RICERCA ITALIANA     

AMYLOID AGGREGATION OF APOMYOGLOBIN: MOLECULAR MECHANISMS AND IDENTIFICATION OF AMYLOIDOGENIC AND CYTOTOXIC POLYPEPTIDE FRAGMENTS

Università degli Studi di Padova

Abstract

Protein aggregation cause the formation of fibrillar aggregates or amyloid precipitates that characterize the group of human diseases known as amyloidoses. Despite a very intensive research of the last years, a detailed understanding of the molecular principles underlying the transformation of soluble proteins into amyloid aggregates is still lacking. The proteins capable of forming amyloid fibrils are very diverse and do not consist only of proteins involved in severe debilitating diseases such as Alzheimer’s and prion diseases, but also of non-pathogenic proteins and even short peptides. However, X-ray fiber diffraction data indicate that all amyloid fibrils share a cross-beta structure, regardless of the native fold or amino acid sequence of the otherwise soluble protein or peptide. Thus, an increasingly adopted view is that the ability to form amyloid fibrils can be a general property of proteins and peptides under suitable experimental conditions.

Considering the generic nature of amyloid structure, it can be proposed that an in-depth study of protein aggregation processes using one model protein system could establish principles that are generally applicable to all other amyloid-forming proteins. For this reason, in this Project we aim to use the 153-residue protein apomyoglobin (apoMb, myoglobin without the heme) as a model protein for unravelling features of protein fibrillogenesis, even if this protein does not seem to be associated with any disease. ApoMb in the last decades has been a paradigm for probing protein structure and dynamics for numerous investigators. More recently, wild-type apoMb was shown to form amyloid precipitates under rather drastic conditions, while a mutant apoMb with Trp7 and Trp14 residues exchanged with Phe was shown to be highly prone to aggregation and amyloid formation at physiological pH and temperature.

Here, we will investigate the conformational features and aggregation processes of some apoMb mutants prepared by recombinant methods by using a combination of biophysical and biochemical approaches, including circular dichroism, fluorescence dynamics, hydrogen/deuterium exchange, as well as limited proteolysis. The intermediate and final protein aggregates will be examined by electron microscopy, Fourier transform infrared spectroscopy (FTIR), thioflavin and Congo Red binding, as well as by assays of cellular cytotoxicity. The intermediates of protein agggregation will be analyzed in-depth, considering that it has been demonstrated that prefibrillar aggregates or protofibrils are more toxic protein species than the final well-ordered amyloid precipitates. Several small molecule inhibitors of protein aggregation will be also examined, considering that eventually these inhibitors can be used for delineating therapeutic strategies for amyloidosis.

A major effort of this Project will be devoted also to the analysis of the conformational features and aggregation phenomena of polypeptide fragments of apoMb prepared by limited proteolysis of the protein or by solid-phase chemical synthesis. The study of aggregation of apoMb fragments (in particular 1–29, 1–55, 1–88 and 89–153) is expected to be relevant, considering that the majority of amyloidogenic diseases (e.g., Alzheimer) involves aggregation and precipitation of protein fragments. Moreover, experimental evidence is compelling in favor of the notion that small regions of a protein chain are those that dictate protein amyloidogenicity and, therefore, small model peptides appear to be suitable for investigating the process of fibrillogenesis and the role of specific amino acid sequences in favoring protein aggregation. Since it has been already shown that the apoMb fragment 1–29 easily forms amyloid precipitates (unpublished), the possibility exists to prepare a variety of relatively short peptide analogs by solid-phase chemical synthesis, thus allowing a systematic analysis of the structural determinants in amyloid formation. <<<

Principal Investigator

Angelo Fontana Università degli Studi di PADOVA

Research Objectives

A considerable body of evidence suggests that amyloid formation generally occurs through a nucleation-dependent polymerisation mechanism. This means that, under destabilising conditions and above a critical protein concentration, a protein species can form a nucleus given by the association of a number of protein molecules (pre-fibrillar state) and then the fibrils are formed by additional protein association and structural rearrangements. Indeed, protofibrils and oligomers are metastable assemblies observed during the growth of amyloid fibrils of a number of proteins and peptides. These oligomeric assemblies are important for at least two reasons. First, it is now believed that such forms, rather than mature fibrils, may be the cytotoxic agents responsible for some amyloid-associated disorders like Alzheimer's and Parkinson's diseases. Second, it has been proposed that these structures may be intimately involved in the amyloid fibril assembly mechanism, both in amyloid nucleation and fibril elongation. It is clear that an understanding of the driving forces involved in the formation of these organized assemblies rich in beta-sheet structure and of the kinetics and molecular features of the overall process of protein fibrillogenesis is required for the identification of therapeutic strategies to prevent or cure the severe diseases associated with amyloidosis. Here, we will be using apomyoglobin (apoMb, myoglobin without the heme) variants and fragments produced by limited proteolysis or selective chemical fragmentation of the protein, or even synthetic peptide analogs, as a model protein system for investigating molecular aspects of the protein fibrillation process, as well as the importance of amino acid sequence on amyloid formation. The main aims of our research will be as follows.

MISFOLDING INTERMEDIATES AND AMYLOID STRUCTURE. The intrinsic heterogeneous, transient and insoluble nature of the amyloid aggregates makes difficult the application of high resolution structural techniques, such as NMR spectroscopy and X-ray diffraction techniques. Useful informations regarding the molecular features of misfolding aggregates and, in particular, amyloid deposits have been achieved by using hydrogen/deuterium (HD) exchange measurements and limited proteolysis. Within this Project, these two techniques will be used to analyze misfolding intermediates formed transiently during the the fibrillogenesis process of apoMb variants. The two approaches are related, since the flexible parts of a polypeptide chain exchange more protons and, moreover, are more easily attacked by a proteolytic probe. Therefore, both techniques can allow the identification of the flexible chain regions devoid of hydrogen-bonded regular secondary structure. In few recent studies, it has been demonstrated that proteolysis experiments can be used to analyze molecular features even of amyloid precipitates. It was found that some amyloid precipitates are given by a core beta-sheet structure, flanked by flexible polypeptide tails. Here, limited proteolysis experiments will be used to analyze molecular features of the whole apoMb molecule after its rearrangement into the amyloid fibrils.

CYTOTOXICITY. In order to identify the nature of the pathogenic species and the mechanisms by which the aggregation process of apoMb variants result in cell damage, cytotoxicity assays will be performed during apoMb aggregation as a function of time and fibril morphology. The cytotoxicity will be examined by adding aliquots of aggregates, obtained at different protein concentrations and at different time intervals, to cell culture media. The aggregate cytotoxicity will be evaluated by MTT reduction inhibition assays, a standard indicator of physiological cellular stress. Cell impairment will be evaluated by measuring the levels of reactive oxygen species (ROS), the intra-cellular free calcium and fluorescein diacetate-propidium iodide test. In addition, cell death will be evaluated by trypan blue internalization and apoptosis molecular markers.

FRAGMENTS OF APOMYOGLOBIN. Given the complexity of the molecular events involved in protein self-association, recently researchers have designed simplified protein fragments and peptide model systems in order to facilitate the discovery of factors that predispose polypeptides to aggregation and amyloid formation. These systems have provided valuable knowledge about the determinants underlying the structural transitions to the polymeric beta-sheet state of the amyloid fibers. The analysis of the aggregation properties of apoMb fragments can provide information about the propensity of specific amino acid sequences or chain regions of apoMb to form amyloid structures and it is particularly useful to understand the physicochemical determinants of protein aggregation. Of interest, the results of these studies on aggregation-prone fragments may be compared with the predicted amyloidogenic regions of apoMb, as given by the use of recently developed algorithms.

INHIBITORS OF PROTEIN AGGREGATION. Within this Project it is aimed to study several potential inhibitors of protein fibrillogenesis using the simplified models of apoMb variants and apoMb fragments. Indeed, amyloidogenesis can be inhibited through the use of compounds that can competitively block the protein–protein interactions that occur during the amyloid formation process. Clearly, the prospects of these studies are to design possible therapeutic strategies for the devastating diseases caused by protein fibrillogenesis. Indeed, the ongoing search for such blocking compounds has thus far yielded several small molecules capable of inhibiting fibril formation of specific proteins. For example, these inhibitors include porphyrins, tetracyclines, amphotericin, flavonoids and polyphenolic compounds. Chemical agents will be also tested for their ability to disaggregate protein fibrils and, in particular, the property of tetracyclines to prevent fibril formation of apoMb variants and redissolve pre-existing fibrils will be tested. <<<

Timescale

24 months

National and international background

Amyloid and amyloid-related diseases result from the deposition of normally soluble proteins into insoluble plaques. Currently, more than 20 proteins are known to be associated with human amyloid diseases [1-6]. In addition, a number of other proteins and peptides with no known disease state have been shown to be capable of producing amyloid-like material [7,8]. Even though the native structures of the known amyloidogenic proteins vary widely, the fibrils they produce exhibit a common cross-beta structure giving rise to a characteristic X-ray fibre diffraction pattern [9-11]. These structural features, together with other assays, including the binding of thioflavin-T (ThT) and the demonstration of red-green birefringence in the presence of Congo red, are normally used to classify fibrillar deposits as amyloid material [10]. The observation that many proteins, even if not all, can assemble into a common cross-beta fibrillar architecture, regardless of their initial structure, suggests that amyloid may form by a common mechanism [12-16].

Considering the generic nature of amyloid structure and the mechanism of its formation, even among proteins unrelated to amyloid diseases [1,7,8,16], nowadays there is a common belief that the fundamentals of protein misfolding and aggregation can be studied with model proteins, not necessarily linked to an amyloid disease [1,16]. For this reason, in this Project we aim to use apomyoglobin (apoMb) as a model protein for unravelling features of protein fibrillogenesis, even if this protein does not seem to be associated with any disease state. Of interest, apoMb has been in the last decades the prototype model protein for studying the protein folding problem, especially since this relatively small 153-residue protein adopts a variety of partly folded states under different solvent conditions of pH, ionic strength and organic solvents [17-20]. More recently, it has been shown that apoMb can aggregate into amyloid precipitates under rather drastic conditions of alkaline pH and temperature [21,22], while a mutant apoMb with Trp7 and Trp14 replaced by Phe residues has the capability to produce fibrils under physiological conditions of pH and temperature [23,24]. Therefore, it seems appropriate to take advantage of the wealth of structural informations regarding apoMb for unravelling features of protein fibrillogenesis.

PROTOFIBRILS AND AMYLOID. The nucleation-dependent polymerisation mechanism of proteins implies the initial formation of protein oligomers or protofibrils, that later on form the final well-ordered fibrils or amyloid [2,8,16]. Nowadays it is generally accepted that the aggregation of soluble globular proteins occurs via partly folded intermediates [12,13,15,16]. Nevertheless, most of the “natively unfolded “ proteins in vivo do not undergo aggregation, indicating that protein unfolding is necessary, but not sufficient, to promote aggregation [15]. Hence, there must be some protein sequence motifs that, once they become exposed to solvent, are more prone to aggregation than others. In fact, experimental evidence has been provided that small regions of a protein can be responsible for its amyloidogenic behaviour [25-30]. Recently, it has been demonstrated that the initial protein aggregates or protofibrils are more toxic than the final well-ordered fibrils or amyloid precipitates [31]. Therefore, it is of utmost importance to analyze the mechanisms of formation and structural properties of these protofibrils.

The molecular features of protein aggregation processes have been intensively investigated in recent years, but a detailed molecular description of intermediate protein aggregates and mature fibrils or amyloids is still lacking [6,10,11]. Indeed, the intrinsic heterogeneous, transient and insoluble nature of the amyloid aggregates makes difficult the application of high resolution structural techniques, such as NMR spectroscopy and X-ray diffraction techniques. Useful information regarding the molecular features of amyloid deposits have been achieved by using X-ray fibre diffraction studies, atomic force microscopy, cryo-electron microscopy, Fourier transform infrared (FTIR) spectroscopy, solid state NMR and H/D exchange analyzed by NMR or mass spectrometry [10]. Recently, the limited proteolysis technique [32,33] has been successfully exploited to analyze aspects of the fibrillogenesis phenomenon of some amyloidogenic proteins, such as Abeta peptide [34], Ure2p [35], alpha-synuclein [36], as well as the model proteins PI3-SH3 [37] and bovine alpha-lactalbumin [38]. The rationale of this approach resides in the fact that the hydrolysis sites by proteases generally occur at flexible chain regions devoid of hydrogen-bonded regular secondary structure, such as alpha-helix and beta-strands [20,32,33].

The Unit Fontana recently has used the limited proteolysis approach even for analyzing conformational features of the amyloid precipitate formed by human lysozyme (Frare et al., J. Mol. Biol., submitted). The results obtained have clearly shown that the entire 130-residue chain of lysozyme is not involved in the beta-sheet fibril core, since pepsin cleaves off the N- and C-terminal segments of the protein embedded in the amyloid fibrils. The conclusion reached in this study was that the lysozyme amyloid is given by a core structure comprising mostly the chain region 32–108, flanked by flexible tails. Therefore, limited proteolysis appears to be a complementary technique that can lead to useful structural information even for high molecular weight protein complexes, such as amyloid aggregates [33,37]. In the ambit of this Project, limited proteolysis experiments will be used to analyze molecular features of the whole apoMb molecule after its rearrangement into the amyloid fibrils. In this way, we can obtain structural data on apoMb in its fibrillar context, therefore taking into account the structural parameters and interactions (long-range, ionic and hydrophobic interactions, beta-sheet propensity) [25] that allow the apoMb molecule to acquire the final, well-structured amyloid precipitate.

FRAGMENTS. The use of protein fragments and peptides have already provided valuable knowledge about the determinants underlying the structural features of protein fibrillation. In a number of cases it has been possible to identify the key regions (hot spots) responsible for the amyloidogenic and aggregating behaviour of a given protein [26-28,39,40]. There is a common belief that the discovery of amyloid-promoting fragments in proteins should have a great impact on the identification of anti-amyloid agents, thus opening the door to the development of therapeutic strategies for the devastating amyloid diseases [30]. Hence, there must be some sequence motifs that, once they become exposed, are more prone to aggregation than others. In fact, experimental evidence is compelling in favor of the hypothesis that small regions of a protein are those that dictate its amyloidogenic behaviour. If amyloid aggregation is actually driven by short fragments or chain regions of a protein, small model peptides should be suitable for investigating those structural elements of amino acid sequences that favor protein aggregation [25]. In particular, the study of aggregation propensities of protein fragments eventually can provide an experimental proof of the recently developed algorithms used to predict the most amyloidogenic regions (hot spots) in proteins [26-28]. In this respect, mention should be given that algorithms were already used to identify the amyloidogenic hot spots in the 153-residue chain of apoMb [41]. Therefore, it will be of interest to correlate the experimentally determined amyloidogenicity of apoMb fragments with the predicted aggregation-prone chain regions of the apoMb molecule.
INHIBITORS. Amyloidogenesis can be inhibited through the use of compounds that can competitively block the protein-protein interactions that occur during the amyloid formation process [30]. Because amyloid usually forms via a nucleation-growth mechanism, consisting of a nuclei formation phase and a fibril elongation phase, blocking agents that interfere with the pathways of either phase provide potentially attractive targets for the development of therapeutic approaches. The ongoing search for such blocking compounds has thus far yielded several small molecules reported to inhibit fibril formation of specific proteins. For instance, in recent studies it has been shown that fibrillation of some pathogenic proteins can be effectively reduced by numerous substances, including amphotericin B, sulphated polyanions, porphyrins, branched polyamines, tetracycline, nicotine, melatonin and rifamycins [see 42 for ref.es]. However, while these initial findings seem to be promising, no currently known amyloid blocking agent has yet proved to be without toxicological consequences in clinical trials. Within this Project, we will assay several small molecules as possible inhibitors of protein aggregation, in particular tetracycline, heme, polyphenols and flavonoids.

SUMMARY. The prospects of this research Project are that variants of apoMb (horse, whale, beef and tuna) and mutants of apoMb, as well as fragments and synthetic peptide analogs, can be used as suitable model systems for studying molecular features of protein amyloidogenesis. It has been already shown that some apoMb mutants (Irace) and, in particular, the N-terminal fragment 1–29 of apoMb (Fontana), aggregate very rapidly under physiological conditions of pH and temperature to form amyloid precipitates. Since these apoMb species produce amyloids with global structural similarities to those observed with the amyloidogenic proteins causing severe diseases, it is our belief that the simpler and easier to obtain polypeptide models such as apoMb and its fragments can be used to replicate the amyloid reactions that occur in some disease states. <<<