Abstract:
Infection of many pastoral grasses by mutualistic fungal endophytes has significant effects on the biology of
these grasses. Thus this fungal-plant interaction is also of very significant economic importance. Many of the
effects are mediated by an array of secondary metabolites produced by the fungi. These secondary metabolites
are in many cases synthesised by a class of enzymes known as the nonribosomal peptide synthetases
(NRPSs). NRPSs are large enzymes that are organised into modules made up of functional domains. NRPSs
are found in many organisms and synthesise a number of medically-important peptides such as antibiotics
and immunosuppressants. This research used structural bioinformatics and experimental structural biology to
further our understanding of the structure and function of NRPSs.
An in silico method of predicting the specificity of NRPSs was investigated. Homology models of the NRPS
domains responsible for substrate recognition, the adenylation domains were built and ligand docking was
used to dock a library of 110 model ligands into the active sites. The docking protocol was refined to ensure that
the mainchain atoms of the amino acid ligands were docked correctly. Testing of the method with adenylation
domains of known specificity revealed a need for further experimental structural data in order to accurately
model the active sites particularly data for domains from fungal sources such as endophytes.
NRPS domains were targeted for structure determination using a structural genomics approach. Twenty four
domain fragments from four fungal endophyte NRPSs and two domains from an Escherichia coli NRPS were
cloned and expressed. Both of the E. coli domains and two of the endophyte NRPSs were solubly expressed. The
purified E. coli NRPS domains were used for preliminary studies into developing a nuclear magnetic resonance-
based specificity assay. Fungal NRPS domains were subject to extensive screening and fine-tuning of expression
constructs, lysis buffer conditions, purification approaches and crystallisation conditions. The structure of the
third adenylation domain from the fungal NRPS, SidN, was solved by X-ray crystallography to 2.0 A resolution.
SidN is involved in synthesising an extracellular Siderophore and has previously been found to be essential for
the maintenance of the mutualistic character of the grass-endophyte relationship.
The structure is made up of two highly-conserved subdomains. The orientation between the two subdomains
was an 'open' conformation previously seen in distantly-related members of the protein superfamily but not
seen in NRPS adenylation domains. This finding has implications for the mechanism of catalysis. The SidNA3
structure revealed a large substrate binding pocket lined by fifteen residues and ligand docking showed that the
pocket would accommodate the putative substrate, Nδ-anhydromevalonyl-Nδ-hydroxyornithine. This is the first
eukaryotic NRPS domain structure to be solved and the first of any protein from a fungal endophyte of grass.
This research provides the tools for a comprehensive analysis of the specificity determinants of fungal NRPSs
and the refinement of a general in silico substrate specificity prediction method.