Eawag_06815
Diss. ETH No. 18869
Bacterial β-Aminopeptidases: Function, Structure and Applications
A dissertation submitted to ETH ZÜRICH
for the degree of Doctor of Sciences
presented by TOBIAS HECK Dipl. Biol. Eberhard Karls Universität Tübingen, Germany
born on April 5, 1980 citizen of Germany
accepted on the recommendation of Prof. Dr. Donald Hilvert, examiner Dr. Hans-Peter E. Kohler, co-examiner Prof. Dr. Sven Panke, co-examiner
Zürich, 2010
Dank
Nach beinahe fünf ereignisreichen Jahren Eawag bietet sich mir an dieser Stelle die passende Gelgenheit, die Zeit seit meiner Diplomarbeit 2005 bis zum Ende der Doktorarbeit 2010 Revue passieren zu lassen. Ich möchte hier die wichtigsten Personen herausheben, die meinen Aufenthalt in der Schweiz in wissenschaftlicher und persönlicher Hinsicht in so unvergesslicher Weise geprägt haben.
Ganz besonders glücklich schätzen konnte ich mich über die grandiose Betreuung und Unterstützung durch Birgit Geueke und Hans-Peter Kohler. Ihr habt es verstanden die richtige Richtung vorzugeben und mir trotzdem genügend Freiraum gelassen, meinen eigenen Weg zu finden und meine Ideen und Interessen in die Tat umzusetzen – für mich DIE PERFEKTE MISCHUNG. Mehr gibt es dazu eigentlich nicht zu sagen, vielen vielen Dank!
Im weiteren möchte ich mich ganz herzlich bedanken bei:
Donald Hilvert für die Übernahme des Referats, seine Hilfsbereitschaft und seine Ideen im Rahmen der Subgroup Meetings.
Sven Panke für das Übernehmen des Koreferats.
Dieter Seebach und der ganzen Gruppe an der ETH, speziell Albert Beck, Michael Limbach, Oliver Flögel, James Gardiner, Stefania Capone, Aneta Lukaszuk und Gildas Deniau für die überaus fruchtbare und enge Zusammenarbeit über die vergangenen fünf Jahre sowie die “Nachhilfe” in der organischen Chemie.
Tobias Merz, Markus Grütter und Beat Blattmann für das Lösen der BapA-Strukturen sowie die Einführung in die Proteinkristallographie und -strukturaufklärung.
Steffen Osswald und Matthijs ter Wiel (Evonik Degussa GmbH), Andreas Schmid mit Gruppe (TU Dortmund), Karl-Heinz Wiesmüller und Renate Spohn (EMC Microcollections GmbH) für die Synthese von Substraten und die erfolgreiche Zusammenarbeit.
Thomi Fleischmann, Hans-Ueli Weilenmann und Colette Bigosch für ihre unermüdliche Hilfsbereitschaft bei technischen Fragen und ihre Unterstützung im Laboralltag.
René Schönenberger und Marc Suter für ihre MS-Kompetenz und das Messen zahlreicher LC-MS Proben.
Sudheer Makam und Artur Reimer für ihren Einsatz und ihre akribische Arbeit während ihren Master-/Diplomarbeiten.
Den aktuellen und ehemaligen Mitgliedern der Umik-Crew, Birgit G. und W., Hanspi, Thomas, Thomi, Colette, Christoph, Hans-Ueli, Marius, Eva S1. und S2., Liz, Frederik, Frédéric, Fränzi, Stefan, Inés, Margarete, Sudheer, Artur, Vishakha, Samir, Miguel, Ruedi, Johannes, Karin, Toni, Silvana, Teresa, Ivo, Michi, Tiffany und Martin für die angenehme Arbeitsatmosphäre, die vielen schönen Momente auf und neben der Arbeit und die Hilfe im Labor.
Den Unihockey-Aktivposten Holger, huwem, chw, Thomas, Anatol, Kim, Jafet, Quang, Judith, Ivo, Régine, Yvonne, etc. für harte aber faire Hockey-Schlachten und so manchen schläfrigen Mittwochnachmittag.
Allen aktiven, zurückgetretenen und langzeitverletzten Filigrantechnikern der Eawag Fussball-Truppe, i.e. Martin F. und S., Stephan, Simon, Sebi, Lukas, Christian A. und S., Aurea, Merle, Jürgen Nr. 1, Matteo, Lars, Fabian, Michael, Miguel, Holger, Tobi, Johannes, Peter, Enriquinho, Robert, Jörg, Thomas, Matheus, Simone, Bouziane und dem Rest der Truppe fürs Draussen und Drinnen Tschute.
Scrätschi, Corine, Nathalie, das Party-Office (Dani, Flavio, Mirjam F. und Régine), Etienne, Christian, Mirjam K., Anne, Daya, Tati, Jannis, Lukas, Nadine, Ilona, Bettina, Etienne, Beat, Anita, Christina, Manu, Damian, Philipp, Susanne, Luba, Simón, Dani P.und R., Maria und dem Empfangsteam und allen weiteren aktuellen und ehemaligen Mitarbeitern, die weder zu den Unihockeyanern noch zu den Fussballern zählen, aber trotzdem dazu beigetragen haben, dass ich mich hier so wohl gefühlt habe.
Der Deutschen Bundesstiftung Umwelt und der Eawag für die Finanzierung der Dissertation.
Meinen Eltern, meinen Freunden und Ania (cmok!) für ihre stete Unterstützung, unzählige Unternehmungen und unvergessliche Reisen.
Table of Contents
Summary...... i
Zusammenfassung ...... iii
1. General Introduction ...... 1
2. Enzyme-Catalyzed Formation of β-Peptides: β-Peptidyl Aminopeptidases BapA and DmpA Acting as β-Peptide-Synthesizing Enzymes ...... 13
3. Kinetic Analysis of L-Carnosine Formation by β-Aminopeptidases...... 33
4. Kinetic Resolution of Aliphatic β-Amino Acid Amides by β-Aminopeptidases...... 51
5. β-Aminopeptidase-Catalyzed Biotransformations of β2-Dipeptides: Kinetic Resolution and Enzymatic Coupling ...... 65
6. Complex Structures of BapA with Penicillin-Derived Inhbitors Reveal New Insights into β-Peptide Conversion by β-Aminopeptidases...... 95
7. Protein Maturation and Activity of BapA and Crystal Structure of a Processing-Deficient BapA Precursor ...... 121
8. General Conclusions and Outlook ...... 137
References...... 143
Curriculum Vitae...... 165
List of Publications...... 167
Summary
Summary
The incorporation of β-amino acids into peptides induces the formation of unique secondary structures and confers resistance to degradation by most proteolytic enzymes. These properties give rise to interesting new biomedical applications for β-peptides as bioactive α- peptide mimics. The β-aminopeptidases 3-2W4 BapA from Sphingosinicella xenopeptidilytica 3-2W4, Y2 BapA from Sphingosinicella microcystinivorans Y2 and DmpA from Ochrobactrum anthropi LMG7991 possess the unusual catalytic ability to cleave backbone-elongated β3- amino acid residues from the N-termini of synthetic β- and mixed β,α-peptides that are otherwise resistant to proteolytic breakdown.
The growing demand for β-amino acid-derived compounds to be used in the pharmaceutical industry and as fine chemicals suggested interesting applications for β-aminopeptidases, such as biocatalytic production of enantiopure β-amino acids by kinetic resolution and synthesis of β-amino acid-containing peptides. The hydrolysis of four racemic β3-amino acid amides with aliphatic side chains was catalyzed by the enzymes 3-2W4 BapA, Y2 BapA and 3 DmpA in a highly L-enantioselective reaction and gave access to L-β -amino acids of high enantiopurity. In addition to their hydrolytic properties, the enzymes were capable of coupling various β3-amino acids to the unprotected N-termini of α-amino acids, β-amino acids, and short peptides. β-Aminopeptidase-catalyzed peptide formation was under kinetic control and required the use of C-terminally activated β3-amino acids in an aqueous reaction system at alkaline pH. The enzymatic synthesis of the naturally occurring β,α-dipeptide L- carnosine was of special interest due to the commercial potential and biological relevance of this compound. A detailed kinetic study indicated the usefulness of the β-aminopeptidases 3-
2W4 BapA and DmpA for the biocatalytic production of L-carnosine, but also pointed out drawbacks of these enzyme-catalyzed reactions, such as competing substrate and product hydrolysis and formation of various peptidic byproducts. In addition to β3-amino acid- containing peptides, the enzymes 3-2W4 BapA and DmpA also converted peptides with N- terminal β2-amino acids, which have been less extensively explored due to their limited availability. The BapA-catalyzed transformation of the diastereomeric mixture of a β2- dipeptide was characterized by simultaneously occurring hydrolysis and coupling reactions.
i Summary
The fact that both peptide hydrolysis and peptide formation were highly (S)-enantioselective indicated that 3-2W4 BapA might be useful as a catalyst for enantioselective preparations of β2-amino acids and β2-peptides.
The crystal structure of the β-aminopeptidase 3-2W4 BapA was determined in collaboration with the group of Prof. Grütter at the University of Zürich. The arrangement of the secondary structure elements in 3-2W4 BapA revealed an αββα-core structure, which resembles the consensus fold of the N-terminal nucleophile (Ntn) hydrolase superfamily. Structures of non-covalent BapA-ligand complexes revealed likely interactions of β-peptidic substrates with active site residues, such as Glu133, which is crucial for the stabilization of the ligand’s free N-terminus. The maturation process of 3-2W4 BapA involves the posttranslational cleavage of an inactive precursor polypeptide into a large α- and a small β- polypeptide chain, which form the active protein. The determination of the crystal structure of a processing-deficient BapA mutant together with results obtained from a comprehensive mutagenesis study pointed out essential amino acid residues for protein maturation and substrate conversion. Point mutations of the active site residues Ser250, Glu133, Ser288 or Glu290 resulted in mutants with greatly reduced processing capability and enzymatic activity. The combination of the obtained structural insights with the knowledge of the biocatalytic properties of β-aminopeptidases provides a strong foundation for further investigations, e.g. aiming for the structure-based rational development of β-peptide-converting catalysts with improved functionality and selectivity.
ii Zusammenfassung
Zusammenfassung
Der Einbau von β-Aminosäurebausteinen ruft in Peptiden einzigartige Sekundärstrukturen hervor und verleiht diesen Peptiden hohe Stabilität gegen enzymatischen Abbau. Diese aussergewöhnlichen Eigenschaften bilden die Grundlage für interessante biomedizinische Anwendungsmöglichkeiten von β-Peptiden als neuartige, biologisch aktive Peptidmimetika. Die β-Aminopeptidasen 3-2W4 BapA aus Sphingosinicella xenopeptidilytica 3-2W4, Y2 BapA aus Sphingosinicella microcystinivorans Y2 und DmpA aus Ochrobactrum anthropi LMG7991 besitzen die ungewöhnliche Fähigkeit, N-terminale β3-Aminosäuren von synthetischen β- und β,α-Peptiden hydrolytisch abzuspalten, die ansonsten resistent gegen proteolytischen Abbau sind.
Aus der steigenden Nachfrage nach β-Aminosäure-basierten Verbindungen in der pharmazeutischen und Feinchemikalien-Industrie ergaben sich neue Einsatzmöglichkeiten für β-Aminopeptidasen als Biokatalysatoren, zum Beispiel zur Herstellung enatiomerenreiner β- Aminosäuren oder zur enzymatischen Synthese von β-Peptiden. Die enzymatische Hydrolyse 3 vier racemischer β -Aminosäureamide mit aliphatischen Seitenketten verlief L-enantioselektiv 3 und ermöglichte so die Herstellung von L-β -Aminosäuren hoher Enantiomerenreinheit. Neben der Peptidhydrolyse katalysierten die β-Aminopeptidasen die Kupplung verschiedener β3-Aminosäuren an die ungeschützten Aminotermini von α-Aminosäuren, β-Aminosäuren und kurzkettigen Peptiden. Voraussetzung für die Ausbildung der Peptidbindung durch β- Aminopeptidasen war die Verwendung von C-terminal aktivierten β3-Aminosäuren in einem alkalischen, wässrigen Reaktionssystem. Von besonderem Interesse war die Anwendung der
Enzyme für die Herstellung von L-Carnosin, einem natürlich vorkommenden und kommerziell erhältlichen β,α-Dipeptid. Eine detailierte kinetische Untersuchung zeigte, dass die Enzyme
3-2W4 BapA und DmpA zwar in der Lage waren L-Carnosin zu bilden, die enzymkatalysierte Reaktion allerdings unerwünschte Nebenreaktionen mit sich brachte wie die Hydrolyse des Substrats und die Bildung verschiedener Nebenprodukte. Neben β3-Peptiden setzten die Enzyme 3-2W4 BapA und DmpA auch Peptide mit N-terminalen β2-Aminosäuren um, die aufgrund ihrer eingeschränkten Verfügbarkeit bislang nur wenig erforscht sind. Bei der Umsetzung des diastereomeren Gemischs eines β2-Dipeptids waren parallel ablaufende
iii Zusammenfassung
Hydrolyse- und Kopplungsreaktionen zu beobachten. Die Tatsache, dass beide konkurrierenden Reaktionen höchst (S)-enantioselektiv abliefen, eröffnet weitere biokatalytische Anwendungsmöglichkeit für 3-2W4 BapA zur Herstellung enantiomerenreiner β2-Aminosäuren und zur enantioselektiven Bildung von Peptiden mit N-terminalen β2- Aminosäuren.
Die Kristallstruktur der β-Aminopeptidase 3-2W4 BapA konnte in Zusammenarbeit mit der Gruppe von Prof. Grütter an der Universität Zürich aufgeklärt werden. Die Anordnung der Sekundärstrukturelemente in 3-2W4 BapA enthält wesentliche Merkmale der αββα- Grundstruktur der Ntn-Hydrolasen (N-terminal nucleophile hydrolases). Zusätzlich wurden die Strukturen von BapA in nicht-kovalenten Komplexen mit drei Liganden aufgeklärt, die mögliche Rückschlüsse auf die Substratbindung im aktiven Zentrum des Enzyms erlaubten. Die Ausbildung einer Salzbrücke zwischen Glu133 und dem freien Aminoterminus des Substrats ist dabei von entscheidender Bedeutung. Die Reifung zum katalytisch aktiven Enzym umfasst die posttranslationale Spaltung eines inaktiven BapA-Vorläuferpeptids. Die Aufklärung der Kristallstruktur einer nicht zur Spaltung befähigten BapA-Mutante in Verbindung mit Ergebnissen umfangreicher Mutagenesestudien lieferte erste Anhaltspunkte für mechanistische Details zur Enzymreifung und zur Substratumsetzung. Der gezielte Austausch der Aminosäuren Ser250, Glu133, Ser288 oder Glu290 beeinträchtigte sowohl die Effizienz der Untereinheitenspaltung als auch die katalytische Aktivität der Mutanten. Die Einblicke in die strukturelle Beschaffenheit von 3-2W4 BapA in Kombination mit dem Wissen über die biokatalytischen Eigenschaften und Anwendungsmöglichkeiten von β- Aminopeptidasen bilden eine fundierte Grundlage für die strukturbasierte Entwicklung veränderter β-Aminopeptidasen mit verbesserten katalytischen Eigenschaften und erhöhter Selektivität.
iv General Introduction
1. General Introduction
β-Amino acids – chemical structure and nomenclature
Extending the backbone of the proteinogenic α-amino acids by an additional methylene group gives rise to the class of β-amino acids (Figure 1.1). Due to the possibility of creating a second stereogenic center in the amino acid backbone, β-amino acids are structurally more diverse than their α-amino acid counterparts. A convenient nomenclature for β-amino acid homologues of the naturally occurring α-amino acids was propsed by Dieter Seebach and coworkers in 2004.[1] Accordingly, β-amino acids with side chains of the proteinogenic α- amino acids are designated as β-homoamino acids (βhXaa). The position of the side chain of a monosubstituted β-amino acid is indicated by a superscript number as β2 or β3. Likewise, disubstituted β-amino acids are designated as β2,2, β3,3 or β2,3. The β-amino acid homologue of glycine, which is commonly named β-alanine, is referred to as β-homoglycine in this nomenclature. Most β3-amino acids with proteinogenic side chains and various derivatives thereof can be readily purchased from different suppliers, whereas the commercial availability of β2-amino acids is still very limited due to their more complex chemical syntheses.[2]
R R O O R1 O H OH H H H N N OH N OH N OH H H H H 2 O n n R n R n
3 2 2,3 α-amino acid β -amino acid β -amino acid β -amino acid
Figure 1.1: Comparison of the chemical structures of differently substituted α- and β-amino acids in peptides.
β-Amino acids in natural compounds
β-Amino acids are ubiquitous in nature and they frequently contribute to the formation of primary and secondary metabolites, some of which are depicted in Figure 1.2. β- Homoglycine is by far the most prevalent naturally occurring β-amino acid. In mammalian and bacterial metabolism, β-homoglycine is a constituent of the essential vitamin
1 Chapter 1
pantothenic acid (vitmain B5, 1), which contributes to the formation of coenzyme A (2). Coenzyme A functions as an acyl group carrier and plays key roles as a cofactor in the tricarboxylic acid cycle and in fatty acid metabolism. Furthermore, β-homoglycine and histidine form the β,α-dipeptide L-carnosine (H-βhGly-L-His-OH, 3). L-Carnosine accumulates to high concentrations in innervated mammalian tissues, such as muscle and brain, and presumably acts as antioxidant and radical scavenger.[3-5] The introduction to Chapter 3 deals with the metabolism and the cellular functions of L-carnosine in a more detailed way. The metabolic level of β-homoglycine in mammals is maintained from two major sources.[6] Firstly, β-homoglycine is formed as a catabolite of the pyrimidine bases cytosine and uracil.
The second major source of β-homoglycine is L-carnosine itself, which is converted to its [7] amino acid building blocks β-homoglycine and L-histidine by tissue and serum carnosinases. Further β-amino acids involved in mammalian and bacterial metabolism are (R)-β3- homovaline, which is formed from (S)-leucine by the enzyme leucine 2,3-aminomutase, and the (S)- and (R)-enantiomeric forms of β2-homoalanine, which are catabolites of (S)-valine and thymine, respectively.[6,8] β3-Homoornithine (β-lysine) was discovered as a catabolite of the anaerobic degradation of lysine to acetate, butyrate and ammonia in clostridia.[9,10]
Moreover, β-amino acid-derived substructures contribute to the formation of a wide variety of highly bioactive secondary metabolites in bacteria, fungi, plants and marine organisms. Five examples of important β-amino acid-containing secondary metabolites of bacterial, plant and fungal origin are depicted in Figure 1.2. Taxol (paclitaxel, 4), which was originally extracted from the bark of the Pacific Yew tree (Taxus brevifolia), is used as an inhibitor of mitosis in cancer chemotherapy.[11] Microcystins, such as microcystin-LR (5), form a class of highly toxic metabolites of cyanobacterial origin.[12,13] The β2,3-amino acid-containing dipeptide bestatin (6) was isolated from a Streptomyces strain and acts as a competitive inhibitor of many aminopeptidases.[14,15] Antibiotics of the β-lactam family, such as penicillin G (7), are of special interest for research on β-amino acids, because the opening of their characteristic β-lactam ring by an oxygen nucleophile or by means of a β-lactamase leads to the formation of an internal β-amino acid unit (8).[16] More comprehensive overviews of the occurrence of β-amino acids as building blocks of natural products were given by Seebach et al. and Juaristi.[1,11] Despite the extraordinary diversity of β-amino acid-derived substructures
2 General Introduction in natural compounds, peptides solely composed of β-amino acid residues are not known in biological systems.
1 NH2 O O O O N N HS OPOPO N N N O O N HN N H H OH O O O OH OH H2N N P OO H O O 23
CO2H O N O HN NH O O O OH OH O O NH HN O NH O H H N N H O O O O OH O O CO2H O OH H2N O O N H2N 4 5
H HH N H S S N OH O O N OH N O O H N N O H 2 O OH OH H O OH O
67 8
Figure 1.2: Chemical structures of some natural compounds with β-amino acid-derived substructures, which are highlighted in red. Pantothenic acid (1), coenzyme A (2) and L-carnosine (3) are of mammalian origin, whereas taxol (4) originates from the Pacific Yew tree, microcystin-LR (5) from cyanobacteria, and bestatin (6) and penicillin G (7) from fungi. β-Lacatm ring opening of 7 leads to the penicilloic acid derivative 8.
3 Chapter 1
From β-amino acids to β-peptides
The first studies on synthetic β-peptides comprised of backbone-elongated β3-amino acids with various proteinogenic side chains indicated that β-peptides adopt unique secondary structures.[17,18] These initial findings encouraged further investigations on the folding properties of β-amino acid-containing peptides. Generally, the conformations found in β- peptides resemble the helix, turn and sheet motifs of conventional α-peptides. However, β- peptidic secondary structures are of larger variety and, even at short chain lengths, of higher stability than their α-peptidic counterparts. Stable helical conformations, such as the 314- helix, were observed in properly designed β-peptides consisting of as few as six β-amino acid residues.[18,19] Synthetic peptides with alternating β2- and β3-amino acid residues as well as mixed β,α-peptides revealed novel hairpin turn-like structures, which resemble the β-turn of conventional α-peptides.[20] A rationally designed hexapeptide comprised of β2,3-amino acid building blocks with proteinogenic side chains was shown to adopt an antiparallel β- peptidic sheet structure with a haripin turn.[21] The β-peptidic conformations mentioned in this section represent only a few selected examples out of numerous secondary structure elements that were found in β-peptides. More comprehensive summaries on the variety of β- peptidic secondary structures can be found in review articles by Seebach et al. and Cheng et al.[1,22]
In 1929, studies by Aberhalden and coworkers suggested for the first time that the incorporation of β-aminobutyric acid (β3-homoalanine) into peptides confers resistance to proteolytic breakdown.[23,24] More recent investigations over the past fifteen years confirmed Aberhalden’s observation and demonstrated that β-peptides comprised of β-amino acids with proteinogenic side chains are stable to degradation by many common proteolytic enzymes in vitro.[18,25-27] Moreover, experiments with microorganisms,[28] plants and insects,[29] and mammals[28,30,31] indicated an unexpectedly high resistance of β-peptide derivatives against metabolic breakdown in vivo.
4 General Introduction
β-Peptide-based peptidomimetics
The increased stability of β-peptides against proteolytic breakdown on the one hand, and the predictability of β-peptidic structural elements on the other hand gave rise to interesting new applications for β-peptides as proteolytically stable mimics of bioactive natural peptides.[22,32-35] Recent examples of rationally designed β-peptidic peptidomimetics include inhibitors of protein-protein interactions and viral cell entry,[36-40] ligands of the somatostatin receptor and the major histocompatibility complex (MHC),[41-43] as well as β-peptides with antifungal and antimicrobial activities.[44-50] Moreover, β-peptides were shown to translocate across natural and synthetic membrane barriers indicating that β-peptides could serve as stable biological carriers for the transport of attached moieties into cells. [51-54] A very recent study suggested mixed α,β-peptides as a new class of water soluble nanoporous materials that adsorb nitrogen gas.[55] Potential applications of β-peptide based pharmaceuticals are inevitably connected to the release of these highly stable compounds into the environment. This is problematic, because the ecotoxicologic impacts of β-peptides on organisms and ecosystems have not yet been addressed.
Stable, but not inert
The proteolytic and metabolic stability of β-peptides raised questions regarding the environmental persistence and biodegradability of these compounds. In 2002, Schreiber et al. described for the first time that β-peptides were amenable to biological breakdown by microbial consortia, which were enriched from environmental samples on a β3-peptidic growth substrate.[56] From these enrichment cultures, Geueke et al. isolated the bacterial strain Sphingosinicella xenopeptidilytica 3-2W4, which was able to utilize β3-peptides as sole source of carbon, nitrogen and energy.[57] The closest relative of S. xenopeptidilytica 3-2W4 was Sphingosinicella microcystinivorans Y2, a previously characterized microcystin-degrading bacterial strain.[58] Like S. xenopeptidilytica 3-2W4, S. microcystinivorans Y2 could be grown on the synthetic β-peptides H-β3hVal-β3hAla-β3hLeu-OH and H-β3hAla-β3hLeu-OH. Both Sphingosinicella strains contained enzymes with β-peptide-hydrolyzing activities that share 89% amino acid identity (Table 1.1). These enzymes were isolated and biochemically characterized by Geueke et al. in 2006.[59] With respect to their origins and their abilities to catalyze the hydrolytic removal of N-terminal β-amino acids from β-peptidic substrates, the
5 Chapter 1 enzymes from S. xenopepticfJJytica 3-2W4 and S. microcystinivorans Y2 were designated as p-peptidyl aminopeptidases 3-2W4 BapA and Y2 BapA, respectively. Both 3-2W4 BapA and Y2 BapA share approximately 40% amino acid identity with two previously characterized bacterial aminopeptidases, namely the L-aminopeptidase D-Ala-esterase/ amidase DmpA from Ochrobactrum anthropi LMG7991 [601 and the p-Ala-Xaa di peptidase BapA from Pseudomonas sp. MCI3434 (Ps BapA) (Table 1.1).[611 The described substrate specificity of DmpA comprised small peptides and amides with N-terminal a-amino acids in L- and D- configuration, whereas Ps BapA was reported to cleave a-amino acid amides and short peptides with N-terminal p-homoglycine residues.
Table 1.1: Identities between the amino acid sequences of BapA from S. xenopeptidi!ytica 3-2W4 (3-2W4 BapA), BapA from S. microcystinivorans Y2 (Y2 BapA), BapA from Pseudomonas sp. (Ps BapA) and DmpA from 0. anthropi.
Amino acid identity
Y2 BapA Ps BapA DmpA
3-2W4 BapA 89% 39% 42%
Y2 BapA 39% 43%
Ps BapA 43%
I n 2006, we investigated the substrate specificities of the recombinantly expressed and purified enzymes 3-2W4 BapA, Y2 BapA and DmpA with 31 short peptidic substrates 3 9 62 comprised of p - and a-amino acids with proteinogenic side chains (Table 1.2). [s , l In these studies we addressed the substrate preferences of the enzymes with respect to (1) the backbone length of the N-terminal amino acid of t he peptide, (it) the side chain of the N- terminal p3-amino acid, and (iii) the maximum length of a p-peptidic substrate for enzymatic conversion.
6 General Introduction
Table 1.2: Substrate specificities of 3-2W4 BapA from S. xenopeptidi/ytica 3-2W4, Y2 BapA from S. microcystinivorans Y2 and DmpA from 0. anthropi, measured at an initial substrate concentration of 2.5 mM at pH 8, 10% DMSO and 37°C. Table modified from Heck et af.C62l n.d. not detected.
Substrate Specific activity [U/ mg] 3-2W4 BapA Y2 BapA DmpA Mixed a,fJ-peptides with the sequence Va/-I/e-Tyr H - ~ 3 hVal-~3 hlle-~3 hTy r-OH 0.51 0.16 n.d. H - ~ 3 hVal-~3 hlle-Tyr-OH 0.69 0.17 n.d. H - ~ 3 hVal-Ile-~ 3 hTy r-OH 0.98 0.45 0.001 H - ~ 3 hVal-Ile-Tyr-OH 3.5 1.0 0.009 H -Va l - ~ 3 h lle- ~ 3 hTy r-OH n.d. n.d. n.d. H -Va l - ~ 3 h lle-Tyr-OH n.d. n.d. n.d. H -Va l -Ile- ~ 3 hTyr-OH n.d. n.d. n.d. H-Val-Ile-Tyr-OH n.d. n.d. n.d. Tripeptides with varying N-terminal /3-homoamino acids H - ~ hG l y-Ile- ~ 3 hTyr-OH 0.009 0.047 0.13 H - ~ 3 hleu -I l e-~3 hTyr-OH 1.9 0.38 0.001 H - ~ 3 hPhe-Ile-~ 3 hTyr-OH [a J 0.68 0.46 n.d. H - ~ 3 hTyr-Ile- ~ 3 hTyr-OH 0.47 0.21 n.d. H - ~ 3 hTrp -Ile- ~ 3 h Tyr-OH[aJ 0.047 0.040 n.d. H - ~ 3 hSer-Ile- ~ 3 hTyr-OH 0.095 0.40 0.006 H - ~ 3 hThr -Ile- ~ 3 h Tyr-OH 0.068 0.05 <0.001 H - ~ 3 hGln-I l e-~3 hTy r-OH 0.007 0.008 n.d. H - ~ 3 hP ro-I l e-~3 hTy r-OH n.d. n.d. n.d. H - ~ 3 h H is-Ile- ~ 3 hTy r-OH 0.008 0.011 n.d. H - ~ 3 hlys-Ile- ~ 3 hTyr-OH 0.017 0.015 n.d. H - ~ 3 hA rg-Ile-~3 hTyr -OH 0.006 0.011 n.d. H - ~ 3 hGlu-Ile-~3 hTy r-OH n.d. <0.001 n.d. H - o-~ 3 hVal-Ile-~3 hTy r-OH 0.028 0.016 n.d. Various other peptides H - ~ 3 hA l a-~3 hleu-OH 1.1 3.1 5.5 H - ~ 3 hVal-~3 hAla- ~ 3 hleu-OH 3.1 0.84 0.019 H-Val-Ala-Leu-OH n.d. n.d. n.d. L-Carnosine (3 )CbJ 0.026 0.063 10 Bestatin (6) n.d. n.d. n.d. H - o-~ 3 hA l a-~3 hVa l - ~ 3 hA l a-~3 hleu-OMe[bl n.d. n.d. n.d. H - ~ 2• 3 hA l a ( a-Me) - ~3 hVa l - ~ 3 hA l a-~3 h leu-OMe[bJ n.d. n.d. n.d. H - ~ 3 hA rg-~3 hArg-~3 hArg-OH [bJ n.d. n.d. n.d. H - ~ 3 hVal-~3 hArg- ~ 3 hA rg-OH [bJ 0.052 0.007 n.d. [aJ The assay mixture contained 40% DMSO. [bl The assay mixture contained no DMSO.
7 Chapter 1
The results presented in Table 1.2 showed that 3-2W4 BapA and Y2 BapA exhibit very similar substrate specificities for a wide variety of small peptides that carry N-terminal β3- amino acid with different proteinogenic side chains.[62] The longest tested peptide completely converted by the BapA enzymes was the β3-hexapeptide H-(β3hAla-β3hLys- 3 β hPhe)2. Interestingly, 3-2W4 BapA and Y2 BapA only cleaved N-terminal β-amino acids from peptides, whereas peptides with N-terminal α-amino acids did not serve as substrates. In contrast, DmpA removed α-amino acids[60] and β-amino acids from the N-termini of peptides, but showed a distinct preference for N-terminal amino acids that are unsubstituted or carry small side chains (e.g. -CH3). Due to their common ability to convert peptidic substrates with N-terminal β-amino acid residues, all four enzymes 3-2W4 BapA, Y2 BapA, DmpA and Ps BapA[61] were functionally assigned to the class of bacterial β- aminopeptidases.[63]
Molecular properties of β-aminopeptidases
In addition to their common substrate specificities for peptides with N-terminal β-amino acids, β-aminopeptidases share characteristic biochemical and physiological properties. The enzmyes consist of two polypeptide chains, a large α-chain of 25–27 kDa and a small β- chain of approximately 13–14 kDa (Figure 1.3, A). Interestingly, all four proteins are encoded by single open reading frames, which are translated into inactive precursor polypeptides (Figure 1.3, B). The α- and β-polypeptide chains are formed by posttranslational cleavage of the precursor polypeptides in front of conserved serine residues (Ser250 in DmpA, 3-2W4 BapA, Y2 BapA and Ser239 in Ps BapA; see Figure 1.4). Size exclusion chromatography experiments suggested that the native β-aminopeptidases adopt homotetrameric quarternary structures consisting of four α- and four β-polypeptide chains.[59-61]
8 General Introduction
A B M 1 2 3 kDa gene 200
100
Ser precursor polypeptide 50
40 processing 30 α-polypeptide 25 Ser
20 α β
15 β-polypeptide
10 active protein α4β4
Figure 1.3: A: SDS-PAGE analysis of 3-2W4 BapA from S. xenopeptidilytica 3-2W4 (1), Y2 BapA from S. microcystinivorans Y2 (2) and DmpA from O. anthropi (3) after purification by three chromatographic steps.[59,62] B: Maturation process of β-aminopeptidases. The position of the cleavage site in front of a conserved serine residue is indicated in green.
Although all four β-aminopeptidases undergo the same maturation process, they differ with respect to the presence of N-terminal signal sequences. The bapA genes from S. xenopeptidilytica 3-2W4 and S. microcystinivorans Y2 encode predicted periplasmatic signal sequences of 29 and 26 amino acid residues, respectively, which are absent in bapA from Pseudomonas sp. and dmpA from O. anthropi (Figure 1.4). This indicates that the enzymes are located in different cell compartments of their bacterial hosts. Nevertheless, processing of the precursor polypeptides and enzymatic activities of 3-2W4 BapA and Y2 BapA were not affected when the proteins were recombinantly expressed in Escherichia coli without their signal sequences.[59]
9 Chapter 1
3-2W4 BapA MTSTORLWSGALPLLTALIVSIAATASLAGllPE-Ell NA ITDV •G VG 30 Y2 BapA NA IT DV ~cJJl vc 30 Ps BapA ~=:~=~~=~:~:~~::~=~~~=~~ I: I~ ~~S ~ NA ITDV · G~VG 30 DmpA ------MTSQTPTRKP • LPE-T P NA ITDV t G e 3-2W4 BapA Y2 BapA Ps BapA DmpA 3-2W4 BapA Y2 BapA Ps BapA DmpA 3-2W4 BapA Y2 BapA Ps BapA DmpA 3-2W4 BapA I IQ IGKEIKGAiEVNGIVAAGPDAGKPQDKi~ LLIVI• • L 266 Y2 BapA S ED I PISKEIKGA • EINGIAALGP DAGKPQD~ LLIVI • L 266 Ps BapA eo • iV PiVGTVLGD- •SPF----KSEKKVGVPGM !VITI• c 255 DmpA lll_ ~ T!lll PiVGQHM~~T• Q SQ----LQER------IIVV • L 266 3-2W4 BapA STAGI EE£1sTSHf----IP~GGKPRLPAI INIDTD S ETMN 322 Y2 BapA • • L VG STAGAL EIJAL • STS ~V----I Pf, GQAPRLPAMINID TDSGTm 322 Ps BapA •. • S GGTEDS DIE!• VGNSNLPAANFGHPGEPTT~LKMVNND~IS 315 DmpA lil)ll• • SI IG TPGG DIE!• STANQ----RPMQHRSAPFLDVEMVNDEPLD 322 3-2W4 BapA 373 LVASETM1 • NNAK--,GI PHDl~~RFPRR- Y2 BapA LVASETM • NNVK-- GIPHD R ~RFPRR- 373 Ps BapA LGADE~ . GNT--- IJ~PE OQVGWKAP 366 DmpA IAAEIDM . PFDRLL ~ID.HE_ v;LB,QYGRLA- 375 Figure 1.4: Alignment of 3-2W4 BapA from S. xenopeptidi!ytica 3-2W4, Y2 BapA from S. microcystinivorans Y2, Ps BapA from Pseudomonas sp. and DmpA from 0. anthropi. Identical amino acids are marked in black, similar amino acids in grey. The posttranslational cleavage site is indicated by an arrow and the N-terminal signal sequences of 3-2W4 BapA (residues 1-29) and Y2 BapA (residues 1-26) are underlined. Residue numbers of 3-2W4 BapA and Y2 BapA do not include the signal sequences. Figure modified from Geueke et a/,l59l Structural classification of DmpA from 0. anthropi - a novel Ntn hydrolase variant The crystal structure of DmpA was the first structure of a 13-aminopeptidase to be 64 65 solved.[ , l It provided interesting insights into the general topology of 13-aminopeptidases and demonstrated that DmpA shares characteristic functional and structural features with 10 General Introduction members of the Ntn (N-terminal nucleophile) hydrolase superfamily, which comprises very diverse enzymes with amidohydrolyzing activities.[66,67] In analogy to Ntn hydrolases, the catalytic core of DmpA consists of two polypeptide chains (α and β) that, as outlined above, are formed by posttranslational processing of an inactive precursor polypeptide. Cleavage of the DmpA precursor is presumably autocatalyzed and occurs N-terminal of Ser250, which then becomes the N-terminal amino acid of the β-polypeptide chain. The catalytic mechanism of Ntn hydrolases presumably involves a single amino acid catalytic center[68] and is hence distinct from conventional serine, cysteine or threonine proteases that contain catalytic triads or diads.[69] DmpA-catalyzed substrate conversion was suggested to follow the proposed mechanism of Ntn hydrolases. Accordingly, the free N-terminal amino group of Ser250 acts as the general base to abstract a proton from its own hydroxyl group in order to enhance its nucleophilicity.[64] The topology of one DmpA subunit resembles the consensus Ntn fold, which is comprised of a central layer of β-sheets flanked by an envelope of α- helices. However, the orientation and connectivity of the secondary structure elements in DmpA are distinct from other members of the Ntn hydrolase family, such as human aspartylglucosaminidase, the proteolytic core of the Saccharomyces cerevisiae proteasome, penicillin G acylase and glutamine phosphoribosyl-pyrophosphate amidotransferase from E. coli.[70] Due to these variations from the classical consensus Ntn fold, DmpA was described as the prototype of a new Ntn hydrolase variant.[60,64] However, it is likely that the functional and topological similarities of DmpA and Ntn hydrolases developed from different evolutionary origins by convergent evolution.[71] The high level of sequence conservation among β-aminopeptidases (Table 1.1 and Figure 1.4) indicated that also 3-2W4 BapA, Y2 BapA and Ps BapA are likely to share the topological and mechanistic properties of DmpA and can hence be structurally assigned to the DmpA-like subclass of the Ntn hydrolase superfamily.[63] 11 Chapter 1 Contents and objectives of this thesis The ability to convert synthetic β-amino acid-containing peptides is a very rare feature among enzymes. This suggested the use of the β-peptide-cleaving β-aminopeptidases 3- 2W4 BapA, Y2 BapA and DmpA for interesting biocatalytic applications. The results presented in Chapter 2 and Chapter 3 demonstrate that β-aminopeptidases do not only catalyze the hydrolysis of β3-peptides, but in fact can also be employed for the formation of peptides with N-terminal β3-amino acids. Chapter 4 describes the production of enantio- enriched β3-amino acids by β-aminopeptidase-catalyzed kinetic resolutions of racemic β3- amino acid amides. Chapter 5 addresses the bioconversion of β2-amino acid-containing dipeptides with the enzymes 3-2W4 BapA and DmpA and points out potential applications of the biocatalysts for the production of enantiopure β2-amino acids and for the formation of β2-peptides. Besides exploring the biocatalytic potential of β-aminopeptidases, a structural analysis of 3- 2W4 BapA was intended to shed light on the unique substrate specificities of the enzyme and give mechanistic insights into the posttranslational maturation process of the protein. In Chapter 6, the crystal structure of 3-2W4 BapA is presented together with three additional structures of non-covalent enzyme-ligand complexes, which give rise to speculations about the interactions of substrates with the BapA active site. Chapter 7 describes the crystal structure of a processing-deficient precursor mutant of 3-2W4 BapA and attributes important functions to active site residues for precursor cleavage and enzymatic activity. 12 Enzyme-Catalyzed β-Peptide Formation 2. Enzyme-Catalyzed Formation of β-Peptides: β-Peptidyl Aminopeptidases BapA and DmpA Acting as β-Peptide- Synthesizing Enzymes Abstract In recent studies, we discovered that the three β-peptidyl aminopeptidases, BapA from S. xenopeptidilytica 3-2W4, BapA from S. microcystinivorans Y2, and DmpA from O. anthropi LMG7991, possess the unique feature of cleaving N-terminal β-amino acid residues from β- and α/β-peptides. Herein, we investigated the use of the same three enzymes for the reverse reaction catalyzing the oligomerization of β-amino acids and the synthesis of mixed peptides with N-terminal β-amino acid residues. As substrates, we employed the β- homoamino acid derivatives H-βhGly-pNA, H-β3hAla-pNA, H-(R)-β3hAla-pNA, H-β3hPhe-pNA, H-(R)-β3hPhe-pNA, and H-β3hLeu-pNA. All three enzymes were capable of coupling the six β-amino acids to oligomers with chain lengths of up to eight amino acid residues. With the enzyme DmpA as the catalyst, we observed very high conversion rates, which correspond to dimer yields of up to 76%. The β-dipeptide H-β3hAla-β3hLeu-OH and the β/α-dipeptide H- βhGly-His-OH (carnosine) were formed with almost 50% conversion, when a five-fold excess of β3-homoleucine or histidine was incubated with H-β3hAla-pNA and H-βhGly-pNA, respectively, in the presence of the enzyme BapA from S. microcystinivorans Y2. BapA from S. xenopeptidilytica 3-2W4 turned out to be a versatile catalyst capable of coupling various β-amino acid residues to the free N-termini of β- and α-amino acids and even to an α- tripeptide. Thus, these aminopeptidases might be useful to introduce a β-amino acid residue as an N-terminal protecting group into a “natural” α-peptide, thereby stabilizing the peptide against degradation by other proteolytic enzymes. This chapter by T. Heck, H.-P. E. Kohler, M. Limbach, O. Flögel, D. Seebach, and B. Geueke was published in Chemistry & Biodiversity 2007, 4, 2016-2030. 13 Chapter 2 Introduction β-Peptides consisting of homologated proteinogenic amino acid residues βhXaa[1] are generally stable to common peptide-cleaving[17,25-28,72] and peptide-metabolizing[29-31,73] enzymes in vitro and in vivo (microorganisms[28], plants[29], insects[29], mammals[28,30,31,73]). Recently, we showed that three structurally related aminopeptidases from Proteobacteria are [59,62,74] able to cleave β-peptides. The L-aminopeptidase D-Ala-esterase/amidase DmpA from O. anthropi LMG7991 was the first described enzyme of this class, and it was isolated due to its ability to hydrolyze both (R)- and (S)-α-amino acids from peptides, esters, and amides.[60,75] The foremost published substrate specificity of this enzyme was extended by the finding that it also cleaves peptides and amides containing small N-terminal β- homoamino acids with high activities.[62] Two further β-peptidyl aminopeptidases originate from the environmental isolates S. xenopeptidilytica 3-2W4 and S. microcystinivorans Y2.[59,74] These bacterial strains were isolated because of their capability to metabolize the artificial β-tripeptide H-β3hVal-β3hAla-β3hLeu-OH (1) and the highly toxic cyanobacterial metabolite microcystin (2), respectively.[56-58,76-78] [A] One distinctive feature of the β-peptidyl aminopeptidases from both Sphingosinicella strains (named 3-2W4 BapA and Y2 BapA, respectively, according to the designation of their host strains) is that they exclusively cleave N-terminal β- but not α-amino acid residues from peptides.[59,62,74] A comprehensive investigation of the substrate specificities of recombinant DmpA, 3-2W4 BapA, and Y2 BapA revealed that the enzymes can hydrolyze a broad range of oligopeptides with N-terminal β-amino acids, yet exhibit different amino acid preferences.[62] The physiological substrates of these aminopeptidases, however, are as yet unknown. For more details about this class of enzymes, we refer to a recent review article on bacterial β- peptidyl aminopeptidases, covering the literature up to the end of 2006.[63] Other enzymes that react with peptides containing β-amino acids are very rare; exceptions are different [A] Note that the cyclo-heptapeptide microcystin-LR contains only two proteinogenic (S)-amino acid residues (Leu, Arg) and is otherwise composed of dehydro-N-Me-Ala, (R)-Ala, (R)-(3Me)-Asp (as β-amino acid residue), (R)-Glu (as γ-amino acid residue), and an α-Me-β-amino acid with a most unusual side chain. For total syntheses of microcystins, see [77,78] . 14 Enzyme-Catalyzed {3-Peptide Formation carnosinases from bacterial[79l and mammal origin} 80l as well as poly(aspartic acid) hydrolases from Sphingomonas sp. KT-1 that degrade the 13-peptidic polymer of L-aspartic 81 84 acid H-[NH-CH(C02H)-CHrCO]n-OH. [ - l H ):JN~NuOH 2 H H 1 'Y (R) OH 2 (microcystin-LR) Although proteases and peptidases are commonly associated with protein digestion, their 85186 use as biocatalysts in the synthesis of peptides was demonstrated a long time ago. [ ] High regio- and stereospecificity, high coupling efficiency, minimal protective group requirements, and mild reaction conditions are advantages associated with the application of proteolytic enzymes that make them interesting alternatives to conventional chemical catalysts. A common approach to protease-catalyzed peptide synthesis, the so-called kinetic approach, is illustrated in Scheme 2.1. 15 Chapter 2 H NR' 2 O RCN R' + HEnzyme aminolysis H O O RCXHEnzyme+ RCEnzyme XH O hydrolysis RCOH + HEnzyme H2O Scheme 2.1: Schematic representation of the kinetic approach to enzyme-catalyzed peptide synthesis. The acylated enzyme intermediate formed from the activated amino acid (X = e.g., pNA, OMe, OEt) and the peptidase undergoes competing aminolysis and hydrolysis reactions. In a kinetically controlled reaction system an “acyl donor” component with a mildly activated carboxy terminus, such as an ester or amide, acylates serine or cysteine proteases to form a characteristic acyl enzyme intermediate. Two nucleophiles, H2O and the amino group of a second amino acid or a peptide (“acyl acceptor”), compete for the acyl moiety bound to the active site of the enzyme, which leads either to hydrolysis or to formation of the desired amide bond (aminolysis). The peptide product temporarily accumulates in the reaction mixture until secondary hydrolysis of the newly formed product outbalances its formation. Thus, it is important for preparative purposes to stop the reaction at the point of maximum product formation. For more detailed information on theory and applications of protease- catalyzed peptide synthesis, we refer to recent review articles.[87-89] Recently, Yokozeki and Hara applied the kinetic approach as an efficient and cost-effective protease-catalyzed method for the production of L-Ala-L-Gln and various other α-di- and oligopeptides in aqueous solution using unprotected starting materials.[90] Apart from one lipase-catalyzed reaction,[91] the synthesis of β-peptides, however, has thus far exclusively been accomplished by nonenzymatic methods. With the three β-peptidyl aminopeptidases, 3-2W4 BapA, Y2 BapA, and DmpA, at our disposal (Table 2.1), we were able to test the enzymes as catalysts for β-peptide formation 16 Enzyme-Catalyzed {3-Peptide Formation in a kinetically-controlled reaction system, and we report herein first, mostly analytica l results. Table 2.1: Enzymes used for peptide coupling in the present study. Enzyme name Abbreviation Origin Ref. p-Peptidyl 3-2W4 BapA Sphingosinicella xenopeptidi/ytica 3-2W4 [59,62,74] aminopeptidase p-Peptidyl Y2 BapA Sphingosinicella microcystinivorans Y2 [59,62] aminopeptidase L-Aminopeptidase o- DmpA Ochrobadrum anthropi LMG7991 [60,62,75] Ala-esterase/ amidase Results and Discussion J3-Amino acid p-nitroanilides (pNA) 3, the substrates for the coupling reactions catalyzed by the enzymes 3-2W4 BapA, Y2 BapA, and DmpA As activated substrates for the enzyme-catalyzed formation of 13-pept ides, we employed the 13-homoamino acid p-nitroanilides 3a, (5)- and (R)-3b, (5)- and (R)-3c, and (S)-3d, derived from glyci ne, alanine, phenylalanine, and leucine, respectively (Scheme 2.2). The use of µ-.JA derivatives of the corresponding 13-amino acids is convenient because of the easy detectability of substrates and reaction products.[BJ The compounds (S)-3b, (S)-3c, and (R)- 3c were prepa red as described in [621. The prepa ration and characterization of the 133hAla derivative (R)-3b and of the leucine-derived 13-amino acid p-nitroanilide (S)-3d is described in the Experimental Section of this chapter. The anilides 3 hydrolyzed very slowly in a nonenzymatic process under the applied experimental condit ions. [BJ The anilides 3 have an absorption maximum at 318 nm, p-nitroaniline itself at 383 nm. 17 Chapter 2 N02 3-2W4 BapA, R 0 Y2 BapA or DmpA n H2N~N D \ IP H (n - 1) H-pNA 3a, R H = 4 (S)-3b and (R)-3b, R =CH 3 (S)-3c and (R)-3c, R =CH2C 5H5 (S)-3d, R =CH2CH (CH3)i 3-2W4 BapA, 3-2W4 BapA, Y2 BapA or Y2 BapA DmpA or DmpA H-pNA n H-pNA R 0 n H2N~OH 5 Scheme 2.2: Formation of p-amino acid oligomers 4 from p-amino acid p-nitroanilides 3 in the presence of the enzymes 3-2W4 BapA, Y2 BapA and DmpA. The amides 3 as well as the formed oligomer p-nitroanilides 4 undergo competing N-terminal cleavage reactions and are finally completely hydrolyzed by the enzymes to the free p-amino acids 5 and p-nitroaniline (H-pNA). Coupling of 3 with the N-terminus of the free amino acids 5 could lead to the formation of free p- 3 oligpeptides H-[p hXaa] 0 -0H; traces of such free oligopeptides were detected in the mass spectra of the samples derived from (.5)-3c and (R)-3c (not described herein). Homo-coupling of the p-nitroanilide.s 3 to oligomer-pNAs 4 To investigate the enzyme-catalyzed oligomerization of 13-amino acid residues, we incubated the corresponding 13-amino acid derivatives 3 (5 mM) with the recombinant enzymes 3-2W4 BapA, Y2 BapA, or DmpA. In all experiments, the formation of oligomer-,d..JAs 4 was detected by LC/ MS ana lyses. As typical examples, the time-concentration curves and HPLC diagrams and for the 133hAla derivatives (S)-3b and (R)-3b are depicted in Figure 2.1 and Figure 2.2, respectively. 18 Enzyme-Catalyzed {3-Peptide Formation (S)-3b (R)-3b 5 5 4 4 3-2W4 BapA (117 µg/ml) ~ 3 ~ 3 .s 3-2W4 BapA .s 0 (4.2 µg/ml) 0 2 2 0 0 0 12 24 36 48 0 24 48 72 96 t [h) t [h) 5 5 4 4 Y2 BapA (383 µg/ml) ~ 3 ~ 3 .s .s 0 Y2 BapA 0 2 (12 µg/ml) 2 0 0 0 12 24 36 48 0 24 48 72 96 t [h) t [h) 5 5 4 4 DmpA (31 µg/ml) ~ 3 ~ 3 .s DmpA .s 0 0 2 (0.47 µg/ml) 2 o ...... ac::;~::;= ==.....;~~---~...-~~+ o L...... _..~=::::;;::::::=:::t:=~~;;;;;;;;;;;;;;;;; o 12 24 36 48 0 24 48 72 96 t [h) t [h) Figure 2.1: Oligomerization of (S)-3b (left-hand side) and (R)-3b (right-hand side) catalyzed by t he enzymes 3-2W4 BapA, Y2 BapA and DmpA. Substrates 3 ( • ), dimer-,d-.JAs 4 , n = 2 (V) and oligomer- ,d-.JAs 4, n > 2 ( .._ ) were detected and quantified at 318 nm, H-,d-.JA (o) at 383 nm. The assay mixtures contained different enzyme concentrations, which are given in the diagrams. 19 Chapter 2 3-2W4 BapA Y2 BapA DmpA 4000 (S)-3b 3000 0 h 0 h 0 h (S)-4b n=2 H-pNA [mAU] 2 h 1 h 1 h 205 2000 n=3 n=4 n=5 n=6 rel. A 8 h 4 h 2 h 24 h 8 h 4 h 1000 48 h 24 h 8 h 96 h 48 h 24 h 0 18 19 20 21 22 18 19 20 21 22 18 19 20 21 22 tR [min] tR [min] tR [min] 3-2W4 BapA Y2 BapA DmpA 4000 (R)-3b 3000 (R)-4b 0 h 0 h n=2 0 h H-pNA [mAU] 2 h 1 h 1 h 205 2000 n=3 rel. A 8 h 4 h 2 h n=4 24 h 8 h 4 h 1000 48 h 24 h 8 h 96 h 48 h 24 h 0 18 19 20 21 22 18 19 20 21 22 18 19 20 21 22 tR [min] tR [min] tR [min] Figure 2.2: Oligomerization of the β3hAla-derived p-nitroanilides (S)-3b (top) and (R)-3b (bottom) in presence of the enzymes 3-2W4 BapA, Y2 BapA and DmpA. The figure shows the HPLC profiles of the assay mixtures after various incubation times. The oligomer-pNAs 4 (n = 2–6) were identified by LC-MS analysis. [M + H]+ for (S)-4b: 309.2 (n = 2), 394.2 (n = 3), 479.2 (n = 4), 564.2 (n = 5), 649.5 (n = 6); [M + H]+ for (R)-4b: 309.3 (n = 2), 394.3 (n = 3), 479.4 (n = 4). 20 Enzyme-Catalyzed {3-Peptide Formation From the figures, we concl ude that 1) all three enzymes generated longer oligomers from (5)-3b (up to six amino acid residues) than from (R)-3b (up to four amino acid residues); ti) for both substrates, oligomer-,d\IAs formed by DmpA were shorter than the corresponding ones generated by 3-2W4 BapA and Y2 BapA; 1/1) DmpA prod uced the highest dimer-,d\IA concentrations (1.2 mM for (5)-4b, n = 2, and 1.9 mM for (R)-4b, n = 2) of all three enzymes and also exhibited the highest specific dimerization activities (cf. Table 2.2). Table 2.2: Comparison of the maximum reached concentrations and the specific activities of dimer- ~A ( 4, n = 2) formation from the ~-amino acid p-nitroanilides 3 catalyzed by the enzymes 3-2W4 BapA, Y2 BapA and DmpA. The initially employed concentration of the substrates 3 was 5 mM. The dipeptide 6b was generated from (S)-3b (5 mM) and ~ 3 -homoleucine Sd (25 mM), carnosine (7a) from 3a (5 mM) and histidine (25 mM) in the presence of the enzymes. One unit (U) is defined as the amount of enzyme that catalyzes the formation of 1 µmol of dimer -~A (4, n = 2) or of the dipeptides 6b and 7a per min. n.d.: not detectable. Target peptide 3-2W4 BapA Y2 BapA DmpA Max. Specific Max. Specific Max. Specific cone. act. cone. act. cone. act. [mM] [U/mg of [mM] [U/mg of [mM] [U/mg of protein] protein] protein] 4a n= 2 0.60 0.18 0.77 0.091 1.8 18.3 n = 3-8 0.51 0.50 0.35 (S)-4b n= 2 0.37 0.54 0.42 0.52 1.2 13.5 n = 3-6 0.13 0.14 0.13 (R)-4b[aJ n= 2 1.5 0.008 1.3 0.005 1.9 0.18 n = 3-4 0.29 0.42 0.40 (S)-4c[b] n= 2 0.08 0.092 0.17 0.006 0.14 < 0.001 n >2 n.d. n.d. n.d. (R)-4c[bJ n= 2 1.1 0.46 0.91 0.007 1.8 0.003 n >2 n.d. n.d. n.d. (S)-4d[a] n= 2 0.28 1.0 0.15 0.015 0.15 0.002 n >2 n.d. n.d. n.d. 6b 1.6 1.3 2.4 1.0 0.41 1.4 7a 1.5 0.13 2.3 0.094 0.64 1.3 [aJ The assay mixture contained 10% DMSO. [bl The assay mixture contained 30% DMSO. 21 Chapter 2 As expected, all oligomers formed were eventually degraded by the enzymes to the β-amino acids 5, so that after long reaction periods the only detectable pNA derivative was p- nitroaniline itself. In general, the degree of oligomerization extended over a range and strongly depended on the employed substrate. The longest oligomer-pNAs 4 were formed from 3a (up to eight βhGly-residues), followed by the β3hAla-oligomers (S)-4b and (R)-4b, whereas only dimer- pNAs were generated from β3hPhe- and β3hLeu-building blocks (Table 2.2, and Figures 2.1 and 2.2). This may be caused by the sterically less demanding side chains (H, Me) of the former two β-amino acid residues or by poor solubility of the oligomer-pNAs containing the hydrophobic β-amino acid residues β3hPhe and β3hLeu. The specific dimerization activities of the enzymes as well as the yields obtained in the dimerization and oligomerization reactions are presented in Table 2.2. The specific activities vary over a range of five orders of magnitude (from less than 0.001 to 18.3 U/mg). The enzyme DmpA acted much faster on the β-amino acid p-nitroanilides 3a, and (S)-3b and (R)-3b than the other enzymes. On the other hand, 3-2W4 BapA and Y2 BapA converted substrates carrying the bulkier side chains of phenylalanine (i.e., (S)-3c and (R)-3c) and leucine (i.e., (S)-3d) faster than DmpA. These observations are in accordance with the expectations from our previous degradation experiments employing the same substrates in a different reaction system.[62] Interestingly, both the (S)- and the (R)-forms of H-β3hAla-pNA ((S)-3b and (R)-3b, respectively) H-β3hPhe-pNA ((S)-3c and (R)-3c, respectively) were accepted as substrates by the enzymes. In the case of the β3hAla derivatives, all three enzymes catalyzed the formation of dimer-pNAs of the (R)-form more slowly than those of the (S)-form; with the β3hPhe derivatives, however, the (R)-form reacted faster than its enantiomer (Table 2.2). In degradation experiments employing H-β3hPhe-pNA as the substrate, all enzymes hydrolyzed the (S)-form faster than the (R)-form.[62] A comparison of the dimer-pNA yields reached with 3-2W4 BapA, Y2 BapA, and DmpA in the coupling of the six β-amino acid p-nitroanilides 3 to oligomer-pNAs 4 reveals the following 22 Enzyme-Catalyzed β-Peptide Formation facts i) the four substrates 3a, (S)-3b, (R)-3b, and (R)-3c were converted in high yields to the corresponding dimer p-nitroanilides 4a, (S)-4b, (R)-4b, and (R)-4c, n = 2, respectively, to reach maximum concentrations between 0.37 and 1.9 mM, which correspond to yields of 15 to 76%; ii) in these four cases, the enzyme DmpA showed the best performance reaching maximum yields from 48 to 76%; iii) the maximum concentrations of the dimer p- nitroanilides (S)-4c and (S)-4d, n = 2, generated from the phenylalanine- and leucine- derived precursors (S)-3c and (S)-3d, respectively, were very small with all three enzymes, i.e., the rate of substrate cleavage exceeded the rate of coupling; iv) the enzymes converted the (R)-enantiomers of 3b and 3c to dimer-pNAs in 1.6- to 4-times, and 5.4- to 14-times higher yields, respectively, than the corresponding (S)-enantiomers. Use of the enzymes 3-2W4 BapA, Y2 BapA, and DmpA for coupling of two different amino acids and for attachment of β-amino acid residues to an α- tripeptide To see whether we can selectively generate dipeptides containing two different β-amino acid residues, or a β- and an α-amino acid residue, we first tested two combinations (Scheme 2.3). Thus, to the aqueous solution containing the β3-homoalanine acid p-nitroanilide (S)-3b (“acyl donor”) and the free amino acid β3-homoleucine 5d (“acyl acceptor”), with a molar ratio of 1:5, we added one of the enzymes 3-2W4 BapA, Y2 BapA, or DmpA (Scheme 2.3a). The time-concentration curves resulting from the HPLC analyses of withdrawn samples are presented on the left-hand side of Figure 2.3, clearly showing formation of the β-dipeptide 6b. 23 Chapter 2 NO2 3-2W4 BapA, O Y2 BapA or O DmpA OO + a) H2NN H2NNOH H H2NOH H-pNA H (S)-3b 5d 6b (1 equiv., 5 mM) (5 equiv., 25 mM) NO RO 2 O 3-2W4 BapA RO O + b) H2NN H2NNOH H H2NOH H-pNA H 3 5d 6a−f (1 equiv., 5 mM) (5 equiv., 25 mM) HN N HN N NO2 3-2W4 BapA, O Y2 BapA or DmpA O + c) H2NN OH OH H H2N H2NN O H-pNA H O 3a H-His-OH 7a (carnosine) (1 equiv., 5 mM) (5 equiv., 25 mM) HN N R O H-His-OH OH H2NN H NO O RO 2 3-2W4 BapA 7a−f d) H2NN H H-pNA R O 3 (1 equiv., 5 mM) OH H-Leu-OH H2NN H O 8a−f NO RO 2 O 3-2W4 BapA O O + H H e) H NN N OH N OH 2 H N N H N N N H 2 2 O H O H-pNA H O H O 3 9 10a−f (1 equiv., 5 mM) (5 equiv., 25 mM) Scheme 2.3: Formation of dipeptides (6–8) and tetrapeptides (10) catalyzed by the enzymes 3- 2W4 BapA, Y2 BapA and DmpA. Due to competing hydrolytic cleavage, the peptides are eventually completely hydrolyzed by the enzymes. For details see Table 2.3. 24 Enzyme-Catalyzed β-Peptide Formation As a second target peptide, we chose the naturally occurring β/α-dipeptide carnosine, H- βhGly-His-OH (7a). Under the same conditions as applied for the formation of the dipeptide 6b, carnosine (7a) was generated from the p-nitroanilide derivative of β-homoglycine (3a) and histidine (Scheme 2.3c and Figure 2.3, right-hand side). The specific enzyme activities as well as the maximum reached concentrations of the dipeptides 6b and 7a are included in Table 2.2. It is interesting to note that the enzyme Y2 BapA, the slowest of the three catalysts, performed best as far as the maximum concentrations of the dipeptides 6b and 7a (corresponding to product yields of 48 and 46%, respectively) are concerned, whereas DmpA was the poorest catalyst in this respect with maximum yields of 8 and 13%, respectively (cf. Figure 2.3). Apart from the desired dipeptides 6b and 7a, we also observed the formation of the oligomer-pNAs (S)-4b and 4a, respectively, as by-products. We made sure that spontaneous, nonenzymatic peptide coupling between the reactants did not occur under the conditions applied in the enzyme-catalyzed reactions. Additionally, we selected the enzyme 3-2W4 BapA to study reactions of the six β-amino acid p-nitroanilides 3 with a five-fold molar excess of free β3-homoleucine (5d; yielding compounds 6a–6f), histidine (yielding 7a–7f), leucine (yielding 8a–8f), or the α-tripeptide H-Val-Ala-Leu-OH (9; yielding 10a–10f). By subjecting withdrawn samples to LC/MS analysis, we could establish the formation of the β-dipeptides 6, of the β/α-dipeptides 7 (apart from 7c, the mother ion of which could not be detected) and 8, as well as of the β/α- tetrapeptides 10 (Table 2.3 and Scheme 2.3). The quantification of the peptide products 6a, 6c–6f, 7b–7f, 8a–8f, and 10a–10f was not possible because authentic standards were not available. These results clearly show that the β-peptidyl aminopeptidase 3-2W4 BapA can be used as a versatile catalyst, coupling a broad range of β-amino acids of (S)- or (R)-configuration to the N-terminus of a second β- or α-amino acid and even to an α-peptide. 25 Chapter 2 6b 7a 5 5 4 4 ~ 3 3-2W4 BapA ~ 3 3-2W4 BapA .s (5.8 µg/ml) .s (29 µg/ml) 0 2 (.) 2 0 0 0 12 24 36 48 0 12 24 36 48 t [h] t (h] 5 5 4 4 Y2 BapA Y2 BapA ~ 3 ~ 3 .s (8.1 µg/ml) .s (81 µg/ml) 0 (.) 2 2 0 0 0 12 24 36 48 0 12 24 36 48 t [h] t (h] 5 5 4 4 ~ 3 DmpA ~ 3 DmpA .s (0.92 µg/ml) .s (0.92 µg/ml) 0 (.) 2 2 o~======~==:=::;:::=-=~ 0 12 24 36 48 36 48 t [h] t (h] Figure 2.3: Formation of the p-dipeptides H-p3hAla-p3hLeu-OH (6b; left-hand side) and of carnosine (7a; right-hand side) ( • ) catalyzed by the enzymes 3-2W4 BapA, Y2 BapA and DmpA. The p-amino acid p-nitroanilides 3a and (.5)-3b ( • ) as well as the oligomer-pNAs 4a and (.5)-4b (V), which were formed as by-products, were detected and quantified at 318 nm, H-pNA (o) at 383 nm. The assay mixtures contained different enzyme concentrations, which are given in the diagrams. 26 VI .j..J • .~ :J Q) Q) ..t:; ·-0 .._Ol.o ro :0 .s .j..J ~ > Q) (§ "'O Q) ..c ~ c £ .j..J ro .._ .!: 0 ~..t:; 2 c "Acyl donor" "Acyl acceptor" E c ~ 0 ·- fil- LJ) :p Cl q: ...._. ro v 3 ~ ,.,- .._ E ro.._ H-p hleu-OH (Sd) H-His-OH H-Leu-OH H-Val-Ala-Leu-OH (9) 'tl VI J2 VI Q) c ~ :Q Q) 0 - ..c ·- 3a Product H-PhGly-p3hLeu-OH (6a) H-PhGly-His-OH (7a) H-PhGly-Leu-OH (Sa) H-PhGly-Val-Ala-Leu-OH (10a) <- I- .j..J .{g ·c:ro c Q) .._0 <(. .._ [M+ Ht 217.0 227.1 203.2 373.1 aI ~ a. :g_ <1' ~ ~ "'O 3 3 3 3 3 V' Q) (.s)-3b Product H-P hAla-P hleu-OH (6b) H-P hAla-His-OH (7b) H-P hAla-Leu-OH (Sb) H-P hAla-Val-Ala-Leu-OH (10b) ~ :g s ~ it] ION .j..J [M+ Ht 231.3 241.0 217.2 387.3 I Q) 0 M°'O c ·- Q) Q) ..c (R)-3b[a] 3 3 3 3 3 roE E .j..J Product H-(R)-p hAla-p hLeu-OH (6c) H-(R)-p hAla-His-OH (7c) H-(R)-p hAla-Leu-OH (Sc) H-(R)-p hAla-Val-Ala-Leu-OH (10c) I >-. • ... CO. N VI r--. c·- [M + Ht 231.3 n.d. 217.3 387.4 N v Q) VI ..c >- .j..J v ro ..cc E .j..J ro (.s)-3c[b] Product H-p3hPhe-p3hLeu-OH (6d) H-p3hPhe-His-OH (7d) H-p3hPhe-Leu-OH (Sd) H-p3hPhe-Val-Ala-Leu-OH (10d) >-(/) ...... e .0 :E [M+ Ht 307.2 317.1 293.1 463.3 "'O 2 -u::?:....J ~ E >- v .0 (R)-3c[b] Product H-(R)-p3hPhe-p3hLeu-OH (6e) H-(R)-p3hPhe-His-OH (7e) H-(R)-p3hPhe-Leu-OH (Se) H-(R)-p3hPhe-Val-Ala-Leu-OH (10e) ClJ"lC Q) N Q) Cl.._..> [M+ Ht 307.2 317.1 293.0 463.2 VI VI 0 Q) "'O .._ "'O c a. ·- :J b_ 0 VI (.s)-3d[a] Product H-p3hLeu-p3hLeu-OH (6f) H-p3hleu-His-OH (7f) H-p3hleu-Leu-OH (Sf) H-p3hLeu-Val-Ala-Leu-OH (10f) v a. ro a. E !: ...... o- Q) [M + Ht 273.3 283.1 259.2 429.4 0 v () .0 .j..J " ro .!!! :::.._ ~ t ....J O CO QJ .j..J I .j..J [aJThe assay mixture contained 10% DMSO. [bJThe assay mixture contained 30% DMSO . •• O. \C) Q) M• vv...._.-o N V VI .j..J ro v o a» "'O c :E >-:p .. ca v a. 'ti t- ~ro a.v "...... Chapter 2 Conclusions The β-peptidyl aminopeptidases BapA isolated from the Gram-negative microorganisms S. xenopeptidilytica 3-2W4 and S. microcystinivorans Y2, and the enzyme DmpA from O. anthropi LMG7991 are able to cleave β3-homoamino acid moieties from the N-terminus of peptides, and were now tested for peptide coupling. All three enzymes were capable of converting p-nitroanilide derivatives of (S)- and (R)-β3-homoamino acids to oligomer-pNAs of various chain lengths, with the dimers generally prevailing. In the case of the smallest substrate (i.e., H-βhGly-pNA (3a)), oligomers containing up to eight β-amino acid residues were detected, whereas only dimers were formed from the sterically demanding substrates (S)-3c, (R)-3c, and (S)-3d, which carry phenylalanyl and leucyl side chains. The BapA enzymes and DmpA differed in their substrate specificities, and the specific dimerization activities of the enzymes varied over a range of five orders of magnitude. In some cases, the maximum concentrations of dimer-pNA product corresponded to high yields of up to 76%. After long reaction times, the coupling products were subsequently cleaved, and the reaction mixtures consisted entirely of the free amino acids and p-nitroaniline, as expected. By using a five-fold molar excess of one β-amino acid or a peptide over the pNA derivative of a β-amino acid, we observed the enzyme-catalyzed formation of mixed β-dipeptides, β/α- dipeptides, and β/α-tetrapeptides. Finally, it is important to note that the enzymes described here are the first peptidases which were applied for coupling of various β-amino acids to peptides. The unique feature of catalyzing reactions with many N-terminal β-amino acid residues of peptides may suggest interesting practical applications for the aminopeptidases 3-2W4 BapA, Y2 BapA, and DmpA, which are not amenable to conventional enzymes. Especially attractive is the enzyme-catalyzed attachment of a β-amino acid to the N-terminus of a “natural” α-peptide because N-terminal β-amino acid residues may be considered as protective groups against exopeptidases in vitro and in vivo.[18,25-31,72,73] 28 Enzyme-Catalyzed β-Peptide Formation Experimental Section General remarks We analyzed substrates and peptide products quantitatively by reversed-phase HPLC on a Dionex HPLC system equipped with a P680 pump, an ASI-100 automated sample injector, and a UVD340U photodiode array detector (Dionex, Sunnyvale, CA, USA). Separation of the compounds was achieved on a Nucleosil 100-5 C18 column (250 × 4 mm; Macherey-Nagel, Düren, Germany), which was equilibrated with 0.1% CF3COOH (TFA) in H2O. Oligomer-pNAs and peptides generated from the amino acid p-nitroanilides 3a, (S)-3b, and (R)-3b were separated with a linear gradient from 0 to 30% MeCN. A gradient from 0 to 60% MeCN was applied for the separation of all compounds derived from (S)-3c, (R)-3c, and (S)-3d. We detected pNA derivatives of β-amino acids and oligomer-pNAs at a wavelength of 318 nm, free p-nitroaniline at 383 nm, and free peptides at 205 nm. Quantification of pNA derivatives, free p-nitroaniline, and of the dipeptides 6b and 7a was achieved on the basis of standard curves with H-βhGly-pNA (3a), p-nitroaniline, H-β3hAla-β3hLeu-OH (6b), and carnosine (7a), respectively. We recorded mass spectra of the products with an API 4000 liquid chromatography/tandem MS system connected to an Agilent 1100 LC system. p-Nitroaniline and carnosine (7a) were purchased from Sigma-Aldrich (Buchs, Switzerland), and compound 3a from Bachem (Bubendorf, Switzerland). The preparation of the β-amino acid p- nitroanilides (S)-3b, (S)-3c, and (R)-3c,[62] and of the peptides 6b and 9[56] had been described previously. For protein determination, we used Bradford reagent (5 ×) from Bio- Rad (Reinach, Switzerland); absorbance measurements were performed at 595 nm with a Specord S 100 spectrophotometer (Analytik Jena, Jena, Germany). General procedure for the preparation of p-nitroanilides of β3-homoamino acids (GP 1)[C] To a soln. of the Boc-protected amino acid (1 equiv.) and p-nitroaniline (1.5 equiv.) in pyridine (0.2 M) at -20°C (MeOH/ice), phosphoroxychloride (1.3 equiv.) was slowly added, and the suspension was stirred under rewarming to 20°C for 6 h. The mixture was diluted with AcOEt (200 ml), ice was added, and the org. phase was extracted with 6 N HCl (3 × 50 [C] The β-amino acid p-nitroanilides were prepared by Michael Limbach and Oliver Flögel, ETH Zürich. 29 Chapter 2 ml), H2O (50 ml), and brine (50 ml), and dried (Na2SO4). The residue was suspended in CH2Cl2 (50 ml), TFA (5 ml) was added at 20°C, and the soln. was stirred for 12 h. Then, the volatiles were removed in the vacuum. The residue was co-evaporated with toluene (3 ×) and with CH2Cl2 (3 ×). The remaining residue was precipitated from THF/pentane to give the desired β-amino acid p-nitroanilides as slightly yellow solids. TFA·H-(R)-β3hAla-pNA ((R)-3b): Reaction of Boc-(R)-β3hAla-OH (1.01 g, 5.0 mmol), p-nitroaniline (1.04 g, 7.5 mmol), and phosphoroxy chloride (0.6 ml, 6.55 mmol) according to GP 1 yielded 877 mg (52%) of (R)- 20 1 3b. M.p. 150–1548. [α]D = - 22.9 (c = 2.0, MeOH). H NMR (300 MHz, CD3OD): 1.39 (d, J = 6.3, Me), 2.70–2.95 (m, CH2), 3.69–3.73 (m, CH), 7.80 (d, J = 8.9, 2 arom. H), 8.16 (d, J 13 = 8.9, 2 arom. H). C NMR (75 MHz, CD3OD): 19.0, 41.1, 45.2, 120.7, 126.0, 144.9, 146.0, 171.1. MALDI-MS: 235 (27), 224 (100, M+), 208 (30). HR-MS: 224.1029 ([M + H]+, + C10H13N3O3 ; calcd. 224.1030). TFA·H-(S)-β3hLeu-pNA ((S)-3d): Reaction of Boc-(S)-β3hLeu-OH (3.83 g, 15.6 mmol), p-nitroaniline (3.23 g, 23.4 mmol), and phosphoroxy chloride (1.86 ml, 6.55 mmol) according to GP 1 yielded 2.61 g (44%) of (S)- 20 1 3d. M.p. 177–1788. [α]D = + 22.6 (c = 2.0, MeOH). H NMR (300 MHz, CD3OD): 0.97 (d, J = 4.1, Me), 0.99 (d, J = 4.1, Me), 1.47–1.70 (m, CH2), 1.70–1.89 (m, CH), 2.74 (dd, J = 16.6, 7.7, CH2), 2.90 (dd, J = 16.6, 3.4, 1 H, CH2), 3.61–3.80 (m, CH), 7.81 (d, J = 9.1, 2 13 arom. H), 8.16 (d, J = 9.1, 2 arom. H). C NMR (75 MHz, CD3OD): 22.7, 22.9, 25.6, 38.3, 43.0, 120.6, 125.9, 144.8, 145.9, 171.2. MALDI-MS: 266 (100, M+), 237 (27). HR-MS: + + 266.1501 ([M + H] , C13H19N3O3 ; calcd. 266.1499). Enzyme preparation and general procedure for the enzyme-catalyzed formation of β-amino acid oligomers and of mixed β- and β/α-peptides The recombinant β-peptidyl aminopeptidases 3-2W4 BapA, Y2 BapA, and DmpA were purified from their E. coli hosts as described in [59,62]. The lyophilized enzyme powders were dissolved in a universal buffer[92] at pH 10, and the protein content of the enzyme solutions was determined according to Bradford[93] on basis of a standard curve with bovine serum albumine (BSA). To study the oligomerization of β-amino acids, we used a reaction mixture 30 Enzyme-Catalyzed β-Peptide Formation 3 containing 5 mM H-β hXaa-pNA (3) in universal buffer (pH 10) at 37°C. For the generation 3 of mixed β- and β/α-peptides, a five-fold molar excess (25 mM) of β -homoleucine (5d), histidine, leucine, or of the α-tripeptide H-Val-Ala-Leu-OH (9) was added to the oligomerization assay mixtures. The reaction was started by the addition of a limiting amount of one of the enzymes 3-2W4 BapA, Y2 BapA, or DmpA. Samples were withdrawn regularly and the enzymatic reaction was stopped by the addition of 0.25% (v/v) 1 M HCl. The compounds present in the samples were analyzed by HPLC, and their masses were confirmed by LC/MS. 31 Kinetic Analysis of L-Carnosine Formation 3. Kinetic Analysis of L-Carnosine Formation by β- Aminopeptidases Abstract The β,α-dipeptide L-carnosine occurs in high concentrations in long-lived innervated mammalian tissues and is widely sold as a food additive. On large scale L-carnosine is produced by chemical synthesis procedures. We established two aqueous enzymatic reaction systems for L-carnosine preparation using the soluble bacterial β-aminopeptidases DmpA from O. anthropi and BapA from S. xenopeptidilytica as catalysts and investigated the kinetics of the enzyme-catalyzed peptide couplings. DmpA catalyzed the formation of L- carnosine from C-terminally activated β-alanine derivatives (acyl donor) and L-histidine (acyl acceptor) in an aqueous reaction mixture at pH 10 with high catalytic rates (vmax = 19.2 -1 -1 μmol min per mg of protein, kcat = 12.9 s ), whereas vmax in the BapA-catalyzed coupling -1 -1 reaction remained below 1.4 μmol min per mg of protein (kcat = 0.87 s ). Although the equilibrium of this reaction lies on the side of the hydrolysis products, the reaction is under kinetic control and L-carnosine temporarily accumulated to concentrations, which correspond to yields of more than 50% with respect to the employed acyl donor. However, competing nucleophiles caused unwanted hydrolysis and coupling reactions that led to decreased product yield and to formation of various peptidic byproducts. The substitution of L-histidine for L-histidine methyl ester as acyl acceptor shifted the pKa of the amino functionality from 9.25 to 6.97, which caused a drastic reduction in the amount of coupling byproducts in an aqueous reaction system at pH 8. This chapter by T. Heck, V. S. Makam, J. Lutz, L. M. Blank, A. Schmid, D. Seebach, H.-P. E. Kohler, and B. Geueke was published in Advanced Synthesis & Catalysis 2010, 352, 407-415. 33 Chapter 3 Introduction β-Peptides and mixed β,α-peptides consisting of homologues of the 20 proteinogenic amino acids are promising compounds for the design of novel peptidomimetics because the incorporation of β-amino acids into peptides enhances proteolytic and metabolic stability.[25,30-35,43] Naturally occurring peptides that exclusively consist of β-amino acids are not known so far. However, compounds containing β-peptidic substructures are common in nature. Prominent examples include pantothenic acid, cyanobacterial microcystins, and the [D] mammalian β,α-dipeptide L-carnosine (β-alanine-L-histidine). As one of the few β-amino acid-containing peptides in mammalian metabolism, L-carnosine accumulates to high concentrations in innervated tissues, such as skeletal muscle and brain. L-Carnosine is not a product of the translational machinery, but is formed by a specific enzyme, carnosine [94-97] synthetase (EC 6.3.2.11), through ATP-dependent coupling of β-alanine and L-histidine. Levels of L-carnosine are controlled by tissue and serum carnosinases (EC 3.4.13.3 and 3.4.13.20, respectively), which hydrolyze L-carnosine into its amino acid constituents, β- [7,80,98] alanine and L-histidine. L-Carnosine exhibits antioxidant, anti-glycating and hydroxyl- radical scavenger properties[3-5,99,100] and has been implicated in neuroprotection and prevention of diabetes-induced nephropathy and artherosclerosis.[101-103] Due to its proposed beneficial biological activities L-carnosine is widely sold as food additive with anti-ageing potential and in complex with zinc as an anti-ulcer drug.[104,105] The chemical bulk production of L-carnosine is well established and different chemical routes are available.[106,107] As alternatives to chemical peptide synthesis in solution or on solid- phase, enzymatic approaches with proteases as catalysts have gained increasing attention. The use of enzymes as biocatalysts is often advantageous due to mild reaction conditions, high stereoselectivity of the biocatalyst and low protective group requirements.[87-89,108,109] However, most proteases and peptidases are limited to reactions with the 20 proteinogenic α-amino acids. Thus, β-amino acid-containing peptides are generally unusual targets for enzymatic conversions. [D] β-Alanine is called β-homoglycine (H-βhGly-OH) according to the nomenclature for β-amino acids with proteinogenic side chains.[1] 34 Kinetic Analysis of L-Carnosine Formation Three members of the N-terminal hydrolase (Ntn) superfamily,[67] namely the β- aminopeptidases BapA from S. xenopeptidilytica 3-2W4, BapA from S. microcystinivorans Y2, and DmpA from O. anthropi were previously shown to cleave N-terminal β-amino acid residues from peptides that are otherwise resistant to proteolytic breakdown.[59,60,63,74] In recent investigations, we characterized the biochemical properties of these enzymes regarding the hydrolysis of β- and mixed β,α-peptides[62] and reported the usefulness of the β-aminopeptidases as catalysts for the efficient kinetic resolution of various aliphatic β-amino acid amides (see Chapter 4).[110] In addition to their hydrolytic properties, we showed that DmpA and BapA were able to catalyze the formation of β-amino acid oligomers and other β- amino acid-containing peptides from C-terminally activated β-amino acid building blocks (see Chapter 2).[111] In another study with the lipases CAL-A and CAL-B from Candida antarctica and with Burkholderia cepacia lipase PS-D, several β-dipeptides and derivatives of L- carnosine were successfully synthesized in organic solvents.[91,112,113] For the peptide couplings catalyzed by the dissolved β-aminopeptidases, we employed β- amino acid p-nitroanilides (H-βhXaa-pNA) or β-amino acid amides (H-βhXaa-NH2) as mildly activated acyl donors. Upon the release of the leaving group (p-nitroaniline or ammonia, respectively) they form a characteristic acyl enzyme complex (Scheme 3.1). In this kinetically controlled reaction system,[114] the free amino group of a second amino acid or a small peptide acts as a nucleophile (acyl acceptor) that accepts the acyl moiety from the active site leading to formation of the desired peptide product. In an aqueous reaction mixture the acyl enzyme can undergo competing hydrolysis resulting in the free amino acid, which is consequently unavailable for enzyme acylation. Since β-aminopeptidases require substrates with a free amino terminus,[63] we used unprotected acyl donors for the coupling reaction. Due to the presence of the free amino terminus the acyl donor itself can behave as a nucleophilic acceptor for an acyl moiety bound to the active site, which leads to dimerization of the acyl donor. The dimer as well as the peptide product may undergo secondary hydrolysis or behave as further nucleophiles that can react with the acyl enzyme. Due to secondary hydrolysis of the newly formed peptide bond the product accumulates to a clear-cut maximum in a kinetically controlled reaction system. To obtain maximum product yields, a kinetic analysis of the complex interplay between product formation and competing coupling and hydrolysis reactions is indispensable. 35 Chapter 3 R1 O X 1 1 H2NN R O R O H X + H Enzyme dimerization H2NNN H H 1 1 2 1 R O R O H2NR R O 2 X + H Enzyme R + H Enzyme H2NN H2N Enzyme product H2NN H formation H XNH2 R1 O hydrolysis + H Enzyme H2NOH H2O Scheme 3.1: β-Aminopeptidase-catalyzed peptide formation under kinetic control. Side reactions caused by competing nucleophiles lead to dimerization and hydrolysis of the acyl donor (Scheme modified from Heck et al.).[111] Possible further coupling reactions and secondary hydrolysis of the newly formed compounds are not depicted in the scheme. Due to its biological and commercial relevance, we chose L-carnosine (H-β-Ala-L-His-OH) as a target compound to investigate β-aminopeptidase-catalyzed peptide couplings in aqueous solution in more detail. As reported earlier, L-carnosine was found to be formed by the β- aminopeptidases DmpA from O. anthropi and BapA from S. xenopeptidilytica 3-2W4 in solution[111] and in recombinant yeast and bacterial whole cell systems producing these enzymes.[115] In the present investigation, we optimized the β-aminopeptidase catalyzed formation of L-carnosine by applying the kinetic model presented in Eq. 3.2. We evaluated the synthesis reaction in terms of product yields and kinetic parameters at different acyl acceptor concentrations. In order to reduce the amount of peptidic byproducts a second approach was followed employing a C-terminally modified acyl acceptor. Results and Discussion Nucleophile reactivity in β-aminopeptidase-catalyzed L-carnosine formation To optimize the product yield of β-aminopeptidase-catalyzed peptide formation and to compare the synthetic performances of DmpA and BapA we adopted a kinetic model for acyl group transfer by proteases that form an acyl enzyme intermediate.[116,117] In previous studies, this model was successfully applied for a kinetic evaluation of the biosynthesis of β- lactam antibiotics by penicillin acylase.[118,119] Accordingly, the synthetic efficiency of peptide 36 Kinetic Analysis of L-Carnosine Formation formation under kinetic control depends on the relative reactivity of the nucleophile and is therefore referred to as “nucleophile reactivity”. In the case of penicillin acylase-catalyzed acyl transfer, the nucleophile reactivity of the acyl acceptor is defined as the ratio between the initial rate of product synthesis (vS) and acyl donor hydrolysis (vH). Improved ratios of (vS/vH) can be directly correlated with increased product yields. The synthetic efficiency of the coupling reaction is quantitatively described by the reaction parameters β0 and γ, and hyperbolically depends on the initial concentration of the employed nucleophile [Nu] (Eq. 3.1). The parameter β0 represents the preference of the acyl enzyme to react with the nucleophilic acyl acceptor instead of water and should be high for good synthetic efficiency.[116] The introduction of the parameter γ takes into account the formation of an acyl enzyme-nucleophile complex between the acyl enzyme and the added nucleophile. This acyl enzyme-nucleophile complex is still prone to hydrolysis. γ characterizes the intensity of hydrolysis of the acyl enzyme-nucleophile complex. In contrast to β0, γ should be low for good synthetic efficiency as under a saturating concentration of acyl acceptor the initial synthesis/hydrolysis ratio reaches a maximum value of 1/γ.[117] ⎛ v ⎞ β ⋅[]Nu ⎜ S ⎟ = 0 (Eq. 3.1) ⎜ ⎟ γβ ⋅⋅+ ⎝ vH ⎠ini 1 0 []Nu The determination of nucleophile reactivity in β-aminopeptidase-catalyzed peptide couplings is more complex, because besides the hydrolysis of the acyl donor the dimerization of the acyl donor occurs as a second undesired side reaction that has to be considered (Scheme 3.1). Hence, we define the nucleophile reactivity of the acyl acceptor for β-aminopeptidase- catalyzed peptide synthesis as the ratio between the initial reaction rate of product formation (vS) and the sum of the initial reaction rates for acyl donor hydrolysis (vH) and dimer formation (vD) (Eq. 3.2). ⎛ v ⎞ β ⋅[]Nu ⎜ S ⎟ = 0 (Eq. 3.2) ⎜ + ⎟ γβ ⋅⋅+ ⎝ vv DH ⎠ini 1 0 []Nu 37 Chapter 3 To compare the synthetic efficiency of L-carnosine formation by the β-aminopeptidases DmpA from O. anthropi and BapA from S. xenopeptidilytica 3-2W4 we determined the nucleophile reactivity of the acyl acceptor L-histidine at two different pH values. Due to easy detectability of substrates and products we employed the chromogenic p-nitroanilide derivative of β-alanine (H-β-Ala-pNA) as activated acyl donor (Scheme 3.2). We measured the initial rates of L-carnosine formation as well as the rates of acyl donor hydrolysis and dimer formation at different concentrations of L-histidine, which served as acyl acceptor. The synthetic efficiency of β-aminopeptidase-catalyzed L-carnosine formation was quantified as the ratio (vS/(vH+vD))ini (Figure 3.1). Table 3.1 summarizes the values for the reaction parameters β0 and γ that were obtained by fitting experimental data to Eq. 3.2. HN N HN N 9.12 O 9.25 9.13 O β-aminopeptidase O O H NNHX+ + H2NX 2 H2N pH 10, 37°C H2N N H O O β-alanine p-nitroanilide L-histidine L-carnosine (X=C6H4NO2) β-alanine amide (X=H) Scheme 3.2: Formation of L-carnosine from H-β-Ala-pNA or H-β-Ala-NH2 (acyl donor) and L-histidine (acyl acceptor) catalyzed by β-aminopeptidases. Only the predominant charge species of the reactants at the given pH are depicted in the scheme. Theoretical pKa values of the protonated amino functionalities are given in frames. As expected from our previously published results, DmpA catalyzed the formation of L- carnosine with much higher initial rates than BapA (see Chapter 2).[111] Despite the low overall activities for L-carnosine formation, the synthetic performance (vS/(vH+vD))ini of the BapA-catalyzed coupling reaction was superior to DmpA (Figure 3.1). 38 Kinetic Analysis of L -Carnosine Formation 0.6 ~ a::. + ::.:r: 0.4 .._...... Ci) ...... ::. 0.2 0 10 20 30 40 50 60 L-histidine [mM] Figure 3.1: Nucleophile reactivity of L-histidine in the formation of L-carnosine from H-~-Ala-,d-.JA (6 mM) and varying concentrations of L-histidine by the ~-aminopeptidases DmpA (circles) and BapA (squares). Reactions were carried out at 37°C and the pH values of the reaction systems were set to pH 8 (filled symbols) or pH 10 (open symbols). Experimental data were fitted to Eq. 3.2. For both enzymes we observed a drastic improvement of the synthetic performance at elevated nucleophile concentrations. Furthermore, the choice of the pH value had a large influence on the nucleophile reactivity of the acyl acceptor. An increase in pH of the reaction system from pH 8 to pH 10 kept L-histidine predominantly in its deprotonated state and thus improved the synthetic performance of both catalysts. This led to a 2.6- and 2.8-fold increase of the reaction parameter j30 for DmpA and BapA, respectively, and to a 2.8- and 3.4-fold decrease of y, which indicates more efficient formation of L-carnosine at pH 10 (Eq. 3.2 and Table 3.1). It is important to mention that the highly alkaline pH of the reaction system did not affect the stability of the peptide product and the enzymes. From these experiments we conclude that (1) an alkaline reaction system is a prerequisite for reaching high product titers and (Ji) the acyl acceptor has to be used in excess in order to favor product formation over acyl donor hydrolysis and dimerization. 39 Chapter 3 Table 3.1: Reaction parameters for ~-aminopeptidase-cata lyzed L-carnosine formation from H-~-Ala ~A (6 mM) and L-histidine (6 to 60 mM) obtained by fitting experimental data to Eq. 3.2. Enzyme pH r DmpA 8 0.72 0.67 10 1.9 0.24 BapA 8 4.6 0.37 10 13 0.11 L-Carnosine formation from H-J3-Ala-NH 2 and L-histidine I n order to avoid the ecologically harmful leaving group trnitroaniline and to enhance the solubility of the acyl donor we employed p-alanine amide (H-p-Ala-NH2) instead of H-p-Ala- µ.JA for all further experiments. As illustrated before, the synthetic performance of DmpA and BapA could be greatly improved by employing high nucleophile concentrations in a basic reaction system. Hence, we investigated the enzyme-catalyzed formation of L-carnosine by varying concentrations of H-p-Ala-NH2 at four concentrations of L-histidine (15, 30, 60 and 120 mM) in a reaction system at pH 10. L-Histidine was always used in at least 1.25-fold molar excess over the activated acyl donor and its concentration was assumed to be constant during the initial phase of product formation. Hence, we determined the initial synthetic rates as well as the maximum accumulation of L-carnosine in the reaction mixture (Figure 3.2). DmpA catalyzed the formation of L-carnosine with high specific initial rates of up to 16 µmol min-1 per mg of protein, whereas the catalytic rates of BapA rema ined below 1 0.5 µmol min- per mg of protein. At elevated concentrations of H-p-Ala-NH2 dimerization of the acyl donor becomes more likely to occur, which led to decreased initial rates of L- carnosine formation in the DmpA-catalyzed reaction. As expected in a kinetically controlled reaction, the coupling product accumulated to a maximum and was subsequently hydrolyzed by the enzymes. The maximum achieved concentrations of L-carnosine strongly depended on the employed excess of L-histidine over H-p-Ala-NH2. We observed maximum L-carnosi ne yields of 40 and 53% with respect to the initial concentration of the rate-limiting acyl donor with the enzymes DmpA and BapA, respectively. The yield was substantially lower when the concentration of the nucleophile was decreased. This result is in agreement with our 40 Kinetic Analysis of L -Carnosine Formation observations on the nucleophile reactivity of L-histidine in the 13-aminopeptidase-catalyzed coupling reactions using the p-nitroanilide derivative of 13-alanine as acyl donor instead of 13- alanine amide (Figure 3.1). Although L-carnosine is formed by DmpA with much higher initial rates than it is formed by BapA, the latter is the more efficient catalyst for t he accumulation of higher product concentrations in the reaction mixtures. c 18 DmpA 0.6 BapA -"iii .....0 15 0.5 -~ ...-0 12 0.4 ' O> ...E 9 0.3 ' c .E 6 0.2 0 E 3 0.1 ~ :::. 0 0.0 0 10 20 30 40 50 0 10 20 30 40 50 50 50 r. 40 40 D • ! •• • D • 0~ c D ~ De 30 D • 30 • 'O -• D • D • Cl> ~ '- D • • 20 • 20 0 D >= 0 • 0 D • COo • 0 0 10 0 • 10 0 0 0 10 20 30 40 50 0 10 20 30 40 50 H-13-Ala-NH2 [mM] H-13-Ala-NH2 [mM] Figure 3.2: Initial rates of L-carnosine formation from H-p-Ala-NH2 in the presence of excess concentrations of L-histidine ( o 15 mM, • 30 mM, o 60 mM, • 120 mM) catalyzed by the p- aminopeptidases DmpA and BapA at 37°C and pH 10. Experimental data were fitted to Eq. 3.3 or Eq. 3.4, respectively (see also Table 3.2). Yields are given in percent of L-carnosine produced per initial H-p-Ala-NH2 supplied. 41 Chapter 3 In order to estimate the kinetic reaction parameters Km and Vmax of the enzymatic formation of L-carnosine, the data sets obtained for both enzymes in the presence of excess L-histidine were fitted to Michaelis Menten models (Eq. 3.3 and Eq. 3.4). L-Carnosine formation catalyzed by BapA at L-histidine concentrations of 15 and 30 mM followed the classical Michaelis-Menten model (Eq. 3.3), whereas the other data sets showed decreased activities at elevated concentrations of H-β-Ala-NH2 and hence were fitted to the substrate surplus inhibition model (Eq. 3.4). The kinetic reaction parameters for β-aminopeptidase catalyzed L- carnosine formation are summarized in Table 3.2. []S Vv max ⋅= (Eq. 3.3) []+ KS m []S Vv max ⋅= (Eq. 3.4) []S 2 [] KS m ++ K s The calculated Km values for the acyl donor differed significantly among the enzymes. DmpA had a high affinity for H-β-Ala-NH2 with Km values below 1 mM, whereas the Km values of BapA were > 11.6 mM (Table 3.2). The catalytic rates for the conversion of H-β-Ala-NH2 and L-histidine to L-carnosine strongly depended on the employed excess of L-histidine. With -1 DmpA the calculated Vmax values for L-carnosine formation ranged between 7.4 μmol min -1 -1 -1 per mg of protein (kcat = 5.0 s ) and 19.2 μmol min per mg of protein (kcat = 12.9 s ), -1 whereas the catalytic rates of BapA remained below 1.4 μmol min per mg of protein (kcat = -1 0.87 s ). When we used initial L-histidine concentrations of 15 and 30 mM and acyl donor concentrations below 12 and 24 mM, respectively, no substrate surplus inhibition was observed for BapA (Figure 3.2). Despite the high synthetic rates in the DmpA-catalyzed reaction, L-carnosine did not accumulate to yields higher than 40% due to competing hydrolysis and coupling reactions. Higher product titers could be reached in the BapA- catalyzed reactions although the enzyme catalyzed the formation of L-carnosine much more slowly than DmpA. This result is supported by the higher nucleophile reactivity of L-histidine observed in the BapA-catalyzed reaction (see Figure 3.1), which may be caused by better accessibility of the enzyme’s active site for the nucleophile. 42 Kinetic Analysis of L -Carnosine Formation Table 3.2: Kinetic parameters of L-carnosine formation by DmpA and BapA determined at varying concentrations of H-~-Ala-NH 2 and excess of L-histidine. Experimental data were fitted to the Michaelis-Menten equation with substrate surplus inhibition (Eq. 3.4), except curves for BapA at 15 and 30 mM L-histidine that followed the classical Michaelis-Menten model (Eq. 3.3). n.d not determined. DmpA BapA L-histidine Km Ks Vmax kcat Km Ks Vmax kcat [mM] [mM] [mM] [µmol [s-1] [mM] [mM] [µmol [s-1] min-1 per min-1 per mg of mg of protein] protein] 15 0.98 15.4 7.4 5.0 11.6 n.d. 0.086 0.055 ± 0.13 ± 2.9 ± 0.48 ± 0.32 ± 2.9 ± 0.01 ± 0.01 30 0.42 28.7 10.6 7.1 18.4 n.d. 0.27 0.17 ± 0.06 ± 4.9 ± 0.50 ± 0.34 ± 0.95 ± 0.01 ± 0.01 60 0.45 29.4 16.7 11.2 19.9 45.8 0.70 0.45 ± 0.07 ± 4.0 ± 0.72 ± 0.48 ± 7.7 ± 25 .1 ± 0.19 ± 0.12 120 0.48 56.4 19.2 12.9 23.0 32.2 1.4 0.87 ± 0.05 ± 7.0 ± 0.59 ± 0.40 ± 8.0 ± 14.4 ± 0.35 ± 0.23 L-Carnosine formation from H-J3-Ala-NH2 and L-histidine methyl ester I n order to reduce the amount of byproducts we followed a second approach for the DmpA- catalyzed formation of L-carnosine using L-histidine methyl ester instead of free L-histidine as acyl acceptor. Esterification of the carboxyl terminus increased its electron-attracting effect and reduced the pKa value of the amino group by more than two log units (from pKa 9.25 to 6.97). Hence, it was possible to lower the pH of the reaction system from pH 10 to pH 7 or 8, respectively. Under these conditions the nudeophilic character of the acyl acceptor was maintained while at the same time the amino termini of the acyl donor and the peptide product were mostly protonated. The enzyme-catalyzed formation of L-carnosine methyl ester was subsequently followed by hydrolysis of the ester bond under alkaline conditions and heating to yield the final product L-carnosine (Scheme 3.3). The stability of the peptide bond of L-carnosine was not affected under these conditions. 43 Chapter 3 HN N HN N 9.12 O 6.97 9.13 O DmpA O O H NNH+ + NH3 3 2 H2N pH 8, 37°C H3N N H O O β-alanine amide L-histidine L-carnosine methyl ester methyl ester alkaline hydrolysis 0.5 M NaOH, 7 min, 95°C HN N O O + MeOH H2N N H O L-carnosine Scheme 3.3: Formation of L-carnosine from H-β-Ala-NH2 (acyl donor) and L-histidine methyl ester (acyl acceptor) catalyzed by DmpA. Only the predominant charge species of the reactants at the given pH are depicted in the scheme. Theoretical pKa values of the protonated amino functionalities are given in frames. In a reaction mixture containing 6 mM H-β-Ala-NH2 and 30 mM L-histidine methyl ester DmpA catalyzed the formation of the methyl ester derivative of L-carnosine. Before HPLC analysis, L-carnosine methyl ester was chemically hydrolyzed under alkaline conditions to the free peptide product L-carnosine. We observed a four-fold increase in the catalytic rates of L- carnosine formation when the pH of the reaction system was raised from pH 7 (2.8 μmol -1 -1 min per mg of protein) to pH 8 (11.9 μmol min per mg of protein). L-Carnosine methyl ester temporarily accumulated to maximum concentrations of 0.4 mM (7% yield of L- carnosine after alkaline hydrolysis) in the reaction system at pH 7 and to 1.2 mM (20% yield of L-carnosine) at pH 8 before enzymatic hydrolysis of the product outbalanced its formation (Figure 3.3). With L-histidine as acyl acceptor at pH 10 we obtained a specific activity for L- -1 carnosine formation of 7.8 μmol min per mg of protein and a maximum accumulation of L- carnosine of 1.1 mM (18% yield). Product titers could be further increased by feeding of an additional 6 mM H-β-Ala-NH2 to the reaction mixtures. The HPLC chromatograms of the analyzed samples reveal that several byproducts were produced during the formation of L- carnosine (Figure 3.3, bottom row). 44 Kinetic Analysis of L-Carnosine Formation 2.5 2.5 2.5 A B C 2.0 2.0 2.0 1.5 1.5 1.5 1.0 1.0 1.0 L-carnosine [mM] 0.5 0.5 0.5 0.0 0.0 0.0 0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6 Time [h] Time [h] Time [h] 10 10 10 L-carnosine L-carnosine 1 L-carnosine 8 8 8 4 6 6 6 2 [%] 205 4 4 4 E 1 4 2 2 2 2 3 0 0 0 3 4 5 6 7 3 4 5 6 7 3 4 5 6 7 Retention time [min] Retention time [min] Retention time [min] Figure 3.3: DmpA-catalyzed formation of L-carnosine (●) from H-β-Ala-NH2 (6 mM) and L-histidine methyl ester (30 mM) in 300 mM MOPS buffer at pH 7 (A) and in 300 mM Tris/HCl buffer at pH 8 (B) after subsequent alkaline hydrolysis of the ester bond (see Scheme 3.3), and from H-β-Ala-NH2 (6 mM) and L-histidine (30 mM) in 100 mM sodium carbonate buffer at pH 10 (C) (see Scheme 3.2). After 1 h an additional 6 mM H-β-Ala-NH2 was fed to the reaction mixtures (○). The lower diagrams show the corresponding HPLC diagrams at the point of maximum product titer in a batch synthesis (A at 2 h, B at 1.5 h, C at 2 h). Peaks are normalized to L-carnosine (100%). The observed byproduct peaks correspond to 1 H-(β-Ala)2-NH2, 2 H-(β-Ala)2-OH, 3 H-(β-Ala)3-NH2, 4 H-(β-Ala)2-L- His-OH (see Table 3.3). All observed byproducts were of peptidic nature and were formed by enzyme-catalyzed couplings between acyl donor molecules (oligomerization) and by the N-terminal extension of the product L-carnosine leading to H-β-Ala-β-Ala-L-His-OH (Table 3.3). The hydrolysis of the acyl donor, which leads to a significant loss of substrate during the reaction and hence to reduced product yields, could not be monitored because H-β-Ala-OH was not detectable with the employed HPLC method. Although similar yields of L-carnosine were obtained in the reaction systems at pH 8 and 10, the accumulation of peptidic byproducts was considerably reduced when L-histidine methyl ester was used as the acyl acceptor at pH 8 instead of L- histidine at pH 10 (Figure 3.3). This observation can be explained by the fact that the free amino groups of H-β-Ala-NH2 and the product L-carnosine have similar pKa-values around 9.1 (Scheme 3.2) and also act as potent nucleophiles at pH 10, which leads to formation of the undesired coupling byproducts. In contrast, the amino termini of H-β-Ala-NH2 and L- 45 Chapter 3 carnosine are mostly in the protonated state when the pH of the reaction system is lowered to pH 8 and their ability to react as nucleophiles is drastically red uced. However, L-histidine methyl ester (pl Table 3.3: HPLC-MS analysis of the byproducts formed during the formation of L-carnosine from H- p-Ala-NH2 and L-histidine or from H-p-Ala-NH2 and L-histidine methyl ester after subsequent alkaline hydrolysis of the ester bonds. Peak numbers correspond to the labels in Figure 3.3. Compound Peak tR [min] Calcd. mass [M+ Hr L-carnosine 3.9 226.1 227.1 H-(p-Ala)rNH2 1 3.3 159.1 160.0 H-(p-Ala)rOH 2 4.6 160.1 161.1 H-(p-Ala)rNH2 3 5.7 230.1 231.3 H-(p-Ala)i-L-His-OH 4 6.4 297.1 298.2 Conclusions I n the present study, the kinetics of 13-aminopeptidase-catalyzed L-carnosine formation from (-terminally modified 13-alanine derivatives and L-histidine or L-histidine methyl ester were investigated in aqueous solution. In this kinetically controlled reaction the transfer of the acyl moiety from the acyl enzyme to competing nucleophiles led to the formation of undesired byproducts. The efficiency of L-carnosine formation was strongly pH dependent and the best results were obtained at pH 10. DmpA catalyzed the formation of L-carnosine with high rates. In contrast, BapA exhibited much lower rates, but it was the catalyst with the higher synthetic performance, wh ich led to the temporary accumulation of prod uct yields above 50%. The formation of peptidic byproducts could be red uced when L-histidine methyl ester was used as acyl acceptor in a reaction system at pH 8, rather than free L-histidine at pH 10. Our approach to enhance the nucelophilic character of the acceptor amino acid by modifying its caboxy terminus cou ld generally serve as a useful strategy to reduce the amount of unwanted byprod ucts in enzyme-catalyzed peptide couplings. 46 Kinetic Analysis of L-Carnosine Formation The formation of L-carnosine can be regarded as a model reaction for the β-aminopeptidase- catalyzed synthesis of peptides with N-terminal β-amino acids. Due to its broad substrate specificity for many N-terminal β-amino acid (see Chapter 2),[62,111] we think that especially BapA could become a versatile catalyst for the formation of various β-amino acid-containing peptides. Experimental Section General remarks L-Histidine, L-histidine methyl ester (H-L-His-OMe·2 HCl), L-carnosine and p-nitroaniline were obtained from Sigma-Aldrich (Buchs, Switzerland), H-β-Ala-pNA·HBr from Bachem (Bubendorf, Switzerland) and H-β-Ala-NH2·HCl from Advanced ChemTech (Louisville, KY, USA). The reference peptides H-(β-Ala)2-NH2, H-(β-Ala)2-OH, H-(β-Ala)3-NH2 and H-(β-Ala)2- L-His-OH were synthesized by solid-phase peptide synthesis and were gifts from EMC Microcollections GmbH (Tübingen, Germany). L-Carnosine and peptide products were analyzed by reversed-phase chromatography on a Dionex HPLC system equipped with a P680 pump, an ASI-100 automated sample injector, an UltiMate 3000 thermostated column compartment and a UVD 340U photodiode array detector (Dionex, Sunnyvale, CA, USA). All compounds were separated on a Nucleodur C18 Pyramid stationary phase (250 × 4 mm, 5 μm particle size; Macherey-Nagel, Düren, Germany), which was equilibrated with 0.1% trifluoroacetic acid (TFA) in water at a column temperature of 10°C. L-Carnosine and peptidic byproducts of the reaction were separated with an isocratic flow of 0.1% TFA at a constant flow rate of 1 ml/min. Product formation was detected at 205 nm and quantified according to a standard curve of L-carnosine. All p-nitroanilide derivatives and free p- nitroaniline were separated with a linear gradient of 0 to 40% acetonitrile and detected at wavelengths of 318 and 383 nm, respectively. Quantification of p-nitroanilide derivatives and of free p-nitroaniline was achieved on the basis of standard curves of H-β-Ala-pNA and p- nitroaniline, respectively. Mass spectra of all compounds were recorded with an API 4000 liquid chromatography/tandem MS system connected to an Agilent 1100 LC system. For protein determination, we used five-fold concentrated Bradford reagent (Bio-Rad, Reinach, Switzerland) and bovine serum albumine (BSA) as a standard; absorbance measurements were performed at 595 nm with a Specord S 100 spectrophotometer (Analytik Jena, Jena, 47 Chapter 3 Germany). For the calculation of pKa values we used the calculator plugins of the software Marvin 5.2 (http://www. chemaxon.com). Enzyme preparation The enzymes DmpA from O. anthropi LMG7991 and BapA from S. xenopeptidilytica 3-2W4 were recombinantly expressed in E. coli and purified according to known procedures.[59,62] The lyophilized enzyme powders were dissolved in a 10 mM Tris/HCl buffer at pH 8 and the protein concentrations of the enzyme stock solutions were determined spectrophotometrically. Standard enzyme assay for the formation of L-carnosine from H-β-Ala-pNA or H-β-Ala-NH2 and L-histidine The reaction mixtures contained H-β-Ala-pNA or H-β-Ala-NH2 and L-histidine in different molar ratios in a 100 mM sodium carbonate buffer at pH 10. Limiting amounts of either DmpA or BapA were added to initiate the reactions at 37°C. Samples were withdrawn at intervals from the mixtures and the enzymatic reaction was quenched by the addition of 25% (v/v) 1 M HCl. The formation of L-carnosine was analyzed by HPLC, and byproducts of the reaction were identified by LC/MS. Standard enzyme assay for the formation of L-carnosine from H-β-Ala-NH2 and L-histidine methyl ester The reaction mixtures contained H-β-Ala-NH2 and L-histidine methyl ester in different molar ratios in a 300 m M Tris/HCl buffer at pH 8. DmpA was added to initiate the reaction at 37°C and samples were withdrawn regularly from the mixtures. The addition of NaOH to a final concentration of 0.5 M (pH 12) and heating of the samples at 95° for 7 min stopped the enzymatic reaction and led to complete hydrolysis of all ester bonds. Before HPLC and LC/MS analysis, samples were acidified by adding HCl to a final concentration of 0.5 M (pH 2). Determination of kinetic parameters The initial rates of DmpA- and BapA-catalyzed L-carnosine formation at different ratios of H- β-Ala-NH2 and L-histidine were calculated from the experimental data obtained by HPLC 48 Kinetic Analysis of L-Carnosine Formation measurements. Experimental data were fitted to Michaelis-Menten models without (Eq. 3.3) and with (Eq. 3.4) substrate surplus inhibition by non-linear regression analysis with the VisualEnzymics software (Softzymics, Princeton, NJ, USA) for the program IGOR Pro (WaveMetrics, Oswego, OR, USA). 49 Kinetic Resolution of β3-Amino Acid Amides 4. Kinetic Resolution of Aliphatic β-Amino Acid Amides by β- Aminopeptidases Abstract The growing demand for enantiomerically pure β-amino acids to be used in the pharmaceutical industry and as fine chemicals requires the development of new strategies for their synthesis. The β-aminopeptidases BapA from S. xenopeptidilytica 3-2W4, BapA from S. microcystinivorans Y2, and DmpA from O. anthropi LMG7991 are hydrolases that possess the unique ability of cleaving N-terminal β-amino acids from peptides and amides. Hydrolysis of racemic β3-amino acid amides catalyzed by these enzymes displays enantioselectivity with a strong preference for substrates with the L-configuration and gives access to various aliphatic β3-amino acids of high enantiopurity. This approach presents a new access to enantiopure β3-amino acids under mild reaction conditions and complements chemical asymmetric synthesis strategies. This chapter by T. Heck, D. Seebach, S. Osswald, M. K. J. ter Wiel, H.-P. E. Kohler, and B. Geueke was published in ChemBioChem 2009, 10, 1558-1561. 51 Chapter 4 Introduction The design of unnatural bioactive peptides has attracted increasing attention over the last years. Peptides containing β-amino acids are especially interesting, since they show enhanced resistance towards proteolytic enzymes. The promising pharmaceutical potential of compounds with β-peptidic substructures goes along with a growing demand for enantiopure β-amino acids as building blocks, which is an incentive to develop innovative approaches for their preparation.[1,32,34,120] Besides well-established chemical methods for the synthesis of enantiopure β-amino acids,[2,11,121-123] enzyme-catalyzed resolutions are an interesting alternative. Unlike kinetic resolution approaches to chiral α-amino acids,[124-128] the biocatalytic synthesis of enantiopure β-amino acids is generally limited by the availability of suitable enzymes that are able to catalyze reactions with β-amino acid-containing compounds. Promising enzymatic approaches that provide access to enantiopure aromatic β3-amino acids include the kinetic resolution of N-terminally modified β3-amino acids by porcine kidney acylase[129] and by penicillin G amidase.[130,131] Furthermore, β3- and β2,3-amino acid esters of high enantiopurity were prepared in organic solvents with the lipases CAL-A and CAL-B from C. antarctica.[132- 134] However, there is still a lack of versatile enzymes that operate under mild reaction conditions and provide direct access to a broad range of enantiomerically pure β3-amino acids by kinetic resolution. The BapA enzymes from S. xenopeptidilytica 3-2W4 (3-2W4 BapA) and S. microcystinivorans Y2 (Y2 BapA) as well as DmpA from O. anthropi LMG7991 are aminopeptidases that catalyze the hydrolysis of N-terminal β-amino acid residues from amides and peptides.[59,60,74] Due to their unique substrate specificities and sequence similarity they are collectively referred to as β-aminopeptidases.[63] Despite being structurally distinct, β-aminopeptidases share the functional properties of the N-terminal nucleophile (Ntn) hydrolase family.[67] In recent investigations, we studied the hydrolytic properties of 3-2W4 BapA, Y2 BapA and DmpA,[62] and also used these three enzymes for the synthesis of β- and mixed β/α-peptides (see Chapters 2 and 3).[111] 52 Kinetic Resolution of β3-Amino Acid Amides Herein, we investigated the enantiodifferentiation of enzyme-catalyzed amide hydrolysis, and employed β-aminopeptidases for the kinetic resolution of four aliphatic β3-amino acid amides (rac-1a–d) to the corresponding β3-amino acids (2a–d) in aqueous solution (Scheme 4.1). RO + NH3 H2NOH RO L-2a-d β-aminopeptidase +H2O H2NNH2 pH 8, 37°C RO rac-1a R = Me rac- R = CH CH Me 1b 2 2 H2NNH2 rac-1c R = C6H11 rac-1d R = t-Bu D-1a-d Scheme 4.1: Kinetic resolution of the β3-amino acid amides rac-1a–d catalyzed by the β- aminopeptidases 3-2W4 BapA, Y2 BapA and DmpA. Results and Discussion Separation of β3-amino acid enantiomers The enzyme-catalyzed hydrolysis of the β3-amino acid amides rac-1a–d was followed by measuring the formation of the corresponding free β3-amino acids 2a–d on a teicoplanin 3 HPLC stationary phase. Under these separation conditions, β -amino acids of D-configuration [135,136][E] elute prior to the L-enantiomers (Table 4.1). [E] In the CIP nomenclature L-2a and b have (S)-, L-2c and d (R)-configuration. 53 Chapter4 Table 4.1: Elution order of the enantiomers of 2a-d on a Chirobiotic T2 teicoplanin HPLC column with a mobile phase composition of 90% MeOH, 10% H20 at 10°C. ~-Amino acid Retention Separation Elution sequencelcJ factor k[aJ factor Jbl ko = 1.79 2a 1.13 D < L (R < S)[dJ HiNUOH l ko = 1.47 2b 1.26 D < L (R < .5) H2»0H l k0 = 0.91 2d 1.22 D < L (5 < R') H2bOHl [aJ Retention factor k = (t R - tM )/t M , where tR is the retention t ime of the compound and t,., is the retention time of an unretained compound. [bl Separation factor a = kD/kL . 135 13 [cJ According to the proposed general elution sequence.l • GJ [dJ According to elution of a standard of L-2a. Kinetic resolution of J33-amino acid amides The three 13-aminopeptidases 3-2W4 BapA, Y2 BapA, and DmpA efficiently resolve the racemic 133-amino acid amides rac-la-cl. The enzyme-catalyzed reactions have L- enantioselectivity and form the respective L- 133-amino acids L-2a-cl in high enantiomeric excess (Table 4.2). With all three enzymes, rac- la and rac- lb were the most efficiently resolved substrates (£ > 200) among the four tested compounds, and yielded amino acids L-2a and L-2b in high enantiomeric excess of over 98%. Employing rac- lb as a substrate, we did not observe hydrolysis of the o-enantiomer until L- lb had been fully converted by the enzymes (Figure 4.1, left side). I n the case of rac- la we could not detect the formation of o-2a over 72 h, 54 Kinetic Resolution of;/ -Amino Acid Amides not even after L-la was completely consumed . Alkal ine hydrolysis of the rema ining amide by the addition of sodium hydroxide and heating led to recovery of o-la. This means that all enzymes completely converted rac- la to enantiopure L- 2a, leaving the amide o-la unreacted. Table 4.2: Kinetic resolution of rac-1a-cl by 3-2W4 BapA, Y2 BapA and DmpA at an initial substrate concentration of 20 mM. Substrate Enzyme Enzyme Initial reaction rate t ; eeof L- e cJ cone. 2a-d [µmol·min-1mg-1 of [h] [%][a] [µg/ml] protein] [%] [b] rac-1a 3-2W4 BapA 6.2 13 8 48 > 98 > 300 Y2 BapA 13 7.3 24 48 > 98 > 200 DmpA 1.6 34 4 49 > 98 > 400 rac-1b 3-2W4 BapA 9.9 16 6 48 > 98 > 300 Y2 BapA 21 4.3 4 48 98 > 300 DmpA 520 0.079 32 so > 98 > 500 rac-1c 3-2W4 BapA 9.9 8.5 1.5 38 > 97 > 100 Y2 BapA 64 0.38 4 28 95 61 DmpA 520 0.0033 120 36 > 97 > 100 rac-1d 3-2W4 BapA 31 2.5 1 24 95 53 Y2 BapA 94 0.75 4 51 92 91 DmpA 520 0.043 8 39 > 97 > 100 [aJ Conversion ,; c~ at t ime t = 1-(c; + )/(c;0 +c fo) [bJ Enantiomeric excess of L-2a-d ee = (c; - c~)/(c; +c~) . 137 [cJ Enantiomeric ratio E = ln[l - ,;(1 +ee ) Vln[l - ,;(1- ee )] . C J To determine the influence of the acyl leaving group, we employed the 13-aminopeptidases 62 111 for the ki netic resolution of rac-3a, the p-nitroa nilide analogue of la (Scheme 4.2). [ , 1 With all three enzymes, we observed full conversion of the L-enantiomer to the acid L-2a, whereas formation of o-2a could not be detected. Alka line hydrolysis of the unreacted p- nitroan ilide led to recovery of o-2a. Ou r data suggest that the enantioselectivity of 13-am ino- 55 Chapter 4 acid amide cleavage catalyzed by the β-aminopeptidases depends on the configuration of the stereogenic center in the β-position and not on the acyl leaving group. In addition to rac-3a, we had enantiopure L-3a and D-3a at our disposal, which allowed a more detailed analysis of the L-enantioselectivity. Hydrolysis experiments with 10 mM solutions of the enantiopure substrates showed that 3-2W4 BapA converts L-3a more than 2000-times faster -1 than D-3a (11.8 and 0.0056 μmol·min per mg of protein, respectively). O NO2 + H NOH 2 H N NO2 2 O L-2a 3-2W4 BapA +H2O H2NN pH 8, 37°C NO H O 2 rac-3a H2NN H D-3a Scheme 4.2: Kinetic resolution of the β3-amino acid p-nitroanilide rac-3a catalyzed by 3-2W4 BapA. The kinetic resolutions of compounds rac-1c (Figure 4.1, right side) and rac-1d, which carry bulky cyclohexyl and tert-butyl substituents at the β-carbon, respectively, yielded the 3 corresponding L-β -amino acids L-2c and L-2d in enantiomeric excess of over 92%. The DmpA-catalyzed reaction displayed the highest enantioselectivity (E > 100 and eeL > 97%) for the resolution of both, rac-1c and rac-1d. Although the substrates 1a–d could not be analyzed on the teicoplanin stationary phase, our results suggest that kinetic resolution of rac-1a–d with the three β-aminopeptidases not only leads to formation of L-2a–d, but also 3 gives access to the unreacted D-β -amino acid amides D-1a–d of high enantiopurity. The reaction rates of the β-aminopeptidases were determined for L-1a–d because in all cases the L-enantiomers were hydrolyzed at much higher rates than the D-enantiomers (Table 4.2). As expected from previous degradation experiments,[62] DmpA rapidly hydrolyzed compound 1a (34 μmol·min-1 per mg of protein). Amides 1b–d, which carry sterically more demanding propyl, cyclohexyl, and tert-butyl substituents, were converted very slowly by DmpA. In contrast, the BapA enzymes showed broad substrate specificity and converted all four of the tested substrates with good catalytic rates (0.38 to 13 mmol·min-1 56 Kinetic Resolution of;/ -Amino Acid Amides per mg of protein). Our results indicate that the nature of the ~-carbon substituent controls the rate of the reaction: The lower the degree of branch ing of the substituent the higher was the rate of enzyme-catalyzed hydrolysis of the respective amide. rac-1b rac-1 c 10 10 8 8 ::::? 6 6 .s (.) 4 3-2W4 BapA 4 3-2W4 BapA (9.9 µ.g/ml) (9.9 µg/ml ) 2 2 0 LJ-11...... ----.----.----.---....- Ou------0 1 2 3 4 0 2 4 6 8 10 8 8 Y2 BapA Y2 BapA (21 µg/ml) (64 µg/ml) ~ 6 6 .s (.) 4 4 2 2 0 r"l--4 ...... _...___.. _e=====;==e====r==~ 0 ~-=:!!:..:;;,.._ _ __,,_~--r----r-- 0 2 4 6 8 0 8 16 24 32 10 8 8 ::::? 6 DmpA 6 DmpA .s (520 µg/ml) (520 µg/ml) (.) 4 4 2 2 0 a.~...... _---.---._ __ ._ 0 ~...... _ ...... __==-=::;:=::::;:::~ 0 8 16 24 32 0 72 144 216 288 t [h] t [h] Figure 4.1: Time-concentration curves for t he kinetic resolution of rac-1b and rac-1c by 3-2W4 BapA, Y2 BapA, and DmpA at an initial substrate concentration of 20 mM. The graphs sow the formation of the hydrolysis producst L-2b,c (o) and o-2b,c ( • ). 57 Chapter 4 Conclusions 3 In summary, we have described a novel and practical procedure for the preparation of L-β - amino acids in high enantiopurity using three β-aminopeptidases as catalysts. The 3 nonhydrolyzed amide enantiomers are welcome precursors to the D-β -amino acids. The enzymes efficiently hydrolyzed all of the tested racemic β3-amino acid amides that carried aliphatic substituents with different degrees of branching. The results show that the β- aminopeptidases are especially useful for the kinetic resolution of methyl- and propyl- 3 substituted β -amino acid amides (E > 200 and eeL ≥ 98%), but also substrates with bulky substituents are efficiently resolved by at least one of the enzymes (E > 100 and eeL > 97%). Our previous investigations on the hydrolytic properties of the enzymes[62] suggest that 3-2W4 BapA, Y2 BapA, and DmpA might also be useful for resolving other β3-amino acid amides with aliphatic, aromatic, or functionalized β-carbon substituents. Experimental Section General remarks The amino acids 2a–d were analyzed by reversed-phase HPLC on a Dionex HPLC system equipped with a P680 pump, an ASI-100 automated sample injector, an UltiMate 3000 thermostatted column compartment and a UVD 340U photodiode array detector (Dionex, Sunnyvale, CA, USA). Enantiomers were separated without further derivatization on the chiral teicoplanin stationary phase Chirobiotic T2 (250 × 4.6 mm; Astec, Whippany, NJ, USA) at a constant temperature of 10°C and detected by measuring the absorbance at 205 nm. The mobile phase was composed of MeOH (90%)/H2O (10%) and the applied flow rate was 1 ml/min. Under the described separation and detection conditions, the detection limit for compounds 2a–d was 0.1 mM. For the determination of protein concentrations, we used Bradford reagent (5 × concentrated) from Bio-Rad (Rheinach, Switzerland) and compared samples to a standard curve of bovine serum albumin; absorbance measurements were performed at 595 nm with a Specord S100 spectrophotometer (Analytik Jena, Jena, Germany). The p-nitroanilide derivatives rac-3a, L-3a and D-3a were synthesized according to known procedures.[62,111] 58 Kinetic Resolution of β3-Amino Acid Amides Enzyme Expression and Purification The recombinant enzymes 3-2W4 BapA, Y2 BapA, and DmpA were purified from their E. coli hosts as described previously.[59,62] The lyophilized enzyme powders were dissolved in an appropriate volume of Tris/HCl buffer (10 mM; pH 8) and the protein content of the enzyme stocks was determined. General procedure for the kinetic resolution of rac-1a–d The reaction mixtures contained 20 mM 3-aminobutanamide (rac-1a), 3-aminohexanamide (rac-1b), 3-amino-3-cyclohexylpropanamide (rac-1c) or 3-amino-4,4-dimethylpentanamide (rac-1d) in Tris/HCl buffer (100 mM; pH 8). Kinetic resolution of the substrates was initiated by the addition of an appropriate amount of one of the enzymes 3-2W4 BapA, Y2 BapA, or DmpA. Samples were withdrawn regularly from the reaction mixtures and the enzymatic reaction was quenched by heating the samples at 95°C for 5 min. The hydrolysis products 2a–d were analyzed by HPLC and quantified by relating the UV205 absorbance to a sample of the respective compound that was fully hydrolyzed at NaOH (1 M) and heating at 80°C for 2 h. General procedure for the enzyme-catalyzed hydrolysis of 3a The reaction mixtures contained rac-3a (20 mM) or enantiopure L- or D-3a (10 mM) in Tris/HCl buffer (100 mM; pH 8) and DMSO (10%). Kinetic resolution of the substrates was initiated by the addition of an appropriate amount of 3-2W4 BapA. Samples were withdrawn regularly and analyzed as described in the previous section. 59 Chapter 4 Supporting Information Preparation of Substrates[F] 3-Aminobutanamide (rac-1a): O O a) i-BuOC(=O)Cl, Et3N, O NH O O NH O b) NH3 (aq.) OH NH2 H ,Pd/C 2 Cl NH3 O ethanol, water NH2 N-3-(Benzyloxycarbonylamino)butanamide. N-3-(Benzyloxycarbonylamino)butanoic acid[138] (50.0 g, 211 mmol) and triethylamine (21.5 g, 213 mmol) were dissolved in MTBE (1480 ml), and cooled to 3°C by an ice-bath. iso-Butylchloroformiate (29.0 g, 213 mmol) was added slowly (exothermic reaction), resulting in the formation of a precipitate. After the addition was complete, the reaction mixture was stirred for an additional 30 min and diluted with MTBE (500 ml). Aqueous ammonia (25%, 80 ml) was added quickly (exothermic) and stirring was continued for an additional 45 min at room temperature. After cooling to 5°C, the precipitated solids were collected and washed with MTBE (250 ml). The moist product was then suspended in water (1.8 l) and the mixture was brought to reflux. Most of the solids dissolved, and a small amount of solvent was removed by distillation. After cooling to room temperature, the precipitated white solid was collected by filtration and washed with water. The desired amide was dried at 60°C in vacuo (72.0 g, 174 mmol, 82.2%). 1H NMR (500 MHz, [D6]DMSO): 1.05 (d, J = 6.8 Hz, 3H), 2.11 (dd, J = 14.2, 8.1 Hz, 1H), 2.27 (dd, J = 14.2, 5.8 Hz, 1H), 3.85 (m, 1H), 5.00 (s, 2H), 6.77 (brs, 1H), 7.13–7.38 (m, 7H). 13C NMR (125 MHz, [D6]DMSO): 20.3, 41.9, 44.1, 65.0, 127.7 (2 ×), 128.3, 137.2, 155.1, 172.1. IR (KBr), λ-1 (cm-1): 3386, 3328, 3191, 1683, 1650, 1537,1275, 1258, 1066, 695. M.p. 166°C. [F] The β-amino acid amides were prepared by Steffen Osswald and Matthijs K. J. ter Wiel, Evonik Degussa GmbH. 60 Kinetic Resolution of β3-Amino Acid Amides 3-Aminobutanamide hydrochloride. N-3-(Benzyloxycarbonylamino)butanamide (31.7 g, 134 mmol) was dissolved in n-butanol (200 ml) and water (200 ml). Pd/C (0.9 g, 10%) was added and the mixture was pressurized with hydrogen gas and heated for 1 h at 67°C. The Pd/C was removed by filtration over celite, and the celite washed with water (100 ml). The phases were separated, and the aqueous phase was concentrated under reduced pressure on a rotary evaporator. The oily residue was taken up in i-propanol and conc. HCl (13.5 g) was added. The solution was concentrated again under reduced pressure and the resulting material was recrystallized from i-propanol to give an off-white material that was dried at 1 50°C in vacuo (14.4 g, 104 mmol, 77.5%). H NMR (400 MHz, [D6]DMSO): 1.20 (d, J = 6.6 Hz, 3H), 2.38 (dd, J = 15.8, 7.3 Hz, 1H), 2.49–2.53 (dd, overlapping with DMSO, 1H), 3.41– 13 3.47 (m, 1H), 7.1 (brs, 1H), 7.7 (brs, 1H), 8.1 (brs, 3H). C NMR (125 MHz, [D6]DMSO): 18.0, 38.7, 44.1, 171.4. IR (KBr), λ-1 (cm-1): 3363, 3178, 2985, 1665, 1640, 1424, 1189, 623. M.p. 141°C. 3-Aminohexanamide (rac-1b): O MeOH NH2 O O NH3 NH2 Methyl hex-2-enoate[139] (38.0 g, 297 mmol) was dissolved in methanol (250 ml). The mixture was filled into an autoclave and pressurized with 20 bar of NH3 overnight at 60°C. After cooling to room temperature, all volatiles were removed from the now yellow solution under reduced pressure. The oil solidified and was taken up in a mixture of MTBE (200 ml) and ethanol (18 ml), to which n-hexane (100 ml) was added dropwise. While being stirred, the mixture was cooled to 0°C in an ice-bath and crystals appeared. The off-white crystals were collected by filtration, washed with a small amount of cold n-hexane and finally dried in vacuo at room temperature to give the desired product (26.4 g, 203 mmol, c.y. 68.4%). 1H NMR (400 MHz, [D6]DMSO): 0.85 (t, J = 7.3 Hz, 3H), 1.16–1.40 (m, 4H), 1.95 (dd, J = 14.5, 8.6 Hz, 1H), 2.10 (dd, J = 14.5, 4.7 Hz, 1H), 2.90–2.96 (m, 1H), 6.70 (brs, 1H), 7.35 (brs, 13 -1 -1 1H). C NMR (125 MHz, [D6]DMSO): 14.0, 18.6, 39.6, 43.6, 48.0, 173.5. IR (KBr), λ (cm ): 3351, 2956, 1668, 1603, 1409, 728. M.p. 67°C. 61 Chapter 4 3-Amino-3-cyclohexylpropanamide (rac-1c): NH2 O NH2 O methanol Cl NH3 O methanol OH O NH2 SOCl2 NH3 (aq.) 3-Amino-3-cyclohexylpropanoic acid (29.7 g, 173 mmol) was suspended in methanol (300 ml) and cooled to 5°C. Thionyl chloride (24.8 g, 208 mmol) was added dropwise to the suspension while keeping the temperature at 5°C by cooling with ice. Stirring was continued for 1 h at 5°C and 1 h at 40°C, after which all volatiles were removed under reduced pressure, yielding the intermediate amino acid ester as a white solid (37.9 g, 171 mmol). The solid was taken up in methanol (180 ml), added to aqueous ammonia (25% in water, 360 ml) and stirred for 24 h at room temperature. All methanol was removed under reduced pressure and the pH of the solution was set to 13. The resulting liquid was concentrated under reduced pressure, leaving a sticky solid after the removal of most of the water, to which toluene (250 ml) was added. The contents of the flask were stripped free of water, and then most of the remaining toluene was removed as well under reduced pressure leaving a white, sticky solid. The solid was taken up in hot MTBE (800 ml), filtered over celite, and the celite was washed with an additional amount of MTBE (300 ml). The combined MTBE solutions were collected and slowly cooled to 0°C. The precipitate was collected, washed with cold MTBE (200 ml) and dried at 60°C under reduced pressure to give the product as a white solid (14.5 g, 85.2 mmol, c.y. 49.2%). 1H NMR (500 MHz, [D6]DMSO): 0.91–1.20 (m, 6H), 1.33 (brs, 2H), 1.59–1.71 (m, 5H), 1.90–1.94 (dd, J = 14.3, 9.4 Hz, 1H), 2.11–2.15 (dd, J = 14.3, 3.9 Hz, 1H), 2.74–2.78 (m, 1H), 6.69 (brs, 1H), 7.36 13 (brs, 1H). C NMR (125 MHz, [D6]DMSO): 25.6, 25.7, 27.8, 28.0, 35.0, 39.9, 52.5, 172.3. IR (KBr), λ-1 (cm-1): 3351, 2956, 1668, 1603, 1409, 729. M.p. 80°C. 62 Kinetic Resolution of β3-Amino Acid Amides 3-Amino-4,4-dimethylpentanamide (rac-1d): O O a) i-BuOC(=O)Cl, Et3N, O NH O O NH O b) NH3 (aq.) OH NH2 H ,Pd/C 2 Cl NH3 O ethanol, water NH2 β3-Cbz-Neopentylglycinamide. To a solution of β3-Cbz-Neopentylglycine[140] (5.0, 17.9 mmol), toluene (50 ml) and triethylamine (1.82 g, 18.0 mmol) iso-butylchloroformiate (2.58 g, 18.0 mmol) was added at -3°C. The gel-like suspension was diluted with a small amount of toluene (30 ml), stirred for an additional 5 min and concentrated aqueous ammonia (6.8 ml, 25%) was added. After stirring for 30 min at 5–10°C, all volatiles were removed under reduced pressure giving a white solid that was taken up in MTBE (100 ml) and a saturated solution of NaHCO3 (25 ml). The biphasic mixture was shortly heated up to reflux and then cooled to room temperature. The precipitated solid was collected by filtration and washed with MTBE (50 ml) and water (50 ml) to give the desired compound as a white solid (3.2 g, 1 11.5 mmol, c.y. 64.2%) after drying at 60°C in vacuo. H NMR (500 MHz, [D6]DMSO): 0.83 (s, 9H), 2.05 (dd, J = 14.3, 10.3 Hz, 1H), ), 2.26 (dd, J = 14.3, 3.5 Hz, 1H), 3.71–3.76 (m, 1H), 4.97 (d, J = 12.0 Hz, 1H), 5.03 (d, J = 12.0 Hz, 1H), 6.72 (brs, 1H), 6.96 (d, J = 9.5 Hz, 13 1H), 7.08 (brs, 1H), 7.29–7.36 (m, 5H). C NMR (125 MHz, [D6]DMSO): 26.2, 35.0, 36.5, 56.4, 64.9, 127.4, 127.5, 128.2, 137.4, 156.0, 172.8. IR (KBr), λ-1 (cm-1): 3464, 3320, 2964, 1693, 1656, 1542, 1282, 1052. M.p. 152°C. 3-Amino-4,4-dimethylpentanamide hydrochloride. β3-Cbz-Neopentylglycinamide (18.1 g, 65.0 mmol) was suspended in water (100 ml) and Pd/C (0.46 g, 10%) was added. The whole was filled into an autoclave and hydrogenated for 4 h at 50°C. The catalyst was removed by filtration of celite and the resulting solution was concentrated to dryness to give a white solid (10.2 g). This material (9.4 g) was dissolved in a mixture of i-propanol (200 ml) and water (10.5 ml) to which were added slowly MTBE (265 ml) and concentrated HCl (7.0 g) at 5°C. Seeding crystals were added followed by additional MTBE (85 ml). The resulting 63 Chapter 4 suspension was stirred for 15 min at 0–5°C and then the product was collected by filtration. After washing with MTBE : i-propanol (1:1, 100 ml), the material was dried in vacuo at 45°C to give the desired amide as a white solid (10.0 g, 55.3, c.y. 85.1%). 1H NMR (500 MHz, [D6]-DMSO): 0.94 (s, 9H), 2.26 (dd, J = 16.3, 8.3 Hz, 1H), 2.55 (dd, J = 16.3, 4.1, 1H), 3.23 13 (dd, J = 8.3, 4.1 Hz, 1H), 7.18 (brs, 1H), 7.69 (brs, 3H). C NMR (125 MHz, [D6]DMSO): 25.6, 33.1, 33.3, 56.5, 172.5. IR (KBr), λ-1 (cm-1): 3388, 3079, 2073, 1681, 1617, 1527, 1426, 1381, 590. M.p. 183°C. 64 Biotransformations of β2-Peptides 2 5. β-Aminopeptidase-Catalyzed Biotransformations of β - Dipeptides: Kinetic Resolution and Enzymatic Coupling Abstract Previously, we could show that the β-aminopeptidases BapA from S. xenopeptidilytica and DmpA from O. anthropi were able to catalyze reactions with non-natural β3-peptides and β3- amino acid amides. Here, we report that these exceptional enzymes are also able to utilize synthetic dipeptides with N-terminal β2-amino acid residues as substrates. The suitability of a β2-peptide as a substrate for BapA or DmpA in aqueous reaction medium was strongly 2 dependent on the size of the Cα-substituent of the N-terminal β -amino acid. BapA was shown to convert a diastereomeric mixture of the β2-peptide H-β2hPhe-β2hAla-OH, but did not act on diastereomerically pure β2,β3-dipeptides containing an N-terminal β2-homoalanine. In contrast, DmpA was only active with the latter dipeptides as substrates. BapA-catalyzed transformation of the diastereomeric mixture of H-β2hPhe-β2hAla-OH proceeded along two highly (S)-enantioselective reaction routes, one leading to substrate hydrolysis and the other to the synthesis of coupling products. The synthetic route predominated even at neutral pH. A rise in pH by three log units shifted the synthesis to hydrolysis ratio (vS/vH) further towards peptide formation. Because the equilibrium of the reaction lies on the side of hydrolysis, prolonged incubation resulted in the cleavage of all peptides that carried an N- terminal β-amino acid of (S)-configuration. After completion of the enzymatic reaction the only detectable enantiomer of β2-homophenylalanine was the (S)-enantiomer (ee > 99% for H-(S)-β2-hPhe-OH, E > 500) confirming the high enantioselectivity of the reaction. Our findings suggest interesting new applications of the enzymes BapA and DmpA for the production of enantiopure β2-amino acids or the enantioselective coupling of N-terminal β2- amino acids to peptides. This chapter by T. Heck, A. Reimer, D. Seebach, J. Gardiner, G. Deniau, A. Lukaszuk, H.-P. E. Kohler, and B. Geueke was published in ChemBioChem 2010, 11, 1129-1136. 65 Chapter 5 Introduction β-Amino acids with proteinogenic side chains are backbone-elongated homologues of the naturally occurring proteinogenic α-amino acids (Figure 5.1).[G] In contrast to their α- peptidic counterparts, peptides comprised of β-amino acids are characterized by high resistance to enzymatic degradation and metabolic breakdown.[25,28,30,31] Properly designed β-peptides composed of β3-amino acid residues with proteinogenic side chains fold into stable secondary structures, such as the 314-helix, which was found to be the predominant secondary structure of β3-peptides.[1,17] β2-Peptides, with side chains attached to the α- instead of the β-carbon (Figure 5.1), are conformationally more flexible than their β3-peptide isomers. Synthetic peptides with alternating β2- and β3-amino acid residues as well as mixed β,α-peptides revealed novel hairpin turn-like structures, which resemble the β-turn of conventional α-peptides.[20,141] These properties give rise to interesting new biomedical applications for β-peptides as proteolytically stable mimics of bioactive natural peptides.[32-35] Some recent examples for β-peptide-based designs of peptidomimetics include inhibitors of protein-protein interactions and viral cell entry,[36-40] ligands of the somatostatin receptor and the major histocompatibility complex (MHC),[41-43,142-145] as well as β-peptides with antifungal and antimicrobial activities.[44-50] Furthermore, a very recent study suggested mixed α,β-peptides as a new class of water soluble nanoporous materials that adsorb nitrogen gas.[55] R R O O H OH H α H α N α N β OH N β OH H H H R O n nn 3 2 α-amino acid β -amino acid β -amino acid Figure 5.1: Structural comparison of α- and β-amino acid residues in peptides. [G] The β-amino acids used in this study are named according the nomencalture for β-amino acids with proteinogenic side chains (βhXaa).[1] 66 Biotransformations of β2-Peptides β2-Amino acids are valuable compounds to extend the repertoire of building blocks for peptide design, because their incorporation into peptides induces unique secondary structure elements and leads to increased stability against proteolytic breakdown.[146] Unlike β3-amino acids, most of which are by now commercially available, enantiopure β2-amino acids with proteinogenic side chains are less easily accessible by chemical synthesis. They cannot be prepared by stereoselective homologation of the naturally occurring α-amino acid analogues, the most commonly applied strategy for the synthesis of enantiopure β3-amino acids (Arndt- Eistert reaction).[1,11,18,146] Instead, the synthesis of enantiomerically pure β2-amino acids frequently involves multi-step transformations with the use of chiral auxiliaries. Classical chemical resolution strategies to obtain enantiopure β2-amino acids from racemic starting materials were so far unsuccessful.[2] The β-aminopeptidases BapA from S. xenopeptidilytica and DmpA from O. anthropi recently aroused interest because they are able to catalyze the degradation of β3-peptides with proteinogenic side chains that are otherwise stable to proteolytic breakdown.[25,28,30,31,59,60,74] Despite sharing an amino acid sequence identity of 42%, BapA and DmpA differ with respect to their substrate specificities. DmpA cleaves both α- and β3-amino acids from the N-termini of peptides, but shows a distinct preference for N-terminal amino acids that are unsubstituted or carry small side chains (e.g. -CH3). In contrast, BapA only catalyzes the removal of N-terminal β3-amino acids and does not cleave peptides with N-terminal α-amino acids. However, BapA has a broader substrate specificity than DmpA as it accepts a wide variety of peptides that carry N-terminal β3-amino acids with different proteinogenic side chains.[62] These findings demonstrated for the first time that the supposed proteolytic stability of β-peptides is not absolute. Further studies with BapA and DmpA revealed their usefulness as biocatalysts for the efficient production of enantiopure aliphatic β3-amino acids by kinetic resolution of racemic β3-amino acid amides (see Chapter 4),[110] and for the enzyme-catalyzed synthesis of various β3-amino acid-containing peptides from C-terminally activated β-amino acid residues under kinetic control (see Chapters 2 and 3).[111,115,147] The promising results obtained from the β-aminopeptidase-catalyzed conversions of β3- peptides encouraged us to conduct enzymatic studies with β2-amino acid-containing peptides, which are less explored due to their limited availability. In fact, only very few enzyme- 67 Chapter 5 catalyzed conversions of β2-amino acid-containing compounds have been reported thus far, most of which describe the enantioselective production of β2-amino acids from racemic starting materials.[134,148-153] To the best of our knowledge, the proteolytic breakdown of β2- peptides has not yet been investigated at all. In the present investigation, we address the potential of the β-aminopeptidases BapA and DmpA to degrade β2-peptidic substrates as well as the stereoselectivity of these reactions. Furthermore, we describe the β- aminopeptidase-catalyzed formation of β2-peptides under kinetic control. Results and Discussion Enzymatic conversion of a diastereomeric mixture of the β2-dipeptide H- β2hPhe-β2hAla-OH (1) The dissolved β-aminopeptidases BapA from S. xenopeptidilytica and DmpA from O. anthropi were tested for their ability to act on substrates composed of non-natural β2-amino acids under aqueous reaction conditions. For this purpose, we synthesized a diastereomeric mixture of the β2-dipeptide H-β2hPhe-β2hAla-OH 1 ((S,S)-1, (S,R)-1, (R,S)-1 and (R,R)-1), which allowed us to investigate the β-aminopeptidase-catalyzed breakdown of the compound and the stereoselectivity of the reaction at the same time. We separated the diastereomers of 1 on a reversed-phase stationary phase and quantified substrates and products of the enzymatic conversions by HPLC-UV. As the diastereomeric mixture of 1 contains two diastereomeric pairs of enantiomers ((S,S)-1, (R,R)-1 and (S,R)-1, (R,S)-1), two peaks were obtained by reversed-phase HPLC. According to the elution of a pure standard of (S,S)-1 (tR = 17.1 min) and a standard of (S,S)-1 that contained traces of (S,R)- 1 (tR = 17.7 min), peak 1 (tR = 17.1 min) could be attributed to (S,S)-1, (R,R)-1 and peak 2 (tR = 17.7 min) to (S,R)-1, (R,S)-1. In our investigation, we detected substrate transformation only with the enzyme BapA, whereas DmpA left the diastereomeric mixture of 1 untouched. During the assay period at pH 7.2 and 37°C, the decrease of 1 leveled off at approximately 57% of the initially employed substrate concentration of 5 mM (Figure 5.2). It is possible that the enzymatic reaction slowed down at the end of the 11-day incubation due a partial inactivation of the enzyme. However, the fact that both peaks of 1 decreased to approximately half of their initial areas indicated a high stereoselectivity of the BapA- catalyzed reaction. The initial decrease of the area under peak 2 ((S,R)-1, (R,S)-1) was 68 Biotransformations ofp2-Peptides about 1.3-times faster than the decrease of the area under peak 1 ((5,5)-1, (R,R)-1). Appropriate control experiments showed that the substrate 1 was chemically stable under the reaction conditions over the assay period. 0 0 48 96 144 192 240 t [h] Figure 5.2: Conversion of H-~ 2 hPhe-~ 2 hAla-OH (1, +)by BapA (0.3 mg of protein per ml) at pH 7.2 and 37°C. Two peaks of the diastereomeric mixture were detected by reversed-phase HPLC, which correspond to (S,.5)-1, (R,R)-1 UR = 17.1 min, ~ ) and to (S,R)-1, (R,.5)-1 UR = 17.7 min, •). Original HPLC traces are shown in the Supporting Information (Figure 51). Previous studies with 133-peptides and 133-amino acid amides showed that BapA and DmpA strongly discriminated between N-terminal 133-amino acids that ca rry side chains of different 52 111 3 sizes. [ , J While BapA accepts a broad range of N-terminal 13 -amino acids with different side chain lengths and functionalities, DmpA cata lyzes reactions with small, sterically undemanding 133-amino acids, such as N-terminal 133-homoalanine and 13-homoglycine.[HJ 2 According to our present results, this observation applies for the degradation of the 13 - dipeptide 1 as well. Structural information on the active site compositions of BapA (see Chapter 6) and DmpA (PDB ID: 1B65)[64J reveal that the different substrate specificities of [HJ ~-Homog l yci ne is commonly referred to as ~-a l an i ne. 69 Chapter 5 BapA and DmpA are possibly caused by different topologies of the enzyme active sites. Whereas the active site pocket of BapA is relatively wide, the DmpA substrate binding site is constricted by a loop region ranging from Gln131 to Trp137, which is likely to hinder the accessibility of the active site for substrates with bulky N-terminal amino acids. Kinetic resolution of the diastereomeric pairs of enantiomers of the β2- dipeptide 1 by BapA 2 2 The BapA-catalyzed release of H-β hPhe-OH (2) from a 5 mM solution of the β -dipeptide 1 at pH 7.2 was analyzed by HPLC on a chiral stationary phase. For the separation of the enantiomers of 2 we used the teicoplanin stationary phase Chirobiotic T2, which provides good separations of many chiral β3- and β2-amino acids.[135,136,154] We achieved baseline separation of the enantiomers of 2, which was a prerequisite for examining the stereoselectivity of the enzyme-catalyzed reaction. Under the applied conditions (R)-2 (tR = 11.3 min) eluted prior to (S)-2 (tR = 15.4 min). The enantiomers of the second hydrolysis product H-β2hAla-OH (3) could not be separated. The four stereoisomers of 1 were partially separated under the applied conditions. According to standards of (S,S)-1 and (S,R)-1, peak 2 (tR = 31.6 min) could be attributed to (S,R)-1 and peak 4 (tR = 36.0 min) to (S,S)-1. Over the reaction period of 11 days (S)-2 accumulated to a concentration of 2.3 mM, which corresponds to a conversion of 46% of the initially employed β2-dipeptide 1 (Figure 5.3). The reaction was highly (S)-enantioselective (ee > 99% and E > 500)[137] and the release of (R)-2 over the assay period remained below the detection limit of 0.01 mM. These results indicate almost quantitative conversion of both diastereomers with an N- terminal β2-homophenylalanine residue of (S)-configuration ((S,S)-1 and (S,R)-1) (Scheme 5.1). The analysis of the same samples by reversed-phase HPLC showed that (S,R)-1 was converted 1.3-times faster than (S,S)-1 (Figure 5.2). Hence, we conclude that the configuration of the N-terminal β2-homophenylalanine residue is crucial for the successful conversion of 1 by BapA, whereas the configuration of the second amino acid in the β2- dipeptide is only of minor importance. 70 Biotransformations ofp2-Peptides 5 t t + 4 + 3 + ...... + + ~ + .s + (.) 2 • • • • • 1 • • 0 A:>• 0 0 0 0 0 0 0 48 96 144 192 240 t [h] Figure 5.3: Formation of H-(S)-p2hPhe-OH ((S)-2, • ) and H-(R)-p2hPhe-OH ((R)-2, o) from the diastereomeric mixture of 5 mM H-p2hPhe-p2hAla-OH (1, +) catalyzed by BapA (0.3 mg of protein per ml) at pH 7.2 and 37°C. Original HPLC traces are shown in the Supporting Information (Figure 52). (S)-2 (S)-3 and (R)-3 BapA and 11 days 1 (R,S)-1 and (R,R)-1 Scheme 5.1: Kinetic resolution of the diastereomeric pairs of enantiomers of the p2-dipeptide 1 by BapA at pH 7 .2 and 37°C. 71 Chapter 5 BapA-catalyzed peptide coupling reactions with the β2-dipeptide 1 The BapA-catalyzed conversion of the β2-dipeptide 1 was characterized by simultaneously occurring hydrolysis and coupling reactions among substrates and products (Scheme 5.2). This complex interplay of competing reactions is based on the postulated general catalytic mechanism of β-aminopeptidases,[63] which was previously described for other enzyme members of the N-terminal nucleophile (Ntn) hydrolase family including DmpA.[64,68] Accordingly, the first step in substrate conversion is the formation of a characteristic acyl enzyme complex through nucleophilic attack of the enzyme’s catalytically active serine residue on the carbonyl carbon atom of the substrate’s amide bond. The acyl moiety is subsequently released from the acyl enzyme by the action of a nucleophile. The nucleophile can be either water, which leads to hydrolysis of the acyl enzyme, or the deprotonated N- terminus of an amino acid or a peptide, which results in peptide bond formation. The creation of a potent nucleophile requires alkaline aqueous conditions to keep the attacking amino group in its deprotonated state. H2O O H2NOH+ HO-BapA hydrolysis (S)-2 OO O HO-BapA H2NNOH H2NO-BapA H 3 1 acyl enzyme O OO coupling H2N N N OH + HO-BapA H H 1 4 Scheme 5.2: BapA-catalyzed routes for the conversion of the β2-dipeptide H-β2hPhe-β2hAla-OH (1) at the starting point of the reaction. All products formed by initial hydrolysis and coupling reactions can further react as nucleophiles. This leads to the formation of additional peptidic coupling products (5–7, see Table 5.1). For the BapA-catalyzed coupling reactions with the β2-dipeptide 1 we used an alkaline aqueous reaction system at pH 10, which was previously shown to support peptide couplings by β-aminopeptidase (see Chapters 2 and 3).[111,115,147] At the initial stage of the reaction we 72 Biotransformations ofp2-Peptides observed substrate hydrolysis and the formation of the 132-tripeptide H-[p2hPhe]rl32hAla-OH (4) (Scheme 5.2). This is due to the fact that water and the substrate itself were the only nucleophiles to release the acyl moiety from the active site of BapA. As the reaction proceeded, all products created by hydrolysis and coupling reactions could further react as nucleophiles themselves, which gave rise to the additional coupling products 5-7 (Table 5.1 and Figure 5.4). Table 5.1: HPLC-MS analysis of the hydrolysis and coupling products that accumulated after 48 h during the BapA-catalyzed conversion of the diastereomeric mixture of H- ~ 2 hPhe- ~ 2 hAla-OH (1) at pH 10 and 37°C (see also Figure 5.4). The detected mass/charge-ratios (ml z) of the hydrogen and sodium adducts of the molecular ions are given. Further fragmentations of the peptides are not shown. n.d. not detected. Compound tR Calcd. mass m/z m/z [min] [M+Hr [M+Nar H-~ 2 hPhe-OH (2) 16.6 179.1 180.2 n.d H-~ 2 hPhe-~ 2 hAla-OH (1) 17.1 264.2 265.1 287.2 17.7 265.1 287.3 H-[~ 2 hPhe]r~ 2 hA l a-OH ( 4) 22.4 425.2 426.2 448.2 22.8 426.1 448.3 23.1 426.2 448.3 H-[~ 2 hPhe]rOH (5) 23.2 340.2 341.4 363.2 H-[~ 2 hPhe]r~ 2 hA l a-OH (6) 25.7 586.3 587.5 609.8 25.9 587.5 609.5 26.1 587.3 609.7 26.4 587.4 609.8 H-[~ 2 hPhe]rOH (7) 26.8 501.3 502.6 524.2 73 Chapter 5 1 2 4 5 6 7 16 18 20 22 24 26 28 t [min] R Figure 5.4: HPLC-UV trace of a sample taken at 48 h from the BapA-catalyzed conversion of the diastereomeric mixture of H-β2hPhe-β2hAla-OH (1) at pH 10 and 37°C. The chromatogram shows the hydrolysis product H-(S)-β2hPhe-OH ((S)-2) as well as the coupling products 4–7 (see also Table 5.1). Since the overall equilibrium of the BapA-catalyzed reaction lies on the side of the hydrolysis products, all observed coupling products that temporarily accumulated in the reaction mixture were finally hydrolyzed to their amino acid constituents β2-homophenylalanine (2) and β2-homoalanine (3) (Figure 5.5). In contrast to 2, the second hydrolysis product 3 could not be analyzed by reversed-phase HPLC as the compound coeluted from the column with the flow-through. Similarly to the reaction at pH 7.2, only (S)-2 was detected by HPLC on chiral column material after completion of the enzymatic reaction at pH 10, and the concentration of (R)-2 remained below the detection limit. After the reaction period of 10 days, the final concentration of (S)-2 (2.3 mM) corresponded to substrate conversion of almost 50%. 74 Biotransformations ofp2-Peptides 5 + + + 4 + + + 3 + + + + .s~ + (.) 2 • • • • 1 • ~ c:P D D D 0 ~· D D 0 48 96 144 192 t [h] Figure 5.5: Conversion of the p2-dipeptide H-p2hPhe-p2hAla-OH (1, +) at a concentration of 5 mM by BapA (0.3 mg of protein per ml) at pH 10 and 37°C. Competing hydrolysis and coupling reactions gave rise to the formation of H-(.5)-p2hPhe-OH ((.5)-2, • )and various peptidic coupling products (4- 7, o), respectively. Due to a lack of standards for the quantification of compounds 4-7, the overall concentration of coupling products was estimated on the basis of the experimentally determined concentrations of 1 and 2, assuming that one equivalent of coupling product is formed from two equivalents of 1. From these data we conclude (1) that the non-converted diastereoisomers correspond to (R,.5)-1 and (R,l<)-1 and (ii) that the acyl enzyme could only be formed when the N-terminal p2-homophenylalanine residue of the peptide had the (.5)-configuration (Scheme 5.2). Consequently, all coupling products 4-7 formed through enzymatic conversion must share the presence of one or more N-terminal p2-homophenylalanine residues of (.5)-configuration. Rega rding the first coupling product H-(.5)-p2hPhe-p2hPhe-p2hAla-OH (4) only three peaks were detected for t he four possible diastereoisomers ((5,5,.5)-4, (S,S,l<)-4, (S,R,S)-4 and (S,R,l<)-4) by reversed-phase HPLC (Figure 5.4 and Table 5.1). However, the appearance of four peaks for the diastereoisomers of the tetrapeptide H-[(.5)-p2hPhe]rP2hPhe-p2hAla-OH (6) implies that the four possible diastereoisomers of the precursor tripeptide 4 must have been present in the reaction mixture. This is supported by the fact that the ratio of the three peak areas of 4 was approximately 1:2: 1, which suggests that two of the four diastereoisomers of 4 coeluted from the column. Only one peak was detected by reversed- 75 Chapter 5 phase HPLC for each of the β2-homophenylalanine oligomers 5 and 7, which temporarily accumulated in the reaction mixture (Figure 5.4). As (R)-2 is not present in the reaction mixture to attack the acyl enzyme, we conclude that 5 and 7 are solely composed of β2- amino acid residues of (S)-configuration. O OO O O OO H2N N N OH H2N N N N OH H H H H H (S,S,S)-4, (S,S,R)-4, (S,R,S)-4 and (S,R,R)-4 (S,S,S,S)-6, (S,S,S,R)-6, (S,S,R,S)-6 and (S,S,R,R)-6 O O O O O H2N N OH H2N N N OH H H H (S,S)-5 (S,S,S)-7 pH-Dependence of BapA-catalyzed hydrolysis and coupling reactions with the β2-dipeptide 1 The influence of pH on the BapA-catalyzed conversion of the diastereomeric mixture of 1 (5 mM) was investigated at pH 7.2, pH 8 and pH 10. From the initial data points obtained by reversed-phase HPLC analysis, we calculated the enzymatic rates for the overall conversion of the substrate 1 (vC) as well as the rates for the formation of the hydrolysis product 2 (vH) and of the coupling product 4 (vS) (Table 5.2). As expected from a previous investigation on [59] the pH profile of BapA, the highest conversion rate was observed at pH 8 (vC = 0.021 -1 μmol·min per mg of protein). The ratio vS/vH was strongly dependent on the pH of the reaction system. Interestingly, peptide synthesis predominated over substrate hydrolysis at the initial stage of the reaction under all pH conditions. A rise in pH from 7.2 to 10 caused a further shift towards coupling product formation, as indicated by an increase in vS/vH from 2.4 to 10.8. Generally, the conversion of 1 catalyzed by BapA was one to two orders of 76 Biotransformations ofp2-Peptides magnitude slower than the conversion of p3-amino acid-containing substrates under similar reaction conditions.[ 62l Table 5.2: Initial rates of the BapA-catalyzed hydrolysis and coupling reactions with the substrate H- ~2hPhe- ~2hA l a-OH (1, 5 mM) at 37°C and at different pH values. The conversion rate ( vc) relates to the initial decrease in substrate, whereas the hydrolysis rate ( vH) and the synthesis rate (Vs) relate to 2 the initial increase of free H- ~ hPhe- OH (2) and coupling product 4, respectively. Cmax, which represents the maximum accumulation of all coupling products 4-7 in the reaction mixture, was estimated under the assumption that one equivalent of the coupling product 4 is formed by the condensation of two equivalents of 1. pH Ve VH Vs [µmol·min-1 per [µmol·min-1 per [µmol·min-1 per [mM] mg of protein] mg of protein] mg of protein] 7.2 0.013 0.0022 0.0053 2.4 0.28 8 0.021 0.0026 0.0094 3.6 0.46 10 0.017 0.00075 0.0081 10.8 0.60 While the accumulation of coupling products 4-7 was very low at pH 7.2 (0.28 mM), we observed maximum coupling product concentrations of 0.46 mM at pH 8 and 0.60 mM at pH 10. These concentrations correspond to conversions of 18.4 and 24%, respectively, of the initially employed dipeptide 1 (Table 5.2). Coupling product accumulation peaked in a clear- cut maximum, after which peptide hydrolysis became the predominant reaction (see Figure 5.5). After completion of the reactions the on ly remaining detectable compounds were (5)-2 and the non-converted stereoisomers of the substrate, which correspond to ( R,S)-1 ( tR = 28.8 min) and (R,R')- 1 UR= 33.5 min) as determined by HPLC on the chiral column material. Enzymatic degradation of J32, J33-dipeptides with N-terminal J32-homoalanine (8 and 9) I n addition to the diastereomeric mixture of the p2-dipeptide 1 we also performed experiments with the p-aminopeptidases BapA and DmpA and four diastereomerically pure p2,p3-dipeptides ca rrying an N-terminal p2-homoalanine and a C-terminal (S)-p3-homoamino acid residue ((5,5)-8, (R,S)-8 and (5,5)-9, (R,5)-9). 77 Chapter 5 OH (S,S)-8 and (R,S)-8 (S,S)-9 and (R,S)-9 (S,S)-10 I nterestingly, we did not detect any conversion of substrates 8 and 9 with BapA, whereas DmpA cleaved both diastereoisomers of 8 and 9. This is in contrast to our observations with substrate 1 and to previous results with H-(5)-133hAla-(5)-133hLeu-OH ((5,5)-10), which is the 3 59 62 13 -dipeptide isomer of (5,5)-8.[ • 1 The results shown in Table 5.3 indicate that the catalytic rates for the DmpA-catalyzed conversion of (S,5)-8 and (R,5)-8 were 70- and 90-times lower than the conversion rate for the corresponding 133-dipeptide (5,5)-10 determined in the present study under the same conditions. An exchange of the (-terminal 133-homoleuci ne residue of 8 by the more hydrophobic 133-homophenylalanine of 9 had an accelerating effect 2 on substrate conversion by DmpA. DmpA converted dipeptides 8 and 9 with N-terminal 13 - homoalanine residues of both (5)- and (R)-configuration with low stereoselectivity, which is in contrast to the previously reported highly (5)-enantioselective kinetic resolution of racemic 133-homoalanine amide and 133-homoalanine trnitroanilide by the same enzyme (see Chapter 4) _[110] Table 5.3: Conversion of the diastereomerically pure substrates 8, 9 and 10 (5 mM) by DmpA at pH 7.2 and 37°C. Substrate Conversion rate [µmol·min-1 per mg of protein] (S,S)-8 0.029 (R,S)-8 0.037 (S,5)-9 0.099 (R,5)-9 0.48 (S,5)-10 2.5 78 Biotransformations of β2-Peptides Conclusions In the present investigation we have studied the conversion of non-natural peptides with N- terminal β2-amino acid residues using the β-aminopeptidases BapA and DmpA as catalysts. As previously observed for the conversion of β3-peptides, the enzymes had distinct substrate preferences depending on the size of the side chain on the N-terminal amino acid. Only BapA acted on substrate 1 with its bulky N-terminal β2-homophenylalanine residue, whereas the conversion of peptides 8 and 9, which carry N-terminal β2-homoalanine and C-terminal β3-homoleucine or β3-homophenylalanine residues, was only catalyzed by DmpA. While DmpA converted substrates with N-terminal amino acids of (S)- and (R)-configuration, the BapA-catalyzed reaction was highly (S)-enantioselective. BapA simultaneously catalyzed hydrolysis and synthesis reactions with the diastereomeric mixture of 1, leading to the hydrolysis of the substrate and to the formation of various coupling products, respectively. The ratio of the rate of peptide formation to the rate of peptide hydrolysis was positively correlated with an increase of the pH. The availability of enantiopure β2-amino acids is limited because chemical preparations of these compounds involve labor- and cost-intensive multi-step reactions.[2] Hence, alternative enzymatic approaches for the enantioselective production of β2-amino acids are highly desirable to complement chemical asymmetric synthesis strategies. The highly (S)- enantioselective reactions of BapA with a model β2-dipeptide indicate that β- aminopeptidases may become useful for the biocatalytic production of enantiopure β2-amino acids by kinetic resolution of racemic β2-amino acid derivatives, e.g. β2-amino acid amides or esters. Furthermore, the enzymes may be used to catalyze the introduction of a β2-amino aicd residue as an N-terminal protecting group into a peptide, thereby stablizing the peptide against degradation by other exopeptidases. To fully asses the biocatalytic potential of the β-aminopeptidases BapA and DmpA for enantioselective conversions of β2-peptidic substrates it will be necessary (i) to optimize the present reaction system with the aim to suppress uncontrolled cross-couplings among substrates and products, and (ii) to carry out a detailed analysis of the enzymes’ substrate specificities for peptides and amides carrying N- terminal β2-amino acids with different side chains. 79 Chapter 5 Experimental Section General Remarks The amino acids and peptides 1–10 were analyzed on a Dionex HPLC system equipped with a P680 pump, an ASI-100 automated sample injector, an UltiMate 3000 thermostatted column compartment and a UVD 340U photodiode array detector (Dionex, Sunnyvale, CA, USA). Enantiomers of the β2-amino acid H-β2hPhe-OH (2) were separated without further derivatization on the chiral teicoplanin stationary phase Chirobiotic T2 (250 x 4.6 mm; Astec, Whippany, NJ, USA) at a constant temperature of 10°C and quantified by relating the absorbance at 205 nm to a standard curve of enantiopure (R)-2. The mobile phase was composed of 90% methanol, 10% H2O and the applied flow rate was 1 ml/min. Under the described separation and detection conditions, the detection limit for 2 was 0.01 mM. Additionally, all samples were analyzed by reversed-phase HPLC on a Nucleodur C18-Pyramid stationary phase (250 × 4 mm, 5 μm particle size; Macherey-Nagel, Düren, Germany), which was equilibrated with 0.1% trifluoroacetic acid (TFA) in water at a column temperature of 20°C. All substrates and products were separated with a linear gradient of 0 to 40% acetonitrile at a constant flow rate of 1 ml/min and detected at a wavelength of 205 nm. Coupling products were quantified on basis of the experimentally determined concentrations of the compounds 1 and 2 under the assumption that one molecule of coupling product is formed from two molecules of substrate. Mass spectra of all compounds were recorded with an API 4000 liquid chromatography/tandem MS system connected to an Agilent 1100 LC system. For protein determination, we used five-fold concentrated Bradford reagent (Bio- Rad, Rheinach, Switzerland) and bovine serum albumin (BSA) as a standard; absorbance measurements were performed at 595 nm with a Specord S 100 spectrophotometer (Analytik Jena, Jena, Germany). Enzyme expression and purification The genes bapA and dmpA coding for the β-aminopeptidases BapA from S. xenopeptidilytica 3-2W4 and DmpA from O. anthropi LMG7991, respectively, were cloned into the expression plasmid pET9c (Novagen, Madison, USA). The plasmid p3BapA[59] was digested with BamHI and NdeI. The bapA-containing fragment was cloned into pET9c that was cut at the same sites yielding pAR116. The dmpA gene was cloned from pODmpA[62] into pET9c according to 80 Biotransformations of β2-Peptides the same procedure with the restriction enzymes BamHI and SalI yielding pAR114. The two obtained expression plasmids were transformed into E. coli BL21(DE3) pLysS (Novagen, Madison, USA) and recombinantly expressed in the presence of kanamycin and chloramphenicol. BapA and DmpA were purified according to established procedures.[62,74] The lyophilized enzyme powders were dissolved in a 10 mM potassium phosphate buffer at pH 7.2, and the protein contents of the enzyme stocks were determined spectrophotometrically. General procedure for the enzyme-catalyzed conversion of the diastereomeric mixture of H-β2hPhe-β2hAla-OH (1) The reaction mixtures contained 5 mM of the diastereomeric mixture of 1 in 100 mM solutions of different buffering systems (potassium phosphate pH 7.2, Tris/HCl pH 8, sodium carbonate pH 10). Enzymatic conversions of the substrates were initiated by the addition of an appropriate amount of one of the enzymes BapA or DmpA. Samples were withdrawn at intervals from the reaction mixtures and the enzymatic reaction was quenched by heating the samples at 90°C for 3 min. The reaction products were analyzed by chiral and reversed- phase HPLC methods. A sample taken after 48 h from the enzymatic assay at pH 10 was subjected to HPLC-MS analysis. General procedure for the enzyme-catalyzed conversion of the enantiopure substrates 8, 9 and 10 2 3 The reaction mixtures contained 5 mM of the diastereomerically pure β ,β -dipeptides H-(S)- β2hAla-(S)-β3hLeu-OH ((S,S)-8), H-(R)-β2hAla-(S)-β3hLeu-OH ((R,S)-8), H-(S)-β2hAla-(S)- β3hPhe-OH ((S,S)-9), H-(R)-β2hAla-(S)-β3hPhe-OH ((R,S)-9) or the β3-dipeptide H-(S)- 3 3 [59] β hAla-(S)-β hLeu-OH ((S,S)-10) in a 100 mM potassium phosphate buffer at pH 7.2 containing 10% DMSO. Enzymatic conversions of the substrates were carried out as described above and samples were analyzed by reversed-phase HPLC. 81 Chapter 5 Supporting Information[I] Preparation of the diastereomeric mixture of H-β2hPhe-β2hAla-OH (1) Preparation of Z-β2hPhe-OH (F1): O O O O HN O O O 1. NaHMDS, THF -78°C N O Cl N O BuLi, THF 2. BrCH2CO2tBu 92% 89% CO2tBu A1 B1 C1 TFA, quant. CH2Cl2 O 1. Et N, DPPA, O O 3 O O LiOH, H2O2 toluene, reflux CbzHN OH 93% N O 2. BnOH, reflux N O 74% NHCbz CO2H F1 E1 D1 PhCH2CH2COCl (A1): Hydrocinnamic acid (10 g, 66.6 mmol, 1 equiv.) was dissolved in CH2Cl2 (100 ml) and cooled to 0°C in an ice bath. Oxalyl chloride (5.914 ml, 69.9 mmol, 1.05 equiv.) was added slowly followed by 3 drops of DMF. The solution was warmed to room temperature overnight and the solvent removed under reduced pressure to give a yellow oil 1 (11.3 g, quant.), which was used without further purification. H NMR (300 MHz, CDCl3): 3.02 (t, J = 7.8, 2H, CH2), 3.22 (t, J = 7.8, 2H, CH2), 7.19-7.34 (m, 5H, arom. H). B1: 2-Oxazolidinone (5.038 g, 57.85 mmol, 1 equiv.) was dissolved in dry THF (100 ml) and cooled to -78°C under argon. BuLi (37.97 ml of 1.6 M in hexane, 60.74 mmol, 1.05 equiv.) was added and the solution stirred at -78°C for 10 min. A solution of A1 (10.24 g, 60.74 mmol, 1.05 equiv.) in THF (20 ml) was added dropwise and the mixture was stirred at -78°C for 15 min, then warmed to room temperature and stirred for a further 30 min. The solution [I] The β-amino acids and β-peptides, which were used as substrates and references in this chapter, were prepared by James Gardiner, Gildas Deniau and Aneta Lukaszuk, ETH Zürich. 82 Biotransformations of β2-Peptides was quenched with satd. aq. NH4Cl (50 ml) at 0°C and the solvent removed under reduced pressure. The residue was diluted with water and extracted with CH2Cl2 (3 × 50 ml). The organic extracts were combined and washed with 1 M NaOH (100 ml), water (100 ml), and satd. aq. NaCl (100 ml), and dried over MgSO4. The solvent was removed under reduced pressure and the product purified by crystallization (EtOAc/hexane) to give B1 as a white solid (11.679 g, 92%). NMR data corresponds with literature.[155] C1: B1 (5 g, 22.83 mmol, 1 equiv.) was dissolved in dry THF (50 ml) and cooled to -78°C under argon. NaHMDS (25.11 ml of 1 M in THF, 25.11 mmol, 1.1 equiv.) was added and the solution was stirred at -78°C for 1 h. A solution of tert-butylbromoacetate (10.11 ml, 68.45 mmol, 3 equiv.) in THF (10 ml) was added slowly and the mixture was stirred at -78°C for 3 h. The reaction mixture was then warmed to room temperature, quenched with satd. aq. NH4Cl (100 ml), and extracted with EtOAc (3 × 50 ml). The extracts were combined, dried over MgSO4, and the solvent was removed under reduced pressure. Crystallization (EtOAc/hexane) of the resulting residue gave C1 as a white solid (6.757 g, 89%). NMR data corresponds with literature [1]. D1: Trifluoroacetic acid (10 ml) was added to a solution of C1 (5.5 g, 16.52 mmol) in CH2Cl2 (100 ml) and the mixture was stirred at room temperature for 4 h. The solvent was removed under reduced pressure and the residue was taken up in EtOAc (100 ml), washed with H2O (100 ml), satd. aq. NaCl (100 ml) and dried over MgSO4. The solvent was removed to give 1 D1 as a white solid (4.61 g, quant.). H NMR (300 MHz, CDCl3): 2.42 (dd, J = 3.8, 17.4, 1H, CH2), 2.58 (dd, J = 9.8, 13.2, 1H, CH2), 2.92 (dd, J = 10.9, 17.4, 1H, CH2), 3.10 (dd, J = 5.5, 13.2, 1H, CH2), 3.89–4.09 (m, 2H, CH2), 4.29 (m, 3H, CH2 and CH), 7.20–7.32 (m, 5H, arom. H). E1: D1 (4.575 g, 16.516 mmol, 1 equiv.) was dissolved in dry toluene (0.1 M, 160 ml) under argon, and Et3N (2.762 ml, 19.819 mmol, 1.2 equiv.) and DPPA (3.599 ml, 16.516 mmol, 1 equiv.) were added. The resulting mixture was stirred for 30 min at room temperature, and subsequently heated at reflux for 1.5 h. BnOH (5.127 ml, 49.55 mmol, 3 equiv.) was then added, and the mixture was refluxed for a further 16 h. After cooling, the mixture was diluted with Et2O (100 ml), washed with satd. aq. NaCl soln. (100 ml), dried (MgSO4), and 83 Chapter 5 concentrated under reduced pressure. Flash chromatography (1:3 Et2O/pentane) of the 1 crude product yielded E1 as a white solid (4.665 g, 74%). H NMR (300 MHz, CDCl3): 2.77 (dd, J = 13.6, 8.0, 1H, CH2), 3.05 (dd, J = 13.6, 7.0, 1H, CH2), 3.37–3.52 (m, 4H, 2 CH2), 3.83 (m, 2H, CH2), 4.19–4.35 (m, 3H, CH2 and CH), 5.07 (dd, J = 13.6 and 12.2, 2H, OCH2), 13 7.18–7.38 (m, 10H, arom. H). C NMR (75 MHz, CDCl3): 35.75, 41.63, 42.61, 45.03, 61.92, 66.60, 126.45, 127.88, 128.28, 128.84, 136.28, 137.70, 152.95, 156.02, 174.03. Z-β2hPhe-OH (F1): E1 (3.705 g, 9.69 mmol, 1 equiv.) was dissolved in a solution of THF/H2O (4:1) and cooled to 0°C. H2O2 (4.392 ml of 30% aq. soln., 38.75 mmol, 4 equiv.) and LiOH.H2O (813 mg in 20 ml H2O, 19.4 mmol, 2 equiv.) were added and the resulting mixture was stirred at 0°C for 2 h. The mixture was diluted with H2O (50 ml) and extracted with EtOAc (3 × 50 ml). The aq. phase was then acidified to pH 2 with 1 M HCl and extracted with EtOAc (3 × 50 ml). The organic extracts were combined, washed with satd. aq. NaCl soln. (50 ml), dried (MgSO4), and concentrated under reduced pressure to give F1 1 as a white solid (2.823 g, 93%). H NMR (300 MHz, CDCl3) rotamers : 2.71–3.49 (m, 5H, 2 CH2 and CH), 5.09 (m, 2H, OCH2), 5.30 (m, 1H, NH), 6.51, 7.13–7.38 (m, 10H, arom. H). Preparation of TFA·H-β2hAla-OMe (I1): O O O LDA, THF, -78°C TFA, CH2Cl2 BocHN OMe TFA.H NOMe BocHN OMe MeI quant. 2 92% G1 H1 I1 Boc-βAla-OMe (G1): HCl·β-alanine methyl ester (8.36 g, 60.14 mmol, 1 equiv.) was suspended in CH2Cl2 (150 ml) and cooled to 0°C. Et3N (16.766 ml, 120.29 mmol, 2 equiv.) was added. A solution of Boc2O (13.126 g, 60.143 mmol, 1 equiv.) in CH2Cl2 (20 ml) was then added dropwise and the mixture was stirred and warmed to room temperature overnight. The solvent was removed and the residue was taken up in EtOAc (150 ml), washed with 5% citric acid, 5% aq. NaHCO3, satd. aq. NaCl, dried over MgSO4, and the 1 solvent was removed to give G1 as a clear oil (11.76 g, quant.). H NMR (300 MHz, CDCl3): 1.43 (s, 9H, tBu), 2.53 (t, J = 6.1, 2H, CH2), 3.39 (dd, J = 12.2, 6.3, 2H, CH2), 3.70 (s, 3H, Me), 5.01 (brs, 1H, NH). 84 Biotransformations of β2-Peptides Boc-β2hAla-OMe (H1): Anhydrous LiCl (3.026 g, 73.8 mmol, 3 equiv.) was suspended in dry THF (50 ml) and cooled to -78°C under argon. DIPA (9.43 ml, 54.12 mmol, 2.2 equiv.) and BuLi (33.825 ml of 1.6 M solution in hexane, 54.12 mmol, 2.2 equiv.) were added and the mixture was stored at -78°C for 10 min. A solution of G1 (5 g, 24.6 mmol, 1 equiv.) in dry THF (50 ml) was added and the mixture was stirred at -78°C for 1 h. MeI (6.124 ml, 98.4 mmol, 4 equiv.) was added and the solution was stirred at -78°C for 1 h and then warmed to room temperature overnight. The reaction was quenched with satd. NH4Cl soln. and successively washed with satd. aq. NaHCO3, NH4Cl, and NaCl solns. The organic phase was dried (MgSO4) and the solvent removed under reduced pressure. Purification of the residue by flash chromatography (1:4 EtOAc/hexane) gave H1 as a colorless oil (4.921 g, 92%). 1H NMR (300 MHz, CDCl3): 1.13 (d, J = 7.2, 3H, Me), 1.40 (s, 9H, tBu), 2.61–2.68 (m, 1H, CH2), 13 3.16–3.29 (m, 2H, CH2 and CH), 3.66 (s, 3H, Me), 4.96 (brs, 1H, NH). C NMR (75 MHz, CDCl3): 14.76, 28.37, 39.97, 42.93, 51.78, 79.19, 155.71, 175.56. 2 TFA·H-β hAla-OMe (I1): H1 (3.140 g, 14.453 mmol) was dissolved in CH2Cl2 (20 ml) and cooled to 0°C. TFA (5 ml) was added and the solution was stirred at room temperature for 2 h. The solvent was removed and the residue dried under h.v. for 2 h to give I1 as a viscous 1 oil (3.18 g, quant.). H NMR (300 MHz, CDCl3): 1.29 (d, J = 7.1, 3H, Me), 2.91–2.94 (m, 1H, CH2), 3.14–3.23 (m, 2H, CH2 and CH), 3.75 (s, 3H, OMe), 4.33 (brs, 1H, NH). 85 Chapter 5 Preparation of HCl·H-β2hPhe-β2hAla-OH (1): HATU DIPEA O O CH2Cl2 F1+ I1 CbzHN N OMe 95% H J1 NaOH, 99% MeOH O O 1. H2/Pd-C, O O MeOH HCl.H2NNOH CbzHN N OH H 2. aq. HCl, H lyophilize 98% 1 K1 Z-β2hPhe-β2hAla-OMe (J1): To a solution of F1 (2.505 g, 7.993 mmol, 1 equiv.) and HATU (3.039 g, 7.993 mmol, 1 equiv) CH2Cl2 (50 ml) under argon, DIPEA (2.784 ml, 15.99 mmol, 2 equiv.) was added and the solution was stirred at room temperature for 1 min. A solution of I1 (1.712 g 7.993 mmol, 1 equiv.) in CH2Cl2 (10 ml) was added slowly and the mixture was stirred at room temperature for 4 h. The solution was then washed with sat aq. NaHCO3 (2 x 50 ml), H2O (50 ml), NaCl (50 ml), dried (MgSO4) and the solvent was removed to give J1 as a white solid (3.123 g, 95%). Mix of 2 sets of enantiomers (4 diastereoisomers). 1H NMR (300 MHz, CDCl3): 0.95, (d, J = 7.2, 3H, Me), 1.05 (d, J = 7.5, 3H, Me), 2.34–2.37 (m, 1H), 2.52–2.88 (m, 7H), 3.32–3.49 (m, 5H), 3.58 (s, 3H, OMe), 3.60 (s, 3H, OMe), 5.04– 5.13 (m, 4H, 2 OCH2), 5.33 (m, 1H, NH), 5.39 (m, 1H, NH), 5.72 (m, 2H, 2 NH), 7.14–7.34 13 (m, 10H, arom. H). C NMR (75 MHz, CDCl3): 14.50, 14.93, 36.41, 36.56, 39.07, 39.37, 41.24, 41.56, 43.10, 49.26, 51.77, 51.86, 66.60, 126.27, 126.38, 127.80, 127.91, 128.32, 128.38, 128.62, 136.33, 138.53, 156.36, 173.04, 173.24, 175.26, 175.36. MALDI-MS 451.2 (40, [M + K]+), 436.2 (22), 435.2 (100, [M + H]+), 413.2 (31, [M + H]+). MALDI-HRMS: + + + + 413.2076/435.1896 (31/100) [M + H] /[M + Na] , calcd. for [C23H29N2O5] /[C23H28N2O5Na] : 413.2057/435.1884. 86 Biotransformations of β2-Peptides 2 2 Z-β hPhe-β hAla-OH (K1): NaOH (15.83 ml of a 1 M aq. soln., 15.83 mmol, 2 equiv.) was added to a solution of J1 (3.265 g, 7.92 mmol, 1 equiv) dissolved in MeOH (60 ml) and the resulting mixture was stirred at room temperature for 4 h. MeOH was removed under reduced pressure and the aq. phase was washed with EtOAc (20 ml). The aq. phase was then acidified to pH 2 with 1 N HCl and extracted with EtOAc (3 × 50 ml). The extracts were combined, dried (MgSO4) and the solvent was removed to give K1 as a white solid (3.149 g, 1 99%). Mix of 2 sets of enantiomers (4 diastereoisomers). H NMR (300 MHz, CDCl3): 0.99, (d, J = 7.2, 3H, Me), 1.08 (d, J = 7.2, 3H, Me), 2.37–2.42 (m, 1H), 2.52–2.64 (m, 1H), 2.66–2.87 (m, 6H), 3.09–3.25 (m, 2H), 3.25–3.45 (m, 6H), 5.00–5.10 (m, 4H, 2 OCH2), 5.60 (m, 2H, NH), 6.20–6.35 (m, 2H, 2 NH), 7.09–7.31 (m, 10H, arom. H). 2 2 HCl·H-β hPhe-β hAla-OH (1): K1 (3.149 g, 7.90 mmol) was dissolved in MeOH under N2. Pd-C (310 mg, 10% w/w) was added and the flask was evacuated and flushed with H2 (3 ×). The solution was then stirred under a H2 atmosphere for 4 h, whereupon it was filtered through celite and the solvent removed under reduced pressure. The residue was then dissolved in 0.1 M HCl (20 ml) and lyophilized to give 1 (2.329 g, 98%) as an HCl salt. A small sample was taken and the diastereoisomers separated by RP-HPLC (acetonitrile/water (0.1% TFA), 25% acetonitrile in 5 min, 25–65% acetonitrile in 30 min, 65–95% acetonitrile in 5 min, C18) to give 1a (10.1 mg) and 1b (9.8 mg) as TFA salts. 1a: Anal. RP-HPLC (5% acetonitrile in 5 min, 5–95% acetonitrile in 30 min, C18, tR = 14.9). 1 H NMR (400MHz, CD3OD): 1.06 (d, J = 7.2, 3H, Me), 2.50 (dd, J =14.2, 7.1, 1H, CH(Me)CO), 2.80 (dd, J = 12.5, 6.3, 1H, CHaPh), 2.89 (m, 1H, CHCH2Ph), 2.97 (m, 2H, CHbPh and H2NCHa), 3.18 (dd, J = 12.8, 8.2, 1H, H2NCHb), 3.27 (d, J = 7.1, 2H, NHCH2), 13 7.19–7.33 (m, 5H, PhH). C NMR (100 MHz, CD3OD): 15.49 (+, Me), 37.77 (-, CH2), 40.54 (+, CH), 41.68 (-, CH2), 43.23 (-, CH2), 46.81 (+, CH), 128.09 (+, arom. C), 129.81 (+, arom. C), 130.06 (+, arom. C), 138.87 (Cq, arom. C), 174.17 (Cq, CO), 178.42 (Cq, CO). MALDI-MS: 287.1 (6, [M + Na]+), 266.2 (12, [M + 2H]2+), 265.2 (100, [M + H]+). HR-MS calcd. (C14H21N2O3, 265.1552); found 265.1547. 1b: Anal. RP-HPLC (5% acetonitrile in 5 min, 5–95% acetonitrile in 30 min, C18, tR = 15.3). 1 H NMR (400MHz, CD3OD): 0.97 (d, J = 7.2, 3H, CH3), 2.54 (dd, J = 13.9, 6.8, 1H, 87 Chapter 5 CH(Me)CO), 2.80 (dd, J = 12.3, 5.9, 1H, CHaPh), 2.89 (m, 1H, CHCH2Ph), 2.96 (m, 2H, CHbPh and H2NCHa), 3.13 (dd, J = 13.5, 6.7, 1H, HNCHa), 3.19 (dd, J = 12.8, 8.0, 1H, 13 H2NCHb), 3.42 (dd, J = 13.5, 6.7, 1H, NHCHb), 7.20–7.33 (m, 5H, PhH). C NMR (100 MHz, CD3OD): 15.22 (+, CH3), 37.74 (-, CH2), 40.30 (+, CH), 41.79 (-, CH2), 43.25 (-, CH2), 46.89 (+, CH), 128.09 (+, arom. C), 129.81 (+, arom. C), 130.11 (+, arom. C), 138.91 (Cq, arom. + C), 174.18 (Cq, CO), 178.50 (Cq, CO). MALDI-MS: 287.1 (6, [M + Na] ), 266.2 (12, [M + 2+ + 2H] ), 265.2 (100, [M + H] ). HR-MS calcd. (C14H21N2O3, 265.1552); found 265.1547. Preparation of H-(S)-β2hPhe-(S)-β2hAla-OH ((S,S)-1)[156] In a SPPS 10 ml plastic reactor, 2-chlorotriyl chloride resin (300 mg, 0.45 mmol, 1.5 mmol/g) 2 was shaken in CH2Cl2 (2 ml) for 1 h, and then the solvent was filtered. Fmoc-(S)-β hAla-OH (2 equiv., 0.293 g, 0.9 mmol), the synthesis of which was described previously,[2] was coupled to the resin in the presence of EtN(i-Pr)2 (4 equiv., 0.313 ml, 1.8 mmol) in CH2Cl2 (3 ml) during 3 h. The solvent was filtered out and the resin was washed with CH2Cl2/MeOH/EtN(i-Pr)2 (17:2:1, 3 × 1 min), CH2Cl2 (3 × 1 min), MeOH (3 × 1 min), followed by drying under vacuum for 2 h. Fmoc deprotection of the first amino acid was carried out using a mixture of 20% piperidine in DMF (3 ml, 2 × 30 min). The resin was filtered and washed with CH2Cl2, DMF, CH2Cl2 (3 ml, 3 × 1 min). Fmoc deprotection was monitored by the Kaiser test[157,158] (the test reagent consists of three solutions: A: 6% ninhydrin in EtOH; B: 80% phenol in EtOH; C: KCN H2O/pyridin). A few resin beads were placed in a small test tube. Three drops of each solution were added and the tube was heated for 5 min at 100°C. Colorless or yellowish color of the resin and solution indicate a negative result (desired after coupling), a blue color indicates the presence of primary amine. The coupling of Fmoc-(S)-β2hPhe-OH[156] (2 equiv., 0.348 g, 0.9 mmol) was completed in the presence of HATU (2 equiv., 0.342 g, 0.9 mmol) and EtN(i-Pr)2 (4 equiv., 0.313 ml, 1.8 mmol) after shaking overnight. The resin was washed and the Fmoc protection was removed as described above. The peptide was cleaved from the resin using a TFA/H2O/Et3SiH (3 ml, 95:2.5:2.5) mixture. Cleavage was performed during 1 h, then the resin was filtered off and the filtrate was evaporated and ice-cold Et2O was added to precipitate the peptide. The resulting gel residue was washed and repeatedly evaporated from Et2O. The peptide was purified by semi-preparative RP-HPLC (acetonitrile/water (0.1% TFA), 5–20% acetonitrile in 50 min; column Nucleosil 100-7 C18: 250 mm, ID: 21 mm, flow 10 ml/min). Two fractions 88 Biotransformations of β2-Peptides were collected containing (S,S)-1 and (S,S)-1 with traces of (S,R)-1, respectively, and lyophilized. (S,S)-1: TFA·H-(S)-β2hPhe-(S)-β2hAla-OH (57 mg, 48% yield), anal. RP-HPLC (acetonitrile/water (0.1% TFA), 3–97% acetonitrile in 20 min; column SP 250/21 Nucleosil + 100-7 C18: 250 mm, ID: 21 mm, flow 1 ml/min, tR = 9.13), LC-MS 265.2 [M + H] , 287.1 [M + Na]+. (S,S)-1 with traces of (S,R)-1: TFA·H-(S)-β2hPhe-(S)-β2hAla-OH + TFA·H-(R)-β2hPhe-(S)- β2hAla-OH (27 mg, 23% yield), anal. RP-HPLC (acetonitrile/water (0.1% TFA), 3–97% acetonitrile in 20 min; column EC 250/4 Nucleosil 100-5 C18: 250 mm, ID: 4 mm, flow 1 ml/min, tR = 9.07 and tR = 9.45). Preparation of H-(R)-β2hPhe-OH ((R)-2) (R)-2 was prepared by catalytic hydrogenation of (R)-2-(benzyloxycarbonylamino-methyl)-3- phenylpropanoic acid (Z-β2hPhe-OH) according to a published procedure[159] with Pd/C in methanol under a positive pressure of H2. After filtration through a pad of celite and evaporation of the solvent under reduced pressure, H-(R)-β2hPhe-OH was obtained as a [160] 22 white solid. Analytical data were in accordance with the literature. [α]D = +22.8 (c 1, 1 35 M HCl), (lit. [α]D = +18.9 (c 0.88, 1 M HCl)). 89 Chapter 5 Preparation of H-β2hAla-(S)-β3hLeu-OH (8) and of H-β2hAla-(S)-β3hPhe-OH (9) HATU DIPEA 1. Pd-C, H2 O O O O O CH2Cl2 2. HCl TFA.H NOBn72% BocHN N OBn 64% HCl.H2N N OH 2 H H + A8 B8 (R,S)-8 O BocHN OH + HATU DIPEA 1. Pd-C, H O O O 2 O O CH2Cl2 2. HCl TFA.H NOBn71% BocHN N OBn 63% HCl.H2N N OH 2 H H A9 B9 (R,S)-9 Preparation of TFA·H-β2hAla-(S)-β3hLeu-OH ((R,S)-8 and (S,S)-8): A8: Boc-(S)-β3hLeu-OBn (1 g, 2.99 mmol) was dissolved in TFA (5 ml) at 0°C and stirred at this temperature for 30 min and then at room temperature for 90 min. The solvent was removed and the residue was dried under h.v. overnight to give A8 (1.18 g, quant.) as a clear oil, which was used without further purification. B8: To a soln. of Boc-(R)-β2hAla-OH[2] (49 mg, 0.24 mmol, 1.2 equiv.) and HATU (84 mg, 0.22 mmol, 1.1 equiv.) in CH2Cl2/DMF (40:1, 4.1 ml) was added DIPEA (75 μl, 0.44 mmol, 2.2 equiv.) and the mixture was stirred for 1 min. A soln. of A8 (70 mg, 0.2 mmol, 1 equiv.) in CH2Cl2 (4 ml) was added and the mixture was stirred for 2 h. The organic layer was washed with satd. aq. NaHCO3 (2 × 10 ml), H2O (2 × 10 ml), NaCl (10 ml), dried (MgSO4) and the solvent was removed. Purification by column chromatography (2:3 EtOAc/hexane) 1 gave B8 as a white solid (60 mg, 72%). H NMR (300 MHz, CDCl3): 0.89 (d, J = 6.3, Me), 0.90 (d, J = 6.6, Me), 1.06 (d, J = 7.1, Me), 1.22–1.32 (m, 1H, CH), 1.42 (s, 9H, tBu), 1.45– 1.67 (m, 2H, CH2), 2.41–2.47 (m, 1H, CHCO), 2.50 (dd, J = 15.7, 5.2, 1H, CH2), 2.60 (dd, J = 15.7, 5.2, 1H, CH2), 3.10–3.30 (m, 2H, CH2), 4.28–4.38 (m, 1H, CH), 5.02 (m, NH), 5.11 13 (dd, J = 15.4, 12.2, 2H, OCH2), 5.98 (d, J = 8.0, NH), 7.34–7.38 (m, 5H, arom. H). C NMR 90 Biotransformations of β2-Peptides (75 MHz, CDCl3): 14.95, 21.59, 22.59, 24.67, 28.01, 38.76, 40.79, 42.68, 43.21, 43.68, 66.01, 78.69, 127.82, 128.04, 128.20, 135.04, 155.52, 170.91, 173.70. (R,S)-8: To a soln. of B8 (50 mg) in MeOH (5 ml) was added Pd-C (5 mg, 10% w/w) and the reaction vessel was evacuated and flushed with H2 (3 ×). The mixture was then stirred vigorously under H2 for 12 h. The mixture was filtered through celite, washed with MeOH, and the solvent was removed under reduced pressure and dried under h.v. for 1 h. The residue was then dissolved in 1 M HCl (2 ml) and lyophilized to give a white solid (32 mg). Purification by RP-HPLC (acetonitrile/water (0.1% TFA), 5% acetonitrile in 5 min, 5–50% acetonitrile in 30 min, 50–95% acetonitrile in 5 min, C18), gave (R,S)-8 as a viscous oil (25 mg, 64%). TFA salt. Anal. RP-HPLC (5% acetonitrile in 5 min, 5–50% acetonitrile in 20 min, 1 C18, tR = 13.73). H NMR (300 MHz, CD3OD): 0.93 (d, J = 6.6, 6H, 2 Me), 1.23 (d, J = 7.5, Me), 1.29–1.38 (m, 1H, CH2), 1.45–1.54 (m, 1H, CH2), 1.59–1.68 (m, 1H, CH), 2.41 (dd, J = 15.0, 7.2, 1H, CH2), 2.48 (dd, J = 15.0, 5.9, 1H, CH2), 2.65–2.71 (m, 1H, CH), 2.98 (dd, J = 12.8, 5.0, 1H, CH2), 3.10 (dd, J = 12.8, 7.8, 1H, CH2), 4.23–4.33 (m, 1H, CH). MALDI-HRMS: + + 231.1703 [M + H] , calcd. for [C11H23N2O3] : 231.1703. The (S,S)-diastereomer (S,S)-8 was also isolated (5.2 mg, 15%) as a viscous oil. TFA salt. 1 Anal. RP-HPLC (5% acetonitrile in 5 min, 5–50% acetonitrile in 20 min, C18, tR = 16.14). H NMR (300 MHz, CD3OD): 0.92 (d, J = 6.5, Me), 0.93 (d, J = 6.5, Me), 1.22 (d, J = 6.8, Me), 1.28–1.37 (m, 1H, CH2), 1.43–1.53 (m, 1H, CH2), 1.58–1.65 (m, 1H, CH), 2.42 (dd, J = 15.3, 7.5, 1H, CH2), 2.48 (dd, J = 15.3, 5.9, 1H, CH2), 2.64–2.71 (m, 1H, CH), 2.96 (dd, J = 12.8, 4.7, 1H, CH2), 3.12 (dd, J = 12.8, 8.1, 1H, CH2), 4.26–4.36 (m, 1H, CH). Preparation of TFA·H-β2hAla-(S)-β3hPhe-OH ((R,S)-9 and (S,S)-9): A9: Boc-(S)-β3hPhe-OBn (1 g, 2.71 mmol) was dissolved in TFA (5 ml) at 0°C and stirred at this temperature for 30 min and then at room temperature for 90 min. The solvent was removed and the residue was dried under h.v. overnight to give A9 (1.1 g, quant.) as a clear oil, which was used without further purification. B9: To a soln. of Boc-(R)-β2hAla-OH[2] (49 mg, 0.24 mmol, 1.2 equiv.) and HATU (84 mg, 0.22 mmol, 1.1 equiv.) in CH2Cl2/DMF (40:1, 4.1 ml) was added DIPEA (75 μl, 0.44 mmol, 91 Chapter 5 2.2 equiv.) and the mixture was stirred for 1 min. A soln. of A9 (77 mg, 0.2 mmol, 1 equiv.) in CH2Cl2 (4 ml) was added and the mixture was stirred for 2 h. The organic layer was washed with satd. aq. NaHCO3 (2 × 10 ml), H2O (2 × 10 ml), NaCl (10 ml), dried (MgSO4) and the solvent was removed. Purification by column chromatography (2:3 EtOAc/hexane) 1 gave B9 as a white solid (64 mg, 71%). H NMR (300 MHz, CDCl3): 0.89 (d, J = 6.3, Me), 0.90 (d, J = 6.6, Me), 1.06 (d, J = 7.1, Me), 1.22-1.32 (m, 1H, CH), 1.42 (s, 9H, tBu), 1.45– 1.67 (m, 2H, CH2), 2.41–2.47 (m, 1H, CHCO), 2.50 (dd, J = 15.7, 5.2, 1H, CH2), 2.60 (dd, J = 15.7, 5.2, 1H, CH2), 3.10–3.30 (m, 2H, CH2), 4.28–4.38 (m, 1H, CH), 5.02 (m, NH), 5.11 13 (dd, J = 15.4, 12.2, 2H, OCH2), 5.98 (d, J = 8.0, NH), 7.34–7.38 (m, 5H, arom. H). C NMR (75 MHz, CDCl3): 15.22, 28.35, 37.36, 39.92, 41.16, 43.46, 47.08, 66.49, 79.13, 126.67, 128.34, 128.39, 128.48, 128.57, 129.17, 135.46, 137.37, 156.07, 171.44, 174.26. (R,S)-9: To a soln. of B9 (50 mg) in MeOH (5 ml) was added Pd-C (5 mg, 10% w/w) and the reaction vessel was evacuated and flushed with H2 (3 ×). The mixture was then stirred vigorously under H2 for 12 h. The mixture was filtered through celite, washed with MeOH, and the solvent was removed under reduced pressure and dried under h.v. for 1 h. The residue was then dissolved in 1 M HCl (2 ml) and lyophilized to give a white solid (33 mg). Purification by RP-HPLC (acetonitrile/water (0.1% TFA), 10% acetonitrile in 5 min, 10–60% acetonitrile in 30 min, 60–95% acetonitrile in 5 min, C18), gave (R,S)-9 as a viscous oil (25 mg, 63%). TFA salt. Anal. RP-HPLC (10% acetonitrile in 5 min, 10–60% acetonitrile in 25 1 min, C18, tR = 12.10). H NMR (300 MHz, CD3OD): 1.18 (d, J = 7.2, 3H, Me), 2.42–2.52 (m, 2H), 2.55–2.64 (m, 1H), 2.83–2.91 (m, 2H), 2.98–3.05 (m, 1H), 4.38–4.45 (m, 1H, CH), 13 7.18–7.31 (m, 5H, arom. H). C NMR (75 MHz, CD3OD): 16.12, 38.74, 39.12, 41.19, 42.88, 48.82, 127.64, 129.47, 130.45, 139.49, 174.79, 175.19. MALDI-HRMS: 265.1550 [M + H]+, + calcd. for [C14H21N2O3] : 265.1552. The (S,S)-diastereomer of (S,S)-9 was also isolated (6 mg, 15%) as a viscous oil. TFA salt. Anal. RP-HPLC (10% acetonitrile in 5 min, 10–60% acetonitrile in 25 min, C18, tR = 14.87). 1 H NMR (300 MHz, CD3OD): 1.02 (d, J = 6.8, Me), 2.4–2.63 (m, 3H), 2.76 (dd, J = 13.7, 8.4, 1H, CH2), 2.87–2.95 (m, 2H), 3.03 (dd, J = 12.7, 8.4, 1H, CH2), 4.47–4.52 (m, 1H, CH), 7.16-–7.29 (m, 5H, arom. H). 92 Biotransformations of β2-Peptides HPLC traces used for the generation of Figure 5.2 (S/S)-1 (S/R)-1 (R/R)-1 (R/S)-1 1000 0 h 800 600 / mAU 400 205 A 200 (S)-2 0 -200 1000 24 h 800 600 / mAU 400 205 A 200 4 0 5 -200 1000 96 h 800 600 / mAU 400 205 A 200 0 -200 1000 336 h 800 600 / mAU 400 205 A 200 0 -200 16 18 20 22 24 / min tR Figure S1: HPLC-UV traces (column C18-Pyramid, Macherey Nagel) of samples taken at 0, 24, 96 and 336 hours of incubation from the BapA-catalyzed conversion of the diastereomeric mixture of H- β2hPhe-β2hAla-OH (1) at pH 7 and 37°C. The peak areas obtained from the chromatograms were used for the generation of Figure 5.2. 93 Chapter 5 HPLC traces used for the generation of Figure 5.3 160 0 h 140 120 100 / mAU 80 205 1 A 60 40 20 0 160 24 h 140 120 100 / mAU 80 (S)-2 205 A 60 40 20 (R)-2 0 160 96 h 140 120 100 / mAU 80 205 A 60 40 20 0 160 264 h 140 120 100 / mAU 80 205 A 60 40 20 0 5 10 15 20 25 30 35 40 45 50 / min tR Figure S2: HPLC-UV traces (column Chirobiotic T2, Astec) of samples taken at 0, 24, 96 and 264 hours of incubation from the BapA-catalyzed conversion of the diastereomeric mixture of H-β2hPhe- β2hAla-OH (1) at pH 7 and 37°C. The peak areas obtained from the chromatograms were used for the generation of Figure 5.3. 94 Crystal Structure and Inhibition of BapA 6. Complex Structures of BapA with Penicillin-Derived Inhbitors Reveal New Insights into β-Peptide Conversion by β-Aminopeptidases Abstract β-Aminopeptidases are unique enzymes that catalyze reactions with synthetic peptides comprised of β-amino acids. The increasing demand for enantiopure β-amino acids and β- peptidic building blocks in biomedical research suggests interesting biotechnological applications for β-aminopeptidases. In the present study, we determined the crystal structure of the β-aminopeptidase BapA from Sphingosinicella xenopeptidilytica 3-2W4 at a resolution of 1.45 Å. The arrangement of the secondary structure elements of the BapA core reveals an αββα-sandwich structure, which resembles the consensus fold of the N-terminal nucleophile (Ntn) hydrolase superfamily. Moreover, three crystal structures of non-covalent BapA-ligand complexes were obtained with the serine protease inhibitor pefabloc SC, the β- lactam antibiotic ampicillin and an ampicillin-derived hydrolysis product. BapA did not catalyze the conversion of the penicillin derivatives ampicillin, penicillin G and carbenicillin, but it was inhibited by ampicillin (Ki = 0.69 mM). The observed BapA-ligand interactions suggest an important function of Glu133 for the stabilization and the positioning of β- peptidic substrates in the active site of BapA. Kinetic analyses with the β-aminopeptidases BapA, DmpA from Ochrobactrum anthropi LMG7991 and a DmpA variant revealed that the substrate specificity of β-aminopeptidases for different β-amino acid-derived substrates is largely influenced by a highly variable region, which defines the width of the substrate binding pocket. The combination of structural information of the BapA-ligand complexes and kinetic investigations of β-aminopeptidase-catalyzed reactions forms a solid basis for the rational design of β-peptide-converting biocatalysts with interesting novel specificities and selectivities. 95 Chapter 6 Introduction The class of β-aminopeptidases so far comprises four functionally characterized hydrolytic enzymes, i.e. BapA from Sphinosinicella xenopeptidilytica 3-2W4,[74] DmpA from Ochrobactrum anthropi LMG7991,[60] BapA from Sphinosinicella microcystinivorans Y2,[59] and BapA from Pseudomonas sp. MCI3434.[61] These enzymes share the exceptional ability to cleave synthetic peptides comprised of backbone-elongated β-amino acid residues, which are not utilized by common proteolytic enzymes (Figure 6.1).[25,30,63] The biochemical and functional properties of BapA from S. xenopeptidilytic and DmpA from O. anthropi are especially well investigated. Despite sharing 42% sequence identity, the substrate specificities of BapA and DmpA differ notably. BapA catalyzes the transformation of a wide variety of peptides with N-terminal β-amino acids, preferably with aliphatic and aromatic side chains. However, BapA reacts only with peptides that carry N-terminal β-amino acids; peptides with N-terminal proteinogenic α-amino acids do not serve as substrates.[59,62] DmpA efficiently converts peptides with small N-terminal β-amino acids, such as β-homoglycine and β-homoalanine, whereas β-amino acids with bulky substituents are removed from the N- termini of peptides with very low catalytic rates.[62] In contrast to BapA, DmpA also utilizes substrates that carry small N-terminal α-amino acids.[60] The presence of a free N-terminal amino group in the peptide was reported to be an essential requirement for substrate conversion by BapA and DmpA; N-acylated or N-terminally protected peptides are not converted by either enzyme.[60,74] Despite their well-studied specificities for synthetic β- peptides, the physiological roles and the natural substrates of β-aminopeptidases are as yet unknown. R R O O R1 O H OH H H H N N OH N OH N OH H H H H 2 O n n R n R n 3 2 2,3 α-amino acid β -amino acid β -amino acid β -amino acid Figure 6.1: Comparison of the structures of α- and β-amino acids in peptides with different substitutions. 96 Crystal Structure and Inhibition of BapA β-Amino acid-containing peptides are valuable building blocks for the design of pharmaceutically relevant peptidomimetics,[32,33,35] as the incorporation of β-amino acids into peptides leads to enhanced resistance against proteolysis and induces unique secondary structures.[1] The growing demand for β-amino acid-based building blocks is an incentive to develop innovative strategies for their syntheses.[134] This suggested the use of the β- aminopeptidases BapA and DmpA for biocatalytic applications, e.g. for the production of enantiopure aliphatic and aromatic β-amino acids by kinetic resolution of racemic starting materials,[110,161] and for the synthesis of β- and mixed β,α-peptides from C-terminally modified β-amino acid derivatives.[111,115,147] In spite of these first successful attempts employing β-aminopeptidases for biocatalysis, a more detailed understanding of the underlying mechanisms of substrate recognition and of the discrimination between α- and β- amino acids on a structural level is essential to fully exploit the interesting biocatalytic potential of β-aminopeptidases. The crystal structure of DmpA from O. anthropi, which was solved at a resolution of 1.82 Å, was the first structure of a β-aminopeptidase.[64] It provided interesting insights into the general topology of β-aminopeptidases and demonstrated that DmpA shares characteristic functional and structural features with members of the Ntn (N-terminal nucleophile) hydrolase superfamily, which comprises very diverse enzymes with amidohydrolyzing activities.[66,67] In analogy to Ntn hydrolases, the catalytic core of DmpA consists of two polypeptide chains (α and β) that are formed by posttranslational processing of an inactive precursor polypeptide. Cleavage of the immature precursor is presumably autocatalyzed and occurs N-terminal of a conserved serine residue (Ser250), which thereupon becomes the N- terminal catalytically active nucleophile of the β-polypeptide chain. The catalytic mechanism of Ntn hydrolases is reported to involve a single amino acid catalytic center[68] and is hence distinct from conventional serine, cysteine or threonine peptidases that contain catalytic triads or diads.[69] DmpA-catalyzed substrate conversion presumably follows the general catalytic mechanism of Ntn hydrolases. Accordingly, the free N-terminal amino group of Ser250 is proposed to act as the general base to abstract a proton from the hydroxyl group in order to enhance its nucleophilicity.[64] The topology of the DmpA subunit resembles the consensus Ntn fold, which comprises two central layers of β-sheets flanked by an envelope of two α-helices on each side. However, the orientation and connectivity of the secondary 97 Chapter 6 structure elements in DmpA is distinct from other members of the Ntn hydrolase family, such as human aspartylglucosaminidase (PDB ID: 1APY), the proteolytic core of the Saccharomyces cerevisiae proteasome (PDB ID: 1RYP), penicillin G acylase (PDB ID: 1PNK), and glutamine phosphoribosyl-pyrophosphate amidotransferase from Escherichia coli (PDB ID: 1ECF).[70] Due to these variations from the classical consensus Ntn fold, DmpA was described as the prototype of a new Ntn hydrolase variant.[60,64] It is however likely that the functional and topological similarities of DmpA and Ntn hydrolases developed from different evolutionary origins by convergent evolution.[71] Since attempts to obtain a crystal structure of the β-aminopeptidase DmpA in complex with a ligand or an inhibitor were unsuccessful, reliable structure-based information on the remarkable substrate specificities of β-peptide cleaving β-aminopeptidases is not available.[64] In the present study, we report the crystal structure of the β-aminopeptidase BapA from S. xenopeptidilytica 3-2W4. Furthermore, we present structures of BapA-ligand complexes with the protease inhibitor pefabloc SC as well as with the β-lactam antibiotic ampicillin and an ampicillin-derived hydrolysis product. The non-covalent BapA-ligand complexes allow us to assign important active site residues for ligand recognition and to speculate on the positioning of a β-peptidic substrate in the BapA active site. A comparison of the BapA complexes with DmpA indicates structural differences in the substrate binding pockets that are shown be responsible for the enzymes’ different substrate specificities. Results and Discussion Overall structure and topology of the catalytic BapA monomer Catalytically active BapA, which is comprised of 373 amino acids per subunit, was expressed and purified without its native 29 amino acid periplasmic signal sequence following a published procedure.[59] The structure of BapA was solved at a resolution of 1.45 Å by molecular replacement using the coordinate file of DmpA from O. anthropi (PDB ID: 1B65) as search model.[64] Except the carboxy terminal residues of the α-polypeptide chain (amino acids 245–249) and of the β-polypeptide chain (amino acids 372–373), all residues could unambiguously be fitted into the electron density map. Refinement statistics are listed in Table 6.1. 98 Crystal Structure and Inh1b1tion of BapA Table 6.1: Data and refinement statistics for the crystal structures of native BapA and of BapA complexed with pefabloc SC 1 (BapA_l), ampicillin 2 (BapA_2), and with the ampicillin-derived hydrolysis product 2hyd (BapA_2hyd). BapA BapA_l BapA_2 BapA_2hyd Space group P21 (No. 4) P21 (No. 4) P21 (No. 4) P21 (No. 4) Cell parameters 87.4 96.7 101.4 87.4 96.8 101.3 86.9 96.4 101.5 88.3 97.1102.2 90 108.2 90 90 108.4 90 90 108.4 90 90 108.7 90 Resolution (A) 50-1.45 50-1.8 50-1.7 50-1.8 (1.47- 1.45) (1.9- 1.8) (1.8- 1.7) (1.9-1.8) Observed reflections 901804 547436 570626 559930 Completeness (%) 98.5 (92.3) 97.6 (9.7) 98.7 (94.4) 98.4 (98.4) Redundancy 3.23 3.78 3.31 3.76 Rsym (%on 1) 4.6 (43.3) 10.0 (43.5) 11.6 (58.1) 9.0 (62.0) I/al 16.8 (2.7) 12.87 (3.07) 9.23 (2.19) 12.24 (2.99) RworJRtree (%) 14.79 I 16.32 17.09 I 20.3 17.50 I 19.64 16.02 I 18.39 Ordered waters 2026 1322 1236 1388 Bond lengths (A) 0.007 0.009 0.011 0.010 0 Bond angles ( ) 1.115 1. 186 1.329 1.148 Average B-factor 18.17 24.30 22.21 20.82 Residues in most 1349 (97 .05%) 1399 (96.35%) 1395 (96.07%) 1396 (96.54%) favored region Residues in generally 35 (2.02%) 47 (3.24%) 49 (3.37%) 44 (3.04%) allowed region Residues in disallowed 6 (0.45%) 6 (0.41%) 8 (0.55%) 6 (0.41%) region 99 Chapter 6 Like DmpA, the β-aminopeptidase BapA shares characteristic functional and topological properties of the Ntn hydrolase superfamily. BapA is comprised of four equal subunits (A, B, C and D) that assemble into a homotetramer of 154 kDa with three perpendicular symmetry axes P, Q and R (Figure 6.2, A).[59] Each subunit of BapA consists of a large α- and a small β-polypeptide chain that range from residues 1–249 and 250–373, respectively (Figure 6.2, B). The two polypeptide chains, which are encoded by a single open reading frame, are created by posttranslational, presumably autocatalyzed cleavage of a 38.6 kDa precursor polypeptide between Asn249 and Ser250. In analogy to DmpA and other Ntn hydrolases, Ser250 becomes the N-terminal catalytic nucleophile of the β-polypeptide chain upon processing of the BapA precursor polypeptide.[64,68] Each catalytic nucleophile Ser250 is located at the interface of three adjacent subunits that contribute to the formation of the four active centers (Figure 6.2, A). The core architecture of one BapA subunit comprises two layers of β-sheets flanked by an envelope of two α- helices on each side (Figure 6.2, B). This αββα-arrangement of the secondary structure elements in BapA resembles the consensus fold of the Ntn hydrolase family. However, the directionality and connectivity of the secondary structure elements of BapA and DmpA are distinct from the common Ntn fold.[70] Based on these variations from the classical Ntn fold, DmpA was described as the prototype of a new Ntn hydrolase variant.[64] Due to its functional and topological similarities to DmpA, BapA represents a new member of the DmpA-like variant of the Ntn hydrolase family. 100 Crystal Structure and Inhibition of BapA Figure 6.2: A: Surface representation of the native homotetramer of BapA from S. xenopeptidilytica 3-2W4 viewed along the P axis. The perpendicular Q and R axes are shown as dotted lines. The four subunits A–D are colored differently, each color representing a heterodimer consisting of one α- and one β-polypeptide. The position of the catalytic nucleophiles Ser250 are colored red and indicated by arrows. B: Cartoon representation of one subunit of BapA. The termini of the α- and β-polypeptide chains are indicated by arrows. Helix elements and β-sheets that contribute to the Ntn fold are depicted in green and yellow, respectively. The catalytic nucleophile Ser250 at the N-terminus of the β-polypeptide chain is shown as cyan sticks with oxygens and nitrogens colored red and blue, respectively. 101 Chapter 6 Crystal structure of BapA complexed with pefabloc SC (1) Biochemical studies with the β-aminopeptidase BapA recently demonstrated the inhibition of the enzyme by the broad-spectrum serine protease inhibitor pefabloc SC (1, see Scheme 6.1).[59] The generally proposed inhibition mechanism of 1 involves the formation of a covalent bond between the hydroxyl group of the enzyme’s active site serine residue and the reactive sulfonyl fluoride group of the inhibitor, resulting in the irreversible inactivation of the enzyme.[162] Upon determination of the native structure of BapA we soaked BapA crystals with compound 1, which was dissolved in mother liquor containing 1.5 M ammonium sulfate in 100 mM HEPES at pH 7.5. The structure of BapA in complex with the inhibitor was refined to a resolution of 1.8 Å (Table 6.1). Interestingly, we observed non-covalent binding of the inhibitor to the four active sites of the BapA homotetramer (Figure 6.3). In fact, the reactive sulfonyl fluoride group of the ligand pointed away from the catalytic nucleophile Ser250. The free amino group of 1 formed a salt bridge with the side chain carboxyl group of Glu133 and was hydrogen bonded to the hydroxyl group of Ser250. Residues of two neighboring subunits, i.e. Leu135 and Leu287 together with Leu84, Leu92, Phe124, Leu127 and Leu128 of the adjacent subunit, formed a hydrophobic cluster, which further supported ligand binding by van der Waals interactions. It is particularly interesting that the aromatic ring of Phe124, which could not be fitted in the non-complexed structure of BapA due to missing electron density, was stacked on the aromatic ring of the ligand in the BapA-ligand complex. H2O O O SF SOH + HF alkaline pH H2N O H2N O 1 Pefabloc SC (AEBSF) 1hyd Scheme 6.1: Hydrolysis of the serine protease inhibitor pefabloc SC (1) in alkaline solution. According to the instructions given by the supplier, the sulfonyl fluoride group of 1 spontaneously hydrolyzes in aqueous solution with a half-life of two hours at pH 7.5 and 37°C (Scheme 6.1). It is hence possible that 1 decomposed to 1hyd under the soaking conditions (1.5 M ammonium sulfate, 100 mM HEPES, pH 7.5) and that the non-covalently linked molecule observed in the active sites of the BapA-ligand complex corresponds to 1hyd 102 Crystal Structure and Inhibition of BapA (Figure 6.3). However, it was not possible to distinguish the fluoro-sulfonyl group of 1 from the sulfonyl group of 1hyd on the basis of the electron density map, because -F and -OH possess the same number of electrons (i.e. 9). Figure 6.3: Active site of BapA (green sticks) non-covalently complexed with the serine protease inhibitor pefabloc SC (1) or its hydrolyzed derivative 1hyd, which is shown in the figure as yellow sticks contoured at 1.6 σ with a 2 Fofc electron density map. Amino acid side chains of residues of the neighboring subunit that contribute to the formation of a hydrophobic cluster are shown in cyan. Oxygen atoms, nitrogen atoms and sulfur atoms are colored red, blue and magenta, respectively. Interactions between atoms are indicated as dotted lines with distances given in Å. Although pefabloc SC is of non-peptidic nature, the compound contains characteristic elements that resemble an N-terminal amino acid of a potential BapA substrate, i.e. a free amino group attached to a hydrophobic region. The observed interactions of pefabloc SC with active site residues of BapA suggest that salt bridge formation with Glu133 and van der Waals interactions with the hydrophobic cluster may as well play important roles for the binding of peptidic substrates. 103 Chapter 6 Inhibition of the β-aminopeptidases BapA and DmpA by penicillin-derived β-lactam antibiotics β-Lactam antibiotics of the penicillin family contain a β-lactam ring as part of their common 6-aminopenicillanic acid (6-APA) core structure (Figure 6.4). The penicillin-derived β-lactam antibiotics ampicillin (2), penicillin G (3) and carbenicillin (4) are distinguished structurally by the D-phenylglycyl, phenylacetyl and D-phenylmalonyl side chains that are, respectively, attached to the 6-APA nucleus. The characteristic β-lactam ring of compounds 2–4 represents an intramolecular cyclized β-amino acid and is therefore of high interest for β- peptide research. Hydrolytic opening of the β-lactam ring of β-lactam antibiotics, e.g. by means of a β-lactamase,[16,163] leads to linear β-amino acid substructures. As an example, the ampicillin-derived hydrolysis product 2hyd is depicted in Figure 6.4. Besides the hydrolysis-prone β-lactam ring, β-lactam antibiotics contain a second amide bond, which is susceptible to enzymatic cleavage by penicillin acylases.[164] These enzymes catalyze a deacetylation reaction, in which the side chain is removed from the β-lactam core. R H H N NH2 H S H S N OH O N N O O H O OH O OH O 2 Ampicillin: R = NH2 2hyd 3 Penicillin G: R = H 4 Carbenicillin: R = COOH Figure 6.4: Structures of the penicillin-derived β-lactam antibiotics ampicillin, penicillin G, carbenicillin (2–4) and the ampicillin-derived penicilloic acid derivative 2hyd. Although BapA and DmpA do not share particularly high sequence identity with members of the penicillin acylase or β-lactamase enzyme families, we considered experiments with the β- aminopeptidases and the β-lactam antibiotics 2–4 as interesting, because β-aminopeptidase- catalyzed biotransformations of cyclic β-amino acids or β-lactams have not been investigated yet. Hence, we examined the potential of the β-aminopeptidases BapA and DmpA to convert the penicillin derivatives 2–4. We observed neither deacylation nor hydrolysis of the β- 104 Crystal Structure and Inh1b1tion of BapA lactam core at pH 7.2 and 37°C with either enzyme, suggesting that BapA and DmpA lack penicillin acylase and p-lactamase activity on any of the tested compounds 2-4. Furthermore, we tested the inhibitory effect of the p-lactam antibiotics 2-4 and of the ampicillin-derived hydrolysis product 2 hyd on BapA and DmpA with the chromogenic reporter substrate H-(5)-p3hAla-,d'JA[62l at different molar ratios of substrate to inhibitor. Despite the structural similarity of the tested p-lactam antibiotics 2-4 with respect to their 6-APA core structure, only 2 inhibited substrate conversion by BapA with a IV value of 0.69 mM (Table 6.2) assuming a competitive inhibition model (Eq. 6.2, Experimental Section). Compound 2 differs from 3 and 4 only in the presence of a free amino group in its D-phenylglycyl side chain attached to the 6-APA core. In contrast to 2, compounds 3 and 4 carry phenylacetyl and D-phenylmalonyl substitutions, respectively, which both lack free amino groups. The structural change associated with conversion of 2 to 2hyd, which was induced by hydrolytic opening of the p-lactam ri ng, did not affect the inhibitory effect of the compound on substrate conversion by BapA (IV = 0.74 mM). I n combination with previously collected 9 62 information on the substrate specificity of BapA for N-terminally unprotected peptides, [s , l these data provide further evidence that the presence of a free amino group plays a key role for ligand recognition by BapA. Interestingly and in contrast to BapA, DmpA was not significantly inhibited by any of the compounds 2-4 and 2 hyd · Table 6.2: Inhibition of the p-aminopeptidase BapA by the penicillin-derived p-lactam antibiotics 2 and 2 hyd· The kinetic parameters Km, Vmax and K; of the reactions were obtained after fitting initial velocities of the p-aminopeptidase-catalyzed conversion of the chromogenic substrate H-(.5)-p3hAla- ,ci'JAC62l (pH 7.2, 37°C) to a competitive inhibition model (Eq. 6.2, Experimental Section). Penicillin (3) and carbenicillin (4) did not inhibit substrate conversion by BapA. Inhibitor K; Km Vmax [mM] [mM] [µmol·min-1 per mg of protein] without - 1.1 ± 0.15 16.4 ± 0.86 2 0.69 ± 0.05 1.1 ± 0.09 16.3 ± 0.50 2hyd 0.74 ± 0.11 0.98 ± 0.16 16.8 ± 1.0 105 Chapter 6 Crystal structures of BapA complexed with the ampicillin-derivatives 2 and 2hyd The identification of penicillin-derived inhibitors of the β-aminopeptidase BapA (Table 6.2) encouraged us to soak BapA crystals with solutions containing 2 or 2hyd. Two crystal structures of BapA in complex with 2 (BapA_2) and 2hyd (BapA_2hyd) were determined to 1.7 Å and 1.8 Å resolution, respectively (Table 6.1). Both 2 and 2hyd were identified in the four substrate binding pockets of the BapA homotetramer, but the ligands showed no covalent γ bond formation with O of the catalytic nucleophile Ser250 (Figure 6.5). The D-phenylglycyl side chains of both 2 and 2hyd were stabilized in a very similar manner by a dense hydrogen- bonding network. The free amino groups of the D-phenylglycyl side chains formed salt bridges with the side chain carboxyl group of Glu133 and were hydrogen bonded to the carbonyl oxygen of Thr76 and the hydroxyl group of Ser250. Furthermore, the backbone nitrogen atom of Leu135, the side chain hydroxyl groups of Ser250, and the carbonyl oxygen atom of Leu287 interacted with the peptide bonds of both ligands. As observed in the crystal structure of BapA complexed with pefabloc SC (Figure 6.3), hydrophobic residues of an adjacent subunit contributed to the formation of a hydrophobic cluster in each of the four BapA substrate binding pockets. These hydrophobic regions stabilized the aromatic rings of the phenylglycyl side chains of 2 and 2hyd by van der Waals interactions. Intramolecular hydrolysis of the β-lactam ring of 2 to the penicilloic acid derivative 2hyd led to a relaxation of the 6-APA nucleus, which caused different interactions of BapA active site residues with the C-terminal parts of 2 and 2hyd. The backbone nitrogen atom of Ser288 hydrogen bonded with the carbonyl oxygen atom of the β-lactam ring of 2 and the side chain of Arg138 interacted with the carboxyl group attached to the thiazolidine ring of the ligand. In contrast to 2, 2hyd did not interact with Arg138, but the free carboxyl group of 2hyd, which was created by hydrolytic cleavage of the β-lactam ring, moved into close proximity to Asn207 Nδ2 and to the free amino terminus of Ser250, thus compensating for the loss of the salt bridge with Arg138. 106 Crystal Structure and Inhibition of BapA Figure 6.5: Active sites of the BapA_2 (A) and BapA_2hyd (B) complexes in line representation. Ser250, Glu133 and Arg138 are highlighted as sticks and shown with 2 Fofc electron density contoured at 1.6 σ. Interactions of BapA active site residues with 2 and 2hyd (yellow sticks) are indicated by dotted lines with distances given in Å. Oxygen atoms, nitrogen atoms and sulfur atoms are colored red, blue and magenta, respectively. 107 Chapter 6 The positioning and the orientation of 2 in the complex structures depicted in Figure 6.5 give structure-based indications as to why BapA exhibits neither penicillin acylase nor β- lactamase activity. Most importantly, the observed distances between Ser250 Oγ and the carbonyl carbon atoms of the phenylglycyl side chain (2.5 Å) and of the β-lactam ring (4.2 Å) are too large to allow direct nucleophilic attack. Moreover, the measured angles between Ser250 Oγ and the planes of the respective carbonyl group of the phenylglycyl side chain (66°) and the β-lactam ring (79°) do not favor nucleophilic attack of Ser250 on either of the two carbonyl carbon atoms. According to the Bürgi-Dunitz trajectory, the specific angle for a nucleophile to approach the carbon-oxygen plane of a carbonyl group during nucleophilic attack should be 105 ± 5°.[165] Implications of ampicillin binding on β-peptide conversion by BapA As depicted in Figure 6.1, β-amino acids contain an additional carbon-carbon bond in the amino acid backbone leading to increased conformational flexibility when compared to their α-amino acid counterparts. Previous investigations showed that peptide conversion catalyzed by the β-aminopeptidase BapA strongly depends on the backbone length of the substrate’s N-terminal amino acid.[59,62] BapA specifically catalyzes the removal of backbone-elongated β-amino acids from the N-termini of peptides, whereas peptides with N-terminal α-amino acids are not converted. Although BapA hydrolyzes peptides with a wide range of differently substituted N-terminal β-amino acids, the enzyme has a clear preference for β3-amino acids with hydrophobic substituents of L-configuration. 3 Based on the crystal structure of the BapA_2 complex, we modeled the free N-terminal L-β - homophenylalanine residue of a peptidic substrate[62] into the BapA active site. Due to the second rotatable C–C bond in the β-amino acid backbone it was possible to place the amide bond of this substrate into an appropriate position for nucleophilic attack by Ser250 Oγ assuming (i) that the salt bridge interaction between Glu133 and the ligand’s free amino terminus is maintained and (ii) that the aromatic substituent of the modeled amino acid residue points into the hydrophobic binding pocket (Figure 6.6). In the modeled enzyme- substrate complex the carbonyl carbon atom of the substrate’s amide bond moved into close proximity to the catalytic nucleophile Ser250 Oγ (1.6 Å). The angle between Ser250 Oγ and 108 Crystal Structure and Inhibition of BapA the plane of the substrate’s carbonyl group was 108°, which would favor nucleophilic attack on the carbonyl carbon atom and thus, allow the formation of the acyl enzyme and the removal of the N-terminal β-amino acid residue from the C-terminal part of the peptide.[165] Furthermore, the proximity of the oxygen atom of the ligand’s carbonyl group to Asn207 Nδ2 (3.6 Å) and the main chain nitrogen atom of Leu135 (3.4 Å) indicates that these residues are likely to contribute to the formation of the oxyanion hole, which stabilizes the negatively charged tetrahedral transition state during substrate conversion by Ntn hydrolases.[70] Figure 6.6: Stick representation of an N-terminal β3-homophenylalanine residue of a peptidic BapA substrate (grey) modeled into the active site of BapA in complex with ampicillin (BapA_2) shown in yellow. The C-terminal rest of the peptide is indicated in orange. Residues of BapA are shown as green lines; Ser250, Glu133, Leu135 and Asn207 are highlighted as sticks. Oxygen atoms, nitrogen atoms and sulfur atoms are colored red, blue and magenta, respectively. Interactions between atoms are indicated by dotted lines with distances given in Å. Comparison of the BapA and DmpA substrate binding pockets Although BapA and DmpA share 42% amino acid identity, they are very distinct with respect to their specificities for β-peptidic substrates. While BapA catalyzes the removal of a wide range of differently substituted N-terminal β-amino acids from peptides, DmpA preferably 109 Chapter 6 hydrolyzes peptides with small N-terminal β-amino acids, such as β-homoglycine or β3- homoalanine.[62] Superposition of the DmpA structure (PDB ID: 1B65) on the structure of the BapA_2 complex reveals that the active sites of both enzymes are highly conserved. The position of Ser250 of BapA is virtually identical with Ser250 of DmpA and the side chain of Glu133 of BapA overlaps with the corresponding residue Glu144 of DmpA (Figure 6.7). Figure 6.7: Superposition of the substrate binding pockets of BapA_2 and DmpA, shown as cartoon representations in pale orange and pale green, respectively. Ampicillin (2) is depicted as yellow sticks. The side chains of the catalytic residues Ser250 and of the salt bridge-forming residues Glu133 and Glu144 of BapA and DmpA, respectively, are shown as sticks. Residues Gln131–Trp137 of DmpA and Glu120–Arg126 of BapA are highlighted in bright green and bright orange, respectively, with side chains shown as lines. Black and green labels refer to amino acid residues of BapA and DmpA, respectively. Oxygen atoms are colored red, nitrogen atoms blue and sulfur atoms magenta. In the previous sections, salt bridge formation between a ligand and Glu133 was shown to be essential for ligand binding by BapA. As suggested earlier,[64] Glu144 presumably fulfills this role in DmpA. Despite the similarity of the enzymes with regard to the positions of these important catalytic residues, the spaciousness of the substrate binding pockets of BapA and DmpA is notably different. The different geometries of the substrate binding pockets of BapA and DmpA are mainly caused by secondary structure elements of the adjacent subunit. A 110 Crystal Structure and Inhibition of BapA 310-helix comprising residues Glu120–Arg126 creates a fairly wide substrate binding pocket in BapA. This 310-helix is replaced by a loop region in DmpA ranging from Gln131 to Trp137, which protrudes towards Glu144 and Ser250 and constricts the active site of DmpA. In the superposed structures of DmpA and the BapA_2 complex the side chain of Trp137 clashes with the aromatic substituent of 2. From the superposition of the substrate binding pockets of BapA and DmpA (Figure 6.7) we assumed that the bulky side chain of Trp137 restricts catalysis by DmpA to the removal of sterically undemanding N-terminal amino acid residues from peptides, such as β- homoglycine and β3-homoalanine.[62] In order to widen the substrate binding pocket of DmpA and thus make it accessible for bulkier substrates we exchanged Trp137 of DmpA for alanine by site-directed mutagenesis creating the mutant enzyme DmpA W137A. It is important to mention that the point mutation W137A negatively affected the maturation of the DmpA precursor polypeptide, which processed only very slowly into the individual α- and β-polypeptides that form the catalytically active protein. Table 6.3 shows a comparison of the kinetic parameters of the conversion of the β3-homoamino acid p-nitroanilides H-βhGly- pNA, H-(S)-β3hAla-pNA, H-(R)-β3hAla-pNA, H-(S)-β3hPhe-pNA and H-(R)-β3hPhe-pNA by BapA, DmpA and DmpA W137A. While BapA converted the β3-homophenylalanine p- -1 -1 nitroanilides of S- and R-configuration with kcat/Km values of 14,500 and 940 M s , respectively, conversion of both enantiomers of H-β3hPhe-pNA by DmpA was negligible -1 -1 (kcat/Km < 10 M s ). The single amino acid exchange of Trp137 of DmpA to alanine drastically altered the substrate specificity of the enzyme leading to more than 280- and 670-fold improved catalytic efficiencies for conversion of the bulky β3-amino acid derivatives H-(S)-β3hPhe-pNA and H-(R)-β3hPhe-pNA, respectively. On the other hand, DmpA W137 converted substrates with small side chains , i.e. H-βhGly-pNA, H-(S)-β3hAla-pNA and H-(R)- β3hAla-pNA less efficiently than wild-type DmpA. (Table 6.3). In summary, these results demonstrate that DmpA could be rationally engineered to create a catalyst with catalytic properties similar to BapA by exchanging a single amino acid residue in the substrate binding pocket. Despite its widened substrate binding pocket, which allowed binding of the bulky β3-homophenylalanine derivatives, neither of the β-lactam-derived compounds 2–4 or 2hyd was converted or had an inhibitory effect on DmpA W137A; this is in contrast to the 111 Chapter6 observed inhibition of BapA by ampicillin (2) and the ampicillin-derived hydrolysis product 2 hyd (Table 6.2 and Figure 6.5). Table 6.3: Kinetic analyses of the conversion of different p3-homoamino acid p-nitroanilidesl62·m1 by the p-aminopeptidases BapA, DmpA and t he point mutant DmpA W137A at pH 7.2 and 37°C. BapA DmpA DmpA W137A Substrate Km Vmax kGJKm Km Vmax kGJKm Km Vmax kGJ Km (mM) [µmol· [M-1s-1] (mM) [µmol· [M-1s-1] (mM) [µmol· [M-1s-1] min-1 min-1 min-1 mg-1) mg-1) mg-1) H-j3 hG l y-~A 6.6 0.65 64 0.016 55.8 2,400,000 2.9 20.1 4,600 ± 2.3 ± 0.15 ± 37 ± 0.001 ± 0.8 ± 200,000 ± 0.2 ± 0.5 ± 400 H-( S)-j33hAla-~A 1.1 16.4 9,600 0.063 109 1,200,000 0.50 40.0 53,800 ± 0.2 ± 0.9 ± 1,800 ± 0.002 ± 1 ± 100,000 ± 0.02 ± 0.3 ± 2,100 H- ( R)- j33 hAl a-~A < 10CaJ 17.0 1.3 52 0.38 0.017 30 ± 2.7 ± 0.2 ± 15 ± 0.13 ± 0.002 ± 13 H-( S)- j33 hPhe-~A 3.5 79.3 14,500 < 10CaJ 0.85 3.5 2,800 ± 1.3 ± 14.5 ± 7,800 ± 0.09 ± 0.1 ± 400 H-( R)-j33hPhe-~A 940[aJ < 10CaJ 2.7 26.7 6,700 ±3 ± 0.4 ± 2.0 ± 1,600 Cal The enzyme activities showed a linear dependency on the substrate concentration between 0.1 and 5 mm. The kca JKm values were calculated according to the equation kcaJKm = v/([E0]·[S]), where v is the rate of the reaction, [S] the concentration of the substrate and [E0] the stoichiometric concentration of active centers (MaaDA: 38,610, MompA: 40,440, MompA W137A: 40,330). Structurally and functionally related enzymes 59 61 4 I n addition to the four charcterized 13-aminopeptidases, C - J l a search using the basic local alignment tool (BLAST) revea led a great number of translated nucleotide sequences that show similarities to the 13-aminopeptidase BapA from S. xenopeptldi/ytica. Most of these gene sequences are retrieved from genome sequencing projects. They originate mainly from Gram-negative bacteria, among them many pathogenic species such as Burkholder/a ma/lei, 8. pseudoma//ei and various strains of Pseudomonas aeruginosa, but also from archaea (e.g. Pyrococcus horikoshil), fungi (e.g. Aspergi//us oryzae) and yeast (e.g. Yarrowia /ipo/ytica) . We al igned the BapA sequence with the sequences of the 13-aminopeptidases DmpA from 0. 112 Crystal Structure and Inhibition of BapA anthropi, BapA from S. microcystinivorans and BapA from Pseudomonas sp. MCI3434 as well as with 18 selected translated nucleotide sequences that showed at least 38% amino acid identity to BapA. Two distinct sections of this alignment covering residues 115–145 and 250–265 of BapA are depicted in Figure 6.8. The presence of β-aminopeptidase-like amino acid sequences in various species of pro- and eukaryotic origin suggests that the ability to utilize synthetic β-peptides as substrates is not restricted to the four described β- aminopeptidases. In fact, it seems that β-peptide-cleaving enzymes are widely distributed among many different organisms. The isolation and characterization of such proteins could lead to the discovery of novel β-aminopeptidase-like enzymes with interesting new properties for biocatalytic conversions of β-amino acid-derived substrates. << 133 //250 >> 1. BapA from S. xenopeptidilytica1 VG---KVPEEALFSRLLPVVAETLDNRLNDVFGH // SLLIVIATDAPLMPHQ 2. DmpA from O. anthropi2 VDRYASTYQTDDFLWIMPVVAETYDGALNDINGF // SIIVVLATDLPLMPHQ 3. BapA from S. microcystinivorans3 IG---KVPEELLFSRVLPVVAETLDSRLNDVFGH // SLLIVIATDAPLMPHQ 4. BapA from Pseudomonas sp. MCI34344 R-EMG----KQHTYWCMPVVLETYDGTLNDIWGQ // SIVITIATDAPLLPHQ 5. Agrobacterium tumefaciens C58 IDRYSATYEGDEQLWIMPVVAETYDGLLNDINGQ // SIIVVIATDVPLAPHQ 6. Burkholderia mallei ATCC 23344 R-AQGGAREREHVYWCMPVVMETFDGLLNDIWGQ // SIVVTLATDAPLLPHQ 7. Burkholderia pseudomallei K96243 R-AQGGAREREHVYWCMPVVMETFDGLLNDIWGQ // SIVVTLATDAPLLPHQ 8. Burkholderia thailandensis E264 R-AQGGARDREHVYWCMPVVMETFDGLLNDIWGQ // SIVVTLATDAPLLPHQ 9. Mycobacterium gilvum PYR-GCK N----RVAPRLGRQWLLPVCAETWDGYLNDINGG // SVIAVIATDAPLLPGQ 10. Mycobacterium smegmatis MC2 155 K-SVRG----DEFFFSLPVVGETCDGILNDLNGF // SIIVLVATDAPLIPTQ 11. Myxococcus xanthus DK 1622 VDRYPG--IGDEHDVIIPVVGECDDSWLNDISGR // SIIAVVATDAPLLSHQ 12. Photorhabdus asymbiotica IDNNLTSPLKDDLFWLLPVVAETWDGILNDINGM // SIIAVVATDAPLAPHQ 13. Pseudomonas aeruginosa PACS2 R-AELG---DSGLYWCMPVVMETFDGLLNDIWGQ // SIVVILATDAPLLPHQ 14. Pseudomonas aeruginosa PAO1 R-AELG---DSGLYWCMPVVMETFDGLLNDIWGQ // SIVVILATDAPLLPHQ 15. Pseudomonas aeruginosa UCBPP-PA14 R-AELG---DSGLYWCMPVVMETFDGLLNDIWGQ // SIVVILATDAPLLPHQ 16. Pseudomonas entomophila L48 R-ERLP---DDAVYWCMPVVMETFDGVLNDIWGQ // SIVVVIATDAPLLPHQ 17. Pseudomonas fluorescens Pf-5 R-EALQ---DPAVYWCMPVVMETYDGLLNDIWGQ // SIVVILATDAPLLPHQ 18. Pseudomonas putida KT2440 R-EQQP--DDGRIYWNMPVVLETYDGLLNDINGF // SIVVCLATDAPLLPHQ 19. Stigmatella aurantiaca DW4/3-1 VERYPG--IGDELDVIIPIVGECDDSYLNDISGR // SIIAVVATDAPLLSHQ 20. Pyrococcus horikoshii OT3 LEENED--IGVTTGSVNPLVLECNDSYLNDIRGR // SIIMIIATDAPLTGRQ 21. Aspergillus oryzae RIB40 VKHLGTNKDGQLEWLMLPVVGETFDGYLHDCTSF // SIIVVLATDAPLHPAQ 22. Yarrowia lipolytica A----KHFPKVAAQWMLPVVGETWDGYLNEINGD // SIIVVIATDAPLLPGQ Figure 6.8: Alignment of the amino acid sequence of BapA from S. xenopeptidilytica (1) with three previously characterized β-aminopeptidases (2–4; references are given as superscript numbers) and 18 selected similar translated nucleotide sequences that were retrieved from a BLAST search (5–22). Two sections (residues 115–145 and 250–65) of the aIignment, which was created with the program ClustalW,[166] are shown in the figure; identical amino acid residues are highlighted in black, similar residues are marked grey. Numbers refer to amino acid residues of BapA from S. xenopeptidilytica. The physiological significance of β-aminopeptidase-catalyzed β-peptide conversion remains elusive since peptides solely composed of β-amino acids are not known in nature. However, β-amino acid-derived substructures occur in a wide range of natural compounds, such as microcystin,[12] taxol,[11] bestatin,[14,15] and in antibiotics of the β-lactam family (see also 113 Chapter 6 Figure 6.4).[16] The presence of these β-amino acid-containing compounds gives rise to speculations about the natural substrates and the physiological roles of β-aminopeptidases, which could participate in the breakdown of these compounds to make them available for further conversions. It is striking that the regions downstream of the N-terminal catalytic nucleophile Ser250 and around the salt-bridge forming residue Glu133 are highly conserved among all compared sequences indicating that these regions are crucial for catalysis (Figure 6.8). In contrast to these highly conserved regions, the compared sequences show great variability upstream of residue 128. As indicated by the superposition of the active sites of BapA and DmpA (see Figure 6.7) and the kinetic data presented in Table 6.3, this variable region defines the width of the enzymes’ substrate binding pockets and hence determines the specificity of the enzymes for substrates carrying N-terminal β-amino acids with side chains of different sizes. Our results obtained with the DmpA mutant W137A (Table 6.3) exemplify that specific amino acid exchanges in this particularly variable region could lead to the creation of new β- aminopeptidase variants with novel properties or improved functionality for biocatalytic applications. Conclusions We solved the crystal structure of the β-aminopeptidase BapA, a β-peptide-converting enzyme from S. xenopeptidilytica 3-2W4. The architecture of the secondary structure elements in BapA resembles the common αββα-sandwich fold of the N-terminal nucleophile (Ntn) hydrolase superfamily. Additional structural information was obtained from non- covalent BapA-ligand complexes with the protease inhibitor pefabloc SC, the β-lactam antibiotic ampicillin and a hydrolyzed ampicillin derivative. Neither ampicillin-derived compound was converted by β-aminopeptidases, but both inhibited the BapA-catalyzed conversion of a β-peptidic reporter substrate. Despite these interesting observations, the physiological relevance of β-lactam binding to BapA remains elusive. The phenylglycyl side chains of the ampicillin-derived ligands resemble the N-termini of conventional α-peptides and binding of these molecules to the BapA active site suggests an essential role for Glu133 (i) to orient the substrate by salt bridge formation with its free N-terminal amino group and 114 Crystal Structure and Inhibition of BapA (ii) to position the peptide bond of the substrate for nucleophilic attack by Ser250 Oγ. In contrast to the narrow substrate binding pocket of DmpA, which restricts DmpA to β-peptidic substrates with small side chain substituents, the BapA active site is spacious enough to accommodate ligands with bulky N-terminal β-amino acids. The exchange of Trp137 of DmpA to alanine opened the substrate binding pocket of the resulting DmpA mutant W137A and conferred specificity for sterically more demanding β3-amino acid derivatives. As the conversion of pharmaceutically interesting synthetic β-amino acid-derived compounds is a rare catalytic feature among enzymes, β-aminopeptidases are valuable biocatalysts for the enzymatic production of building blocks for β-peptide-based drug design. The structures of the BapA-ligand complexes together with kinetic data of wild-type and mutant β- aminopeptidases presented herein provide a starting point for structure-based enzyme engineering approaches to create β-peptide-converting catalysts with novel properties and improved selectivities. Experimental Section General remarks Pefabloc SC (1), ampicillin (2), penicillin G (3) and carbenicillin (4) were purchased from Sigma-Aldrich (Buchs, Switzerland). H-βhGly-pNA was obtained from Bachem (Bubendorf, Switzerland), the β3-homoamino acid p-nitroanilides H-(S)-β3hAla-pNA, H-(R)-β3hAla-pNA, H- (S)-β3hPhe-pNA and H-(R)-β3hPhe-pNA were prepared chemically following established procedures.[62,111] The enzymatic conversions of compounds 2–4 were followed by HPLC analysis on a Dionex HPLC system equipped with a P680 pump, an ASI-100 automated sample injector, an UltiMate 3000 thermostatted column compartment and a UVD 340U photodiode array detector (Dionex, Sunnyvale, CA, USA). 10 μl of the samples were loaded onto the reversed-phase HPLC stationary phase Nucleodur C18-Pyramid (250 × 4 mm, 5 μm particle size; Macherey-Nagel, Düren, Germany), which was equilibrated with 0.1% formic acid in water for 3 min at a column temperature of 20°C. The compounds were separated with a linear gradient of 0 to 60% acetonitrile within 15 min at a constant flow rate of 1 ml/min and detected at a wavelength of 205 nm. Mass spectra of all compounds were recorded with an API 4000 liquid chromatography/tandem MS system connected to an 115 Chapter 6 Agilent 1100 LC system. For protein determination, we used five-fold concentrated Bradford reagent (Bio-Rad, Rheinach, Switzerland) and bovine serum albumin (BSA) as a standard; absorbance measurements were performed at 595 nm with a Specord S 100 spectrophotometer (Analytik Jena, Jena, Germany). Proteins were analyzed by SDS-PAGE using pre-cast 10% Novex tricine gels (Invitrogen AG, Basel, Switzerland) according to the manufacturer’s instructions. The protein gels were stained with Coomassie Brillant Blue and the purities of the enzymes were estimated by calculating the relative intensities of the protein bands with a GS-800 calibrated imaging densitometer and the software Quantity One (Bio-Rad, Rheinach, Switzerland). Enzymatic production of 2hyd 2hyd was produced from 2 by enzymatic conversion with the β-lactamase that was encoded on the pET3c plasmid (Novagen, Madison, WI, USA), constitutively expressed in E. coli BL21(DE3) pLysS and partially purified by anion exchange and hydrophobic interaction chromatography. 50 mg of compound 2 were dissolved in 5.2 ml H2O and the enzymatic reaction was started by the addition of the β-lactamase containing solution to yield a final protein concentration of 0.015 mg/ml. After 16 h of incubation at 30°C, the enzymatic reaction was stopped by removing all proteins > 10 kDa by centrifugation with Centricon YM-10 centrifugal devices (10 kDa MWCO; Millipore Corp., Billerica, CA, USA). The formation + of 2hyd was verified by HPLC-MS analysis (tR = 10.3 min, [M + H] : 368.0 (calcd. M: 367.1), purity after peak integration > 90%). The solution containing 2hyd was frozen to -80°C and freeze-dried. Expression and purification of BapA for protein crystallization The β-aminopeptidases BapA from S. xenopeptidilytica 3-2W4 was expressed without its N- terminal 29 amino acid signal sequence in a pET3c-expression system and purified from the E. coli host.[59] After the two-step chromatographic procedure the BapA-containing fractions were collected and concentrated to a volume of 1 ml with YM-3 Centriprep and Centricon centrifugal devices (3 kDa MWCO; Millipore AG, Zug, Switzerland). As an additional purification step, the concentrated protein sample was loaded onto a Superdex 200 size exclusion column (1.6 × 60 cm; GE Healthcare, Uppsala, Sweden) at a flow rate of 1 ml/min 116 Crystal Structure and Inhibition of BapA with a 50 mM Tris/HCl buffer (pH 8) containing 150 mM KCl. Throughout the purification procedure the temperature was kept constant at 4°C. The BapA-containing fractions were collected and dialyzed three times with a Spectra/Por dialysis membrane (6–8 kDa MWCO; Spectrum Laboratories Inc., Rancho Dominguez, CA, USA) against 2 L of a 0.5 mM Tris/HCl buffer (pH 8). The dialyzed protein sample was again concentrated by centrifugation to a protein concentration of 15 mg/ml. Crystallization, data collection, and structure determination[J] Crystals were grown in 2 μl drops with a 1:1 ratio of mother liquor (1.8 M ammonium sulfate and 100 mM Tris/HCl, pH 8) to purified BapA (15 mg/ml) in 24-well Cryschem sitting drop plates (Hampton Research, Aliso Viejo, CA, USA) using the vapor diffusion method at 20°C. Within one day, crystals of BapA grew as stacked layers of thin plates (Figure 6.9, A). Sequential macroseedings with crystal fragments, which were transferred to new wells containing 1.5 M ammonium sulfate, 100 mM Tris/HCl (pH 8), yielded single three- dimensional, mostly plate-shaped crystals of up to 300 × 300 × 30 μm in size (Figure 6.9, B and C). The crystals were soaked for 5 min in a solution of 1.5 M ammonium sulfate, 100 mM HEPES (pH 7.5) containing 50 mM pefabloc SC (1) and for 30 min in 1.5 M ammonium sulfate, 100 mM HEPES (pH 7.5) containing saturating concentrations of ampicillin (2) or the ampicillin- derived hydrolysis product 2hyd. After cryo-protection in a solution containing 1.5 M ammonium sulfate, 100 mM HEPES (pH 7.5) and 30% glycerol, the crystals were flash frozen in liquid nitrogen and measured at the Swiss synchrotron light source (SLS, PSI Villigen, Switzerland; beamline PX06). Diffraction data up to 1.45 Å resolution on a total of 720 frames were recorded with an oscillation range of 0.5° per frame. The data was processed with the program XDS.[167] Initial phases were obtained by molecular replacement using the program PHASER[168] and the structure of DmpA (PDB ID: 1B65) as search model. The asymmetric units each contained one molecule of the BapA homotetramer, which was refined using the program PHENIX.[169] The model building and superpositions for figures [J] Crystal structures were solved and refined by Tobias Merz, University of Zürich. 117 Chapter 6 were performed with the programs COOT[170] and SSM[171], respectively. Figures were created with the program Pymol.[172] Figure 6.9: BapA crystals grown in 2 μl drops before (A; 1.8 M ammonium sulfate, 100 mM Tris/HCl, pH 8) and after macroseeding (B and C; 1.5 M ammonium sulfate, 100 mM Tris/HCl, pH 8) at a protein concentration of 15 mg/ml. Expression of BapA, DmpA and DmpA W137A for enzymatic activity assays To exclude the possibility of β-lactamase cross-contamination from enzyme expression in a β-lactamase-encoding pET3c expression system, the genes bapA from S. xenopeptidilytica 3- 2W4 and dmpA from O. anthropi LMG7991 were cloned into pET9c vectors yielding pAR116 and pAR114, respectively.[161] DmpA W137A was directly generated from the expression vector pAR114 with the QuikChange multi site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA) using the primer 5’-GACCGACGATTTCCTCGCGATCATGCCGGTTGTC-3’; the codon change that introduced the point mutation is underlined. The enzymes were recombinantly expressed in E. coli BL21(DE3) pLysS and purified according to established procedures.[59,62] The lyophilized enzyme powders were dissolved in a 10 mM potassium phosphate buffer at pH 7.2, and the protein contents of the enzyme stocks were determined spectrophotometrically. The purities of the dissolved enzymes were estimated by digital image analysis of the protein gels. 118 Crystal Structure and Inhibition of BapA Penicillin acylase and β-lactamase activity assay The reaction mixtures contained 1 mM of the β-lactam antibiotics 2, 3 or 4 in 100 mM potassium phosphate buffer at pH 7.2 and 37°C. The enzymatic reactions were initiated by the addition of one of the enzymes BapA or DmpA. Samples were withdrawn at intervals from the reaction mixtures and the enzymatic reactions were quenched by heating at 90°C for 3 min. The samples were analyzed by reversed-phase HPLC and HPLC-MS. Enzymatic activity and inhibition assay The β-aminopeptidase-catalyzed conversion of the β3-homoamino acid p-nitroanilides H- βhGly-pNA, H-(S)-β3hAla-pNA, H-(R)-β3hAla-pNA, H-(S)-β3hPhe-pNA and H-(R)-β3hPhe-pNA was determined spectrophotometrically by monitoring the release of free p-nitroaniline at a -1 -1 [62] wavelength of 405 nm (ε405 = 8,800 M cm ). The assay mixtures contained different molar concentrations of the substrates in a 100 mM potassium phosphate buffer at pH 7.2 an 37°C; 10% DMSO was added as a co-solvent. The enzymatic reactions were initiated by the addition of one of the enzymes BapA, DmpA or DmpA W137A and followed for one minute. The initial rates of the enzyme-catalyzed release of p-nitroaniline were calculated from the initial slopes of the obtained curves. The kinetic parameters Km and Vmax of the enzymatic reactions were estimated by fitting experimental data to the Michaelis-Menten model presented in Eq. 6.1 by regression analysis with the VisualEnzymics software (Softzymics, Princeton, NJ, USA) for the program IGOR Pro (WaveMetrics, Oswego, OR, USA). ⋅[]SV v = max (Eq. 6.1) m + []SK The inhibitory effect of the compounds 2–4 and 2hyd on the β-aminopeptidase-catalyzed conversion of the reporter substrate H-(S)-β3hAla-pNA was determined spectrophotometrically as described above. The assay mixtures contained the reporter substrate and the inhibitors in different molar ratios. The kinetic parameters Km and Vmax of the reaction as well as the inhibitory constants Ki were estimated by fitting experimental data to the competitive inhibition model presented in Eq. 6.2. ⋅[]SV v = max (Eq. 6.2) []I Km )1( ++⋅ []S Ki 119 Protein Maturation, Activity and Precursor Crystal Structure of BapA 7. Protein Maturation and Activity of BapA and Crystal Structure of a Processing-Deficient BapA Precursor Abstract β-Aminopeptidases are Ntn hydrolases that require posttranslational cleavage of an immature precursor polypeptide to form the catalytically active protein. As the formation of the enzyme active site directly depends on successful processing of the precursor, it is difficult to study the influence of single amino acid residues on precursor processing and enzymatic activity independently. The separate, but concurrent expression of the BapA α- and β-polypeptide chains on a single expression vector uncoupled catalytic activity from precursor processing allowing for independent investigations of these two processes. The introduction of point mutations in wild-type (WT) BapA and in the coexpressed polypeptide chains demonstrated that the active site residues Ser250, Glu133, Ser288 and Glu290 have important functions for both precursor cleavage and enzymatic activity. The crystal structure of the unprocessed BapA precursor mutant S250A revealed that the immature BapA precursor mutant shares characteristic structural features of the WT protein, i.e. (i) a homotetrameric quarternary structure and (ii) an Ntn-like αββα-sandwich fold. The designated active site of the unprocessed precursor is blocked by a loop comprising residues Ala238 to Asn249, which translocates to the periphery of the protein upon precursor cleavage. 121 Chapter 7 Introduction As outlined in Chapter 6, β-aminopeptidases share characteristic functional and topological properties with enzymes of the N-terminal nucleophile (Ntn) hydrolase superfamily.[63,67,70] Mutagensis studies with the β-aminopeptidase DmpA from O. anthropi revealed that the exchange of the catalytic nucleophile Ser250 for alanine, cysteine or threonine interfered with protein maturation and trapped the protein in an unprocessed, proteolytically inactive precursor state (see also Figure 1.3).[60] Recently, the crystal structure of an unprocessed precursor mutant of E. coli L-asparaginase EcAIII, another member of the Ntn hydrolase family, gave interesting insights into the molecular mechanism of posttranslational precursor cleavage.[173] However, the fact that the formation of the catalytically active protein directly depends on the successful processing of the premature protein makes it difficult to study the mechanistic details of precursor cleavage and enzymatic activity independently. An elegant approach to uncouple enzymatic activity from precursor processing was recently shown by Boanca et al., who expressed the two polypeptide chains of the Ntn hydrolase γ- glutamyltranspeptidase from Helicobacter pylori separately but concurrently from a single expression vector.[174] Coexpression of the α- and β-polypeptides resulted in active protein that allowed independent investigations of the mechanisms of precursor cleavage and substrate conversion. In this chapter, the construction of a coexpression system for the β-aminopeptidase BapA from S. xenopeptidilytica 3-2W4 is described. Point mutations of residues in proximity to Ser250 were introduced in the separately expressed subunits and in the WT protein in order to study their effect on precursor cleavage and catalytic activity. Moreover, the crystal structure of a processing-deficient 3-2W4 BapA precursor mutant is presented herein. 122 Protein Maturation, Activity and Precursor Crystal Structure of BapA Results and Discussion Crystal structure of a processing-deficient BapA precursor mutant The maturation process of wild-type (WT) BapA involves the posttranslational, presumably autocatalyzed cleavage of a 38.6 kDa precursor polypeptide into a large α-polypeptide chain of 25.4 and a small β-polypeptide chain of 13.2 kDa (Figure 7.1, A). As shown in Chapter 6, four α- and four β-polypeptide chains assemble into the homotetrameric conformation of the native BapA protein. The exchange of the catalytic nucelophile Ser250 for alanine by site- directed mutagenesis trapped the BapA mutants in an unprocessed, catalytically inactive precursor state (Figure 7.1, B). A TCG B GCG gene 5’ 3’ bapA 5’ 3’ bapA S250A Ser250 Ala250 protein precursor precursor Ser250 α β WT BapA αβ BapA S250A α4β4 ( )4 Figure 7.1: Schematic representation of protein maturation in WT BapA (A) and in the processing- deficient BapA mutant S250A (B). Crystals of the partially purified BapA mutant S250A were grown under very similar conditions as described for WT BapA (see Chapter 6), and the crystal structure of the 123 Chapter 7 processing-deficient precursor was solved at 1.8 Å resolution.[K] Like WT BapA (see Figure 6.2, A), BapA S250A adopted a homotetrameric quarternary structure. A structural comparison of the secondary structure elements of the unprocessed BapA mutant S250A with processed WT BapA in complex with ampicillin (see Chapter 6) is shown in Figure 7.2. It reveals that the unprocessed precursor polypeptide adopts essentially the same conformation as the correctly folded WT protein. A loop region ranging from Ala238 to Asn249 in the precursor polypeptide (Figure 7.2, A) spans the designated substrate binding pocket of the native protein (Figure 7.2, B). Upon processing of the precursor polypeptide between Asn249 and Ser250, this loop becomes the C-terminal region of the newly created α-polypeptide chain. It translocates from the cleavage site to the periphery of the protein allowing the substrate binding pocket to adopt a catalytically active geometry. Mutagenesis of WT BapA Following determination of the crystal structure of the processing-deficient BapA mutant S250A (Figure 7.2), it was interesting to study the effects of additional amino acid exchanges in the vicinity of Ser250 on precursor processing. A comparison of the molecular environments of the cleavage site in the unprocessed BapA mutant S250A (A) with the active site in the native BapA protein complexed with ampicillin (B) is presented in Figure 7.3. Due to their immediate proximity to Ala250 in the unprocessed precursor mutant (Figure 7.3 A), residues Glu133, Asn249, Ser288 and Glu290 were chosen as targets for further mutagenesis. Point mutations at the indicated positions were introduced by site-directed mutagensis to give the BapA mutants E133A, N249A, S288A and E290A, respectively. In order to study the effects of these mutations on precursor processing, the mutant proteins were partially purified and incubated at a protein concentration of 0.6 mg/l in 50 mM Tris/HCl buffer (pH 8) at 37°C. Samples were taken after 0, 54, 120 and 192 h and the maturation state of the proteins was analyzed by denaturing SDS-PAGE. [K] The crystal structure of BapA S250A was solved and refined by Tobias Merz, University of Zürich. 124 Protein Maturation, Activity and Precursor Crystal Structure of BapA Figure 7.2: Three-dimensional structures of an unprocessed subunit of the BapA mutant S250A (A) and of a processed subunit of WT BapA complexed with ampicillin (B; see also Chapter 6). Ala250 (A) and Ser250 (B) are highlighted in yellow with oxygen and nitrogen atoms colored red and blue, respectively. The loops from Ala238 to Asn249 are colored magenta, ampicillin (B) is shown as cyan sticks. 125 Chapter 7 Figure 7.3: Stick representations of the cleavage site of the unprocessed BapA mutant S250A (A) and of the active site of WT BapA in complex with ampicillin (B, see also Chapter 6). Oxygen and nitrogen atoms are colored red and blue, respectively, ampicillin (B) is depicted in cyan. The position of a conserved water molecule in A is indicated by a brown star. Interactions between atoms are shown as dotted lines with distances given in Å. 126 Protein Maturation, Activity and Precursor Crystal Structure of BapA The protein gels depicted in Figure 7.4 show that all introduced mutations (S250A, S250C, E133A, S288A, E290A and N249A) negatively affected the maturation of the respective precursor polypeptides. In contrast to WT BapA, which processed rapidly into its α- and β- polypeptide chains during protein expression and purification (see also Figure 1.3), the unprocessed (αβ)-precursor polypeptide was detected in all BapA mutants. In the case of the BapA mutants S250A, S288A and E290A the only detectable protein band after 192 h of incubation corresponded to the unprocessed protein subunit (αβ), and no bands corresponding to the single α- and β-polypeptide chains were detected. However, crystallization of the BapA mutants S288A and E290A was not feasible, as the premature polypeptide chains were not stable under the crystallization conditions (1.6 M ammonium sulfate, 100 mM Tris/HCl, pH 8) and underwent processing into the α- and β-subunits in the crystallization wells. This observation indicates that precursor cleavage was induced in these mutants at high concentrations of salt and/or protein (15 mg/ml). When Ser250 was exchanged for cysteine, we observed very slow processing of the precursor over the assay period. Hence, the exchange of Ser250 for cysteine as an alternative nucleophile does not abolish protein maturation completely, but drastically reduces its efficiency. From these experiments we conclude that Ser250, Ser288 and Glu290 are essential for efficient cleavage of the (αβ)-precursor polypeptide and that an exchange of these residues leads to mutants with greatly reduced processing capability. The precursor polypeptides of the BapA mutants E133A and N249A processed slowly into the α- and β-polypeptide chains. According to densitometric analysis of the protein bands, the yield of processed BapA N249A increased steadily with time to 83% after 192 h of incubation (Figure 7.4). This observation shows that the presence of Asn249 is important, but not indispensible for the precursor cleavage process. In the case of the BapA mutant E133A, 51% of the protein was present in the cleaved form after 54 h, but no further processing of the precursor was observed over the residual assay period. It is hence possible that the sample of BapA E133A contained two different protein fractions: one, which has the ability to undergo precursor cleavage, and another one, which lacks this ability. Upon precursor processing after 192 hours of incubation, BapA N249A converted the chromogenic substrate H-β3hAla-pNA[62] with a specific activity of 23 μmol min-1 mg-1 of protein, which is in the same range as the specific activity of purified WT BapA (25 μmol min-1 mg-1 of 127 Chapter 7 protein). This indicates that Asn249 reduces the efficiency of precursor cleavage, but has no influence on the catalytic mechanism of substrate conversion. In contrast to BapA N249A, BapA E133A remained completely inactive, even after partial cleavage of the precursor polypeptide. These observations are consistent with the crystal structures of WT BapA, which suggested an essential role for Glu133 in ligand binding (see Chapter 6). In contrast to Glu133, which remains at the active site after precursor processing, Asn249 translocates from the cleavage site to the periphery of the BapA subunit and hence does not interact with other amino acid residues of the active site (Figures 7.2 and 7.3). S250A S250C E133A 0 1 2 3 0 1 2 3 0 1 2 3 (αβ) α β S288A E290A N249A 0 1 2 3 0 1 2 3 0 1 2 3 (αβ) α β Figure 7.4: SDS-PAGE analysis of six processing-deficient or slow-processing BapA point mutants. The partially purified mutant proteins (0.6 mg/ml) were incubated in 50 mM Tris/HCl buffer (pH 8) at 37°C. Samples were taken after 0 h (0), 54 h (1), 120 h (2) and 192 h (3). 128 Protein Maturation, Activity and Precursor Crystal Structure of BapA Uncoupling enzymatic activity from precursor proce.ssing As shown in the previous section, the exchange of amino acid residues in proximity to Ser250 abolished or decelerated the maturation process of the BapA precursor polypeptide into t he catalytically active protein (Figure 7.4). To study the effects of point mutations on the enzymatic activity independently from precursor processing, we expressed the a- and 13- polypeptide chains separately, but concurrently on a pETDuet dual-expression vector (Figure 7.S).l174l The same point mutations that were previously investigated for their effects on BapA precursor processing were introduced into the coexpressed a - and 13-polypeptide chains, yielding the recombinant Duet-BapA mutants S250A, S250C, E133A, S288A and E290A. The point mutation N249A was not included, because Asn249 translocates to the periphery of the protein after precursor processing and was shown in the previous sections not to contribute to the catalysis of substrate cleavage (Figure 7.3). A B I ATGTCG I I ATGGCG I gene 5' ______:: =Iii iiiiiiiiiiii~ 3' duet 5, -----....'.:=Iii iiiiiiiiiiii~ 3, duet bapA bapA S250A protein a ~ a ~ ' p ' p a ~ Duet-BapA a ~ 4 4 4 4 Duet-BapA S250A Figure 7.5: Schematic representation of protein maturation in Duet-BapA (A) and in the Duet-BapA mutant S2SOA (B). The additional inserted start codon is underlined. The N-terminal methionine residue is posttranslationally removed from the ~-polypeptide chain by E coli methionyl aminopeptidase (MAP). To allow the separate expression of the a- and 13-polypeptide chains, an additional start codon had to be introduced into the gene sequence encoding the 13-polypeptide. Hence, we 129 Chapter 7 had to ensure that the N-terminal methionine residue, which was encoded by the newly inserted start codon, was removed after recombinant expression of the Duet-BapA enzymes. The cleavage of N-terminal methionine residues from proteins commonly occurs in a significant fraction of cytosolic E. coli proteins and is catalyzed by the enzyme methionyl aminopeptidase (MAP). The efficiency of MAP-catalyzed methionine removal was shown to depend on the penultimate amino acid, with the extent of N-terminal methionine excision negatively correlating with the side chain length of the penultimate amino acid.[175] For all investigated proteins containing serine and alanine residues in the second position, the N- terminal methionine residue was almost quantitatively removed by MAP upon translation. Partial purification of the Duet-BapA mutant S250A and N-terminal sequence analysis of the primary structure revealed alanine as the N-terminal amino acid residue in > 97% of the β- polypeptide fraction. This result indicates that N-terminal methionine is completely excised by E. coli MAP from the β-polypeptide chains of the Duet-BapA constructs. The separately expressed polypeptide chains were able to form catalytically active enyzme, as demonstrated by measuring β-aminopeptidase activity with the chromogenic substrate H- β3hAla-pNA[62] in the cell-free extracts. However, the specific activity of the cell-free extract containing Duet-BapA (0.18 μmol min-1 mg-1 of protein) was 40-times lower than the specific activity determined for the WT BapA-containing sample (7.0 μmol min-1 mg-1 of protein). SDS-PAGE analysis of the crude extracts revealed clear protein bands for the overexpressed α- and β-polypeptide chains in the WT BapA-containing sample, whereas no bands corresponding to the two separately expressed polypeptide chains of Duet-BapA were visible (gel not shown). These results indicate that the individually expressed α- and β-polypeptide chains were produced at much lower levels than the WT protein. Moreover, it is likely that the individual α- and β-polypeptides cannot adopt a stable conformation, but require one another to fold into the ordered, conformationally stable quarternary structures of the αβ- monomer or of the catalytically active α4β4-homotetramer (see Figure 7.2, A). Hence, the separately expressed, non- or misfolded α- and β-polypeptide chains presumably undergo rapid degradation by cytosolic proteases in the E. coli host.[176-178] All investigated point mutations further reduced or completely abolished the enzymatic activity of the Duet-BapA mutants S250A, S250C, E133A, S288A and E290A. No β- 130 Protein Maturation, Activity and Precursor Crystal Structure of BapA aminopeptidase activity was detected in the cell-free extracts containing Duet-BapA S250A and Duet-BapA E133A. These results further show the importance of Ser250 for nucleophilic attack on the substrate and Glu133 for substrate binding at the active site (see Chapter 6). The cell-free extracts containing Duet-BapA S250C, Duet-BapA S288A and Duet-BapA E290A maintained residual levels of β-aminopeptidase activity, but these were too low to allow accurate quantification. Conclusions The objective of the experiments presented in this chapter was to study the importance of single amino acid residues for precursor cleavage and enzymatic activity of BapA comparatively, but independently. Point mutations in WT BapA gave a direct indication of the affects of exchanging each residue on posttranslational cleavage of the precursor polypeptide. The same point mutations introduced into the separately expressed α- and β- polypeptide chains revealed the relevance of these amino acids for the catalytic mechansim of BapA-catalyzed β-peptide conversion. The effects of the inserted point mutations on precursor processing and on enzymatic activity of BapA are summarized in Table 7.1. In general, mutations that negatively affected precursor cleavage also caused a drastic loss of enzymatic activity. Ser250 not only acts as the catalytic nucleophile in the correctly folded protein, but it is also crucial for precursor processing. Precursor cleavage in Ntn hydrolases is suggested to be autocatalytic involving nucleophilic attack of the hydroxyl group of the catalytic nucleophile on the preceeding carbonyl carbon atom.[67,173] Even the exchange of Ser250 for cysteine, which is the most closely related amino acid, resulted in mutants that possessed greatly reduced processing capability and enzymatic activity. As expected from the results obtained in Chapter 6, Glu133 was also crucial for enzymatic activity as it promotes substrate binding to the BapA active site. Additionally, the exchange of Glu133 for alanine decreased the efficiency of precursor processing drastically, but not completely. On the other hand, Ser288 and Glu290 turned out to be essential for precursor processing, but the exchange of these residues for alanine still maintained low levels of enzymatic activity. The obtained results reveal interesting molecular insights into the molecular details of 131 Chapter 7 precursor processing and substrate cleavage by p-am inopeptidases and may contribute to the formulation of the underlying catalytic mechansims. Table 7.1: Effects of point mutations in the vicinity of Ser250 of BapA and Duet-BapA on precursor processing and enzymatic activity. n.i. not investigated. Expression as precursor polypeptide Expression as individual polypeptide chains Point mutation Precursor cleavage Enzymatic Formation of enzymatically activity[a] active protein[bJ - + 100% + S250A - - - S250C (-) - (+) E133A slow, partial - - N249A slow, complete 100% n.i. S288A - - (+) E290A - - (+) [aJ Measured after 192 hours of incubation of the partially purified precursor polypeptides. [bJ Measured in the cell-free extracts containing the Duet-BapA mutants. As shown in Chapter 6, three neighboring subunits, each consisting of one a- and one p- polypeptide chain, contribute to the formation of each of the four active site pockets in the BapA homotetramer (Figure 6.2 A). It is hence likely that the enzyme needs to adopt its native a4p4-quarternary structure to cata lyze substrate conversion. The fact that four unprocessed BapA precursor polypeptides folded into the same homotetrameric conformation as the WT protein suggests that the formation of the homotetramer precedes the cleavage of the precursor polypeptides. 132 Protein Maturation, Activity and Precursor Crystal Structure of BapA Experimental Section Generation of the Duet-BapA expression construct The gene sequences coding for the individual a- and 13-polypeptide chains of BapA were amplified by PCR from the plasm id p3BapA containing bapA from S. xenopeptidilytica 3-2W4 without its periplasmatic signal sequence. [59l The forward and reverse primers designed for PCR contained an additional stop codon for the gene sequence encoding the a-polypeptide and an additional start codon for the gene sequence encoding the 13-polypeptide as well as appropriate restriction sites (Table 7.2). The two PCR products were directly cloned into pGEM-T easy (Promega GmbH, Mannheim, Germany) yielding pGB-a and pGB-13. The gene construct coding for the 13-subunit was isolated by restriction digest with Ndel and Xhol from pGB-13 and cloned into pETDuet-1 (Novagen, Madison, USA) that was previously cut with the same enzymes yielding pEB-13. Subsequently, the a-subunit was cut from pGB-a and cloned into pEB-13 via the Ncol and BamHI restriction sites. The novel plasmid that encoded both polypeptide chains separately was named p3DuetBapA. E. coli NovaBlue and E. coli BL21(DE3) plysS strains (Novagen, Madison, USA) were used for cloning and protein expression, respectively. Sequences of all gene products that were obtained after PCR were confirmed by DNA sequencing at Microsynth (Balgach, Switzerland). Table 7.2: Primers and restriction enzymes used for the cloning of genes encoding the individual u- and ~-polypeptide chains of bapA. Restriction sites are indicated by bold letters. Native and additionally inserted start and stop codons are underlined once and twice, respectively. fwd: forward; rev: reverse. Gene construct Restriction Primers enzyme bapA a-chain Ncol fwd 5'-GGAA TTCCADi,GGGCCGCGCGCTCGCGATCT-3' BamHI rev 5'-GGAA TTGGATCCTAATTCTTGTCCTGCGGCTTGCCT-3' bapA ~-chain Ndel fwd 5'-GGAA TTCCCATATGTCGCTGCTGATCGTGATCGCT-3' Xhol rev 5'-GGAA TTCTCGAG~CCGGCGCGGAAACCGCGCCT-3 ' Generation of point mutants Poi nt mutations were directly generated from the expression vectors p3BapA and p3DuetBapA with the QuikChange multi site-directed mutagenesis kit (Stratagene, La Jolla, 133 Chapter 7 USA) following the manufacturer's protocol. The primers that were used to introduce the point mutations are listed in Table 7.3. Table 7.3: Primers designed for the generation of point mutants of WT BapA and Duet-BapA. The codon changes that introduced the mutations are underlined. Point mutation Primer BapA S250A 5'-CCGCAGGACAAGAA TGCGCTGCTGATCGTG-3' BapA S250C 5'-CCGCAGGACAAGAA TIGCCTGCTGATCGTG-3' BapA E133A 5'-CCGGTGGTCGCCGCAACGCTCGACAAC-3' BapA N249A 5'-CAAGCCGCAGGACAAGGCTICGCTGCTGATCGTG-3' BapA S288A 5'-GCGGGGGCGCTIGCGGGTGAGTICGCG-3' BapA E290A 5'-GCGCTTICGGGTGCGTICGCGCTCGCC-3' Duet-BapA S250A 5'-GGAGATATACATATGGCGCTGCTGATCGTG-3' Duet-BapA S250C 5'-GAAGGAGATATACATATGTGCCTGCTGATCGTGATC-3' Duet-BapA E133A 5'-CCGGTGGTCGCCGCAACGCTCGACAAC-3' Duet-BapA S288A 5'-GCGGGGGCGCTIGCGGGTGAGTICGCG-3' Duet-BapA E290A 5'-GCGCTTICGGGTGCGTICGCGCTCGCC-3' Protein expre.ssion, purification and activity measurements E coli BL21 (DE3) plysS cells harboring p3BapA and p3DuetBapA were cultivated in Luria- Bertani (LB) medium (10 g/ I tryptone, 5 g/I yeast extract, 10 g/ I NaCl, pH 7) containing ampicillin (100 µg/ ml) and chloramphenicol (25 µg/ ml) at 37°C and constant shaking. The cultures were induced at an ODsso of 0.5 by the addition of 0.5 mM isopropyl-13-0- thiogalactopyranoside (IPTG) and grown for an additional 4 h. The cells were harvested by centrifugation and the cell pellets were frozen at -80°C. Cell-free extracts were prepared by ultrasonication and the cell debris was removed by centrifucation. The proteins were purified 2 59 from the crude extracts according to previously described chromatographic procedures[s , l 134 Protein Maturation, Activity and Precursor Crystal Structure of BapA and the presence of the target proteins was verified by SDS-PAGE analysis of the eluted fractions. Size exclusion chromatography was added as an additional purification step as described in the Experimental Section of Chapter 6. β-Aminopeptidase activity in the cell-free extracts and in the purified protein samples was determined with the chromogenic substrate 3 [62] H-β hAla-pNA (5 mM) at pH 8 following a previously published procedure. SDS-PAGE, N-terminal protein sequencing and enzyme activity tests The cell-free extracts and the partially purified proteins were analyzed by SDS-PAGE using pre-cast 10% Novex tricine gels (Invitrogen AG, Basel, Switzerland) according to the manufacturer’s instructions. The protein gels were stained with Coomassie Brillant Blue and the relative intensities of the protein bands were determined with a GS-800 calibrated imaging densitometer and the software Quantity One (Bio-Rad, Rheinach, Switzerland). Partially purified Duet-BapA S250A was transferred from the gel onto a PVDF membrane by electroblotting and the N-terminal primary structure of the β-polypeptide chain was sequenced at the Functional Genomics Center Zürich (Switzerland). 135 General Conclusions and Outlook 8. General Conclusions and Outlook Exploiting the biocatalytic potential of β-aminopeptidases The β-aminopeptidases investigated in the course of this thesis belong to the very few known enzymes that are capable of reacting with peptides comprised of β-amino aicds with proteinogenic side chains. The growing demand for β-amino acids and β-peptides in biomedical research encouraged us to use these unique enzymes for biocatalytic applications aiming (i) for the enzyme-catalyzed synthesis of small peptides with N-terminal β-amino acid residues and (ii) for the production of enantiopure β-amino acids from racemic starting materials. The results presented in Chapters 2 to 5 demonstrate the general usefulness of β- aminopeptidase-catalyzed bioconversions on a laboratory scale. Potential strategies to overcome drawbacks of the existing reaction systems and ideas for future investigations are discussed in the following sections. Enzymatic production of enantiopure β-amino acids Chapter 4 describes the β-aminopeptidase-catalyzed kinetic resolution of four racemic β3- amino acid amides with aliphatic side chains. The reactions were L-enantioselective and led 3 to the formation of the respective L-β -amino acids of high enantiomeric excess. The broad substrate specificities of the β-aminopeptidases for peptides with N-terminal β3-amino acids (Table 1.2)[62] suggest that these enzymes may not be limited to resolve aliphatic β3-amino acid amides, but could as well be useful for kinetic resolutions of β3-amino acid derivatives with functionalized side chains. Moreover, the results presented in Chapter 5 indicate that β- aminopeptidases may also be used for the enantioselective production of β2-amino acids. Kinetic resolution approaches generally suffer from the fact that, even in the optimal scenario of an enantiospecific reaction, they are limited to maximum theoretical product yields of 50%. Furthermore, an additional workup procedure is required after completion of the reaction to separate the formed enantiopure product from the unreacted substrate. These limitations can be overcome by dynamic kinetic resolution (DKR) strategies, which involve the combination of an enantioselective transformation with a simultaneously occurring substrate racemization process.[179,180] Continuous substrate racemization makes both enantiomers of the starting material available to enantioselective conversion. It is 137 Chapter 8 hence possible to achieve quantitative conversion of a racemic substrate to an enantiopure product. Substrate racemization can be catalyzed by means of a chemical catalyst or an enzyme.[179,181] It is generally caused by initial deprotonation of the substrate’s chiral carbon atom followed by reprotonation on the opposite face of the planar intermediate. Convenient chemoenzymatic procedures for the production of various α-amino acids by DKR were described by Chen et al. and Schichl et al., who coupled alcalase-catalyzed L-enantioselective resolutions of racemic α-amino acid esters with racemization of the non-hydrolyzed esters by catalytic amounts of the naturally occurring racemization catalyst pyridoxal-5-phosphate[182] or different salicylaldehydes[183]. Purely enzyme-based DKR processes were established for the biocatalytic production of many natural and non-natural α-amino acids. Successful examples giving access to various enantiopure α-amino acids include the hydantoinase/carbamoylase process in combination with a hydantoin racemase and the acylase-catalyzed enantioselective deacylation of N-acylated α-amino acids in combination with an N-acylamino acid racemase.[184] A recent investigation by Asano et al. showed that α-amino-ε-caprolactam (ACL) racemase catalyzes the pyridoxal-5-phosphate-dependent racemization of α-amino acid amides, but not of α-amino acids.[185,186] In continuation of their studies, the authors used ACL racemase in combination with the D-aminopeptidase DAP from O. anthropi for the DKR of racemic α-amino acid amides to yield D-α-amino acids of high enantiomeric excess.[187,188] The finding of similar enzymatic or chemical racemization catalysts acting on β-amino acid amides or esters would pave the way for the development of β-aminopeptidase-catalyzed DKR processes, which could enable the production of β- amino acids of high enantiomeric excess and in high yield. However, it has to be considered that the acidity of the hydrogen atom at the chiral centers of β2- and especially β3-amino acids is remarkably decreased when compared to the α-hydrogen atom of an α-amino acid. This fact complicates deprotonation of the chiral centers of β-amino acid derivatives and hence makes racemization less likely to occur. β-Aminopeptidase-catalyzed β-peptide formation As outlined in Chapters 2, 3 and 5, β-aminopeptidases catalyze the coupling of a broad variety of β-amino acids to the free N-terminus of α-amino acids, β-amino acids and peptides in aqueous solution. In these experiments, β-aminopeptidase-catalyzed peptide 138 General Conclusions and Outlook formation was under kinetic control and required the use of C-terminally activated substrates. However, unwanted side reactions occurred, such as the hydrolysis of substrate and product and the formation of various peptidic byproducts. In fact, byproduct formation is a major limitation of the presented reaction system for larger scale peptide production, because (i) it drastically reduces product yields, and (ii) it requires the development of an additional workup procedure to separate the product from the undesired byproducts. Approaches to suppress competitive reactions in protease-catalyzed peptide synthesis are well described in literature,[87-89] but they have not been investigated in detail in the course of the present thesis. Some of these approaches to optimize proteolytic enzymes for peptide synthesis and their practicability for β-aminopeptidase-catalyzed β-peptide synthesis are briefly discussed in this section. Changing the composition of the reaction medium, e.g. by the addition of organic co-solvents, is a common strategy to prevent competing hydrolysis reactions during protease-catalyzed peptide synthesis. Unfortunatelly, the addition of organic co-solvents often negatively affects catalyst activity and stability. The stability issue can sometimes be overcome by immobilizing the enzyme on a solid support, which also allows easy process recovery of the catalyst.[189] Besides hydrolysis, β-aminopeptidase-catalyzed cross-couplings among substrates and product resulted in the formation of various byproducts with N- terminal β-amino acids. The formation of peptidic byproducts is a commonly observed problem in protease-catalyzed peptide synthesis under kinetic control. Undesired cross- couplings can generally be minimized by the removal of competing nucleophiles from the reaction mixture. In this respect, N-terminal protection of the acyl donor would be favorable to avoid the oligomerization of the acyl donor component. However, the introduction of N- terminal protecting groups is not feasible in the case of β-aminopeptidase-catalyzed reactions, because the enzymes require a free N-terminus for substrate binding and acyl enzyme formation (see Chapter 6). The fact that the product itself can act as a nucleophile, could be prevented by continuously removing the product from the reaction mixture during the course of the transformation. In the case of the β-aminopeptidase-catalyzed formation of L-carnosine, which is presented in Chapter 3, product removal could for instance be [190,191] achieved by L-carnosine precipitation with zinc or other complexing metals. Although common strategies to improve the synthetic activity of proteolytic enzymes are available, they are not generally applicable and have to be optimized to meet the requirements of each respective reaction. Hence, the presented β-aminopeptidase-catalyzed formations of β- 139 Chapter 8 peptides should be regarded as the initial step towards biocatalytic synthesis of β-peptides, but they are as yet far too inefficient in terms of product yields to serve as an efficient alternative to conventional chemical β-peptide synthesis procedures.[192] Each enzymatic reaction system has to be studied carefully and optimized with respect to the above mentioned points in order to assess its usefulness as a biocatalytic process. Combining structure with biocatalysis The redesign of biocatalytically useful enzymes is a common strategy to create catalysts with novel functions, new substrate specificities or improved selectivities. Modifications can be achieved (i) by directed evolution, (ii) by rational catalyst design based on the three- dimensional structure of the catalyst, and (iii) by structure-inspired semi-random mutagenesis as a combination of the above mentioned.[193-195] Directed evolution is a powerful random approach, which does not require structural knowledge of the catalyst. It involves random mutagenesis of the target gene in combination with gene expression and high-throughput screening methods to search for enzyme mutants with improved enzymatic properties. The availability of high-resolution crystal structures is the requirement for structure-based catalyst redesign. Recent investigations reported the structure-inspired semi-random mutagenesis of penicillin G acylase. Saturation mutagenesis of selected amino acid residues, which were chosen based on the crystal structure of penicillin G acylase, yielded mutant enzymes with highly improved catalytic properties for the synthesis of penicillin antibiotics under kinetic control.[118,196] The determination of high-resolution crystal structures of 3-2W4 BapA (see Chapter 6) and DmpA[64] paves the way for similar strategies to create novel β-aminopeptidase variants with improved catalytic functions or altered substrate specificities by structure-based catalyst redesign. Amino acid exchanges in the BapA helix comprising residues Glu120 to Arg126 or in the corresponding loop of DmpA comprising residues Gln131 to Trp137 (see Figure 6.7) are especially interesting since these highly variable regions (see Figure 6.8) define the width of the substrate binding pockets and thus determine the specificities of the enzymes for peptides with N-terminal β-amino acids of different sizes. 140 General Conclusions and Outlook The quest for the natural substrate Earlier investigations on the biocatalytic breakdown of β-peptides[59,62] and results presented in this thesis show that the β-aminopeptidases 3-2W4 BapA from S. xenopeptidilytica and Y2 BapA from S. microcystinivorans possess the ability to cleave a wide variety of linear synthetic peptides with N-terminal β-amino acids. In contrast, natural peptides with N- terminal proteinogenic α-amino acids do not serve as substrates for the enzymes. The substrate specificities of β-aminopeptidases for non-natural β-peptides are astonishing, since pure β-peptides composed of β-amino acid residues with proteinogenic side chains do not exist in nature. However, β-amino acid-derived substructures occur in a very wide range of natural compounds (Table 1.2).[1,11] The existence of these β-amino acid-containing compounds give rise to speculations about the natural substrates and the physiological roles of the investigated enzymes. Thus, β-aminopeptidases could degrade or participate in the conversion of β-amino acid-containing compounds to make them available for further breakdown by the catabolic machinery of their bacterial hosts. Despite the fact that only four β-aminopeptidases have thus far been isolated and characterized, alignment searches revealed the presence of similar gene sequences in various bacterial species.[63] It is therefore quite likely that the ability to utilize β-peptides as substrates is not restricted to the four described bacterial strains. In fact, it seems that β-peptide-cleaving enzymes are widely distributed among microorganisms (see Figure 6.8), but their physiological relevance remains to be unraveled. The isolation and characterization of further proteins from the pool of β-aminopeptidase-like sequences may lead to the discovery of novel catalysts with interesting properties and selectivities for biotechnological applications. 141 References References [1] D. Seebach, A. K. Beck, D. J. Bierbaum. The world of β- and γ-peptides comprised of homologated proteinogenic amino acids and other components. Chem. Biodiversity 2004, 1, 1111-1239. [2] D. Seebach, A. K. Beck, S. Capone, G. Deniau, U. Grošelj, E. Zass. Enantioselective preparation of β2-amino acid derivatives for β-peptide synthesis. Synthesis 2009, 1, 1-32. [3] K. Bauer. Carnosine and homocarnosine, the forgotten, enigmatic peptides of the brain. Neurochem. Res. 2005, 30, 1339-1345. [4] A. Guiotto, A. Calderan, P. Ruzza, G. Borin. Carnosine and carnosine-related antioxidants: a review. Curr. Med. Chem. 2005, 12, 2293-2315. [5] A. R. Hipkiss. Carnosine, a protective, anti-ageing peptide? Int. J. Biochem. Cell Biol. 1998, 30, 863-868. [6] O. W. Griffith. β-Amino acids − mammalian metabolism and utility as α-amino acid analogs. Ann. Rev. Biochem. 1986, 55, 855-878. [7] A. Pegova, H. Abe, A. Boldyrev. Hydrolysis of carnosine and related compounds by mammalian carnosinases. Comp. Biochem. Physiol., B: Comp. Biochem. 2000, 127, 443-446. [8] J. M. Poston. Leucine 2,3-aminomutase, an enzyme of leucine catabolism. J. Biol. Chem. 1976, 251, 1859-1863. [9] R. C. Bray, T. C. Stadtman. Anaerobic degradation of lysine. III. 15N Studies on conversion of lysine to 3,5-diaminohexanoate. J. Biol. Chem. 1968, 243, 381-385. [10] J. Rétey, F. Kunz, D. Arigoni, T. C. Stadtman. Zur Kenntnis der β-Lysin-Mutase- Reaktion: Mechanismus und sterischer Verlauf. Helv. Chim. Acta 1978, 61, 2989- 2998. 143 References [11] E. Juaristi. Enantioselective synthesis of β-amino acids. John Wiley & Sons, Hoboken, 2005. [12] R. M. Dawson. The toxicology of microcystins. Toxicon 1998, 36, 953-962. [13] R. Luukkainen, K. Sivonen, M. Namikoshi, M. Färdig, K. L. Rinehart, S. I. Niemelä. Isolation and identification of 8 microcystins from thirteen Oscillatoria agardhii strains and structure of a new microcystin. Appl. Environ. Microbiol. 1993, 59, 2204-2209. [14] H. Umezawa, T. Aoyagi, H. Suda, M. Hamada, T. Takeuchi. Bestatin, an inhibitor of aminopeptidase-B, produced by actinomycetes. J. Antibiot. 1976, 29, 97-99. [15] H. Suda, T. Takita, T. Aoyagi, H. Umezawa. Structure of bestatin. J. Antibiot. 1976, 29, 100-101. [16] A. D. Deshpande, K. G. Baheti, N. R. Chatterjee. Degradation of β-lactam antibiotics. Curr. Sci. 2004, 87, 1684-1695. [17] D. Seebach, P. E. Ciceri, M. Overhand, B. Jaun, D. Rigo, L. Oberer, U. Hommel, R. Amstutz, H. Widmer. Probing the helical secondary structure of short-chain β-peptides. Helv. Chim. Acta 1996, 79, 2043-2066. [18] D. Seebach, M. Overhand, F. N. M. Kühnle, B. Martinoni, L. Oberer, U. Hommel, H. Widmer. β-Peptides: synthesis by Arndt-Eistert homologation with concomitant peptide coupling. Structure determination by NMR and CD spectroscopy and by X-ray crystallography. Helical secondary structure of a β-hexapeptide in solution and its stability towards pepsin. Helv. Chim. Acta 1996, 79, 913-941. [19] D. H. Appella, L. A. Christianson, I. L. Karle, D. R. Powell, S. H. Gellman. β-Peptide foldamers: robust helix formation in a new family of β-amino acid oligomers. J. Am. Chem. Soc. 1996, 118, 13071-13072. [20] D. Seebach, S. Abele, K. Gademann, G. Guichard, T. Hintermann, B. Jaun, J. L. Matthews, J. V. Schreiber, L. Oberer, U. Hommel, H. Widmer. β2- and β3-peptides with proteinaceous side chains: synthesis and solution structures of constitutional 144 References isomers, a novel helical secondary structure and the influence of solvation and hydrophobic interactions on folding. Helv. Chim. Acta 1998, 81, 932-982. [21] D. Seebach, S. Abele, K. Gademann, B. Jaun. Pleated sheets and turns of β-peptides with proteinogenic side chains. Angew. Chem., Int. Ed. 1999, 38, 1595-1597. [22] R. P. Cheng, S. H. Gellman, W. F. DeGrado. β-Peptides: from structure to function. Chem. Rev. 2001, 101, 3219-3232. [23] E. Aberhalden, R. Fleischmann. Weitere Beobachtungen über die spezifische Einstellung von Erepsin und von Trypsin-Kinase. Fermentforschung 1929, 10, 195- 212. [24] E. Aberhalden, F. Reich. Vergleichende Studien über den Einfluss von n-Alkali, von Erepsin und von Trypsin-Kinase auf polypeptidartige Verbindungen, an deren Aufbau β-Alanin beteiligt ist. Fermentforschung 1929, 10, 173-178. [25] J. Frackenpohl, P. I. Arvidsson, J. V. Schreiber, D. Seebach. The outstanding biological stability of β- and γ-peptides toward proteolytic enzymes: an in vitro investigation with fifteen peptidases. ChemBioChem 2001, 2, 445-455. [26] D. F. Hook, P. Bindschädler, Y. R. Mahajan, R. Sebesta, P. Kast, D. Seebach. The proteolytic stability of 'designed' β-peptides containing α-peptide-bond mimics and of mixed α,β-peptides: application to the construction of MHC-binding peptides. Chem. Biodiversity 2005, 2, 591-632. [27] D. F. Hook, F. Gessier, C. Noti, P. Kast, D. Seebach. Probing the proteolytic stability of β-peptides containing α-fluoro- and α-hydroxy-β-amino acids. ChemBioChem 2004, 5, 691-706. [28] D. Seebach, S. Abele, J. V. Schreiber, B. Martinoni, A. K. Nussbaum, H. Schild, H. Schulz, H. Hennecke, R. Woessner, F. Bitsch. Biological and pharmacokinetic studies with β-peptides. Chimia 1998, 52, 734-739. 145 References [29] R. Lind, D. Greenhow, S. Perry, T. Kimmerlin, D. Seebach. Comparative metabolism of α- and β-peptides in the insect Heliothis virescens and in plant cells of black Mexican sweet maize. Chem. Biodiversity 2004, 1, 1391-1400. [30] H. Wiegand, B. Wirz, A. Schweitzer, G. P. Camenisch, M. I. Rodriguez Perez, G. Gross, R. Woessner, R. Voges, P. I. Arvidsson, J. Frackenpohl, D. Seebach. The outstanding metabolic stability of a 14C-labeled β-nonapeptide in rats − in vitro and in vivo pharmacokinetic studies. Biopharm. Drug Dispos. 2002, 23, 251-262. [31] H. Wiegand, B. Wirz, A. Schweitzer, G. Gross, M. I. Rodriguez Perez, H. Andres, T. Kimmerlin, M. Rueping, D. Seebach. Pharmacokinetic investigation of a 14C-labelled β3/α tetrapeptide in rats. Chem. Biodiversity 2004, 1, 1812-1828. [32] D. Seebach, J. Gardiner. β-Peptidic peptidomimetics. Acc. Chem. Res. 2008, 41, 1366-1375. [33] M.-I. Aguilar, A. W. Purcell, R. Devi, R. Lew, J. Rossjohn, A. I. Smith, P. Perlmutter. β- Amino acid-containing hybrid peptides − new opportunities in peptidomimetics. Org. Biomol. Chem. 2007, 5, 2884-2890. [34] J. A. Patch, A. E. Barron. Mimicry of bioactive peptides via non-natural, sequence- specific peptidomimetic oligomers. Curr. Opin. Chem. Biol. 2002, 6, 872-877. [35] D. L. Steer, R. A. Lew, P. Perlmutter, A. I. Smith, M.-I. Aguilar. β-Amino acids: versatile peptidomimetics. Curr. Med. Chem. 2002, 9, 811-822. [36] R. Akkarawongsa, T. B. Potocky, E. P. English, S. H. Gellman, C. R. Brandt. Inhibition of Herpes simplex virus type 1 infection by cationic β-peptides. Antimicrob. Agents Chemother. 2008, 52, 2120-2129. [37] E. P. English, R. S. Chumanov, S. H. Gellman, T. Compton. Rational development of β-peptide inhibitors of human cytomegalovirus entry. J. Biol. Chem. 2006, 281, 2661- 2667. [38] J. A. Kritzer, J. D. Lear, M. E. Hodsdon, A. Schepartz. Helical β-peptide inhibitors of the p53-hDM2 interaction. J. Am. Chem. Soc. 2004, 126, 9468-9469. 146 References [39] J. A. Kritzer, O. M. Stephens, D. A. Guarracino, S. K. Reznik, A. Schepartz. β-Peptides as inhibitors of protein-protein interactions. Bioorg. Med. Chem. 2005, 13, 11-16. [40] O. M. Stephens, S. Kim, B. D. Welch, M. E. Hodsdon, M. S. Kay, A. Schepartz. Inhibiting HIV fusion with a β-peptide foldamer. J. Am. Chem. Soc. 2005, 127, 13126-13127. [41] K. Gademann, M. Ernst, D. Hoyer, D. Seebach. Synthesis and biological evaluation of a cyclo-β-tetrapeptide as a somatostatin analogue. Angew. Chem., Int. Ed. 1999, 38, 1223-1226. [42] K. Gademann, T. Kimmerlin, D. Hoyer, D. Seebach. Peptide folding induces high and selective affinity of a linear and small β-peptide to the human somatostatin receptor 4. J. Med. Chem. 2001, 44, 2460-2468. [43] A. I. Webb, M. A. Dunstone, N. A. Williamson, J. D. Price, A. de Kauwe, W. Chen, A. Oakley, P. Perlmutter, J. McCluskey, M.-I. Aguilar, J. Rossjohn, A. W. Purcell. T cell determinants incorporating β-amino acid residues are protease resistant and remain immunogenic in vivo. J. Immunol. 2005, 175, 3810-3818. [44] E. A. Porter, X. Wang, H.-S. Lee, B. Weisblum, S. H. Gellman. Non-haemolytic β- amino-acid oligomers. Nature 2000, 404, 565. [45] E. A. Porter, B. Weisblum, S. H. Gellman. Mimicry of host-defense peptides by unnatural oligomers: antimicrobial β-peptides. J. Am. Chem. Soc. 2002, 124, 7324- 7330. [46] E. A. Porter, B. Weisblum, S. H. Gellman. Use of parallel synthesis to probe structure- activity relationships among 12-helical β-peptides: evidence of a limit on antimicrobial activity. J. Am. Chem. Soc. 2005, 127, 11516-11529. [47] P. I. Arvidsson, J. Frackenpohl, N. S. Ryder, B. Liechty, F. Petersen, H. Zimmermann, G. P. Camenisch, R. Woessner, D. Seebach. On the antimicrobial and hemolytic activities of amphiphilic β-peptides. ChemBioChem 2001, 2, 771-773. 147 References [48] P. I. Arvidsson, N. S. Ryder, H. M. Weiss, G. Gross, O. Kretz, R. Woessner, D. Seebach. Antibiotic and hemolytic activity of a β2/β3 peptide capable of folding into a 12/10-helical secondary structure. ChemBioChem 2003, 4, 1345-1347. [49] A. J. Karlsson, W. C. Pomerantz, B. Weisblum, S. H. Gellman, S. P. Palecek. Antifungal activity from 14-helical β-peptides. J. Am. Chem. Soc. 2006, 128, 12630-12631. [50] M. A. Schmitt, B. Weisblum, S. H. Gellman. Interplay among folding, sequence, and lipophilicity in the antibacterial and hemolytic activities of α/β-peptides. J. Am. Chem. Soc. 2007, 129, 417-428. [51] J. Gardiner, A. V. Thomae, R. I. Mathad, D. Seebach, S. D. Kramer. Comparison of permeation through phosphatidylcholine bilayers of N-dipicolinyl-α- and -β- oligopeptides. Chem. Biodiversity 2006, 3, 1181-1201. [52] B. Geueke, K. Namoto, I. Agarkova, J.-C. Perriard, H.-P. E. Kohler, D. Seebach. Bacterial cell penetration by β3-oligohomoarginines: indications for passive transfer through the lipid bilayer. ChemBioChem 2005, 6, 982-985. [53] M. Rueping, Y. Mahajan, M. Sauer, D. Seebach. Cellular uptake studies with β- peptides. ChemBioChem 2002, 3, 257-259. [54] N. Umezawa, M. A. Gelman, M. C. Haigis, R. T. Raines, S. H. Gellman. Translocation of a β-peptide across cell membranes. J. Am. Chem. Soc. 2002, 124, 368-369. [55] S. Guha, T. Chakraborty, A. Banerjee. Water soluble synthetic dieptide-based biodegradable nanoporous materials. Green Chem. 2009, 11, 1139-1145. [56] J. V. Schreiber, J. Frackenpohl, F. Moser, T. Fleischmann, H.-P. E. Kohler, D. Seebach. On the biodegradation of β-peptides. ChemBioChem 2002, 3, 424-432. [57] B. Geueke, H.-J. Busse, T. Fleischmann, P. Kämpfer, H.-P. E. Kohler. Description of Sphingosinicella xenopeptidilytica sp. nov., a β-peptide-degrading species, and emended descriptions of the genus Sphingosinicella and the species Sphingosinicella microcystinivorans. Int. J. Syst. Evol. Microbiol. 2007, 57, 107-113. 148 References [58] T. Maruyama, H. D. Park, K. Ozawa, Y. Tanaka, T. Sumino, K. Hamana, A. Hiraishi, K. Kato. Sphingosinicella microcystinivorans gen. nov., sp. nov., a microcystin-degrading bacterium. Int. J. Syst. Evol. Microbiol. 2006, 56, 85-89. [59] B. Geueke, T. Heck, M. Limbach, V. Nesatyy, D. Seebach, H.-P. E. Kohler. Bacterial β- peptidyl aminopeptidases with unique substrate specificities for β-oligopeptides and mixed β,α-oligopeptides. FEBS J. 2006, 273, 5261-5272. [60] L. Fanuel, C. Goffin, A. Cheggour, B. Devreese, G. van Driessche, B. Joris, J. van Beeumen, J.-M. Frère. The DmpA aminopeptidase from Ochrobactrum anthropi LMG7991 is the prototype of a new terminal nucleophile hydrolase family. Biochem. J. 1999, 341, 147-155. [61] H. Komeda, Y. Asano. A DmpA-homologous protein from Pseudomonas sp. is a dipeptidase specific for β-alanyl dipeptides. FEBS J. 2005, 272, 3075-3084. [62] T. Heck, M. Limbach, B. Geueke, M. Zacharias, J. Gardiner, H.-P. E. Kohler, D. Seebach. Enzymatic degradation of β- and mixed α,β-oligopeptides. Chem. Biodiversity 2006, 3, 1325-1348. [63] B. Geueke, H.-P. E. Kohler. Bacterial β-peptidyl aminopeptidases: on the hydrolytic degradation of β-peptides. Appl. Microbiol. Biotechnol. 2007, 74, 1197-1204. [64] C. Bompard-Gilles, V. Villeret, G. J. Davies, L. Fanuel, B. Joris, J.-M. Frère, J. van Beeumen. A new variant of the Ntn hydrolase fold revealed by the crystal structure of L-aminopeptidase D-Ala-esterase/amidase from Ochrobactrum anthropi. Structure 2000, 8, 153-162. [65] C. Bompard-Gilles, V. Villeret, L. Fanuel, B. Joris, J.-M. Frère, J. van Beeumen. Crystallization and preliminary X-ray analysis of a new L-aminopeptidase-D- amidase/D-esterase activated by a Gly-Ser peptide bond hydrolysis. Acta Crystallogr., Sect. D: Biol. Crystallogr. 1999, 55, 699-701. [66] P. J. Artymiuk. A sting in the (N-terminal) tail. Nat. Struct. Biol. 1995, 2, 1035-1037. 149 References [67] J. A. Brannigan, G. Dodson, H. J. Duggleby, P. C. E. Moody, J. L. Smith, D. R. Tomchick, A. G. Murzin. A protein catalytic framework with an N-terminal nucleophile is capable of self-activation. Nature 1995, 378, 416-419. [68] H. J. Duggleby, S. P. Tolley, C. P. Hill, E. J. Dodson, G. Dodson, P. C. E. Moody. Penicillin acylase has a single-amino-acid catalytic center. Nature 1995, 373, 264-268. [69] Ö. D. Ekici, M. Paetzel, R. E. Dalbey. Unconventional serine proteases: variations on the catalytic Ser/His/Asp triad configuration. Protein Sci. 2008, 17, 2023-2037. [70] C. Oinonen, J. Rouvinen. Structural comparison of Ntn-hydrolases. Protein Sci. 2000, 9, 2329-2337. [71] H. Cheng, N. V. Grishin. DOM-fold: a structure with crossing loops found in DmpA, ornithine acetyltransferase, and molybdenum cofactor-binding domain. Protein Sci. 2005, 14, 1902-1910. [72] G. Lelais, D. Seebach. Synthesis, CD spectra, and enzymatic stability of β2- oligoazapeptides prepared from (S)-2-hydrazino carboxylic acids carrying the side chains of Val, Ala, and Leu. Helv. Chim. Acta 2003, 86, 4152-4168. [73] H. M. Weiss, B. Wirz, A. Schweitzer, R. Amstutz, M. I. Rodriguez Perez, H. Andres, Y. Metz, J. Gardiner, D. Seebach. ADME investigations of unnatural peptides: distribution of a C14-labeled β3-octaarginine in rats. Chem. Biodiversity 2007, 4, 1413-1437. [74] B. Geueke, K. Namoto, D. Seebach, H.-P. E. Kohler. A novel β-peptidyl aminopeptidase (BapA) from strain 3-2W4 cleaves peptide bonds of synthetic β-tri- and β-dipeptides. J. Bacteriol. 2005, 187, 5910-5917. [75] L. Fanuel, I. Thamm, V. Kostanjevecki, B. Samyn, B. Joris, C. Goffin, J. Brannigan, J. van Beeumen, J.-M. Frère. Two new aminopeptidases from Ochrobactrum anthropi active on D-alanyl-p-nitroanilide. Cell. Mol. Life Sci. 1999, 55, 812-818. [76] H.-D. Park, Y. Sasaki, T. Maruyama, E. Yanagisawa, A. Hiraishi, K. Kato. Degradation of the cyanobacterial hepatotoxin microcystin by a new bacterium isolated from a hypertrophic lake. Environ. Toxicol. 2001, 16, 337-343. 150 References [77] B. M. Gulledge, J. B. Aggen, A. R. Chamberlin. Linearized and truncated microcystin analogues as inhibitors of protein phosphatases 1 and 2A. Bioorg. Med. Chem. Lett. 2003, 13, 2903-2906. [78] J. M. Humphrey, J. B. Aggen, A. R. Chamberlin. Total synthesis of the serine- threonine phosphatase inhibitor microcystin-LA. J. Am. Chem. Soc. 1996, 118, 11759-11770. [79] K. F. Vongerichten, J. R. Klein, H. Matern, R. Plapp. Cloning acid nucleotide-sequence analysis of pepV, a carnosinase gene from Lactobacillus delbrueckii subsp. lactis DSM- 7290, and partial characterization of the enzyme. Microbiology (Reading U. K.) 1994, 140, 2591-2600. [80] M. Teufel, V. Saudek, J.-P. Ledig, A. Bernhardt, S. Boularand, A. Carreau, N. J. Cairns, C. Carter, D. J. Cowley, D. Duverger, A. J. Ganzhorn, C. Guenet, B. Heintzelmann, V. Laucher, C. Sauvage, T. Smirnova. Sequence identification and characterization of human carnosinase and a closely related non-specific dipeptidase. J. Biol. Chem. 2003, 278, 6521-6531. [81] K. Tabata, M. Kajiyama, T. Hiraishi, H. Abe, I. Yamato, Y. Doi. Purification and characterization of poly(aspartic acid) hydrolase from Sphingomonas sp. KT-1. Biomacromolecules 2001, 2, 1155-1160. [82] K. Tabata, K.-I. Kasuya, H. Abe, K. Masuda, Y. Doi. Poly(aspartic acid) degradation by a Sphingomonas sp. isolated from freshwater. Appl. Environ. Microbiol. 1999, 65, 4268-4270. [83] T. Hiraishi, M. Kajiyama, K. Tabata, H. Abe, I. Yamato, Y. Doi. Biochemical and molecular characterization of poly(aspartic acid) hydrolase-2 from Sphingomonas sp. KT-1. Biomacromolecules 2003, 4, 1285-1292. [84] T. Hiraishi, M. Kajiyama, K. Yamato, Y. Doi. Enzymatic hydrolysis of α- and β-oligo(L- aspartic acid)s by poly(aspartic acid) hydrolases-1 and 2 from Sphingromonas sp. KT- 1. Macromol. Biosci. 2004, 4, 330-339. 151 References [85] M. Bergmann, H. Fraenkel-Conrat. The enzymatic synthesis of peptide bonds. J. Biol. Chem. 1938, 124, 1-6. [86] M. Bergmann, J. S. Fruton. Some synthetic and hydrolytic experiments with chymotrypsin. J. Biol. Chem. 1938, 124, 321-329. [87] F. Bordusa. Proteases in organic synthesis. Chem. Rev. 2002, 102, 4817-4867. [88] D. Kumar, T. C. Bhalla. Microbial proteases in peptide synthesis: approaches and applications. Appl. Microbiol. Biotechnol. 2005, 68, 726-736. [89] C. Lombard, J. Saulnier, J. M. Wallach. Recent trends in protease-catalyzed peptide synthesis. Protein Pept. Lett. 2005, 12, 621-629. [90] K. Yokozeki, S. Hara. A novel and efficient enzymatic method for the production of peptides from unprotected starting materials. J. Biotechnol. 2005, 115, 211-220. [91] X.-G. Li, L. T. Kanerva. Lipases in β-dipeptide synthesis in organic solvents. Org. Lett. 2006, 8, 5593-5596. [92] T. Teorell, E. Stenhagen. Ein Universalpuffer für den pH-Bereich von 2,0 bis 12,0. Biochem. Z. 1938, 299, 416-419. [93] M. M. Bradford. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 1976, 72, 248-254. [94] G. D. Kalyankar, A. Meister. Enzymatic synthesis of carnosine and related β-alanyl and γ-aminobutyryl peptides. J. Biol. Chem. 1959, 234, 3210-3218. [95] R. E. Winnick, T. Winnick. Carnosine-anserine synthetase of muscle. III. Experiments on the mechanism of peptide bond formation. Biochim. Biophys. Acta 1960, 37, 214- 220. [96] H. Horinishi, M. Grillo, F. L. Margolis. Purification and characterization of carnosine synthetase from mouse olfactory bulbs. J. Neurochem. 1978, 31, 909-919. 152 References [97] S. D. Skaper, S. Das, F. D. Marshall. Some properties of a homocarnosine-carnosine synthetase isolated from rat-brain. J. Neurochem. 1973, 21, 1429-1445. [98] H. Unno, T. Yamashita, S. Ujita, N. Okumura, H. Otani, A. Okumura, K. Nagai, M. Kusunoki. Structural basis for substrate recognition and hydrolysis by mouse carnosinase CN2. J. Biol. Chem. 2008, 283, 27289-27299. [99] M. A. Babizhayev, M. C. Seguin, J. Gueyne, R. P. Evstigneeva, E. A. Ageyeva, G. A. Zheltukhina. L-Carnosine (β-alanyl-L-histidine) and carcinine (β-alanylhistamine) act as natural antioxidants with hydroxyl-radical-scavenging and lipid-peroxidase activities. Biochem. J. 1994, 304, 509-516. [100] R. Kohen, Y. Yamamoto, K. C. Cundy, B. N. Ames. Antioxidant activity of carnosine, homocarnosine, and anserine present in muscle and brain. Proc. Natl. Acad. Sci. U. S. A. 1988, 85, 3175-3179. [101] B. Janssen, D. Hohenadel, P. Brinkkoetter, V. Peters, N. Rind, C. Fischer, I. Rychlik, M. Cerna, M. Romzova, E. de Heer, H. Baelde, S. J. L. Bakker, M. Zirie, E. Rondeau, P. Mathieson, M. A. Saleem, J. Meyer, H. Köppel, S. Sauerhoefer, C. R. Bartram, P. Nawroth, H.-P. Hammes, B. A. Yard, J. Zschocke, F. J. van der Woude. Carnosine as a protective factor in diabetic nephropathy: association with a leucine repeat of the carnosinase gene CNDP1. Diabetes 2005, 54, 2320-2327. [102] I. Rashid, D. M. van Reyk, M. J. Davies. Carnosine and its constituents inhibit glycation of low-density lipoproteins that promotes foam cell formation in vitro. FEBS Lett. 2007, 581, 1067-1070. [103] R. Tabakman, P. Lazarovici, R. Kohen. Neuroprotective effects of carnosine and homocarnosine on pheochromocytoma PC12 cells exposed to ischemia. J. Neurosci. Res. 2002, 68, 463-469. [104] M. Hirano, S. Katayama (Zeria Pharmaceutical Company, Hamari Chemicals Company). Preventives/remedies for stomatitis. US Patent 6169083 B1, 2001. 153 References [105] T. Matsukura, T. Takahashi, Y. Nishimura, M. Takaya, H. Fujimura (Hamari Chemicals Company). Crysatlline L-carnosine zinc complex having antiulcer activity. US Patent 4981846, 1991. [106] S. Hildbrand, T. Ruppen, D. Veghini (Lonza AG). Method for producing β- alanineamides. US Patent 6878829 B2, 2005. [107] F. J. Vinick (Pfizer Inc.). Process for preparing L-carnosine. US Patent 4359416, 1982. [108] K. Faber. Biotransformations in organic chemistry: a textbook. Springer, Berlin, 2004, 370-377. [109] M. Yagasaki, S. Hashimoto. Synthesis and application of dipeptides; current status and perspectives. Appl. Microbiol. Biotechnol. 2008, 81, 13-22. [110] T. Heck, D. Seebach, S. Osswald, M. K. J. ter Wiel, H.-P. E. Kohler, B. Geueke. Kinetic resolution of aliphatic β-amino acid amides by β-aminopeptidases. ChemBioChem 2009, 10, 1558-1561. [111] T. Heck, H.-P. E. Kohler, M. Limbach, O. Flögel, D. Seebach, B. Geueke. Enzyme- catalyzed formation of β-peptides: β-peptidyl aminopeptidases BapA and DmpA acting as β-peptide-synthesizing enzymes. Chem. Biodiversity 2007, 4, 2016-2030. [112] P. D'Arrigo, L. T. Kanerva, X.-G. Li, C. Saraceno, S. Servi, D. Tessaro. Enzymatic synthesis of carnosine derivatives catalysed by Burkholderia cepacia lipase. Tetrahedron: Asymmetry 2009, 20, 1641-1645. [113] X.-G. Li, M. Lähitie, L. T. Kanerva. Burkholderia cepacia lipase and activated β-lactams in β-dipeptide and β-amino amide synthesis. Tetrahedron: Asymmetry 2008, 19, 1857-1861. [114] V. Kasche. Mechanism and yields in enzyme catalysed equilibrium and kinetically controlled synthesis of β-lactam antibiotics, peptides and other condensation products. Enzyme Microb. Technol. 1986, 8, 4-16. 154 References [115] J. Heyland, N. Antweiler, J. Lutz, T. Heck, B. Geueke, H.-P. E. Kohler, L. M. Blank, A. Schmid. Simple enzymatic procedure for L-carnosine synthesis: whole-cell biocatalysis and efficient biocatalyst recycling. Microb. Biotechnol. 2010, 3, 74-83. [116] M. Y. Gololobov, I. L. Borisov, V. M. Belikov, V. K. Švedas. Acyl group transfer by proteases forming acyl enzyme intermediate: kinetic-model analysis. Biotechnol. Bioeng. 1988, 32, 866-872. [117] M. Y. Gololobov, I. L. Borisov, V. K. Švedas. Acyl group transfer by proteases forming an acylenzyme intermediate: kinetic-model analysis (including hydrolysis of acylenzyme-nucleophile complex). J. Theor. Biol. 1989, 140, 193-204. [118] S. A. W. Jager, I. V. Shapovalova, P. A. Jekel, W. B. L. Alkema, V. K. Švedas, D. B. Janssen. Saturation mutagenesis reveals the importance of residues αR145 and αF146 of penicillin acylase in the synthesis of β-lactam antibiotics. J. Biotechnol. 2008, 133, 18-26. [119] M. I. Youshko, G. G. Chilov, T. A. Shcherbakova, V. K. Švedas. Quantitative characterization of the nucleophile reactivity in penicillin acylase-catalyzed acyl transfer reactions. Biochim. Biophys. Acta, Proteins Proteomics 2002, 1599, 134-140. [120] G. Cardillo, L. Gentilucci, A. Tolomelli. Unusual amino acids: synthesis and introduction into naturally occurring peptides and biologically active analogues. Mini- Rev. Med. Chem. 2006, 6, 293-304. [121] A. Berkessel, F. Cleemann, S. Mukherjee. Kinetic resolution of oxazinones: an organocatalytic approach to enantiomerically pure β-amino acids. Angew. Chem., Int. Ed. 2005, 44, 7466-7469. [122] J.-A. Ma. Recent developments in the catalytic asymmetric synthesis of α- and β- amino acids. Angew. Chem., Int. Ed. 2003, 42, 4290-4299. [123] N. Sewald. Synthetic routes towards enantiomerically pure β-amino acids. Angew. Chem., Int. Ed. 2003, 42, 5794-5795. 155 References [124] A. S. Bommarius, B. R. Riebel. Biocatalysis: fundamentals and applications. Wiley-VCH, Weinheim, 2004. [125] B. Geueke, W. Hummel. A new bacterial L-amino acid oxidase with a broad substrate specificity: purification and characterization. Enzyme Microb. Technol. 2002, 31, 77- 87. [126] A. Schmid, F. Hollmann, J. B. Park, B. Buhler. The use of enzymes in the chemical industry in Europe. Curr. Opin. Biotechnol. 2002, 13, 359-366. [127] M. Breuer, K. Ditrich, T. Habicher, B. Hauer, M. Keßeler, R. Stürmer, T. Zelinski. Industrial methods for the production of optically active intermediates. Angew. Chem., Int. Ed. 2004, 43, 788-824. [128] W. Leuchtenberger, K. Huthmacher, K. Drauz. Biotechnological production of amino acids and derivatives: current status and prospects. Appl. Microbiol. Biotechnol. 2005, 69, 1-8. [129] H. Gröger, H. Trauthwein, S. Buchholz, K. Drauz, C. Sacherer, S. Godfrin, H. Werner. The first aminoacylase-catalyzed enantioselective synthesis of aromatic β-amino acids. Org. Biomol. Chem. 2004, 2, 1977-1978. [130] D. C. Li, S. Cheng, D. Wei, Y. Ren, D. Zhang. Production of enantiomerically pure (S)- β-phenylalanine and (R)-β-phenylalanine by penicillin G acylase from Escherichia coli in aqueous medium. Biotechnol. Lett. 2007, 29, 1825-1830. [131] V. A. Soloshonok, N. A. Fokina, A. V. Rybakova, I. P. Shishkina, S. V. Galushko, A. E. Sorochinsky, V. P. Kukhar, M. V. Savchenko, V. K. Švedas. Biocatalytic approach to enantiomerically pure β-amino acids. Tetrahedron: Asymmetry 1995, 6, 1601-1610. [132] S. Gedey, A. Liljeblad, L. Lázár, F. Fülöp, L. T. Kanerva. Preparation of highly enantiopure β-amino esters by Candida antarctica lipase A. Tetrahedron: Asymmetry 2001, 12, 105-110. 156 References [133] S. Gedey, A. Liljeblad, L. Lázár, F. Fülöp, L. T. Kanerva. Structural effects on chemo- and enantioselectivity of Candida antarctica lipase B − resolution of β-amino esters. Can. J. Chem. 2002, 80, 565-570. [134] A. Liljeblad, L. T. Kanerva. Biocatalysis as a profound tool in the preparation of highly enantiopure β-amino acids. Tetrahedron 2006, 62, 5831-5854. [135] A. Péter, A. Árki, E. Vékes, D. Tourwé, L. Lázár, F. Fülöp, D. W. Armstrong. Direct and indirect high-performance liquid chromatographic enantioseparation of β-amino acids. J. Chromatogr. A 2004, 1031, 171-178. [136] A. Péter, L. Lázár, F. Fülöp, D. W. Armstrong. High-performance liquid chromatographic enantioseparation of β-amino acids. J. Chromatogr. A 2001, 926, 229-238. [137] A. J. J. Straathof, J. A. Jongejan. The enantiomeric ratio: origin, determination and prediction. Enzyme Microb. Technol. 1997, 21, 559-571. [138] N. I. Martin, J. J. Woodward, M. B. Winter, W. T. Beeson, M. A. Marletta. Design and synthesis of C5 methylated L-arginine analogues as active site probes for nitric oxide synthase. J. Am. Chem. Soc. 2007, 129, 12563-12570. [139] S. G. Alcock, J. E. Baldwin, R. Bohlmann, L. M. Harwood, J. I. Seeman. On the conjugative isomerizations of β,γ-unsaturated esters. Stereochemical generalizations and predictions for 1,3-prototropic shifts under basic conditions. J. Org. Chem. 1985, 50, 3526-3535. [140] M. K. J. ter Wiel, M. Arnold, S. Peter, I. Troltsch, S. Merget, F. Glaser, M. Schwarm, H. S. Bhatti, B. Kuriakose, S. S. Pol, M. Balamurugan, V. V. Joshi. Enantiopure β3- neopentylglycine: synthesis and resolution. Tetrahedron: Asymmetry 2009, 20, 478- 482. [141] D. Seebach, B. Jaun, R. Sebesta, R. I. Mathad, O. Flögel, M. Limbach, H. Sellner, S. Cottens. Synthesis, and helix- or hair-pin-turn secondary structures of "mixed" α/β- 157 References peptides consisting of residues with proteinogenic side chains and of 2-amino-2- methyl-propanoic acid (Aib). Helv. Chim. Acta 2006, 89, 1801-1825. [142] S. Krebs, J. R. Lamas, S. Poenaru, G. Folkers, J. A. López de Castro, D. Seebach, D. Rognan. Substituting nonpeptidic spacers for the T cell receptor-binding part of class I major histocompatibility complex-binding peptides. J. Biol. Chem. 1998, 273, 19072-19079. [143] D. Seebach, E. Dubost, R. I. Mathad, B. Jaun, M. Limbach, M. Löweneck, O. Flögel, J. Gardiner, S. Capone, A. K. Beck, H. Widmer, D. Langenegger, D. Monna, D. Hoyer. New open-chain and cyclic tetrapeptides, consisting of α-, β2-, and β3-amino acid residues, as somatostatin mimics − a survey. Helv. Chim. Acta 2008, 91, 1736-1786. [144] S. Dedier, S. Krebs, J. R. Lamas, S. Poenaru, G. Folkers, J. A. López de Castro, D. Seebach, D. Rognan. Structure-based design of nonnatural ligands for the HLA-B27 protein. J. Recept. Signal Transduction 1999, 19, 645-657. [145] S. Poenaru, J. R. Lamas, G. Folkers, J. A. López de Castro, D. Seebach, D. Rognan. Nonapeptide analogues containing (R)-3-hydroxybutanoate and β-homoalanine oligomers: synthesis and binding affinity to a class I major histocompatibility complex protein. J. Med. Chem. 1999, 42, 2318-2331. [146] G. Lelais, D. Seebach. β2-amino acids − syntheses, occurrence in natural products, and components of β-peptides. Biopolymers 2004, 76, 206-243. [147] T. Heck, V. S. Makam, J. Lutz, L. M. Blank, A. Schmid, D. Seebach, H.-P. E. Kohler, B. Geueke. Kinetic analysis of L-carnosine formation by β-aminopeptidases. Adv. Synth. Catal. 2010, 352, 407-415. [148] G. M. Salamonczyk, K. Han, Z.-w. Guo, C. J. Sih. Total synthesis of cryptophycins via a chemoenzymatic approach. J. Org. Chem. 1996, 61, 6893-6900. [149] M. Solymar, A. Liljeblad, L. Lázár, F. Fülöp, L. T. Kanerva. Lipase-catalysed kinetic resolution in organic solvents: an approach to enantiopure α-methyl-β-alanine esters. Tetrahedron: Asymmetry 2002, 13, 1923-1928. 158 References [150] B. Kaptein, Q. B. Broxterman, J. G. T. Kierkels, T. Milcent (DSM IP Assets B.V.). Enzymatic preparation of an enantiomerically enriched β2-amino acids. WO 2005/085462 A1, 2005. [151] D.-Y. Ma, D.-X. Wang, Q.-Y. Zheng, M.-X. Wang. Nitrile biotransformations for the practical synthesis of highly enantiopure azido carboxylic acids and amides, 'click' to functionalized chiral triazoles and chiral β-amino acids. Tetrahedron: Asymmetry 2006, 17, 2366-2376. [152] M. Fitz, E. Forró, E. Vigóczki, L. Lázár, F. Fülöp. Lipase-catalysed N-acylation of β2- amino esters. Tetrahedron: Asymmetry 2008, 19, 1114-1119. [153] A. I. Martínez-Gómez, S. Martínez-Rodríguez, J. Pozo-Dengra, D. Tessaro, S. Servi, J. M. Clemente-Jiménez, F. Rodríguez-Vico, F. J. L. Heras-Vazquez. Potential application of N-carbamoyl-β-alanine amidohydrolase from Agrobacterium tumefaciens C58 for β- amino acid production. Appl. Environ. Microbiol. 2009, 75, 514-520. [154] Z. Pataj, I. Ilisz, R. Berkecz, A. Misicka, D. Tymecka, F. Fülöp, D. W. Armstrong, A. Pétér. Comparison of performance of Chirobiotic T, T2 and TAG columns in the separation of β2- and β3-homoamino acids. J. Sep. Sci. 2008, 31, 3688-3697. [155] J. Wang, M. Uttamchandani, J. Q. Li, M. Y. Hu, S. Q. Yao. Rapid assembly of matrix metalloprotease inhibitors using click chemistry. Org. Lett. 2006, 8, 3821-3824. [156] M. Mergler, J. P. Durieux. The Bachem practice of SPPS. 2005. [157] E. Kaiser, R. l. Colescott, C. D. Bossinger, P. I. Cook. Color test for detection of free terminal amino groups in solid-phase synthesis of peptides. Anal. Biochem. 1970, 34, 595-598. [158] V. K. Sarin, S. B. H. Kent, J. P. Tam, R. B. Merrifield. Quantitative monitoring of solid- phase peptide synthesis by the ninhydrin reaction. Anal. Biochem. 1981, 117, 147- 157. 159 References [159] J. C. D. Müller-Hartwieg, L. La Vecchia, H. Meyer, A. K. Beck, D. Seebach. (R)-2- (Benzyloxycarbonylamino-methyl)-3-phenylpropanoic acid (Z-β2hPhe-OH). Org. Synth. 2008, 85, 295-306. [160] E. Arvanitis, H. Ernst, A. A. Ludwig, A. J. Robinson, P. B. Wyatt. Enantioselective synthesis of 2-substituted 3-aminopropanoic acid (β-alanine) derivatives which are β- analogues of aromatic amino acids. J. Chem. Soc., Perkin Trans. 1 1998, 3, 521-528. [161] T. Heck, A. Reimer, D. Seebach, J. Gardiner, G. Deniau, A. Lukaszuk, H.-P. E. Kohler, B. Geueke. β-Aminopeptidase-catalyzed biotransformations of β2-dipeptides: kinetic resolution and enzymatic coupling. ChemBioChem 2010, 11, 1129-1136. [162] J. C. Powers, J. L. Asgian, Ö. D. Ekici, K. E. James. Irreversible inhibitors of serine, cysteine, and threonine proteases. Chem. Rev. 2002, 102, 4639-4750. [163] A. Matagne, A. Dubus, M. Galleni, J.-M. Frère. The β-lactamase cycle: a tale of selective pressure and bacterial ingenuity. Nat. Prod. Rep. 1999, 16, 1-19. [164] A. K. Chandel, L. V. Rao, M. L. Narasu, O. V. Singh. The realm of penicillin G acylase in β-lactam antibiotics. Enzyme Microb. Technol. 2008, 42, 199-207. [165] H. B. Bürgi, J. D. Dunitz, J. M. Lehn, G. Wipff. Stereochemistry of reaction paths at carbonyl centers. Tetrahedron 1974, 30, 1563-1572. [166] M. A. Larkin, G. Blackshields, N. P. Brown, R. Chenna, P. A. McGettigan, H. McWilliam, F. Valentin, I. M. Wallace, A. Wilm, R. Lopez, J. D. Thompson, T. J. Gibson, D. G. Higgins. ClustalW and ClustalX version 2. Bioinformatics 2007, 23, 2947-2948. [167] W. Kabsch. Automatic processing of rotation diffraction data from crystals of initially unknown symmetry and cell constants. J. Appl. Crystallogr. 1993, 26, 795-800. [168] A. J. McCoy, R. W. Grosse-Kunstleve, P. D. Adams, M. D. Winn, L. C. Storoni, R. J. Read. Phaser crystallographic software. J. Appl.Crystallogr. 2007, 40, 658-674. [169] P. D. Adams, R. W. Grosse-Kunstleve, L. W. Hung, T. R. Ioerger, A. J. McCoy, N. W. Moriarty, R. J. Read, J. C. Sacchettini, N. K. Sauter, T. C. Terwilliger. PHENIX: building 160 References new software for automated crystallographic structure determination. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2002, 58, 1948-1954. [170] P. Emsley, K. Cowtan. Coot: model-building tools for molecular graphics. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2004, 60, 2126-2132. [171] E. Krissinel, K. Henrick. Secondary-structure matching (SSM), a new tool for fast protein structure alignment in three dimensions. Acta Crystallogr., Sect. D: Biol. Crystallogr. 2004, 60, 2256-2268. [172] W. L. De Lano, DeLano Scientific, Palo Alto, CA, USA, 2002. [173] K. Michalska, A. Hernandez-Santoyo, M. Jaskolski. The mechanism of autocatalytic activation of plant-type L-asparaginases. J. Biol. Chem. 2008, 283, 13388-13397. [174] G. Boanca, A. Sand, J. J. Barycki. Uncoupling the enzymatic and autoprocessing activities of Helicobacter pylori γ-glutamyltranspeptidase. J. Biol. Chem. 2006, 281, 19029-19037. [175] P.-H. Hirel, J.-M. Schmitter, P. Dessen, G. Fayat, S. Blanquet. Extent of N-terminal methionine excision from Escherichia coli proteins is governed by the side-chain length of the penultimate amino acid. Proc. Natl. Acad. Sci. U. S. A. 1989, 86, 8247- 8251. [176] A. L. Goldberg. Protein degradation and protection against misfolded or damaged proteins. Nature 2003, 426, 895-899. [177] M. Groll, M. Bochtler, H. Brandstetter, T. Clausen, R. Huber. Molecular machines for protein degradation. ChemBioChem 2005, 6, 222-256. [178] F. Baneyx, M. Mujacic. Recombinant protein folding and misfolding in Escherichia coli. Nat. Biotechnol. 2004, 22, 1399-1408. [179] H. Pellissier. Recent developments in dynamic kinetic resolution. Tetrahedron 2008, 64, 1563-1601. 161 References [180] N. J. Turner. Enzyme catalysed deracemisation and dynamic kinetic resolution reactions. Curr. Opin. Chem. Biol. 2004, 8, 114-119. [181] M. E. Tanner. Understanding nature's strategies for enzyme-catalyzed racemization and epimerization. Acc. Chem. Res. 2002, 35, 237-246. [182] S.-T. Chen, W.-H. Huang, K.-T. Wang. Resolution of amino acids in a mixture of 2- methyl-2-propanol/water (19:1) catalyzed by alcalase via in situ racemization of one antipode mediated by pyridoxal 5-phosphate. J. Org. Chem. 1994, 59, 7580-7581. [183] D. A. Schichl, S. Enthaler, W. Holla, T. Riermeier, U. Kragl, M. Beller. Dynamic kinetic resolution of α-amino acid esters in the presence of aldehydes. Eur. J. Org. Chem. 2008, 3506-3512. [184] O. May, S. Verseck, A. Bommarius, K. Drauz. Development of dynamic kinetic resolution processes for biocatalytic production of natural and nonnatural L-amino acids. Org. Proc. Res. Dev. 2002, 6, 452-457. [185] Y. Asano, S. Yamaguchi. Discovery of amino acid amides as new substrates for α- amino-ε-caprolactam racemase from Achromobacter obae. J. Mol. Catal. B: Enzym. 2005, 36, 22-29. [186] S. Okazaki, A. Suzuki, T. Mizushima, T. Kawano, H. Komeda, Y. Asano, T. Yamane. The novel structure of a pyridoxal 5 '-phosphate-dependent fold-type I racemase, α- amino-ε-caprolactam racemase from Achromobacter obae. Biochemistry 2009, 48, 941-950. [187] Y. Asano, S. Yamaguchi. Dynamic kinetic resolution of amino acid amide catalyzed by D-aminopeptidase and α-amino-ε-caprolactam racemase. J. Am. Chem. Soc. 2005, 127, 7696-7697. [188] S. Yamaguchi, H. Komeda, Y. Asano. New enzymatic method of chiral amino acid synthesis by dynamic kinetic resolution of amino acid amides: use of stereoselective amino acid amidases in the presence of α-amino-ε-caprolactam racemase. Appl. Environ. Microbiol. 2007, 73, 5370-5373. 162 References [189] R. A. Sheldon. Enzyme immobilization: the quest for optimum performance. Adv. Synth. Catal. 2007, 349, 1289-1307. [190] E. J. Baran. Metal complexes of carnosine. Biochemistry (Moscow) 2000, 65, 789-797. [191] M. Förster, H. Vahrenkamp. Zinc complexes of histidine-containing di- and tripeptides. Chem. Ber. 1995, 128, 541-550. [192] T. Kimmerlin, D. Seebach. '100 years of peptide synthesis': ligation methods for peptide and protein synthesis with applications to β-peptide assemblies. J. Pept. Res. 2005, 65, 229-260. [193] M. D. Toscano, K. J. Woycechowsky, D. Hilvert. Minimalist active-site redesign: teaching old enzymes new tricks. Angew. Chem., Int. Ed. 2007, 46, 3212-3236. [194] C. Jäckel, P. Kast, D. Hilvert. Protein design by directed evolution. Ann. Rev. Biophys. 2008, 37, 153-173. [195] K. Hult, P. Berglund. Engineered enzymes for improved organic synthesis. Curr. Opin. Biotechnol. 2003, 14, 395-400. [196] E. M. Gabor, D. B. Janssen. Increasing the synthetic performance of penicillin acylase PAS2 by structure-inspired semi-random mutagenesis. Protein Eng., Des. Sel. 2004, 17, 571-579. 163 Curriculum Vitae Curriculum Vitae Tobias Heck Born on April 5, 1980 in Stuttgart, Germany Citizen of Germany Sep. 2006 – Feb. 2010 Doctoral thesis at the Swiss Federal Institute of Aquatic Science and Technology (Eawag, Department of Environmental Microbiology, Dübendorf, Switzerland) and at the Swiss Federal Institute of Science and Technology (ETH, Department of Chemistry and Applied Biosciences, Zürich, Switzerland) Oct. 2000 – Mar. 2006 Studies in Biology, Eberhard Karls University, Tübingen, Germany (Diplom 2006) May 2005 – Mar. 2006 Diploma thesis at Eawag, Switzerland July 2003 – Nov. 2003 Victoria University of Wellington, New Zealand July 1999 – Aug. 2000 Alternative civilian service, Jugendhilfe Böblingen, Germany 1990 – 1999 Gymnasium Renningen, Germany (Abitur 1999) 165 List of Publications List of Publications T. Heck, A. Reimer, D. Seebach, J. Gardiner, G. Deniau, A. Lukaszuk, H.-P. E. Kohler, B. Geueke. β-Aminopeptidase-catalyzed biotransformations of β2-dipeptides: kinetic resolution and enzymatic coupling. ChemBioChem, 2010, 11, 1129-1136. T. Heck, V. S. Makam, J. Lutz, L. M. Blank, A. Schmid, D. Seebach, H.-P. E. Kohler, B. Geueke. Kinetic analysis of L-carnosine formation by β-aminopeptidases. Adv. Synth. Catal. 2010, 352, 407-415. J. Heyland, N. Antweiler, J. Lutz, T. Heck, B. Geueke, H.-P. E. Kohler, L. M. Blank, A. Schmid. Simple enzymatic procedure for L-carnosine synthesis: whole-cell biocatalysis and efficient biocatalyst recycling. Microb. Biotechnol. 2009, 3, 74-83. T. Heck, D. Seebach, S. Osswald, M. K. J. ter Wiel, H.-P. E. Kohler, B. Geueke. Kinetic resolution of aliphatic β-amino acid amides by β-aminopeptidases. ChemBioChem 2009, 10, 1558-1561. T. Heck, H.-P. E. Kohler, M. Limbach, O. Flögel, D. Seebach, B. Geueke. Enzyme-catalyzed formation of β-peptides: β-peptidyl aminopeptidases BapA and DmpA acting as β-peptide- synthesizing enzymes. Chem. Biodiversity 2007, 4, 2016-2030. T. Heck, H.-P. E. Kohler, B. Geueke. Enzymatisches Syntheseverfahren zur Herstellung β- aminosäurehaltiger Peptide. Swiss Patent Application 01007/07, 2007. T. Heck, M. Limbach, B. Geueke, M. Zacharias, J. Gardiner, H.-P. E. Kohler, D. Seebach. Enzymatic degradation of β- and mixed α,β-oligopeptides. Chem. Biodiversity 2006, 3, 1325-1348. B. Geueke, T. Heck, M. Limbach, V. Nesatyy, D. Seebach, H.-P. E. Kohler. Bacterial β- peptidyl aminopeptidases with unique substrate specificities for β-oligopeptides and mixed β,α-oligopeptides. FEBS J. 2006, 273, 5261-5272. J. Dittmann, K. Keller-Matschke, T. Weinschenk, T. Kratt, T. Heck, H. D. Becker, S. Stevanovic, H. G. Rammensee, C. Gouttefangeas. CD8+ T-cell response against MUC1- derived peptides in gastrointestinal cancer survivors. Cancer Immunol. Immunother. 2005, 54, 750-758. 167