| dc.description.abstracteng | NMR spectroscopy plays an important role in all stages of the drug discovery
process. Namely in the structure elucidation of natural products, synthetic
ligands and metabolites, as well as a high-throughput screening technique.
Yet, the application of NMR in structure-based drug design, e.g. in finding
the binding mode of a small molecule drug to a macromolecular protein receptor
is far from making the most from the opportunities available to it.
Structure-based drug design is a powerful and widely used tool for the optimization
of low molecular weight compounds that should be turned into
highly efficient drugs. The method mainly relies on high-resolution crystal
structures of the receptor-ligand complex to obtain the required information
for optimizing target binding of small molecules. However, obtaining crystals
and structures of sufficient quality cannot be achieved for nearly the
half of pharmaceutically relevant protein targets. For those target proteins
that cannot be crystallized, NMR spectroscopy is an alternative and structures
of protein-ligand complexes can be determined, provided the protein
can be labelled with stable isotopes such as 13C or 15N. However, pharmaceutically
relevant non-crystallizable target proteins are often non-tractable
by NMR, because they are too large and result in overcrowded spectra or
they cannot be expressed in bacteria and therefore cannot be labelled with
stable isotopes enabling heteronuclear NMR. In such cases one can employ
INPHARMA (Inter-Ligand NOEs for PHARmacophore MApping). It utilizes
two ligands that bind competitively to the same binding pocket of a
protein. INPHARMA peaks in a NOESY spectrum emerge from the magnetization
transfer from the protons of one ligand to the protons of the other
ligand via the protein protons, provided the ligands dissociate from the protein
several times during the NOESY mixing time. The method is further
developed and it is investigated whether the methodology can be improved by
inclusion of Saturation Transfer Difference (STD) restraints and transferred
NOE (trNOE) restraints in addition to the INPHARMA restraints. STD is
a frequently used technique in NMR spectroscopy and NMR-based screening
for protein binders. The technique is developed and tested on protein kinase
A, where crystal structures of the protein/ligand complexes are known. The
results show that the combination of the NMR methods INPHARMA, tr-
NOE and STD results in a precise scoring function for docking modes and
therefore the determination of ligand binding modes. It is demonstrated
that the method is superior to docking scoring functions alone and can lead
to the correct result by using a molecular dynamics simulation driven refinement,
even if the initial conformation of the protein side chains is not
correct. Multiplexing of several ligands improves the reliability of the scoring
function further. Then the technique is extended the G-protein coupled
receptor GPR40, a membrane protein, for which only homology models exist
and which is an interesting drug target in on-going research. For this system,
the ligand binding mode found is supported by SAR data. The binding
mode of epothilone to tubulin, an important interaction for cancer therapy
is reinvestigated using STD data. The binding mode found by INPHARMA
is confirmed and further optimized, while the electron crystallography derived
structure does not fit to the experimental NMR data. The NMR-based
ligand binding mode determination method is presented to derive binding
modes of ligands based on simple NMR experiments (NOESY and STD). It
is demonstrated on the examples of PKA, GPR40 and the tubulin-epothilone
complex, that based on a crystal structure or homology model of the protein,
binding modes can be determined that can be used for pharmacophore
mapping and drug optimization.
In the second part the drug metabolism of anle138b, a modulator of toxic
protein oligomers in prion and Parkinson’s disease is investigated. A methodology
is developed to extract the drug from organs and to determine its concentration
in the brain. It was confirmed that anle138b is the only active
compound in the brain, while metabolites are only formed in liver and kidney.
With combined HPLC, mass spectrometry and NMR techniques, the
structures of the metabolites were determined and the drug metabolism of
anle138b in the mice and rat model was revealed.
In the last part NMR spectroscopy is applied to reinvestigate the structural
and stereochemical features of arthrofactin, a potentially antibiotic natural
product. Arthrofactin was initially reported in 1993 as a bioactive cyclic
lipopeptide from the bacterium Pseudomonas sp. The structure of arthrofactin
and its derivatives was reassigned on the basis of extensive NMR experiments
and chiral HPLC analysis. A new approach of phylogenetic structure
prediction is tested and was successfully approved with NMR data.
In conclusion, NMR spectroscopy is applied and further developed in this
thesis to several challenges of the drug discovery process. | de |