of the Max - MDC

Proteomics and Molecular
Mechanisms of
Neurodegenerative Disorders
Erich E. Wanker
Cells are made of macromolecules and metabolites that interact
to form highly complex networks. Protein-protein, RNAprotein
and DNA-protein interactions are critical for the formation
of molecular machines, and contribute to global transcriptional
networks, positive and negative circuits and other regulatory
mechanisms. Macromolecular networks appear to govern
all fundamental cellular processes, and perturbations of these
networks obviously underlie many human diseases.
The main objective of our work is to understand the cell’s
functional organization and to link individual proteins to
signalling cascades and disease processes. For the systematic
identification of protein-protein-interactions (PPIs)
and the analysis of gene regulatory networks, we have
established an automated yeast two-hybrid (Y2H) system
and other high-throughput technologies. We have applied
these methods to create comprehensive, static maps of the
human proteome. Recently, we established an online database
termed UniHI (Unified Human Interactome) that contains
the 3300 Y2H interactions we identified as well as
other computationally and experimentally derived interaction
datasets. It includes more than 220,000 distinct interactions
between over 20,000 unique human proteins and is
available at http://www.mdc-berlin.de/unihi. Closely linked
to our work on the identification and functional characterization
of protein-protein interactions is our research to elucidate
the pathomechanisms of Huntington’s disease (HD)
and Alzheimer’s disease (AD). Recently, a drug development
pipeline for the identification and validation of small molecules
that modulate protein misfolding and aggregation in
HD and AD was established. Hit compounds have been
derived from several screenings that employed our membrane
filter technology and other assays we recently established.
These compounds are currently investigated pharmacologically
and tested for their activity in cell-based
assays and different transgenic disease models.
From interaction networks to disease modifiers
Using Y2H screens, we previously generated a protein-protein
interaction network for Huntington’s disease that contains
188 mostly novel interactions between 86 different
proteins (Goehler et al., 2004). In the last two years, the
identified protein-protein interactions (PPIs) were systematically
validated with independent pull-down assays, coimmunoprecipitations
or co-localisation studies. We developed
a novel membrane-based co-immunoprecipitation
assay and were able to demonstrate, e.g., that huntingtin
and GASP2 (G protein-coupled receptor associated sorting
protein 2) form a complex in mammalian cells. We could
show that the two proteins co-localise under physiological
conditions in SH-SY5Y cells, which indicates that huntingtin
and GASP2 may interact in neurons. As the GASP protein
family plays a role in G protein-coupled receptor sorting,
our data suggest that huntingtin might influence receptor
trafficking via the interaction with GASP2 (Horn et al.,
2006).
In collaboration with Prof. Hitoshi Okazawa from Tokyo
Medical and Dental University, Japan, a proteomics analysis
of soluble nuclear proteins in HD cell model systems was
performed. This analysis led to the identification of the proteins
HMGB1 and 2 that both bind to mutant huntingtin and
influence its toxicity as well as aggregation behaviour in
vitro and in vivo. Immunohistochemistry showed that the
proteins HMGB1/2 are reduced in the nuclear region outside
of the huntingtin inclusion bodies in affected neurons.
Compensatory expression of the proteins ameliorated
polyglutamine (polyQ)-induced pathology in primary neurons
and in Drosophila polyQ models. Furthermore,
HMGM1/2 repressed genotoxic stress signals induced by
mutant huntingtin expression in neurons. We conclude that
HMGB proteins may be critical regulators of polyQ disease
pathology and could be targets for further therapy development
(Qi et al., 2007).
Towards a human interactome project
By systematic interaction mating with our automated Y2H
technology, we have previously identified about 3,300
mostly novel human PPIs among 1705 proteins (Stelzl et al.,
2005). In the last two years, we further expanded our
efforts to identify human PPIs on a large scale. Using about
200 bait proteins of the MAPK as well as Akt signalling pathways,
∼2000 novel PPIs were identified with repeated Y2H
screens. A significant fraction of these interactions was validated
with functional assays. Among other experiments, we
performed ELISAs in order to monitor ERK and Akt phosphorylation
for about 100 selected proteins. This allowed the
identification of ∼20 proteins that modulate MAPK and/or
Akt signalling in mammalian cells. These proteins are currently
further validated with cell-based assays. Through
164 Function and Dysfunction of the Nervous System