Drug Target Crystallography and Structure Based Drug Design

Program development and management experience

Program Development. Experience form my own research projects and my participation in and review of large collaborative programs strongly suggests that a mark of a successful research focus is effective utilization of synergies between discovery based, technology driven components and clearly defined, hypothesis driven interests. An important, and in my opinion indispensable, third component is formed by advanced bioinformatics and in silico screening and prediction efforts. Genomics, HT proteomics, or large scale virtual ligand screening are fundamental discovery components, which need substantial bioinformatics support to create knowledge from data. Together with my colleague Prof. Kantardjieff, CSUF, I recently reviewed the role of structural bioinformatics in drug discovery, and an example for in silico based discovery may serve the mapping of protein-protein interaction pathways in the TB genome which now provides an opportunity to develop novel classes of antibiotic drugs and chemotherapeutics.  

Program Management Experience. I founded in 1993 the drug target crystallography group at the U. of California Lawrence Livermore National Laboratory. After startup funding, I supported my team for the last 7 years through extramurally funded grants for which I was fully responsible as the signing principal investigator (volume $1.2M/year). As I have reached the limits of expansion possible in the present environment, I seek to pursue my interests in drug discovery and to establish a dedicated research team in a collaborative environment spanning all aspects of modern biomedical science and drug discovery. I act under non-disclosure as a consultant to several US drug discovery ventures and instrument manufacturers. I have filed multiple patents, licensed technologies, and I am familiar with intellectual property protection. Presently establishing a venture employing an innovative and unique combination of genetic screening and directed evolution with structure based drug design methods and systems analysis to the design of antimicrobial and antiviral therapeutics.     

Drug target structures and drug discovery

MTB. The scientific goal on the TB consortium project and one of my personal research interests was to provide drug target structures and lead compounds for TB, a serious, re-emerging disease receiving little interest from commercial ventures. I have guided several students in the determination of a number of drug target complex structures, mostly in the fatty acid and amino acid pathways, unique and thus interesting drug target pathways in Mycobacterium tuberculosis (see LysA paper). Two more structures in the rhamnose synthetic pathway (rmlC) and pyrimidine nucleotide synthetic pathway (pyrR) have been independently determined in my lab, and with Dr Kantardjieff, CSFU, we have evaluated drug leads using virtual ligand screening (Acta D cover page), and have recently reviewed the role of structural bioinformatics in the development of new antimycobacterial drugs. In an extension of our successful collaboration with Jim Sacchettini, Texas A&M, together with Bill Jacobs, an authority in TB function and genomics, we were awarded a project grant focusing specifically on MDR TB drug target biochemistry utilizing my structural based drug discovery and ligand screening experience. Several compounds designed in collaboration with CSUF based on my target structures are presently in in-vivo trials.

Calcium signalling and allergens. In an independent project, Mark Wilson (Brandeis) and I have recently discovered a novel, effective conformational switch inhibitor of Ca signalling in a near atomic resolution structure I solved in my lab, emphasizing the exciting element of discovery inherent in every structure determination (EBI-14440, previous structure in Nature Reviews poster). Surprisingly, the structure of the collapsed form of the CAM complex exhibits features and binding pockets nearly identical to a domain swapped dimer of a pollen allergen, providing potential leads for effective small molecule anti-allergens. The electron density of the new chemical entity has been featured on the cover of Drug Discovery Today 11 (13-14), 2006.

Botulinum neurotoxins. Another drug target structure from my lab that provided evidence for a novel, noncanonical protease mechanism for the action of these most powerful natural neurotoxins has been recently published in the Proceedings on the National Academy of Sciences, after I published critical remarks in Nature SB earlier. This study had a wide impact on small molecule peptidomimetic drug design against botulinum. Another, related interesting project I pursued was the elucidation of the mechanism of a Superantigen Vaccine.     

Neuroreceptors. I enjoy a very exciting collaboration with Prof. Werner Sieghart, Institute for Brain Research of the Medical University in Vienna, involving structure based drug design and structural biology of gamma-aminobutyric acid type A (GABA-A) receptors.  GABA-A receptors are ligand-gated chloride channels and are the major inhibitory transmitter receptors in the central nervous system. They are modulated by a variety of clinically important drugs such as benzodiazepines, barbiturates, steroids, anesthetics, and convulsants. Currently available drugs, however, cannot distinguish between the large number of GABA-A receptor subtypes, and thus, exhibit significant side effects. Prof. Sieghart and his team are leaders in this field, and can produce intact receptors in Sf-9 cells as well as express extracellular subunits. We hope that combined with my experience in small-scale, high throughput expression and solubilization screening, we can determine a functional subunit complex structure with the drugs binding located at specific subunit interfaces. As a long term goal, I proposed to determine the intact, pentameric membrane receptor structure. I plan to build on my substantial structure based drug discovery expertise and expand to target membrane proteins and large protein-protein complexes, which in fact form the majority of proteins presently targeted by pharmaceuticals, as well as targets for entirely new classes of drugs.

GPCR (taste) receptors. During my sabbatical at Texas A&M in a collaboration with the Sacchettini group, we determined co-crystal structures of the murine monoclonal IgG2b(κ) antibody NC6.8 Fab fragment complexed with high-potency sweetener compound SC45647 and non-tasting high affinity antagonist TES. The crystal structures show how sweetener potency is fine-tuned by multiple interactions between specific receptor residues and the functionally different groups of the sweeteners. Comparative analysis with the structure of NC6.8 complexed with the super-potency sweetener NC174 reveals that although the same residues in the antigen binding pocket of NC6.8 interact with the zwitterionic, tri-substituted guanidinium sweeteners as well as TES, specific differences exist and provide guidance for the design of new artificial sweeteners. In case of the non-sweetener TES, the interactions with the receptor are indirectly mediated through a hydrogen bonded water network, while the sweeteners bind with high affinity directly to the receptor. The presence of a hydrophobic group interacting with multiple receptor residues as a major determinant for sweet taste has been confirmed. The nature of the hydrophobic group is likely a discriminator for super- versus high-potency sweeteners, which can be exploited in the design of new, highly potent sweetener compounds. Overall similarities and partial conservation of interactions indicate that the NC6.8 Fab surrogate is representing crucial features of the T1R2 taste receptor VFTM binding site. We plan to determine the structure of the human T1R2 GPCR receptor VFTM domain complexed with sweeteners.

Superantigen Vaccines. In my lab we have also answered a structural question in immunology by determining the structure of a Staphylococcal superantigen vaccine. The SEA mutant vaccine revealed a cascade of structural rearrangements located in three loop-regions essential for binding the alpha-subunit of major histocompatibility complex (MHC) class II molecules, thus breaking the T-cell receptor - MHC-II cross linking via the superantigen. As multi-drug resistance of S. aureus (MRSA) is a major cause of often lethal nosocomial infection, with only vancomycin left as last resort, an effective vaccine is of great interest.     

Structural bioinformatics and computing

Quality Assessment. Work on drug target structures has also clearly indicated to me that quality assessment of protein-ligand structures is an extremely important aspect of modern drug target crystallography, and is often ignored by beginning as well as advanced crystallographers. After commenting critically on a C. botulinum drug target structure in Nature Structural Biology and being instrumental in the retraction of another incorrect drug target structure, I implemented a bias removal protocol server which is now publicly accessible. Structures of drug targets and their ligand complexes need not only to be accurate to allow conclusive insights for drug discovery and mechanism, but there is also a need to rapidly identify and discover yet unknown binding sites and key features relevant to biological activity and ligand binding. In addition to present sequence and structural homology based methods, electron density and structure based feature identification, implemented as an extension of the TB consortium validation service, provides new tools allowing identification of structural hot spots. Examples are backbone torsion angle outliers in well defined electron density, validated cis-peptide locations, and deviations form geometry targets in structurally well defined regions. Automatic graphic presentation and annotation of these features during validation provides additional insight into structure-function relations in an unbiased fashion. This information itself is data-minable and enables discovery of new structure-function determining patterns.

Machine learning. High throughput screening and sequencing projects produce enormous amounts of data, and even in hypothesis driven research, much information is created that needs to be transformed into knowledge by statistical methods and machine learning. In a recent paper (Maximum Likelihood Crystallization) and a invited review on Predictive Models for Protein Crystallization I critically evaluate the status of machine learning for predictive frameworks for protein crystallization and point out that statistical frameworks need to be embedded with practical implementation to optimally make use of the wealth of crystallization data accumulated in high throughput efforts.

Instrumentation development

High-throughput crystallography. During the past years I have developed the high throughput drug target crystallography facility of the TB Structural Genomics Consortium and shifted the focus to structure based drug discovery. Initially proposed throughput targets were exceeded by a factor of ten. This is largely a result of my skills in developing and efficiently integrating novel automation technologies in a research environment. In a number of recent reviews I have outlined the caveats of implementing high throughput technology in drug discovery and I emphasize the challenges of robotic automation and integration with data handling in structural proteomics and the need for a process driven view of high throughput drug target crystallography. With my coworkers, I have also implemented in my previous lab small scale auto-induction protocols for rapid parallel protein expression screening. In collaboration with Square One Systems Design, Jackson, WY, I am presently challenging the final frontier in high throughput drug target crystallography, the design of automated crystal harvesting robots through a NIH STTR Phase-II grant through my company.   

Outlook

Novel therapeutic drugs and drug target pathways. As a result of the methodological and technical developments in protein production, crystallization, high throughput drug target crystallography and screening techniques I have contributed to, structure guided drug discovery is at a stage where a synergistic view of biological processes becomes feasible. Protein-protein interactions, membrane and complex association, context sensitive conformation, and the role of low complexity regions are now emerging as interacting research areas that will allow to construct a picture informing about whole pathways and cellular processes in unprecedented scale and detail, providing new and exciting avenues for the design and discovery of novel pharmaceutical drugs. I expect to contribute such exciting developments and discoveries by establishing an interdisciplinary and unique structure based drug discovery core integrated in a highly interactive and top rated research environment. My latest research focuses on establishing a venture employing an innovative and unique combination of genetic screening and directed evolution with structure based drug design methods and systems analysis to the design of resistance-resistant chemotherapeutics.   

Bernhard Rupp