C30 IDENTIFICATION OF PROTEINS THAT BIND WILD‐TYPE (WT) AND/OR FAMILIAL AMYOTROPHIC LATERAL SCLEROSIS (FALS)‐LINKED MUTANT CU/ZN SUPEROXIDE DISMUTASE (MTSOD1) USING PHAGE DISPLAY
Ghadge GD1, Kay BK2, Scholle MD3, Monti AL1, Fu R1, Roos RP1
1The University of Chicago Pritzker School of Medicine, Chicago, IL, USA, 2University of Illinois at Chicago, Chicago, IL, USA, 3Argonne National Laboratory, Argonne, IL, USA
E‐mail address for correspondence: [email protected]
Background: There is convincing evidence that FALS‐linked MTSOD1 kills motor neurons because of toxicity rather than a deficiency of dismutase activity; however, the basis for this toxicity remains unclear as does effective treatment for this devastating fatal disease. Studies have suggested that MTSOD1 misfolds, and that the mutant protein's aggregation interferes with normal SOD1‐protein interactions, sequestering proteins that are important for cell survival.
Objective: To identify peptides that interact with WT and MTSOD1 in order to better understand FALS pathogenesis and to generate reagents that can be used in FALS treatment.
Methods: We used phage display technology to identify peptides that interact with WT and/or MT SOD1. Affinity selection was carried out using phage‐displayed combinatorial peptide libraries with purified SOD1 that had been expressed and biotinylated in E. coli as the target. Phage particles displaying peptides were isolated and tested for binding affinity to SOD1 by ELISA. High affinity binding peptides were sequenced, and the predicted amino acid sequence of the peptide ligand was aligned to identify a consensus motif.
Results: We have isolated a number of peptide clones by affinity selection using WT or MTSOD1 as a target. Following sequencing, peptide ligands have been identified that are presumed to be involved in SOD1 binding. Blast search of these peptide ligands identified a number of proteins with a similar consensus sequence. Some of these proteins have been implicated in protein transport and misfolding.
Discussion and conclusions: The pathogenesis of FALS and other neurodegenerative diseases has been hypothesized to involve the formation of aggregates containing the mutant protein relevant to the disease (e.g. SOD1), and the sequestration of proteins, key to the viability of the cell, in these aggregates. For this reason, the identification of SOD1‐binding proteins is of importance.
We chose to identify SOD1‐binding proteins using phage display technology because of the great power of this method. Our studies identified a number of binding peptides. Sequencing of these peptides demonstrated similar consensus sequences from a number of the peptides. The peptide sequences that we identified are present in proteins involved in protein transport and misfolding. We are presently confirming binding of these proteins to SOD1 by immunoprecipitation. The identification of these proteins may clarify FALS pathogenesis.
We are also testing the ability of the peptides that we have identified to perturb MTSOD1‐induced aggregation and/or death in cell culture models. Peptides that interfere with the pathogenicity of MTSOD1 may be useful in FALS therapy.
Acknowledgement: This work was supported by the ALS Association.
C31 INTERMOLECULAR DISULPHIDE BONDS‐MEDIATED AGGREGATION OF SOD1 IN AMYOTROPHIC LATERAL SCLEROSIS
Deng H‐X, Shi Y, Furukawa Y, Zhai H, Fu R, Liu E, Gorrie G, Dal Canto MC, O'Halloran T, Siddique T
Northwestern University, Chicago, IL, USA
E‐mail address for correspondence: [email protected]
Background: Twenty per cent of the familial form of amyotrophic lateral sclerosis (ALS) is caused by mutations in the Cu/Zn superoxide dismutase gene (SOD1) through the gain of a toxic function. The nature of this toxic function of mutant SOD1 has remained largely unknown. SOD1 aggregates are a pathological hallmark in ALS patients with SOD1 mutations and ALS mouse models that overexpress ALS‐associated SOD1 mutants. The relevance of these SOD1 aggregates to development of ALS and the fundamental molecular mechanism by which the SOD1 mutants form aggregates are not clear.
Objectives: The objectives of this study were to elucidate 1) the possible causal relationship between SOD1 aggregates and development of ALS; and 2) the molecular mechanism of how the SOD1 forms aggregates.
Methods: Single and double transgenic mice overexpressing wild‐type and various SOD1 mutants were developed. SOD1 aggregates were analysed using biochemical, immunological and pathological methods.
Results: We found that wild‐type SOD1 exacerbates the ALS phenotype in double transgenic mouse models overexpressing SOD1G93A and SOD1L126Z. This phenomenon is similar to that previously reported in prion disease. We also found that w‐t SOD1 can convert an unaffected phenotype to an ALS phenotype in mutant SOD1 transgenic mouse model overexpressing SOD1A4V in a dose‐dependent manner. Further analyses of the single and double transgenic mice revealed that conversion of mutant SOD1 from soluble form to an aggregated and detergent‐insoluble form was associated with the development of an ALS phenotype in transgenic mice. Conversion of w‐t SOD1 from soluble form to an aggregated form correlates with exacerbation of the disease or conversion to a disease phenotype in the double transgenic mice. We found that this conversion, observed in the mitochondria of the spinal cord, involved formation of insoluble SOD1 dimers and multimers that are cross‐linked through inter‐molecular disulphide bonds via oxidation of cysteine residues in SOD1.
Discussion and conclusions: The aggregated and detergent‐resistant form of SOD1, either mutant or wild‐type, is associated with ALS and is apparently the pathogenic form of the protein. These aggregates are predominantly observed in mitochondria of the spinal cord of the ALS mouse models and are formed by intermolecular linked SOD1 dimers and multimers via oxidation of the cysteine residues in SOD1. Thus, our findings provide evidence of direct links between oxidation, protein aggregation, mitochondrial damage and SOD1‐mediated ALS with possible applications to the ageing process and other late‐onset neurodegenerative disorders, such as prion disease and Alzheimer's disease.
Acknowlegements: These studies were supported by grants from the National Institutes of Health (NS040308, NS050641, NS046535), Les Turner ALS Foundation, ALSA, Playing to Win 4 Life Foundation, V. E. Schaff ALS Research Fund, H. Post Research Professorship, Wenske Foundation, Falk Medical Research Trust, The Les Turner ALS Foundation/Herbert C. Wenske Foundation Professorship, and The David C. Asselin MD Memorial Fund
C32 STRUCTURE BASED DESIGN OF SOD‐1 AGGREGATION INHIBITORS: IMPLICATIONS FOR DEVELOPMENT OF A NEW CLASS OF ALS THERAPEUTICS
Ray S, Brown R, Lansbury P, Nowak R
1Harvard Medical School, Cambridge, MA, USA, 2Brigham & Women's Hospital, Boston, MA, USA, 3Massachusetts General Hospital, Charlestown, MA, USA
E‐mail address for correspondence: [email protected]
Background: Aggregation of mutant forms of the housekeeping gene SOD‐1 has been implicated in motor neuron cell death in a subset of the familial form of ALS. The downstream mechanisms following aggregation that lead to death of motor neurons are poorly understood and have proven to be difficult targets for drug discovery. We propose a novel approach for development of ALS therapeutics based on the concept of ‘native state protein stabilization’.
Objective: In the absence of any consensus for the mechanism of SOD‐1 mediated toxicity, stabilization of the native state of the protein using small molecules is a viable strategy for developing ALS therapeutics. We propose to design molecules that would be structure dependent rather than mechanism dependent. Stabilizing the SOD‐1 dimer would automatically eliminate all downstream events that eventually lead to neurotoxicity.
Methods: Based on this concept, we initiated an in silico screening programme to find drug‐like molecules that would stabilize the SOD‐1 dimer. About 1.5 million drug‐like molecules from commercial databases were screened against a dimer interface cavity. 100 molecules with the highest predicted binding affinity were tested experimentally in vitro and 15 out of the 100 molecules significantly inhibited aggregation and denaturation of a number of common FALS mutants such as A4V, G85R, S134N, G93A and H46R. All of the mutants exhibited near wild‐type‐like stability in the presence of these 15 molecules.
Results: Close examination of these 15 hits led to the identification of an aza‐uracil pharmacophore around which a structure‐activity relationship (SAR) was designed using a set of 50 related compounds with drug‐like features (Lipinski based). The compounds were tested both in vitro (biochemical assays) and in mouse neuroblastoma (N2A) mutant SOD‐1 cell lines. A second assay was designed to improve binding‐selectivity for these compounds in serum and CSF to generate highly target‐specific molecules (designer drugs). A set of compounds with drug‐like properties were obtained from the secondary screens. The molecules were also found to be fairly non‐toxic in cell culture and preliminary mice studies. Currently, some of these molecules are being tested in a FALS transgenic mouse model.
Conclusions: We have obtained set of molecules capable of blocking SOD‐1 aggregation, which also have drug‐like properties and have low levels of toxicity in cells. While there are no viable ALS drugs and not much progress is expected from the private sector since the market is too small to justify an intensive effort, we propose a model where at least part of the drug development procedure can be undertaken in an academic environment. We hope such an effort will promote collaboration with industry towards a much‐needed cure.
C33 ULTRA‐HIGH AND ATOMIC RESOLUTION STRUCTURES OF CU/ZN SOD1 AND ITS FALS MUTANTS
Antonyuk SV, Strange RW, Hasnain SS
Daresbury Laboratory, Warrington, Cheshire, UK
E‐mail address for correspondence: [email protected]
About 10% of all ALS cases are familial (FALS). Dominant inheritance of point mutations in Cu/Zn superoxide dismutase (SOD1) are involved in about 20% of FALS. This well characterized subset of FALS offers the best chance of understanding the origin of the disease and arriving at a general strategy for a cure. SOD1 is a crucial component of the cellular response to oxidative stress, catalysing the dismutation of the superoxide radical to hydrogen peroxide and water. Recently, the toxic gain‐of‐function properties of SOD1 including misfolding, unfolding and/or aggregation have been implicated as a probable cause for at least a fraction of the sporadic cases of ALS. Hence, it is essential to not only gain detailed insight into the molecular behaviour of the FALS mutant molecules but also the wild type under different conditions.
Knowledge of protein structures and how they are affected by, for example mutations and/or loss of metals, is a prerequisite for developing suitable compounds and drugs which may inhibit the disease‐causing character. X‐ray crystallography provides the most detailed picture of the atomic arrangement within a protein molecule. Through ours and others' efforts, there is clear evidence that 1) mutations impair the global and/or local stability of SOD1 dimer molecules; and 2) that metal loss in SOD1 results in protein aggregation. We are looking to identify potential ‘drug binding pockets’ using the catalogue of structures we have determined; we are also searching for compounds with corrective properties. To this end we have continued to improve the resolution of the structures of the wild‐type and FALS mutant molecules. Thus, for example, we have obtained the structures of recombinant wild‐type SOD1 to 0.8 Å1 (known as ultra‐high resolution where even hydrogen atoms can be seen) and A4V to 1.15 Å (known as atomic resolution), to examine the structures of the wild‐type enzyme and its A4V mutant. At such resolutions, it is possible to observe protein conformational substates that are occupied for only a small fraction of time. It is also possible to distinguish between regions in the molecule that adopt conformations along a continuous range, or that jump between two or more distinct conformations. Such information would contribute towards identifying ‘binding pockets’ and aid the search for compounds with corrective properties. These and early results for ligand/compound binding will be presented.
1There are currently 36,531 structures of protein molecules and their derivatives in the protein data bank. Of these only 16 structures are known to a resolution at or better than this resolution.