Index | Round
Tables | Selected
Abstracts | News
Presenilin Binding Proteins Abound
by Kenneth S. Kosik
July 22, 1998 Because presenilin was discovered in positional cloning experiments involving large familial Alzheimer's disease kindreds, there were no previous data concerning its biological function. As a result many groups have used a variety of techniques to garner clues regarding the function of this novel gene, which bears a key relationship to Alzheimer's disease.
Since the cloning of presenilin, one of the important functional clues to emerge was the critical role that presenilin plays in development. This insight derived from presenilin 1 "knock-out" animals. The next step in further understanding the function of presenilin is the identification of the verification of binding partners. The lengthy procedures involved in discovering these interactors and more importantly in authenticating them accounts for the two-year interval between the discovery of presenilin and the reporting of several interactors. Before this meeting the putative interactive PS1 proteins were beta-catenin, delta-catenin, actin-binding protein, and the amyloid precursor protein. All of these interactors have proponents and nay-sayers. Described in the Molecular Pathology II session, the new interactors were CLIP-170 (with PS1) from the Robakis group, and calsenilin (with PS1 and PS2) from the Wasco group. Other presentations not in this session described an interaction between Notch and PS1 by Alison Goate's group and an interaction between sorcin and PS2 by Tae-Wan Kim in the Tanzi group. Together these proteins constitute a thicket of new clues with few if any clear directions concerning which of the interactions, if any, will play a role in the disease pathogenesis. All of them are worthy of further pursuit.
Tezapsidis N et al. (abstract 1257) reported that one PS1 mutation (E280A) showed an increased binding to CLIP170. This protein, first discovered by Tomas Kreis in the early nineties, has 1390 amino acids with a large coiled coil alpha helical domain in the central part of the protein. 57 amino acids at the amino terminus bind to vesicles and an acidic carboxy terminus binds to microtubules. Thus CLIP170 lies at a most interesting interface between endocytosis and the microtubules. In addition to the in vitro interaction, the authors described co-localization of a myc-CLIP 170 transfectant and PS1 in 5YSY cells. This cell line is a human neuroblastoma line.
Choi E et al. (abstract 1258) described how calsenilin was discovered in a two-hybrid screen using the carboxy terminus of PS2 as the bait. Confirmation of the interaction was by co-immunoprecipitation both ways in transfected COS cells. The authors were also able to co-immunoprecipitate calsenilin and PS1. The high degree of homology between the PS1 and PS2 amino termini probably accounts for the ability of both proteins to form a complex with calsenilin. In contrast the lack of homology between the PS1 and PS2 loop regions has made the identification of common interactors with this region more problematic. When transfected, calsenilin shows a cytoplasmic distribution, but following co-transfection with PS2 the complex gets recruited to the endoplasmic reticulum and the Golgi. Although a novel gene, calsenilin is a member of the recoverin family. One characteristic of the family is the presence of EF hand motifs; there are four in calsenilin. The authors demonstrated that calsenilin binds calcium by showing a mobility shift with and without EGTA and calcium 45 overlay techniques. Interestingly, the protein is expressed specifically in the brain by Nothern blots and in both neurons and glia by in situ hybridization.
Together these new interactors all offer the potential for new insights into presenilin biology and possibility.
Comment from Nikolaos Tezapsidis, PhD,
Mount Sinai School of Medicine, New York
Dear Dr. Kosik,
It was an honor that our work, presented at the 6th International
Conference on AD, was described in your briefing about
presenilin-binding proteins. Similarly to your team and the
identification of the catenins as possible modulators of PS-1 function,
we have pursued this avenue of research in an attempt to identify, on
one hand, a normal function for the presenilins and on the other,
mechanisms by which this is perturbed by two of the chromosome 14
I would like to refer to a mistake in the description of the
functional domains of CLIP-170 and take the opportunity to expand a
little. CLIP-170, which is essentially identical to restin, binds to
vesicular structures through its acidic C-terminal domain and to the
microtubules (MTs) through its basic N-terminal domain (from work by
Thomas Kreis et al). This would be consistent with a model where
CLIP-170 acts as the bridging protein of vesicle attachment to the MTs
and where PS-1 acts as the CLIP-170 anchor/ receptor on those vesicles.
The binding occurs between the large cytoplasmic loop of PS-1 and the
C-terminal domain of CLIP-170. The E280A and L286V mutations, both cause
an increase in binding, as determined by in vitro binding assays and the
corresponding Kds in the presence of Ca (required for binding). Thus
the increased binding due to the mutations may be associated with an
exacerbated normal function, consistent with a gain of function
scenario. An aspect however that cannot be adequately explained with the
available data is how do other mutations scattered throughout the length
of PS-1 precipitate in the same phenotype. It may be that PS-1/2 are
multifunctional proteins and that the mutations cause dysfunctions, all
of which are associated with early-onset AD. For example, more than 700
mutations have been identified in the cystic fibrosis (CF) transmembrane
conductance regulator that are linked to CF through different
mechanisms. Further, a mutation on SCAP (SREBP cleavage activating
enzyme) at a site distant from its binding domain to SREBPs (sterol
regulatory element binding proteins) activates cleavage of SREBP at site
1, possibly through a tighter binding between SCAP and SREBP.
We have shown by immuno-confocal microscopy that in MycCLIP-170
transfected SY5Y cells that CLIP-170 (detected by anti-CLIP monoclonal
antibodies) co-localizes to endogenous PS-1 in subcellular structures
that resemble ER-Golgi. This is currently being confirmed with anti-Myc
antibodies in Myc-CLIP-170-transfected SY5Y cells undergoing neuronal
differentiation (endogenous c-Myc drops).
An interesting aspect of our studies is the prediction that PS-1
will also bind to restin, an intermediate filament associated protein.
Therefore, vesicular trafficking and consequently APP processing may
depend on the nature and duration of the molecular interactions between
the presenilins and cytoskeleton-associated proteins.
Thanks for your attention.