The pathologies of many serious human diseases are thought to develop from the effects of intra‐ or extracellular aggregates of non‐native proteins. Inside cells, chaperone and protease systems regulate protein folding; however, little is known about any corresponding mechanisms that operate extracellularly. The identification of these mechanisms is important for the development of new disease therapies. This review briefly discusses the consequences of protein misfolding, the intracellular mechanisms that control folding and the potential corresponding extracellular control processes. Finally, a new speculative model is described, which proposes that newly discovered extracellular chaperones bind to exposed regions of hydrophobicity on non‐native, extracellular proteins to target them for receptor‐mediated endocytosis and intracellular, lysosomal degradation.
It has been estimated that about 400 g of protein are synthesized and degraded each day in the human body. Individual proteins are degraded at extremely varied rates, with half‐lives ranging from several minutes to many hours. Intracellularly, this variation in half‐life has been attributed to differences in the intrinsic stability of proteins and the recognition of non‐native structures by highly selective and precisely regulated protein quality‐control systems.
Protein misfolding and disease
Correct protein folding is essential for the assembly of the functional machinery required by all living organisms. Mistakes in transcription or translation that result in errors as small as an amino‐acid substitution can destabilize, and thus prevent, normal folding. In addition, environmental conditions such as macromolecular crowding, inappropriate ionic strength, oxidative stress and extremes of pH and temperature are known to promote the formation of misfolded states or slowly reacting intermediates that are trapped on the folding pathway. They also promote the partial unfolding of fully folded native proteins. Exposed hydrophobic regions on non‐native proteins bind to similar regions on nearby proteins by highly specific self‐association (Rajan et al, 2001). The aggregating species can continue to bind exposed hydrophobic regions of neighbouring proteins and thus has the potential to form various stable structures, including soluble aggregates and insoluble deposits (Fig 1). All of these protein structures have been found to be associated with disease states. For example, soluble oligomeric protein aggregates have been implicated as cytotoxic species in various neurodegenerative disorders (Demuro et al, 2005). Insoluble protein deposits contribute to the pathology of a variety of serious human diseases (Table 1), including the amyloidoses that are characterized by insoluble, fibrillar protein aggregates with a β‐sheet‐rich structure.
Intracellular defences against non‐native proteins
Prokaryotic and eukaryotic cells use post‐translational quality‐control systems, which include molecular chaperones and proteases and are designed to repair or remove damaged proteins. Consequently, misfolded intracellular proteins have three possible fates: rescue by chaperones, destruction by proteases and aggregation. Chaperones are a diverse group of proteins that selectively recognize and bind to exposed hydrophobic surfaces of non‐native proteins in a stable non‐covalent interaction. Several families of chaperones—defined by size, cellular compartment and function—work together to prevent protein aggregation and facilitate the correct folding of non‐native proteins. Through regulated binding and release, chaperones can directly facilitate the folding of proteins by a mechanism that is ATP‐dependent (as exemplified by heat‐shock protein 90 (Hsp90), Hsp70 and Hsp60 in the cytosol) or independent (as exemplified by calnexin and calreticulin in the endoplasmic reticulum (ER); Fink, 1999; Hartl & Hayer‐Hartl, 2002). Other types of chaperones—for example, the small Hsp25 and α‐crystallin—bind reversibly to exposed hydrophobic regions on unfolded polypeptides but are generally unable to independently facilitate refolding. They are thought to protect non‐native proteins from aggregation until “folding‐helper” chaperones are available (Bross et al, 1999).
Misfolded proteins are degraded by a variety of proteases in different cell compartments. However, the main intracellular site of degradation is the proteasome, a barrel‐shaped cytosolic protein complex that selectively degrades ubiquitin‐tagged proteins (Hohfeld et al, 2001). Polyubiquitin chains can be attached to misfolded proteins in the ER lumen (Tsai et al, 2002), but ubiquitination predominantly occurs through the action of ubiquitin ligases and cofactors in the cytosol (Welchman et al, 2005). Proteins destined for secretion into the extracellular environment undergo a rigorous quality‐control process within the ER lumen, which includes interaction with several ER‐resident chaperones. Some of the ER chaperones, such as calnexin and calreticulin, are lectins that have a specialized role in the quality control of glycoprotein folding. Many quality‐control substrates cycle between the ER and the Golgi during their maturation (Trombetta & Parodi, 2003). When misfolded proteins originally destined for secretion cannot be refolded, they are degraded by ER proteinases, directed to the lysosome or, more commonly, retro‐translocated into the cytosol for proteasomal degradation (Fig 2). Under circumstances when the chaperone/protease machinery is overwhelmed, misfolded proteins may form insoluble aggregates inside the ER (Trombetta & Parodi, 2003) or in the cytosol (an aggresome). Once proteins are secreted, they escape these characterized intracellular quality‐control processes and enter an environment in which the mechanisms of folding control—if they exist—have yet to be discovered.
Are extracellular proteins subject to quality control?
There is, on average, 15 litres of extracellular fluids in a 70 kg human, including five litres of blood. Compared with intracellular fluid, extracellular fluids have a lower protein concentration—6% in plasma and 2% in interstitial fluid, as opposed to 30% in cytosol—but are more oxidizing (Sitia & Braakman, 2003). Unlike intracellular fluids, extracellular fluids are continuously subjected to shear stress—for example, through the pumping of plasma around the body—which is known to induce protein unfolding and aggregation (Ker & Chen, 1998). Intuitively, the relatively harsh extracellular conditions suggest that active protein quality‐control mechanisms are required. Studies from past decades have provided substantive support for this hypothesis. For example, in the 1970s and 1980s it was shown that denatured plasma proteins were catabolized in vivo more rapidly than their native counterparts (Margineanu & Ghetie, 1981), and that the catabolic rates of plasma proteins correlate with their in vitro susceptibility to proteases (Dice & Goldberg, 1976). In addition, liposomes with greater surface‐exposed hydrophobicity were cleared from the circulation more quickly than those coated with a hydrophilic polymer; it was proposed that clearance was enhanced by plasma factors that could bind to exposed hydrophobic surfaces on the liposomes (Senior et al, 1991). These data are consistent with the hypothesis that exposed hydrophobicity might be the structural change that identifies individual extracellular proteins as needing quality‐control intervention, as occurs intracellularly. Further support for this hypothesis has come from the demonstration that a recently discovered extracellular chaperone, clusterin (see below), binds to its substrates by hydrophobic interactions (Poon et al, 2002).
What protein quality‐control mechanism(s) actually operate in extracellular space? Substantial recent advances in understanding intracellular protein quality‐control mechanisms highlight our near‐complete lack of understanding of corresponding extracellular processes. While not excluding other possibilities, the potential involvement of chaperones and selective protease systems are clearly worth considering.
Extracellular chaperones and proteasome‐like systems
Intracellular folding‐helper chaperones—for example, Hsp70 & Hsp90—might be released from necrotic cells or during viral cell lysis. Indeed, they have been discovered in human plasma and associated with cell surfaces, in particular cancer cells. Several extracellular roles have been postulated for these chaperones, such as cancer cell invasiveness (Eustace et al, 2004) and immune presentation (Becker et al, 2002). If established, these specialized functions are obviously important; however, in the context of generic mechanisms for extracellular protein quality control, it is pertinent to note that these ‘normally intracellular’ chaperones are present extracellularly at very low (ng/ml) levels and require ATP to effect protein refolding. The abundance of extracellular ATP is at least 1,000 times lower than intracellular ATP (Farias et al, 2005). In the event of a large‐scale presentation of extracellular non‐native proteins—such as might occur during a chronic infection—the capacity of the low levels of Hsp70 and similar chaperones that are present extracellularly would be quickly exceeded. Thus, although a role for these chaperones in extracellular protein quality control cannot be excluded, it is far more likely that large‐scale ‘handling’ of non‐native proteins is dealt with by much more abundant extracellular chaperones (ECs). Two such ECs have recently been identified.
Clusterin, a ubiquitous and highly conserved secreted protein, is an efficient EC (Humphreys et al, 1999; Wilson & Easterbrook‐Smith, 2000). Clusterin potently inhibits stress‐induced protein aggregation by ATP‐independent binding to non‐native proteins to form soluble, high‐molecular‐weight complexes (Humphreys et al, 1999; Poon et al, 2000). More recently, it was established that haptoglobin (Hp; previously best known for its high‐affinity binding to haemoglobin) is also an abundant EC with a chaperone action similar to clusterin (Yerbury et al, 2005). Both ECs are highly conserved, widely distributed glycoproteins that are present in most physiological fluids, including plasma and cerebrospinal fluid (Bowman & Kurosky, 1982; Jenne & Tschopp, 1992). They are composed of variably sized oligomers made up of different numbers of disulphide‐linked α‐ and β‐subunits, bind to a broad range of ligands, and have been detected in association with clinical amyloid deposits in vivo (Calero et al, 2000; Powers et al, 1981). In human plasma, clusterin is present at about 100 μg/ml (Humphreys et al, 1999) and Hp at 0.3–1.9 mg/ml (Bowman & Kurosky, 1982). Immunoaffinity depletion of either clusterin or Hp from human serum renders proteins in whole serum more susceptible to aggregation and precipitation (Poon et al, 2000; Yerbury et al, 2005).
Neither clusterin nor Hp has the ability to independently refold heat‐stressed, non‐native enzymes (Poon et al, 2000; Yerbury et al, 2005). However, like the small heat‐shock proteins, clusterin is able to preserve heat‐inactivated enzymes in a state competent for subsequent ATP‐dependent refolding by heat‐shock chaperone 70 (Hsc70; Poon et al, 2000). Although it is possible that specifically extracellular folding‐helper chaperones exist, they remain to be convincingly demonstrated. It has been reported that the plasma pentraxin serum amyloid P component (SAP) has some ATP‐independent refolding activity in vitro. However, the physiological relevance of this is uncertain because, to effect significant refolding, SAP was required at a tenfold molar excess to the only substrate tested (lactate dehydrogenase; Coker et al, 2000). Further studies of the chaperone action of SAP using other substrates, and the investigation of potential interactions between SAP and other ECs, would be enlightening. SAP is a main plasma acute‐phase protein in mice, but in humans, this role is fulfilled by the related pentraxin C‐reactive protein (CRP; Mold et al, 2001); it is therefore tempting to speculate that CRP might also be an EC. Regardless, it is likely that other abundant ECs will be discovered in the future.
Previous studies have indicated that 50%–90% of plasma proteins are degraded in the liver and reticuloendothelial system (Bouma, 1982). This suggests that a principal component of any extracellular protein quality‐control system comprises mechanisms for cellular uptake and intracellular degradation of aged/non‐native proteins. However, this does not exclude the existence of selective mechanisms for the proteolysis of extracellular non‐native proteins. Both ubiquitin and the proteasome are found in human plasma and their concentrations increase during disease, possibly reflecting their release from dead and dying cells. The plasma concentrations reported range from 7.7–200 ng/ml (Akarsu et al, 2001; Savas et al, 2003) and 2.1–2.4 μg/ml (Lavabre‐Bertrand et al, 2001; Stoebner et al, 2005), respectively. At these levels, ubiquitin and the proteasome are probably at least 300 times less abundant in human plasma than inside cells (Born et al, 1996; Lightcap et al, 2000). In addition, extracellular proteasome‐mediated protein degradation would require extracellular ATP, which is scarce, as noted above. Therefore, current evidence does not support a major role for the mammalian ubiquitin/proteasome system in extracellular protein quality control. Nevertheless, the existence of other extracellular protease systems that might perform a similar function, perhaps independently of ATP, remains a possibility.
A control mechanism for extracellular protein folding
There are four, not mutually exclusive, theoretical models for extracellular quality control. Non‐native extracellular proteins could be:
Assisted to refold in the extracellular space. As discussed above, at present there is no evidence to substantiate this.
Selectively endocytosed through receptors and degraded intracellularly. This seems feasible and will be discussed in more detail below.
Randomly endocytosed and selected for degradation at the level of, for example, sorting endosomes, with native extracellular proteins recycled back to the extracellular space. This model cannot be excluded but at present there is no evidence to support it.
Randomly endocytosed and non‐selectively degraded intracellularly. This model can be excluded because it cannot account for the more rapid catabolism of denatured versus native plasma proteins observed in vivo (described above).
Although models (i) and (iii) cannot be excluded, recent advances support model (ii). In this model, a quality‐control mechanism ‘senses’ increased hydrophobicity on non‐native proteins to select them for ‘processing’ Obvious candidates for this sensing role are the recently identified ECs clusterin and Hp; other, yet to be identified ECs may also be involved. The formation of non‐covalent complexes between ECs and non‐native extracellular proteins would effectively label the latter for quality‐control processing. The suggestion that ECs complex with misfolded proteins in vivo is strongly supported by the fact that in vitro, ECs inhibit the aggregation of purified proteins subjected to chemical and physical stresses by forming stable non‐covalent complexes with them (Humphreys et al, 1999; Yerbury et al, 2005). In addition, the same (endogenous) ECs in unfractionated human serum inhibit heat‐induced aggregation of serum proteins (Poon et al, 2000; Yerbury et al, 2005). It will be important to verify experimentally the formation of complexes between ECs and non‐native proteins in vivo. There are precedents for the idea that by complexing with proteins, ECs can mediate the uptake of their protein ligands via cell‐surface receptors and direct them towards intracellular degradation. For example, complexes formed in vivo between clusterin and the Alzheimer's β‐peptide (amyloid‐β, Aβ) are taken up by neural epithelial cells by way of megalin (LRP2)‐mediated endocytosis and degraded in lysosomes (Hammad et al, 1997). Furthermore, Hp–haemoglobin complexes are quickly bound and internalized by CD163 expressed on human macrophages and are subsequently degraded (Kristiansen et al, 2001). We propose that by virtue of their interactions with both extracellular non‐native proteins and cell surface receptors, ECs identify and deliver non‐native proteins for cell uptake and degradation (Fig 3).
If this model is correct, then it would be expected that manipulation of EC expression may affect the clearance and/or aggregation of proteins in animal models. Recent work with clusterin‐knockout mice supports the hypothesis that clusterin exerts important effects on protein clearance and aggregation in vivo. For example, a long‐term study of clusterin‐knockout mice showed that ageing mice developed a progressive glomerulopathy characterized by the accumulation of insoluble protein (containing immunoglobulins) in the kidney mesangium (Rosenberg et al, 2002). Another study of double‐knockout mice with ablated expression of both clusterin and apolipoprotein E genes concluded that these two proteins worked together to inhibit the deposition of fibrillar Aβ in a model for Alzheimer's disease (DeMattos et al, 2004). Studies of Hp‐knockout mice have not examined the effects of Hp on the aggregation/clearance of non‐native extracellular proteins; it is, however, interesting to note that ablation of Hp expression caused a small but significant reduction in post‐natal viability (Lim et al, 1998).
There are several cell‐surface molecules that might be involved in receptor‐mediated endocytosis of EC‐(non‐native) protein complexes. For example, clusterin interacts with several members of the low‐density lipoprotein receptor (LDLR) family. Clusterin binds to chicken lipoprotein receptor 8 and an LDLR‐related protein (Mahon et al, 1999), and uptake of clusterin–leptin complexes by apoER2 and very LDLR (VLDLR) has been proposed to facilitate leptin clearance (Bajari et al, 2003). Furthermore, clusterin and LRP1/megalin have been implicated in the clearance of cellular debris by non‐professional phagocytes (Bartl et al, 2001). Aside from CD163, another potential candidate receptor for Hp–(non‐native) protein complexes is the CD11b/CD18 integrin (Mac‐1/CR3), which binds to Hp, denatured proteins and the iC3b fragment of complement (Ross, 2000). SAP and CRP are also known to bind and be internalized by Fc receptors (Mold et al, 2001). These observations have provided leads to pursue in future studies of extracellular protein quality control.
If the hypothesis is correct that a quality‐control system exists to oversee the folding of extracellular proteins, then it follows that the dysfunction of this system might contribute to the development of diseases associated with inappropriate extracellular protein aggregation and deposition. Therefore, elucidation of the mechanisms controlling extracellular protein folding could lead to the development of new treatments for serious human diseases.
M.R.W. and E.M.S. are supported by an Australian Research Council Discovery‐Project grant (DP0211310). J.J.Y. and A.R.W. are supported by Australian Postgraduate Awards.
- Copyright © 2005 European Molecular Biology Organization