Systematische Analyse von Epoxidhydrolasen : Erstellung einer familienspezifischen Datenbank, Entwicklung einer strukturbezogenen Klassifizierung und Amplifikation unbekannter Epoxidhydrolasen

Abstract

Im Laufe dieser Arbeit wurde eine systematische Analyse an 93 Epoxidhydrolase (EH) Sequenzen durchgeführt. Ein Vergleich der drei bekannten EH Strukturen aus Agrobacterium radiobacter, Mus musculus und Aspergillus niger zeigte die hohe Konservierung des a/b Hydrolase Folds und die Existenz zweier variabler Loop-Regionen, dem NC-und dem Cap-Loop. Aus dem Multisequenz-Alignment der 93 EHs wurden die Loop-Längen aller EHs bestimmt und in einem Diagramm gegeneinander aufgetragen. Hierbei bildeten sich drei nach Superfamilien und Organismengruppen getrennte Cluster. Jedes Cluster enthält eine der drei bekannten Strukturen. Unter der Annahme, dass ähnliche Loop-Längen eine ähnliche Struktur bedingen, konnten Homologiemodelle von fünf EHs erstellt werden. Vergleiche der Substratspektren mit der Cluster Zugehörigkeit führten zu der Annahme, dass Loop-Länge und Substratspezifität korrelieren, da z.B. nur für EHs mit langen Cap-Loops der Umsatz epoxidierter Fettsäuren beschrieben wurde. Mit dem Programm CODEHOP wurden degenerierte familienspezifische Primer erstellt und an zwei bakteriellen Organismen getestet. Aus Streptomyces antibioticus Tü4, welcher eine EH Aktivität gegen Styroloxid aufweist und Rhodococcus ruber LU760 konnten jeweils Fragmente mit signifikanter Homologie zu EHs amplifiziert werden.

Der zweite Teil dieser Arbeit befaßt sich mit der Erstellung und Auswertung einer familienspezifischen Datenbank auf der Grundlage der Lipase Engineering Database (LED). Die Epoxidhydrolase/Haloalkandehalogenase (EH/HD) Datenbank enthält 397 Sequenz-einträge, 305 Proteineinträge und für 19 dieser Proteineinträge 51 Strukturdatensätze. Neben EHs und HDs enthält die Datenbank weitere Enzymfamilien, die in 14 homologe Familien und zwei Superfamilien eingeteilt wurden. Untersuchungen der Strukturen aus EH/HD Datenbank und LED zeigten, dass trotz massiver struktureller Unterschiede, zwischen den Strukturen dieser beiden Datenbanken, das katalytische Zentrum strukturell in allen a/b Hydrolasen erhalten ist. Auch das für Lipasen ermittelte Anker-Konzept für die Stabilisierung des Oxyanionholes greift bei den Proteinen der EH/HD Datenbank. Für das hochkonservierte GXGXS-Motiv konnte eine mögliche strukturrelevante Funktion entwickelt werden, die seine konservierte Struktur und Lage erklärt.

Epoxide hydrolases (EHs) from eukaryotic and prokaryotic sources were found to be applicable for the kinetic resolution of racemic mixtures of epoxides, and high enantiomeric accesses have been reported. Enantiopure epoxides and vicinal diols are extensively employed as useful chiral building blocks for synthesis of various bioactive products in the pharmaceutical and agrochemical industry. Aim of this work was to identify similarities and differences in this highly divers enzyme family by systematic analysis of available EH sequences and structures. 93 different EHs were identified by keyword and BLAST search. A phylogenetic analysis showed the clear separation into the two separate superfamilies cytosolic (cEHs) and microsomal (mEHs) EHs and several homologous families within each superfamily. In contrast to eukaryotic EHs, bacterial EHs show a high variability and barley form any well defined homologous families. EHs belong to the family of a/b hydrolases. Analysis of the three known structures from Agrobacterium radiobacter, Mus musculus und Aspergillus niger lead to division of the modular structures into N-terminal domain, N-terminal catalytic domain, NC-loop, cap-domain, cap-loop und C-terminal catalytic domain. Besides obvious differences in molecule size of the three structures, caused by the variable N-terminal domain, EHs contain a highly conserved a/b hydrolase fold only interrupted by two variable loops: NC- (connects N-terminal catalytic domain and cap-domain) and cap-loop (located within the cap-domain). The bacterial EH from Agrobacterium radiobacter contains short loops, the EH from Mus musculus contains a medium sized NC-loop and a long cap-loop and the microsomal EH from Aspergillus niger contains a long NC-loop and a short cap-loop. Loop-lengths of all EHs were determined using the multisequence alignment of 93 EHs and were plotted against each other. Three clusters ordered by superfamily and organisms were formed, each containing one known structure. Although EHs contain a highly conserved structure, they show low sequences identities, what complicates reliable homology modelling. Assuming that EHs of one cluster, and thus with similar loop-lengths posses a similar structure, the loop-based classification into clusters and the multisequence alignment of 93 EHs was used to perform homology models of five representative EHs. Of 78% of all 93 EHs the complete structure can be predicted using the insights resulting from the systematic analysis. Another application for the loop-based assignment into clusters, is the prediction of substrate specificity of EHs. Looking at the substrate spectra of EHs, it is remarkable that EHs with long cap-loops are able to convert epoxy fatty acids, while EHs containing long NC-loops accept epoxy polycyclic aromatic hydrocarbons as substrates. The assignment to homologous families received from the phylogenetic analysis was used for the design of family specific primers. Tests of these primers were performed with two bacterial organisms. Streptomyces antibioticus Tü4 shows EH activity towards styrene oxide. An already amplified fragment of this species was located in the branch of the bacterial EHs of cluster I. With primers specific for this group, a fragment with significant homology to EHs was amplified. After purification, N-terminal sequencing and tryptic digestion, it was possible to complement the fragment N-terminal. Using the same set of primers, amplification of a fragment from genomic DNA from Rhodococcus ruber (LU760)was also possible. The second part of this work is concerned with building up and interpretation of a family specific database. The epoxide hydrolase/haloalkane dehalogenase database (EH/HD) is an extension to the 93 EH alignment with inclusion of further enzyme families and structure information. A set of representative sequences was used for BLAST searches. The collected data was ordered hierarchically by the database system DWARF in protein entry, homologous family and superfamily. The database consists of 397 sequence entries, 305 protein entries and contains for 19 protein entries 51 structure data sets, which were combined into 14 homologous families and two superfamilies. The GXGXS-Motif is highly conserved and forms a structurally also highly conserved loop. The GXGXS-Motif occurs only in the structures of the EH/HD database and is located beneath the first two b-strands of the a/b hydrolase fold and the cap-helice a4 and a5, which are connected by hydrogen-bonds. The positions of the five residues involved in hydrogen bonding of the strands b1 and b2 and the GXGXS-motif are structurally conserved. Hydrogen bonding between GXGXS-Motif and cap-domain is more variable, but all residues involved are located at the loop between the cap-helices a4 and a5. The GXGXS-Motif stabilizes the strands b1 and b2 and is acting like a lock, connecting cap domain and N-terminal catalytic domain.