Dissertations / Theses on the topic 'Folded Peptides'

Thesis (Ph. D.)--University of North Carolina at Chapel Hill, 2009.
Title from electronic title page (viewed Oct. 5, 2009). "... in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Chemistry. " Discipline: Chemistry; Department/School: Chemistry.

Since ancient times, substances derived from amphibian skins have been recognised to possess various medicinal properties, In recent years, the peptides from amphibian skin have attracted extensive attention for their profound significance in providing clues directed toward novel drug development, for better understanding of miscellaneous physiological and pathological processes, for elucidation of phylogenetic relationships and for improving taxonomy. Previously, the chemical complexity of the defensive skin secretion of the broad-folded frog, Hylarana latouchii, has not been studied in detail. In this thesis, parallel transcriptomic and peptidomic analyses of the electrically-stimulated skin secretion have enabled the identification and characterisation of seven, biologically-active peptides. Five of these peptides are antimicrobial peptides displaying differential growth-inhibitory activity toward test microorganisms and human cancer cell lines; two of which belong to the previously-identified brevinin-l and temp orin families, while the others show little structural similarity with other antimicrobial peptides and represent the prototypes of a novel peptide family - the hylaranins. Hylaranin-Ll and hylaranin-L2 (two 18-mer structurally-related but distinct peptides) are described in Chapter 3 and hylaranin-L3 (a unique l3-residue peptide) is described in Chapter 5. In Chapter 4, a novel Bowman-Birk-type trypsin inhibitory peptide, pLR-HL, belonging to the pLRlranacyclin family, was identified. By substitution of its Lys8 residue with Phe, the Phe8-pLR-HL analogue was found to be transformed into a chymotrypsin inhibitor. The II -residue canonical reactive loops within each peptide could exert corresponding protease inhibitory activity independently, In Chapter 6, a peptide structurally-related to bombesin was isolated and found to possess contractile activity on rat urinary bladder and uterus smooth muscles. These data illustrate that amphibian skin continues to provide numerous novel peptides for furthering research in the fields of pharmaceutical science, biological chemistry, medical science, and systematics. It is expected that many more novel amphibian skin peptides are awaiting discovery.

Thesis (Ph. D.)--Massachusetts Institute of Technology, Dept. of Chemistry, 2001.
Includes bibliographical references.
This thesis describes the development of iterative and combinatorial methods for identifying small peptide scaffolds able to support catalytic function. The incorporation of thiamine coenzyme functionality into small peptide scaffolds is achieved through the use of a coenzyme amino acid chimera (Taz). This thiazole amino acid can be alkylated to generate the active thiazolium species. A series of designed peptides containing the Taz coenzyme chimera revealed that an increasingly structured and hydrophobic peptidyl environment can dramatically enhance the acidity of the thiazolium C2 methine, the initial step in the catalytic pathway. A combinatorial approach is developed to permit a more rapid screening of large libraries of ppa peptide sequences for unique functional properties. The method used a fluorescent 13-diketone probe to identify a family of similar sequences that incorporate a functional primary amine at the center of the hydrophobic core of the PPa peptide structure. The function of the amine is enhanced by the provision of the hydrophobic peptidyl environment and is able to modestly enhance aldol condensation and retro-aldol reactions. Finally, a new trimeric Apa oligomer structure is evaluated as a functional peptide scaffold. Initially, the structural requirements of the motif are explored through a series of sequence modifications. These studies lead to the incorporation of a primary amine at a unique site in the scaffold that greatly enhances the reactivity of the amine. It is further demonstrated that this increased functional ability is directly related to the folded stability of the trimer.
by Kevin A. McDonnell.
Ph.D.

Die Pektat-Lyasen gehören zu einer Proteinfamilie, die meistens von pflanzenpathogenen Mikroorganismen sekretiert werden. Die Enzyme katalysieren den Abbau von Polygalakturonsäure, einem Hauptbestandteil in
pflanzlichen Mittellamellen und Primärzellwänden. Der Abbau der alpha-1,4-verbrückten Galakturonsäurereste erfogt durch eine beta-Eliminierungsreaktion, dabei entsteht ein Produkt mit einer ungesättigten C4-C5 Bindung am nicht reduzierenden Ende, das durch spektroskopische Messungen beobachtet werden kann. Für die enzymatische Reaktion der Pektat-Lyasen ist Calcium nötig und das pH-Optimum der Reaktion liegt bei pH 8.5. Alle bis jetzt bekannten Strukturen der Pektat- und Pektin-Lyasen haben das gleiche Strukturmotiv - eine rechtsgängige parallele beta-Helix. Die Struktur der Pektat-Lyase aus Bacillus subtilis (BsPel) ist im Komplex mit Calcium gelöst worden. BsPel ist ein monomeres Protein mit einer ungefähren Molekularmasse von 43 kDa, das keine Disulfidbrücken enthält. Dies erlaubte sowohl eine effiziente rekombinante Expression des Wildtypproteins, als auch von destabilisierten Mutanten im Cytoplasma von E. coli. Parallele beta-Helices sind relativ große, jedoch verhältnismäßig einfach aufgebaute Proteine. Um detailliertere Informationen über die kritischen Schritte bei der in vitro-Faltung von parallelen beta-Helices zu erhalten, sollte in der vorliegenden Arbeit versucht werden, den Faltungsmechanismus dieses Proteins näher zu charakterisieren. Dabei sollte vor allem die Frage geklärt werden, welche Wechselwirkungen für die Stabilität dieses Proteins einerseits und für die Stabilität von essentiellen Faltungsintermediaten andererseits besonders wichtig sind.

Rückfaltung von BsPel, ausgehend vom guanidiniumchlorid-denaturierten Zustand, war bei kleinen Proteinkonzentrationen und niedrigen Temperaturen vollständig möglich. GdmCl-induzierte Faltungsübergänge waren aber nicht reversibel und zeigten eine apparente Hysterese. Kinetische Messungen des Fluoreszenz- und CD-Signals im fernen UV ergaben eine extreme Denaturierungsmittelabhängigkeit der Rückfaltungsrate im Bereich des Übergangmittelpunktes. Der extreme Abfall der Rückfaltungsraten mit steigender Denaturierungsmittelkonzentration kann als kooperative
Entfaltung eines essentiellen Faltungsintermediats verstanden werden. Dieses Faltungsintermediat ist temperaturlabil und kann durch den Zusatz Glycerin im Renaturierungspuffer stabilisiert werden, wobei sich die Hysterese verringert, jedoch nicht vollständig aufgehoben wird. Durch reverse Doppelsprungexperimente konnten zwei transiente Faltungsintermediate nachgewiesen werden, die auf zwei parallelen Faltungswegen liegen und beide zum nativen Zustand weiterreagieren können. Fluoreszenzemissionsspektren der beiden Intermediate zeigten, daß beide schon nativähnliche Struktur aufweisen. Kinetische Daten von Prolin-Doppelsprungexperimenten zeigten, daß Prolinisomerisierung den geschwindigkeitsbestimmenden Schritt in der Reaktivierung des denaturierten Enzyms darstellt. Desweiteren konnte durch Prolin-Doppelsprungexperimenten an Mutanten mit Substitutionen im Prolinrest 281 gezeigt werden, daß die langsame Renaturierung von BsPel nicht durch die Isomerisierung der einzigen cis-Peptidbindung an Prolin 281 verursacht wird, sondern durch die Isomerisierung mehrerer trans-Proline. Die beiden beobachteten transienten Faltungsintermediate sind somit wahrscheinlich zwei Populationen von Faltungsintermediaten mit nicht-nativen X-Pro-Peptidbindungen, wobei sich die Populationen durch mindestens eine nicht-native X-Pro-Peptidbindung unterscheiden.

Der Austausch des Prolinrestes 281 gegen verschiedene Aminosäuren (Ala, Ile, Leu, Phe, Gly) führte zu einer starken Destabilisierung des nativen Proteins und daneben auch zu einer Reduktion in der Aktivität, da die Mutationsstelle in der Nähe der putativen Substratbindetasche liegt. Die Rückfaltungskinetiken der Prolinmutanten war bei 10°C annähernd gleich zum Wildtyp und die geschwindigkeitsbestimmenden Schritte der Faltung waren durch die Mutation nicht verändert. Die durch die Mutation verursachte drastische Destabilisierung des nativen Zustands führte zu einem reversiblen Entfaltungsgleichgewicht bei pH 7 und 10°C. GdmCl-induzierte Faltungsübergänge der Mutante P281A zeigten bei Messungen der Tryptophanfluoreszenzemission und der Aktivität einen kooperativen Phasenübergang mit einem Übergangsmittelpunkt bei 1.1 M GdmCl. Durch die Übereinstimmung der Faltungsübergänge bei beiden Messparametern konnten die Faltungsübergänge nach dem Zwei-Zustandsmodell ausgewertet werden. Dabei wurde eine freie Sabilisierungsenthalpie der Faltung für die Mutante von - 64.2 ± 0.4 kJ/mol und eine Kooperativität des Übergangs
von - 58.2 ± 0.3 kJ/(mol·M) bestimmt.


BsPel enthält, wie die meisten monomeren rechtsgängigen parallelen beta-Helix-Proteine, einen internen Stapel wasserstoffverbrückter Asparagin-Seitenketten. Die Mehrheit der erzeugten Mutanten mit Substitutionen im Zentrum der Asn-Leiter (N271X) waren als enzymatisch aktives Protein zugänglich. Die Auswirkung der Mutation auf die Stabilität und Rückfaltung wurde an den Proteinen BsPel-N271T und BsPel-N271A näher analysiert. Dabei führte die Unterbrechung des Asparaginstapels im Inneren der beta-Helix zu keiner drastischen Destabilisierung des nativen Proteins. Allerdings führten diese Mutationen zu einem temperatur-sensitiven Faltungsphänotyp und die Hysterese im Denaturierungsübergang wurde verstärkt. Offenbar wird durch die Unterbrechung des Asparaginstapel ein essentielles, thermolabiles Faltungsintermediat destabilisiert. Der Asparaginstapel wird somit bei der Faltung sehr früh ausgebildet und ist wahrscheinlich schon im Übergangszustand vorhanden.
Pectate lyases belong to a family of proteins secreted by plant pathogenic microbes. The enzymes cleave alpha-1,4 linked galacturonic acid by a beta-elimination that results in an unsaturated product, which can be quantified spectrophotometrically. Calcium is essential for the activity and the pH-optimum is near 8.5. All known structures of pectate and pectin lyases have the same structural motif - a right handed parallel beta-helix. The structure of pectate lyase from Bacillus subtilis (BsPel) has been solved in complex with calcium. It is a monomeric protein, with a molecular mass of about 43 kDa and without disulfide bonds. This allows its high-yield recombinant expression in the cytoplasm of Escherichia coli. Parallel beta-helices are relative large proteins, however with a simple folding topology. The objective of this work was to characterize the folding mechanism of BsPel. In particular we investigated the role of the interactions of certain residues in the parallel beta-helix for the stability of the native protein and the stability of essential folding intermediates.

Refolding of BsPel was possible at low protein concentrations and low temperature. However, denaturation of BsPel was not freely reversible. De- and renaturation curves showed a large apparent hysteresis. Furthermore, the folding rate constant deduced from fluorescence and circulardichroism measurements showed a very strong dependence on denaturant concentrations near the midpoint of the renaturation transition. This can be explained with a cooperative unfolding of an essential folding intermediate. Upon stabilisation of the temperature-sensitive intermediate by addition of glycerol in the renaturation buffer, the hysteresis is reduced, but does not disappear. Reverse double mixing kinetic experiments have shown that two transient folding intermediates are on the folding pathway. These intermediates are on parallel pathways and both can fold to the native state. Fluorescence emission spectra have shown the native-like structure of both intermediates. Furthermore, data from proline double mixing kinetic experiments revealed that isomerization of peptidyl-prolyl bonds was responsible for the slow kinetics in the reactivation of the enzyme. However, the isomerization of the single cis-peptidyl-prolyl bond at Pro281 was not responsible for the slowest folding phase observed, but rather the isomerization of other trans-peptidyl-prolyl bonds. Thus, both transient folding intermediates observed probably represent two populations of folding intermediates with non-native X-Pro-peptide bonds. The difference of the two populations is at least one non-native X-Pro-peptide bond.

Mutations of the proline 281 against various residues (Ala, Ile, Leu, Phe, Gly) resulted in a strong destabilization of the native protein. Also, the activity of the mutant proteins was strong reduced due to the position of the mutation site near the putative active center of the protein. At 10°C the kinetic folding behavior of the proline mutants was not significant changed. However, the strong destabilization of the native state in the proline mutants resulted in a reversible folding equilibrium at pH 7 and 10°C. The unfolding of the P281A mutant was reversible as determined by fluorescence emission and enzyme activity measurements. The coincidence of these detected transitions is consistent with a two-state equilibrium transition. At pH 7 and 10°C the delta G°(H₂O) for folding of P281A was - 64.2 ± 0.4 kJ/mol, with a midpoint of the transition at 1.1 M GdmCl and a cooperativity of - 58.2 ± 0.3 kJ/(mol·M).

BsPel has an asparagine ladder in turn 2 of the parallel beta-helix with extensive network of side-chain hydrogen bonds between the Asn residues. Such an Asn-ladder is a conserved feature of many monomeric beta-helices crystallized so far. The middle Asn residue (271) was selected and exchanged for various residues. Most of the mutants were expressed at 25°C as soluble and active proteins but with a significant reduction in yield. Mutants N271T and N271A were selected to study the stability and refolding of these proteins in comparison with the wild-type protein. The substitution in the Asn-ladder did not drastically destabilize the native protein, but caused a temperature-sensitive-folding (tsf) phenotype with an increased hysteresis in the de- and renaturation transition curves. In addition, the disruption of the Asn-ladder resulted in destabilization of an essential, thermosensitive folding intermediate. Thus, the Asn-ladder is formed very early during the folding, probably well before the transition state of folding.

In silico investigation on protein domains structure and linear/structural motifs can strongly boost functional analyses and technological design. Protein surface features study is crucial to understanding Protein-Protein Interactions (PPI); in particular, surface and pockets conservation and variation, in terms of hydrophobicity, steric hindrance and electrostatics can act as driving forces in protein evolution and functional specialization. Therefore, molecular modeling and structure comparison techniques play an important role in shedding light on “protein behavior” and this PhD work took advantage from integrating computational approaches based on some known molecular modeling methods, such as e.g. Homology Modeling, Fold Recognition, Ab initio Modeling, PBE (Poisson-Boltzmann Electrostatics), Protein-peptide Docking and Hydropathy Analysis with structure and sequence comparison and scanning tools and, of course, with feedback from wet lab analyses performed by co-workers. Such an integrative approach was followed along investigations on a number of different biological systems: • Surface determinants in H5N1 type A Influenza viruses: Here, an analysis of surface determinants from H5N1 haemagglutinin, involved in host-viral interaction, was completed and then published. Genomic variation is very high in influenza A viruses. However, viral evolution and spreading are strongly influenced by immunogenic features and capacity to bind host cells, depending in turn on the two major capsidic proteins (haemagglutinin and neuraminidase). Current analyses of viral evolution are based on serological and primary sequence comparison; however, comparative structural analysis of capsidic proteins can provide functional insights on surface regions possibly crucial to antigenicity and cell binding. We performed molecular modeling and extensive structural comparison of influenza virus haemagglutinin and of their domains and sub-regions to investigate type- and/or domain specific variation. We found that structural closeness and primary sequence similarity are not always tightly related; moreover, type-specific features could be inferred when comparing surface properties of haemagglutinin subregions, monomers and trimers, in terms of electrostatics and hydropathy. Focusing on H5N1, we found that the variation at the receptor binding domain (RBD) surface intriguingly relates to branching of still circulating clades from those ones that are no longer circulating. Recent evidence on the association between electrostatic fingerprints at the haemagglutinin receptor binding surface and the evolutionary success and spreading of H5N1 avian influenza clades prompted us to perform further integrated phylogenetic and structural bioinformatic analysis in H9N2 viruses. In fact, influenza A virus is a zoonotic agent with a significant impact both on public health and poultry industry and switch to human host has been reported for both H5N1 and H9N2 viruses. We performed the evolutionary analysis of a large and non-redundant viral strain dataset, leading to clustering of H9N2 viruses in five groups. Then and according to recent evidence on H5N1, congruence resulted among phylogenetic data and surface electrostatic fingerprints from structural comparison. In particular, surface feature fingerprints could be inferred that relate group specific variation in electrostatic charges and isocontour to well-known hemagglutinin sites involved in modulation of immune escape and host specificity. Results from this second work strengthen suggestion that when integrating up-to-date phylogenetic analyses with sequence-based and structural investigation of surface features may represent a front-end strategy for inferring trends and relevant mechanisms in influenza virus evolution. • Domain architecture variation in mammalian protein trafficking: Human VAMP7b is the most interesting variant among those produced by alternative splicing of the encoding gene SYBL1. Production of VAMP7b variants is determined by skipping of exon 6 which in turn results in coding sequence frameshift. We found that this event is conserved in other mammalian species. VAMP7b shares with the main isoform the N-terminal, inhibitory longin domain and the first half of the SNARE motif. In mammals, VAMP7b is a truncated protein in which the C-terminal half of the SNARE motif and the transmembrane region are replaced by short and variable peptides. Intriguingly instead, only in human and apes sequence frameshift determined by exon 6 skipping results in the creation of a novel unique domain of unknow function, hence human VAMP7b is not truncated but even 40 residues longer than the main isoform. Since existence of such “long” isoform and of its unique domain at protein level were confirmed by specific antibodies, we embarked on in silico dissection of the novel domain by position specific matrix sequence analysis and by ab initio structural modeling. Moreover, since the N-terminal region of the SNARE motif is conserved and it is known to mediate intramolecular binding to the Longin domain, we investigated both in vivo (by two-hybrid in yeast analysis) and in vitro (by NMR analysis) on conservation of the closed conformation. Furthermore, SCL of both VAMP7b and Ykt6b was investigated using GFP and RFP chimeras. Last but not least, b isoforms of the longin genes were analyzed by qPCR and found to be developmentally regulated. • Binding motif regulating neurite outgrowth and guidance: Fine tuning of PPIs by variation in domain architecture or by changing local motifs by surface features modulation can regulate both extracellular and intracellular signaling pathways. Extracellular PPIs can play a central role in heterologous recognition (e.g. host-pathogen) as well as in homologous signaling among cells from the same organism. Proteins exposed at the plasma membrane (PM) can interact each other and with the extracellular matrix (ECM) to provide positional information and guidance cues. Cell adhesion molecules (CAMs) are PM proteins mediating either attractive or repulsive signals by homo- and heterophilic interactions of their extracellular domains (Eds). CAM EDs are most often composed by Ig-like or Fibronectin type III fold repeats. Current evidence suggests that the four N-terminal Ig 1-4 domains of CAM EDs play a major role in such homo- or heterophilic interactions and in particular an important interaction motif is contributed by repeat Ig2. In our lab, biomimetic peptides have been developed by reproducing the known or predicted interaction motifs from the Ig2 domain of human L1CAM and the single Ig domain of human LINGO1, i.e. two proteins that play a crucial role in neurite outgrowth and guidance and in neuronal differentiation. Based on the somehow surprising structural and sequence conservation of the motif region (even when proteins show very different ED architectures), we started investigating on variation and conservation of the putative motif region by means of homology search, regular expression and finally by structural modeling and comparison. Preliminary results highlighted strong conservation of the central Arg residue in the interaction motif, while in other positions of the motif residue properties rather than specific residues are conserved. Such evidence is in agreement with finding that mutation of such residue in L1CAM is responsible for a severe neurological disorder, while mutations at other residues of the motif, results in less severe phenotype. This suggests the motif is an epitope positionally conserved around the central Arg allowing limited, but significant structural variability in surrounding sequence. In order to check such a hypothesis, a structural superposition of the Ig domains containing the interaction motif was performed, confirming that the peptide motif itself is positionally conserved but the highest positional and structural conservation concerns the central Arg residue. Experiments with peptides mutated in the central Arg showed biological activity of these peptides in terms of neuritogenesis signalling. These works carry out a bioinformatic protocol for the characterization of interaction determinants and their functional modulation, easily transportable to other proteins.
Gli studi in silico aventi per oggetto la struttura di domini proteici e di motif sia strutturali che lineari, sono in grado di fornire un importante apporto in termini di comprensione di funzione e nelle biotecnologie. Lo studio delle caratteristiche a carico della superficie proteica si rivelano essenziali nella comprensione delle Interazioni Proteina-Proteina (PPI); in particolare, la conservazione e variazione della superficie proteica e delle relative cavità in termini di idrofobicità, ingombro sterico e caratteristiche elettrostatiche, possono essere considerate come la forza in grado di guidare l’evoluzione e la specializzazione funzionale delle proteine stesse. Alla luce di quanto sopra esposto, tecniche come la Modellistica Molecolare ed il confronto tra strutture giocano un ruolo importante nel chiarire il modus operandi delle proteine e questo progetto di Dottorato ha proprio sfruttato l’approccio integrato di alcune ben note tecniche di biologia computazionale basate sulla Modellistica Molecolare come, ad esempio, Homology Modeling, Fold Recognition, Ab initio Modeling, PBE (Poisson-Boltzmann Electrostatics), Protein-peptide Docking e Hydropathy Analysis con confronto di sequenze e strutture. Elemento indispensabile e prezioso, ovviamente, il feedback ottenuto dagli esperimenti al banco effettuati dai nostri collaboratori. Questo approccio integrato è stato dunque applicato a differenti sistemi biologici: • Individuazione di determinanti di superficie in virus influenzali di tipo A H5N1: è stata effettuata un’analisi dei determinanti di superficie a carico dell’emoagglutinina proveniente dal virus influenzale H5N1, coinvolta nell’interazione virus-ospite. Questo lavoro ha già condotto ad una pubblicazione. La variazione genomica è elevata nei virus influenzali di tipo A. L’evoluzione e la diffusione dei virus sono molto influenzate dalle caratteristiche immunogeniche e dalla capacità del virus, di interagire con le cellule dell’ospite tramite le due più importanti proteine presenti sul capside virale: l’emoagglutinina e la neuraminidasi. Le analisi oggi a disposizione sono basate sul confronto dell’attività sierologica e di sequenze primarie; alla luce di ciò, l’analisi strutturale di queste proteine capsidiche può essere in grado di svelare delle conoscenze a riguardo di certe regioni presenti sulla superficie proteica che possono essere cruciali per l’antigenicità e per il legame alle cellule dell’ospite. L’emoagglutinina, sezionata nei suoi domini e subdomini, è stata da noi studiata con metodi di Modellistica Molecolare e sottoposta a confronti strutturali fini, per individuare quelle variazioni che potessero risultare tipo/dominio specifiche. Abbiamo evidenziato che la vicinanza strutturale e la similarità di sequenza primaria non sempre sono correlate; in più, caratteristiche tipo-specifiche di sottoregioni dell’emoagglutinina, monomeri e trimeri, possono essere rivelate grazie al confronto delle loro proprietà di superficie, (in termini di elettrostatica ed idrofobicità) appartenenti a sottoregioni dell’emoagglutinina, monomeri e trimeri. In questo lavoro ci siamo focalizzati sul virus H5N1 e abbiamo scoperto che il dominio di legame recettoriale dell’emoagglutinina (RBD) presenta delle variazioni tra clade circolanti e non più circolanti. Le recenti scoperte riguardanti l’associazione tra la disposizione delle cariche al RBD ed il successo in termini evolutivi e di diffusione del virus H5N1 ci hanno spinto ad eseguire analisi integrate di filogenesi e biologia strutturale a carico dei virus H9N2. Infatti, l’influenza A è un agente zoonotico in grado di produrre un grosso impatto sia sulla salute pubblica che sull’industria del pollame, avendo la capacità di effettuare il salto d’ospite, come riportato proprio per H5N1 ed H9N2. Abbiamo effettuato un’analisi evoluzionistica su un grande dataset non ridondante di ceppi virali e questo ci ha consentito di individuare cinque gruppi di virus H9N2. In accordo con le precedenti analisi effettuate per H5N1, abbiamo ottenuto accordo tra i dati filogenetici con quelli ottenuti dalle analisi di confronto strutturale. In particolare, emerge che la variazione della disposizione delle cariche coincide con quella di siti noti dell’emoagglutinina coinvolti nell’evasione al sistema immunitario e nella specificità d’ospite. I risultati ottenuti da questo secondo lavoro pongono l’accento sull’importanza dell’integrazione tra analisi di tipo filogenetiche e di biologia strutturale nella scoperta di nuovi meccanismi evolutivi dei virus dell’influenza. • Variazione dell’architettura di domini in proteine di mammifero coinvolte nel traffico vescicolare: la proteina umana VAMP7b è la più interessante tra quelle prodotte per splicing alternativo del gene SYBL1. La produzione di VAMP7b è causata dal salto dell’esone 6 che si traduce in uno slittamento della sequenza codificante. Abbiamo scoperto che questo evento è conservato in altre specie di mammiferi. VAMP7b condivide con l’isoforma principale il dominio inibitorio longin N-terminale e la prima metà dello SNARE motif. Nei mammiferi, VAMP7b è una proteina tronca in cui al C-terminale metà dello SNARE motif e la regione transmembrana sono sostituite da peptidi corti e variabili. È molto interessante notare come negli uomini e nelle scimmie antropomorfe lo slittamento della regione codificante determinato dal salto dell’esone 6 abbia prodotto un nuovo dominio di funzione sconosciuta: proprio per questo VAMP7b umana non è tronca, ma addirittura 40 residui più lunga rispetto all’isoforma principale. Dal momento che l’esistenza di questa isoforma “lunga” ed il suo nuovo dominio sono stati confermati a livello proteico grazie all’ausilio di specifici anticorpi, abbiamo effettuato una dissezione in silico del nuovo dominio adoperando un’analisi di sequenza di tipo matrice posizione-specifica (PSI-BLAST), seguita da da Modellistica Strutturale di tipo ab initio. In più, dal momento che la regione N-terminale dello SNARE motif è conservata ed è nota nel mediare il legame intramolecolare al dominio Longin, abbiamo appurato la conservazione della conformazione chiusa sia in vivo (saggio del doppio ibrido in lievito) che in vitro (analisi NMR). Inoltre, la localizzazione subcellulare (SCL) di VAMP7b e Ykt6b è stata studiata adoperando chimere contenenti GFP e RFP. Non ultimo, le isoforme b dei geni longin sono stati analizzati tramite qPCR e si è scoperto essere regolate durante lo sviluppo. • Motif di legame con azione regolatoria sulla crescita e l’indirizzamento neuronale: La regolazione fine delle interazioni proteina-proteina che avviene grazie alle variazioni nell’architettura dei domini o dal cambiamento di motif locali indotto dalla modulazione di caratteristiche di superficie, è in grado di regolare i percorsi di segnalazione sia a livello intra- che extracellulare. Le interazioni proteina-proteina extracellulari possono giocare un ruolo fondamentale nel riconoscimento eterologo (es. ospite-patogeno) come nella segnalazione omologa tra cellule appartenenti allo stesso organismo. Le proteine esposte in membrana plasmatica (PM) possono interagire le une con le altre e con la matrice extracellulare (ECM) per consentire informazioni posizionali e segnali di indirizzamento. Le molecole di adesione cellulare (CAMs) sono proteine della membrana plasmatica in grado di mediare segnali sia di natura attrattiva che repulsiva grazie ad interazioni omo- ed eterofiliche a carico dei loro domini extracellulari (EDs). Questi ultimi sono composti per la magior parte da domini ripetuti aventi fold di tipo Ig o Fibronectina di tipo III. Le attuali conoscenze suggeriscono che i 4 domini extracellulari N-terminali di tipo Ig siano importanti nelle interazioni omo- o eterofiliche ed in modo particolare il dominio Ig2 è provvisto di un importante motif di interazione. Nel nostro laboratorio abbiamo sviluppato dei peptidi biomimetici che riproducono i motif di interazione conosciuti o predetti appartenenti al dominio Ig2 di L1CAM umana e al singolo dominio Ig di LINGO1 umana, proteine, queste, che giocano un ruolo fondamentale nella crescita, nell’indirizzamento e nel differenziamento neuronale. Sulla base della conservazione strutturale della regione del motif (anche tra proteine con architetture molto diverse dei loro EDs), abbiamo iniziato a studiarne la variazione di sequenza mediante analisi per omologia e per espressioni regolari, per infine tornare al livello strutturale mediante Modellistica Molecolare. I risultati preliminari indicano una forte conservazione dell’Arginina centrale presente nel motif d’interazione, mentre nelle altre posizioni del motif si osserva la conservazione di proprietà dei residui piuttosto che la presenza di specifici residui. Questa evidenza è in accordo con il dato di fatto che la mutazione dell’Arginina in L1CAM è responsabile di un serio disordine neurologico, mentre mutazioni a carico di altri residui del motif causano un fenotipo meno grave. Questo suggerisce che il motif è un epitopo posizionalmente conservato attorno all’Arginina centrale in grado di consentire una variabilità limitata ma significativa nella sequenza circostante. Per verificare quest’ipotesi è stata effettuata una superimposizione strutturale dei domini Ig contenenti il motif d’interazione: il risultato ha confermato che il peptide contenente il motif è di per sé conservato posizionalmente e che la conservazione maggiore sia a livello posizionale che struturale è a carico del residuo centrale di Arginina. Esperimenti con peptidi mutati nell’Arginina centrale hanno dimostrato un’attività in termini di segnalazione nella neuritogenesi. Questi lavori hanno consentito di sviluppare un protocollo bioinformatico per la caratterizzazione di determinanti d’interazione e della loro modulazione funzionale, facilmente trasportabile su altre proteine.

X-ray crystallographic studies of designed peptides provide definitive proof of the success of a design strategy and yield essential structural information that can be utilized in the future design of biologically and structurally important polypeptides. The ability to create locally folded, hydrogen bonded structures in short peptide oligomers is important for peptide design strategies, which rely on the use of folding nuclei in the construction of protein secondary structure modules like helices and β-hairpins. The realization that backbone homologated amino acids, specifically  and  residues can be incorporated into folded polypeptide structures, has stimulated considerable recent interest in the areas of peptide mimetic and foldamer design. Polypeptides with unnatural, modified backbones provide an entry to novel classes of intramolecularly hydrogen bonded helical structures, which are without precedent in conventional peptides composed of  amino acid residues. The design of hybrid sequences, which contain ,  and  amino acid residues, greatly expands the diversity of peptide structure space. The insertion of additional backbone atoms into amino acid residues enhances the number of torsional variables that describe local residue conformation. For example, in the case of  residues, four torsional variables describe the conformational space,  = (Ci-1-Ni-C i-C i), 1 = (Ni-C i-C i-C i), 2 = (C i-C i-C i-Ci), and  = (C i-C i-Ci-Ni+1), as compared to two Ramachandran angles , , for an  residue. In developing the structural chemistry of  residues, it has proved fruitful to explore the conformational properties of residues which are conformationally constrained, either by backbone cyclisation or by the use of gem-dialkyl substitution. In these approaches, conformational choices at selected positions are biased, using local stereochemical constraints that limit the range of accessible backbone torsion angles. Recent trends in peptide research focus on the incorporation of β-, γ- and higher homologues of the α-amino acid residues in designed peptides as they confer significant proteolytic stability. X-ray crystallographic studies of such modified peptides containing non-protein residues are essential, since information on the geometric and stereochemical properties of modified amino acids can only be gathered from systematic structural studies of synthetic peptides incorporating them. This thesis reports a study of the structures and conformations of designed peptides containing stereo chemically unconstrained γ-amino acid residues, derived from naturally occurring proteinogenic α-amino acid residues by backbone homologation. The structures described in this thesis contain the backbone homologated, unconstrained γ-amino acid residues, 4Val, 4Leu and 4Ile and the 3-monosubstituted γ-amino acid residue pregabalin (Pgn). The crystal structure determination of peptides permitted the characterization of intramolecularly hydrogen bonded helices in hybrid sequences and homooligomeric γ- peptides. The studies enabled the precise determination of conformational and geometric parameters of three backbone homologated γ-amino acid residues, 4Val, 4Leu and 4Ile and one monosubstituted unconstrained γ-amino acid residue, pregabalin (Pgn). A detailed analysis of the backbone conformations and intramolecular hydrogen bond parameters of γ- amino acid residues in different hetero- and homooligomeric helices is also provided. This thesis is divided into 7 chapters. Chapter 1 provides a brief introduction to the stereochemistry of polypeptide chains, description of backbone torsion angles of α-, β- and γ- amino acid residues and the major secondary structures of α-, β-, γ- and hybrid peptides. Some previous studies on unconstrained backbone homologated γ-residues are briefly reviewed, followed by a concise introduction to X-ray diffraction and the methods of structure solution. Chapter 2 describes the characterization of the  C12 helix in oligopeptides containing the unconstrained  residue, 4(R)Val. The crystal structure determinations of [Aib-4(R)Val]n oligomers viz. Boc-[Aib-4(R)Val]2-OMe, Boc-[Aib-4(R)Val]4-OMe, Boc-[Aib-4(R)Val]5- OMe and Boc-[Aib-4(R)Val]8-OMe reveal the formation of C12 helical structures in the solid state. The structures of ()n peptides ranging in length from 4 to 16 residues permit characterization of the conformational parameters for -residues. A comparison of the structures of 4 and 8 residue peptides containing repeat ()n, (αβ)n and ()n sequences is also described. The structure determination of the ()n sequences, Boc-[Aib-(S)Val]n-OMe (n = 2 and 4) and the (αβ)n sequences, Boc-[Aib-β3(R)Val]n-OMe (n = 2 and 4), permitted a comparison of the intramolecular hydrogen bonding patterns in peptide helices. The ()n sequences formed 310 helical structures stabilized by C10 hydrogen bonding rings, while the corresponding ()n sequence yielded the backbone expanded C12 helix. In the case of the (αβ)n series the tetrapeptide yielded two consecutive C11 hydrogen bonds corresponding to a C11 helix. In contrast, the (αβ)4 octapeptide yielded a mixed C14/C15 helical structure generated by successive three residue turns of the type αβα and βαβ. Chapter 3 presents the crystallographic characterization of an ()4 dodecapeptide Boc-[Aib-γ4(R)Val-γ4(R)Val]4-OMe. This study was intended to establish whether helical folding could be maintained in sequences containing contiguous γ4(R)Val residues. Specifically, an (αγγ)n sequence composed of successive two residue, 4→1 hydrogen bonded turn can result in a helical structure in which the αγ and γα sequences form C12 hydrogen bonded turns while the γγ segment forms a C14 turn. The crystal structure of the 12-residue peptide Boc-[Aib-γ4(R)Val-γ4(R)Val]4-OMe described in this chapter does indeed demonstrate formation of a mixed C12/C14/C12 helix. Three independent molecules are observed in the crystallographic asymmetric unit permitting an analysis of the conformational and hydrogen bond parameters in this hybrid helical structure. Chapter 4 addresses the question whether ()n helices can be characterized in crystals in sequences composed of unconstrained  and  residues. The crystal structures of 4  hybrid peptides, Boc-Leu-γ4(R)Val-Val-OH, Boc-[Leu-γ4(R)Val-Val]2-OH, Boc-[Leu- γ4(R)Leu]2-OMe and Boc-[Leu-γ4(R)Val]5-OMe are presented. In all four cases folded intramolecularly hydrogen bonded structures are obtained. C12 intramolecular hydrogen bonding is observed in all 4 peptides, suggesting that helical folding is readily obtained in()n and ()n sequences even in the absence of local backbone constraints. Preorganization of  residues does not appear to be a necessary condition for initiating helical folding in hybrid peptides containing -residues. Chapter 5 follows up the results obtained in the preceding chapters and examines the conformational properties of homooligomeric ()n sequences. The structures of the - peptides Boc-[4Val]n-OMe (n = 2-6, 8), Boc-[4Leu]n-OMe (n = 4) and Boc-[4Ile]n-OMe (n = 6, 10) determined in crystals by X-ray diffraction, are described. The tetrapeptides Boc- [4(R)Leu]4-OMe and Boc-[4(S)Leu]4-OMe reveal incipient C14 helical structures of opposite handedness. The hexapeptide Boc-[4(R)Ile]6-OMe and decapeptide Boc-[4(R)Ile]10-OMe fold into C14 helices stabilized by 4 and 8 intramolecular hydrogen bonds respectively, which are backbone expanded analogues of the -peptide 310 helices. The foldability of unconstrained homooligomeric -peptides, is in sharp contrast to the tendency of their - counterparts to form extended, sheet like structures. The structures presented in this chapter constitute the first reported characterization of C14 helix in homooligomers of unconstrained -residues. Chapter 6 describes the crystal structures of two pregabalin zwitterions, (S) Pgn and (R, S) Pgn, five derivatives, one γγ dipeptide, Boc-Gpn-(R, S) Pgn-OH (Gpn, gabapentin) and two hybrid isomeric pentapeptides with the template Boc-Aib-Xxx-Leu-Phe-Val-OMe [where, Xxx = (S)Pgn and γ4(R)Leu]. Five derivatives of pregabalin, Boc-(R, S) Pgn-OH, Boc-(R, S) Pgn-Cyclohexylamide, Boc-(R, S) Pgn-NHMe, Piv-(S) Pgn-NHMe and Piv-(R, S) Pgn-NHMe, reported in this chapter do not possess any intramolecular hydrogen bonds. The γγ dipeptide, Boc-Gpn-(R, S) Pgn-OH forms a C9 hydrogen bond, involving the Gpn residue. Pregabalin (Pgn) is a γ3 substituted analogue and is isomeric with γ4Leu, described in Chapters 4 and 5. Two model pentapeptide sequences were examined to compare the conformational characteristics of γ3 and γ4 monosubstituted γ-residues. The crystal structures of two isomeric pentapeptides, Boc-Aib-(S)Pgn-Leu-Phe-Val-OMe and Boc-Aib- γ4(R)Leu-Leu-Phe-Val-OMe are reported. In the case of the Pgn containing pentapeptide two polymorphic crystal forms are obtained, containing three independent peptide molecules. All three independent molecules fold into short helical segments, with C12 hydrogen bonds at αγ and γα segments and C10 at αα segments. The γ4Leu pentapeptide yielded a conformation almost identical to that observed in the γ3 (Pgn) pentapeptides. Chapter 7 presents a summary of the results obtained and highlights the major, conclusions. Stimulated by the structural results obtained in this study a detailed analysis of γ residue containing helical turns was also carried out by extracting the structures from the CCDC (Cambridge Crystallographic Data Centre) database. Appendix presents a series of tables listing the backbone torsion angles and hydrogen bond parameters of the peptides relevant for the present study. The structures reported in this thesis are listed below (for the structures deposited in the Cambridge Crystallographic Data Centre, the respective CCDC numbers are indicated in parentheses): 1. Boc-[Aib-α(S)Val]2-OMe [C24 H44 N4 O7] (881183) 2. Boc-[Aib-α(S)Val]4-OMe [C42 H76 N8 O11] (882909) 3. Boc-[Aib-β3(R)Val]2-OMe [C26 H48 N4 O7. 2H2O (2O)] (881179) 4. Boc-[Aib-β3(R)Val]4-OMe [C46 H84 N8 O11] (881180) 5. Boc-[Aib-4(R)Val]2-OMe [C28H52N4O7] (881181) 6. Boc-[Aib-4(R)Val]4-OMe [C50 H92 N8 O11] (881182) 7. Boc-[Aib-4(R)Val]5-OMe [C61 H112 N10 O13] (881177) 8. Boc-[Aib-4(R)Val]8-OMe [C78H144N12O15. 2H2O] (881178) 9. Boc-[Aib-4(R)Val-4(R)Val]4-OMe [C94 H172 N16 O19] (909490) 10. Boc-Leu-4(R)Val-Val-OH [C23H43N3O6] (881185) 11. Boc-[Leu-4(R)Val-Val]2-OH [C41H76N6O9] (881186) 12. Boc-[Leu-4(R)Leu]2-OMe [C34H64N4O7. 2H2O (O)] (930729) 13. Boc-[Leu-4(R)Val]5-OMe [C71H132N10O13. 2H2O (2O). 2CH3OH (CO)] (922799) 14. Boc-[4(S)Leu]4-OMe [C38H72N4O7 . 0.5 H2O (0.5 O)] (922802) 15. Boc-[4(R)Leu]4-OMe [C38H72N4O7 . 0.5 H2O (0.5 O)] (922800) 16. Boc-[4(R)Ile]6-OMe [C54H102N6O9] (930731) 17. Boc-[4(R)Ile]10-OMe [C86H162N10O13] (930730) 18. Pregabalin Zwitterion (S) Pgn [C8H17NO2] 19. Pregabalin Zwitterion (R, S) Pgn [C8H17NO2] 20. Boc-(R,S) Pgn-OH [C13H25NO4] 21. Boc-(R, S) Pgn-Cyclohexylamide [C19H36N2O3] 22. Boc-(R, S) Pgn-NHMe [C14H28N2O3] 23. Piv-(S) Pgn-NHMe [C14H28N2O2] 24. Piv-(R, S) Pgn-NHMe [C14H28N2O2] 25. Boc-Gpn-(R, S) Pgn-OH [C22H40N2O5] 26. Boc-Aib-(S) Pgn-Leu-Phe-Val-OMe (Form-I) [C38H63N5O8. H2O] (971416) 27. Boc-Aib-(S) Pgn-Leu-Phe-Val-OMe (Form-II) [C38H63N5O8. H2O (O)] (971417) 28. Boc-Aib-4(R)Leu-Leu-Phe-Val-OMe (Form-I) [C38H63N5O8] (971418)

Understanding the origin of protein folds, and the mechanism by which evolution has generated them, is a critically important step on a path towards rational protein design. Modifying existing proteins and designing our own novel folds and functions is a lofty but achievable goal, for which there are many foreseeable rewards. It is believed that modern proteins may have arisen from a primordial set of peptide precursors, which were initially only pseudo-stable or stable only as complexes with RNA, and later were able to self-assemble into multimeric complexes that resembled modern folds. In order to experimentally examine the feasibility of this theory, an attempt was made at reconstructing the evolutionary path of a beta-trefoil. The beta-trefoil is a naturally abundant fold or superfold, possessing pseudo-threefold symmetry, and usually having a sugar-binding function. It has been proposed that such a fold could arise from the triplication of just one small peptide on the order of 40-50 amino acids in length. The evolutionary path of a ricin, a family within the beta-trefoils known to possess a carbohydrate binding function was the chosen template for evolutionary modelling. It was desirable to have a known function associated with this design, such that it would be possible to determine if not only the fold, but also the function, could be reconstructed. A small peptide of 47 amino acids was designed and expressed. This peptide not only trimerized as expected, but possessed the carbohydrate binding function it was predicted to have. In an evolutionary model of the early protein world, the gene for this peptide would undergo duplication and later, triplication, eventually resulting in a completely symmetrical beta-trefoil, which would represent the first modern beta-trefoil fold. Such a completely symmetrical protein was also designed and expressed by triplicating the gene for the aforementioned small peptide. This hypothetical first modern beta-trefoil is: well folded, stable, soluble, and appears to adopt a beta-trefoil fold. Together these results demonstrate that an evolutionary model of early life: that proteins first existed as self-assembling modular peptides, and subsequent to gene duplications or fusions, as what we now recognize as modern folds, is experimentally consistent and not only generates stable structures, but those with function, which of course is a prime requisite of evolution. Moreover the results show that it may be possible to use this modular nature of protein folding to design our own proteins and predict the structure of others.

The PDZ domain family is one of the most abundant peptide recognition modules in metazoan proteomes. Characterization of natural PDZ domains has provided insight into the structural basis and diversity of peptide recognition by this fold. In order to test the limits of the current model, I evolved synthetic PDZ domains. Based on the Erbin PDZ domain and selected for binding to pep- tides with different position-2 residues, synthetic variants were characterized using high throughput peptide profiling. This approach generated insight into subclass specificities in the most common natural specificity classes (I [ST]-2 and II Φ-2), demonstrated an alternative basis for a rare specific- ity (Class III D-2) and predicted that some natural domains may exhibit an easily-evolved novel specificity (Class IV R-2). These results also emphasize that some non-contact residues may have a disproportionate effect on position-2 specificity, contrary to the predictions of the original model.

碩士
國立清華大學
生命科學系
90
Fold assignment directly from sequences is valuable in the prediction of protein structures. Unlike secondary structure prediction, where a local coding scheme of sequence information will usually suffice, fold identification calls for global protein descriptors as well local descriptors for the whole protein sequences. Previous studies have shown that machine learning methods can yield reasonable prediction accuracy of fold assignment directly from sequences by a variety of global sequence coding schemes. In this thesis, using global protein descriptors based on -peptide distribution, we apply the support vector machine method (SVM) to the 27 most populated folds that contain 386 representative proteins in the Structural Classification of Protein (SCOP) database. Our approach achieved a prediction accuracy 69.6% on an independent set, and 55.5% in the ten-fold cross validation, both of which are an order of magnitude higher than the current methods. Our results show that SVM using suitable global sequence coding schemes can significantly improve prediction in fold recognition from sequences, and should offer a useful tool in structure modeling.