Kirchmair Group | Application of Computational Methods in Chemical Biology

Application of computational methods in chemical biology

Our current projects in computational chemical biology focus on the identification and optimisation of bioactive small molecules that

  • Disrupt protein-protein interactions
  • Recover the activity of mutated, dysfunctional proteins or stabilise their structure
  • Are active on drug-resistant microorganisms
  • Target highly flexible biomacromolecules.

We employ a wide range of computational methods to

  • Analyse the structure and function of biomacromolecules, in particular also the implications of mutations on protein structure, function, flexibility and ligand binding:
  • Protein structure modelling
  • Molecular dynamics simulations
  • Screen large virtual libraries for bioactive synthetic compounds and natural products:
  • Docking, pharmacophore modelling, alignment-based methods, etc.
  • Identify the likely mode of action (i.e. the biomolecular targets) of small molecules:
  • Machine learning, 2D and 3D similarity-based approaches, structure-based approaches, etc.
  • Design innovative ligands based on novel molecular scaffolds:
  • De novo design
  • Elucidate and analyse quantitative structure-activity and structure-property relationships (QSARs/QSPRs):
  • Classical QSAR and QSPR models
  • Machine learning
  • Predict the metabolic fate and toxicity of small molecules:
  • Machine learning
  • Alignment-based and structure-based approaches
  • Identify compounds prone to trigger false readouts in biological assays:
  • Machine learning
  • Analyse the physicochemical property and scaffold space covered by large molecular libraries, and for diversity analysis and compound selection:
  • Clustering, etc.

 

The molecular libraries that we compiled in our laboratory for virtual screening include:

  • 10 million compounds readily purchasable from reliable commercial sources
  • 370 million compounds synthesised on demand by CROs within a short period of time
  • 250k natural products (to our knowledge, this is the largest ready-to-screen database of natural products available to date)
  • 25k readily purchasable natural products

 

Selected research outcomes

We successfully applied a pharmacophore-based approach to identify thienoquinolines as a novel class of compounds disrupting the protein-protein interaction of PKCepsilon and RACK2, a result of a long-term collaboration with pharmacologists and chemists. We also identified novel chemical chaperones that correct phenylketonuria in mice with an alignment-based approach. These success stories encouraged us to look more deeply into this interesting topic, and to extend our collaborations with academic partners in Germany, Austria and the UK.

Influenza neuraminidase is the primary target of anti-influenza drugs and is known for its significant conformational flexibility. In recent years influenza strains resistant to marketed drugs have emerged, and hence the development of effective drugs is of utmost importance to public health. Together with researchers at the University of Jena, University of Innsbruck and University of Vienna, we have identified synthetic and natural product inhibitors of influenza neuraminidase, and our lab also contributed to the understanding of the underlying inhibitory mechanisms by observations from molecular dynamics simulations. One of our most recent successes is the discovery of natural products and synthetic molecules inhibiting both viral and bacterial neuraminidases, hence disrupting the lethal synergism between Streptococcus pneumoniae and influenza virus.

We have also found new leads inhibiting the viral coat protein 1 of coxsackievirus B3 and revealed the mechanism of drug resistance caused by mutations of coat protein 1 that are located distant from the ligand (i.e. pocket factor) binding site. Further works include the identification and design of inhibitors for several anti-viral targets, e.g. HIV-1 integrase and reverse transcriptase.

Structural model of influenza neuraminidase bound with oseltamivir.