibility is essential for the function of many proteins. Some of these, together with some that do not possess a fixed structure--at least not before they engage in a biological function--but reside in a so-called natively unfolded conformation (Uversky, 247), possess an intrinsic propensity to form stable β-sheet structures, which are prone to aggregation and protease resistance. More than 20 naturally occurring unrelated native proteins may have this propensity (Ellis, Pinheiro, 75); they probably represent the extreme of a continuum. Actually, all proteins will sooner or later be trapped under adverse conditions (low pH, high temperature, high salt, macromolecular crowding) in conformations that promote the formation of aggregates (Ellis, 74; Minton, 165; Perutz et al, 186; Takase, 237). The ease by which the aggregation process proceeds and the nature of the aggregate can, however, be quite different (Dobson, 66; Minton, 165). Properties of importance in this context may include free energy differences as well as activation energy barriers between a large number of conformations within the total conformational space of the given protein, which may define a certain flexibility or softness of the protein. The protein may jump between conformational states separated by activation energies, which are comparable to the energy of thermal movements (Zaccai, 279). Among the possible conformational states are β-sheet structures, which contain stretches of amino acids possessing special interactive properties (Aggeli et al, 2; Aggeli et al, 3) and which promote the formation of stable aggregates/fibrils. The mechanism by which the β-sheeted molecules form the fibrils is still debated, but a seeding mechanism is probable (Michalik, Van, 163; Scherzinger et al, 210; Selkoe, 217; Soto, 227). The process is initiated by a stochastic conformational change to an unfolded state with β-sheet propensity, followed by oligomerization and, eventually, further formation of polymers and aggregates by seeding. This simple mechanism is attractive because it is compatible with the hypothesis that the PQC systems in the healthy and young cell eliminate the molecules with nonnative conformations before they can embark in oligomer assembly, and thereby prevent the oligomers and aggregate formation (Stefani, Dobson, 66).
Again, although the general concept that special interactive amino acids with high β-sheet formation propensity are necessary for the formation of stable aggregates and fibrils, the diversity of mechanisms by which the aggregates form is large (Milewski et al, 164; Perutz et al, 186; Uversky, Fink, 248). Furthermore, there may be interactions from other parts of the molecule that modify the aggregation tendency (Perlmutter, 185). Although amino acids 90-112 in the prion protein are essential for the transformation of its soluble to toxic form, sequence alteration in other parts of the gene enhances the transformation. The reason is probably not a lowering of the β-sheet energy, but rather an elevation of the global minimum of the mutant proteins in general, which may then enhance the rate of transformation (Huang et al, 111; Prusiner et al, 193; Riek et al, 202). Furthermore, Chiti and colleagues (41) have shown that mutational changes in simple physicochemical properties like hydrophobicity, charge, and secondary structure propensity correlate with changes in rates of aggregation of the unfolded peptides.
The message here is that all proteins--some more than others--are flexible and vulnerable to intramolecular (internal) perturbations, e.g., amino acid substitutions and other sequence variations, as well as external (extramolecular) stresses, such as high temperature or oxidative stress, may generate modified proteins (see below); cellular PQC systems may try to eliminate the consequences of perturbations and stresses. In addition, the same conceptual framework is applicable to our understanding of acquired, age-related and many inherited conformational disorders.
Thus, specific amino acids can be decisive for the susceptibility of a given protein to aggregation and formating fibrils. The effect cannot be predicted, but when the protein loses control of the native folds, the tendency toward aggregation increases drastically (Perutz et al, 186), regardless of the reason for the aberration. Variant proteins, even if they have slipped through the folding quality control, may therefore aggregate under stress, but whether the misfolded conformations concentrate in fibril-forming β-sheets or in temporary amorphous aggregates, they must depend on the actual amino acid substitution as well as on the efficiency of the PQC systems.
Before discussing the efficiency of the PQC systems, it is necessary to introduce yet another of their functions: the attempt to eliminate proteins that have been damaged, for example by carbonylation due to oxidative stress. This cleaning function or damage control is rather effective in young and healthy cells, but in cells containing protein variants and in aged cells (where the functional efficiency of the PQC systems decreases, perhaps because of overload), the modified proteins will accumulate (Nardai et al, 170; Verbeke et al, 251).
Modification of Misfolded Proteins
During their lifetime, all proteins, particularly during stress, experience posttranslational events that may damage their functional structure. Proteins with a long lifetime and those with a loose and flexible structure are especially vulnerable. Relevant for misfolding or conformational diseases are modifications exerted by reactive carbonyl species (RCS), reactive oxygen species (ROS), and reactive nitrogen species (RNS), all of which are known to damage proteins (as well as lipids and DNA) (for references see Dalle-Donne et al, 54; Hayashi et al, 106; Roberts et al, 203). Modifications of proteins by these species have been associated with the development of a variety of diseases, notably those where misfolding of proteins has been proved or indicated to be pathogenetic, such as atherosclerosis, diabetes, Alzheimer disease, Parkinson disease, and the collagen diseases (Butterfield, 27; Butterfield, Kanski, 28; Castegna et al, 33; Ischiropoulos, 116; McCarthy et al, 158; Stadtman, 228; Ursini et al, 246; Xu, 272; Ye et al, 275).
The extent to which oxidative modification has any biological significance in the development of the pathologies, especially in the brain, in other types of conformational diseases, such as cystic fibrosis, α1-antitrypsin deficiency, PKU, and the fatty acid oxidation defects, is not known. However, since chaperone complexed or "free" misfolded proteins have been detected/indicated in all these disorders (Gamez et al, 88; Gelman, Kopito, 91; Gregersen et al, 96; Pedersen et al, 183; Perlmutter, 185), it is of interest to note that aberrant proteins are more susceptible than native proteins to oxidative modification (Butterfield, Kanski, 28; Nystrom, 176). Without increase in the oxidative load (Dukan et al, 70) and after introducing defects in a small fraction of cellular proteins in E. coli by small amounts of streptomycin, Dukan and colleagues measured the levels of oxidative defense proteins (superoxide dismutase, catalase, and increases in disulfide bond formation) as well as carbonylated proteins and GroEL. Although the level of oxidative defense proteins remained constant, carbonylated proteins and GroEL increase in amounts in the presence of streptomycin, indicating that misfolding, which induced the production of the chaperonin, exposed the proteins to oxidative modifications. If these results can be translated to human cells they could indicate, even in situations where oxidative stress is not applied, that cells from patients with conformational diseases other than the traditional aggregation disorders may accumulate oxidatively modified proteins, either in chaperone complexes or as soluble polymers, thus contributing to the pathology. If this is true, there must exist a continuum from diseases in which accumulation of the aberrant protein does not occur (because the protein is rapidly degraded in all circumstances); to those disorders where the mutant proteins accumulate only during certain cell conditions; and ultimately to the classic diseases in which the effects of the accumulated aberrant protein are decisive for the pathology.
Within this concept, the efficiency of the PQC systems will be of decisive importance for the relationship between genotype and disease expression at the cellular level--and probably at the clinical level as well. In the predominantly loss-of-function diseases, constitutively efficient and induced PQC systems will aggravate the phenotype by eliminating possible amounts of functional mutant protein, while they may alleviate the consequences of protein oligomerization/aggregation by the attempt to dissolve, refold, or degrade the aggregates or the misfolded precursors.
It is therefore relevant to discuss in general the consequences of varying efficiencies of the PQC systems as well as the consequences for the genotype-phenotype relationship. The aim is to present a unifying view, the overall structure of which can be applied in discussions of conformational diseases in general.
Efficiency of the PQC Systems; up- and Downregulation; Genotype-Phenotype
The efficiency of the PQC systems may be one of the determining factors in the pathogenesis of conformational diseases, substantiating the emerging paradigm that only a minority of genetic disorders can be explained by a monogenic molecular pathogenesis; rather, the molecular etiology is complex and influenced by sequence variations in modifier genes as well as by cellular and environmental factors (Badano, Katsanis, 13; Dipple, McCabe, 2000a; Dipple, McCabe, 2000b; Nadeau, 168; Nadeau, 169; Sakahira et al, 209; Scriver, Waters, 216; Slavotinek, Biesecker, 222). Although there exists a few examples in which genetic variations in components of the PQC systems are associated with disease (see Protein Folding, above), but a single disease, α1-antitrypsin deficiency, in which sequence variations in genes involved in the degradation of accumulated ER proteins have been indicated (Wu et al, 269), no published cases exist where such modifications have been documented in genetic diseases in humans. Therefore, we will concentrate on the variation in the efficiencies of the PQC systems affected by cellular and environmental factors, as it has been indicated by genotype-phenotype studies, often in combination with in vitro experiments, and by experiments involving overexpression and downregulation of components of the PQC systems. This focus on the PQC systems is rationalized by the statement "chaperones act between the genotype and phenotype to affect the expression of genetic variation" (Rutherford, 205). This statement applies to the role of chaperones in masking genetic variation as an evolutionary force, but it can easily be translated to the concept of genetic disorders, which in general are not part of a possible positive evolution, but rather represent nature’s unfortunate mistakes, which were not eliminated before they came to expression. The alternative concept could be stated as follows: the PQC systems act between the genotype and phenotype to affect the expression of disease-associated missense sequence variation.
Figures 13.1-11, 13.1-12, and 13.1-13 illustrate paradigms for the prototypes of conformational diseases: by loss-of-function or by gain-of-function, the latter involving a dominant negative aggregation scenario.
Simplified genotype-phenotype relationship in many loss-of-function diseases. Sequence variations may be mild or severe with respect to the intrinsic folding ability of the variant protein, resulting in varying degrees of degradation of the misfolded conformations. The effect of varying efficiencies of the protein quality control (PQC) systems, e.g., due to elevation of the temperature, is represented by the shift from high residual to low residual protein function. The clinical phenotype may further be influenced by non-PQC system factors, such as genetic modifiers, cellular stress factors, or metabolic flux, all of which contribute to the penetrance and clinical expression of the disease.
Simplified hypothetical genotype-phenotype relationship in diseases with negative dominant pathogenesis. Severe sequence variations are those that are not degraded, while mild variations may be degraded to a varying degree, according to the type and location of the variation, as well as to the efficiency of the protein quality control (PQC) systems. Consequently, varying amounts of full-length toxic protein, which may exert negative dominance, may be present in the cell. In addition, other factors, such as genetic modifiers or cellular stress, may determine the penetrance and clinical expression of the disease.
Simplified genotype-phenotype relationship in many "aggregation" diseases. Some intracellular wild-type proteins possess an intrinsic aggregation tendency; others require sequence variations to acquire such a tendency. Depending on the type and location of the particular sequence variation, the tendency to aggregate may be high or low. At the first level, the efficiency of the protein quality control (PQC) systems may determine the amount of aggregate (or toxic preaggregate). In addition, genetic modifiers as well as cellular stress may be determinants in the development of clinical symptoms.
Diseases with Loss-of-Function Pathogenesis
Figure 13.1-11 shows the influence of the PQC systems on the biochemical phenotype in typical loss-of-function disorders. The upper bar illustrates the situation where missense sequence variations may affect the folding landscape differently, creating local minima, which may trap the folding intermediates. "Mild" sequence variations display intrinsic folding abilities, which in the normal cell under normal conditions are rather similar to wild-type. "Severe" variations represent missense variant proteins, which are not able to fold properly and are trapped early in the landscape. They may be compared with null mutations, e.g., PTCs, which result in elimination of the mRNA and do not produce any functional protein. In between these extremes, all degrees of severity from sequence variation may exist. Although the "true" intrinsic foldability (and phenotype severity) of a given sequence variation cannot be assessed without studying the folding in vitro--or at some time in the future by in silico computation of the folding pathway--a large number of cellular model studies as well as genotype-phenotype investigations have shown that such a continuum of severity exists. Earlier in this chapter, we discussed some of these investigations with regard to PKU (Gamez et al, 88; Pey et al, 189; Waters, 262) and the fatty acid oxidation defects (Andresen et al, 7; Andresen, Olpin et al, 8; Bross et al, 24; Bross et al, 23; Gregersen et al, 96; Jensen et al, 119; Pedersen et al, 183). From these studies, it has become clear that some missense variant proteins completely abolish enzyme activity as well as detectable protein, while others do show residual activity in detectable amounts of protein. Interestingly, we and others have identified a number of variant proteins, e.g., certain PAH and SCAD variants, where there does not exist a simple correlation between the amount of cellular protein and the actual function of the protein. Possible explanations for this phenomenon have been discussed in detail elsewhere (Gregersen et al, 2001b), where it was suggested that--at least concerning the soluble part of the total SCAD and PAH variant proteins--the proteins not found in the fully folded and assembled conformations are bound to chaperones and therefore are inactive. This hypothesis has been corroborated by transcription/translation studies in vitro of a number of SCAD variant proteins (Pedersen et al, 183), which showed that the variant proteins can be detected in complex with Hsp60 as well as in tetrameric complexes, which represent the native and active variant SCAD. However, the distribution between the Hsp60-bound and tetrameric SCAD was not invariable for a given variant protein. By changing the temperature, it was possible to change the distribution; at low temperature the amounts of tetrameric enzyme produced were much higher than the production at higher temperature (Pedersen et al, 183). A number of other studies indicate the same for PKU (Gamez et al, 88) and for cystic fibrosis (Gelman, Kopito, 91; Qu et al, 196). This phenomenon suggests that the efficiency of the folding and of the PQC systems’ ability to guide the proteins through the folding landscape varies with temperature. Efficiency of the folding process can be enhanced by overexpression of chaperones (Bross et al, 22; Corydon et al, 45; Farinha et al, 78). Cellular conditions, including the amount of PQC system components, may thus be decisive for the molecular phenotype or--in the case of enzyme deficiencies--for the enzymatic phenotype.
These results are the reason why the second bar in Fig. 13.1-11 represents the functional amounts of the variant proteins. In this model, we have not considered the amounts of proteins that are bound to chaperones nor the possible aggregated forms. In the liver phenotype form of α1-antitrypsin deficiency, aggregated variant protein is abundant and decisive for the phenotype; in the predominantly pulmonary form, which is considered here, aggregates--if they occur--are degraded rapidly and do not contribute significantly to the pathology. The liver form will be discussed below. In cystic fibrosis, the most common mutant CFTR protein containing the Δ-508 amino acid deletion is prone to aggregation and does aggregate when overexpressed in vitro or under cellular stress, where a stress response has been shown (Gelman, Kopito, 91). The significance of this is still unclear in vivo. For both PAH and SCAD, aggregates of some of the disease-associated variant proteins are found in vitro (Gamez et al, 88; Pedersen et al, 183; Pey et al, 189; Waters, 262). It is not yet known whether such aggregates contribute to the pathology of these diseases.
The relationship between the upper and middle bar in Fig. 13.1-11 recognizes that the functional amounts of variant proteins are dependent on cellular conditions such as elevated temperature; this influences the efficiency of the PQC systems with regard to the particular protein and sequence variation. The ability of the PQC systems to eliminate a particular aberrant protein is also determined by the amounts of, and balance between, the components of the PQC systems.
The lower bar and the arrows between the middle and lower bars in Fig. 13.1-11 may not be representative of all loss-of-function diseases. They may generally represent the influence from other--non-PQC system--modifier genes as well as other internal and external conditions. With regard to inborn errors of intermediate metabolism, they illustrate that the activity through the particular metabolic pathway may have decisive importance for whether the functional deficiency results in clinical disease. It thus emphasizes that even sequence variations that abolish the function completely may be not be expressed clinically under normal conditions, but come to clinical attention only during metabolic stress, such as high fatty acid oxidation rate during fasting. The rationale for the fact that many patients with fatty acid oxidation deficiencies can remain asymptomatic until they experience a fever, which is often accompanied with fasting, is thus twofold. First, many of the sequence variations encountered in patients with these disorders code for temperature-sensitive variant proteins, which become eliminated from the pool of functional conformations at higher temperatures, while they possess some residual activity at lower--physiologic--temperatures (the upper part of Fig. 13.1-11). Second, at the same time when the activity is low, the demand for the protein is enhanced, and the deficiency becomes apparent (the lower part of Fig. 13.1-11). In these types of diseases, it is predominantly product deficiency and substrate accumulation that become pathogenic. An unfortunate paradox is that higher PQC system efficiency, which in the healthy individual acts to clean the cells of aberrant protein for the good of the individual, can "unmask" a "mild" variation as pathogenic under stress conditions.
In essence this situation for the individual is not different from the situation at the population level when polymorphisms that have been "masked" by buffering with chaperones develop new expression patterns under certain environmental stress conditions. In most cases these patterns are deleterious, even more than in the inherited diseases in which the individuals survive, but in certain fortunate cases they become the starting point of more adaptive populations (Queitsch et al, 197; Rutherford, 205).
Diseases Involving Dominant Negative Pathogenesis
In these types of diseases, the situation is quite different, but also more hypothetical since substantial documentation is lacking, because the pathogenic mechanism is determined by the amount of aberrant protein that slips through the PQC systems, and not by the amount that is eliminated. The hypothetical situation is shown in Fig. 13.1-12.
The upper bar in Fig. 13.1-12 represents the degradability of the variant protein in question. Depending on the efficiency of the PQC systems and the actual sequence variation in a specific gene, the balance between stabilized aberrant toxic conformations and degraded protein (lower bar in Fig. 13.1-12), as well as aggregated protein, may vary considerably; the pathologic consequences may likewise vary drastically from negative dominance to recessive or dominant inheritance because of aggregation. A single case from model studies could illustrate the situation (Driscoll, Gerstbrein, 68). The mec-4 gene in Caenorhabditis elegans encodes a subunit of a sodium channel that, when carrying an amino acid with a large side-chain instead of a small conserved residue, becomes hyperactive and toxic. Interestingly, the neurotoxicity exerted by the stable variant subunit is alleviated in a null-allelic C. elegans mutant of calreticulin. Although other effects, probably on calcium homeostasis, are important for the change, it is indicated that folding/assembly of the toxic subunit contributes to the neurotoxicity. Given current knowledge, the aberrant subunit probably exerts a dominant negative effect, which is transformed to a recessive milder pathogenesis by degradation. To our knowledge, no report exists on human diseases where similar effects have been indicated; however, current knowledge about the mechanism of protein quality control suggests that cases will emerge in the near future.
Diseases with Aggregation Pathology
We are aware that by constructing the conceptual models in Figs. 13.1-11, 13.1-12, and 13.1-13, we simplify the biology. This is especially true for the model in 13.1-13, where our concept of pathogenesis and pathology in the aggregation type of conformational diseases is illustrated.
In the upper bar, it is necessary to include wild-type in the continuum of amino acid substitution severity, simply because the wild-type sequences in many of the known aggregation diseases do not need a sequence variation to form insoluble aggregates.
The most important factor is susceptibility of the polypeptide, wild-type or aberrant, to form β-sheets or other structures, which may be trapped in semistable conformations with the ability to develop into insoluble aggregates. Indeed, an increasing number of studies have questioned the belief that the aggregates/fibrils as such are toxic; they may be a rescue mechanism by which the cell sequesters misfolded proteins that are too stable to be eliminated immediately (Driscoll, Gerstbrein, 68; Michalik, Van, 163; Sherman, Goldberg, 220). More specifically, two important general investigations illustrate this emerging concept (Bucciantini et al, 25; Walsh et al, 254). Bucciantini and colleagues showed that oligomeric protein assemblies formed early in the aggregation process of two non-disease-associated proteins, the SH3 domain of phosphatidyl-inositol-3′-kiniase and E. coli HypF protein, affect cell viability and exert cytotoxicity, while the fully aggregated form does not. Similarly, Walsh and colleagues microinjected Aβ-oligomer, derived from the AAP proteins implicated in Alzheimer disease, into rat brain and found that the oligomers inhibited synaptic plasticity, as measured by the long-term-potentiation (LTP), are important component for memory and learning. In contrast, monomers and aggregated Aβ-peptide did not participate in the toxicity.
Although their results are important with regard to the nature of the diseases associated with toxic molecular species producing amyloid/fibrils, neither Bucciantini and colleagues nor Walsh and colleagues directly address the extent to which soluble complexes of misfolded proteins and chaperone/protease components of the PQC systems contribute to the toxicity. Both research groups mention the importance of effective folding (Bucciantini et al, 25) and elimination of monomeric forms of aggregation-prone proteins (Walsh et al, 254). Bucciantini and colleagues argue that "toxicity could primarily arise because, on the surface of disordered aggregates, there is likely to be a combinational display of amino acids enabling these species to interact inappropriately with a wide range of cellular components." This is a central hypothesis and applies directly to the concept of this chapter. If we add the results of a large number of investigations both in vitro and in vivo showing that the aggregates related to conformational diseases are likely to contain ubiquitin, chaperones, proteases, and components of the proteasome, and a number of other cellular constituents such as transcription factors (Adachi et al, 1; Cummings et al, 51; Driscoll, Gerstbrein, 68; Fonte et al, 84; Garcia-Mata et al, 89; Ito et al, 117; Lee et al, 141; Lindersson et al, 144; Macario, Conway de, 152; Sherman, Goldberg, 220; Shinder et al, 221; Suhr et al, 233; Winklhofer et al, 266; Wood et al, 268), it is highly probable that complexes between monomeric/oligomeric aberrant proteins and the components of the PQC systems are created in the soluble phase before the aggregates have been formed. The toxicity should then be exerted either directly from the accumulated complexes or indirectly through sequestration of PQC apparatus components or other important cellular factors, such as transcription factors. The deterministic effects from these various mechanisms, which are probably much more diverse than indicated here, may be quite different in the various diseases. However, a possible contributing sequestration of PQC system components is of interest and, in this regard, represents another "sequestration model of toxicity," which is implied in the ideas of Bucciantini and colleagues, and which is discussed in relation to polyQ disease (Driscoll, Gerstbrein, 68; Hughes, 113) as well as amyotrophic lateral sclerosis (SOD1 deficiency) (Okado-Matsumoto, Fridovich, 177). The relevance here is high, since it is compatible with the chaperone sequestration model discussed earlier in this chapter. Although the model discussed by Hughes and colleagues and Driscoll and colleagues does not include other cellular factors containing polyQ stretches, it may be more general. The idea concerning polyQ (trinucleotide repeat) diseases is that the tendency of polyQ regions--such as those in huntingtin implicated in Huntington disease--to form aggregates could be extended to other cell constituents with polyQ tracts. Actually, the transcription cofactor CREB binding protein (CBP), which contains a polyQ, has been found to co-precipitate after co-overexpression of another protein with a polyQ tract (Kazantsev et al, 127; Michalik, Van, 163; Wyttenbach et al, 271). This might represent an extreme case. However, there are studies--mentioned above--that show that other cellular components, particularly chaperones and proteolytic complexes, such as the proteasome do co-precipitate with the misfolded protein and are constituent of the inclusions. The structure known as the aggresome (Kopito, Sitia, 132), an inclusion in the cytosol seen in cells containing variant CFTR protein or presenilin mutant proteins, contains chaperones, proteases, and ubiqiutinated variant protein (Cummings et al, 51; Driscoll, Gerstbrein, 68; Macario, Conway, 153; Sherman, Goldberg, 220). Likewise, an investigation by Fonte and colleagues (84), in which they identified Hsp70 and sHSPs in complex with Aβ-peptide, indicates that chaperone interaction with the aberrant Aβ-protein plays an early role in the toxic process of Alzheimer disease. The fact that Hsp70 overexpression can reverse polyQ toxicity in cell cultures and in a mouse model (Cummings et al, 52; Jana et al, 118), and can decrease the alpha-synuclein toxicity in a Drosophila Parkinson disease model--together with investigations (Driscoll, Gerstbrein, 68) that show enhanced neurotoxicity with mutations in chaperones or with inhibition of the proteasome--strengthens the concept that components of the PQC systems play a role in phenotypic expression.
The conclusion is that many missense variant proteins and some wild-type proteins under certain circumstances, which may be exacerbated during cellular stress (see below), may be trapped in semistable conformations and complex with PQC system components and other cellular factors; they may then develop into β-sheet structures or promiscuous aggregated states, to be disposed of in the degradation systems (Perlmutter, 185).
Another determinant of cell toxicity is efficiency of the PQC systems. With decreased efficiency or overloading, the probability increases that a folding intermediate or an unfolded or damaged protein will be trapped kinetically in a local conformational space minimum, and that it will be trapped as oligomers/aggregates before folding/refolding/degradation can occur. There is evidence that induction and artificial elevation of the amounts of the chaperone components of the PQC systems can alleviate the toxicity of misfolded proteins (Adachi et al, 1; Bailey et al, 14; Chai et al, 35; Chan et al, 36; Cummings et al, 51; Jana et al, 118; Krobitsch, Lindquist, 136; Lindersson et al, 144; Wyttenbach et al, 270)--and proteasome inhibitors can enhance the amount of aggregates (Chai et al, 35; Gelman et al, 90; Ito, et al, 117; Warrick et al, 259). Accordingly, efficiency of the PQC systems is decisive for the trapping of aggregation-prone structures.
The extent to which other oligomeric/aggregated structures other than the precursors to β-sheet composed fibrils/aggregates should be represented in the middle bar of Fig. 13.1-13 is not really known, but one can visualize a continuum of more or less toxic conformations, which in addition to the misfolded or damaged protein itself contains chaperones, proteases, and other cell constituents that--depending on the particular disease--contribute to the toxicity by sequestration or by damaging cellular responses. The relationship between the middle bar and the lower bar in Fig. 13.1-13 is therefore extremely complex and has been subjected to many studies and speculations. Although it is difficult to generalize the cellular pathology in the aggregation diseases, there are some mechanisms that seem to be general and probably contribute to the pathology of many conformational diseases. Collectively, these mechanisms can be referred to as stress responses, either directly affected by the internal stress associated with the misfolded protein or elicited by other cellular and environmental factors, such as aging and oxidative stress.
We will now describe cellular responses to the unfolded protein, ER overloading, mitochondrial stress, and cytosolic stress. All these stress responses are relevant to the conformational diseases and may constitute a framework to understand a particular disease.