Our laboratory is primarily interested in the mechanisms underlying neuronal injury and degeneration in neurodegenerative disorders such as Alzheimer’s disease. Recent evidence from a variety of sources has suggested that these diseases arise from the accumulation of misfolded proteins (typically in β-pleated sheet conformations). The specific neurotoxic misfolded protein that accumulates and the location of these deposits differ between different neurodegenerative disorders. Nevertheless, the concept that many neurodegenerative diseases are disorders of protein folding seems well established.
Our current work focuses on the role of the presenilin proteins in the generation of the amyloid β-peptide (Aβ peptide) and in the pathogenesis of Alzheimer's disease, inclusion body myositis, and cerebral amyloid angiopathy, which are all characterized by pathological deposits of extracellular Aβ peptide.
The overall sequence of proteolytic events that lead to the generation of Aβ has been well worked out. Thus, the amyloid precursor protein (APP) is sequentially cleaved, first by a membrane-bound aspartyl protease termed β-secretase or BACE. The membrane-bound C-terminal stub of APP is then cleaved within its transmembrane domain by the presenilin complex to generate the Aβ peptide fragment and a soluble cytoplasmic C-terminal derivative termed amyloid intracellular domain (AICD). The latter protein seems to be involved in signal transduction and the transcriptional upregulation of several genes including neprilysin, an enzyme involved in degradation of Aβ. We are focusing on the role of the presenilin complexes in this cascade.
We initially cloned the presenilin proteins PS1 and PS2 as the site of mutations causing autosomal dominant familial Alzheimer's disease (FAD). We showed that the presenilins form high molecular weight complexes with nicastrin, TMP21, δ-catenin/NPRAP, aph-1, and pen-2. We and others have shown that the presenilin complexes support an aspartyl protease activity that cleaves several Type 1 proteins (including APP, Notch, p75, Irep1, Delta, Cadherin) within their TM (transmembrane) domains. These cleavages occur at at least three distinct sites (γ,ζ,ε) within the TM domains of these substrates. As described above, cleavage of APP C-terminal stubs at the γ-site generates the neurotoxic Aβ peptide in AD. Cleavage of APP, Notch, p57, and so forth, at the ζ- and ε-sites generates the intracellular C-terminal signaling moieties (e.g., Notch Intracellular Domain or NICD and AICD).
We and others have shown that nicastrin (likely via its N-terminal ectodomain, which contains a conserved DYIGS motif resembling the transferrin receptor) selects substrates by binding to the N-terminal ectodomains of the substrate Type-1 membrane proteins that have already been cleaved by enzymes such as β-secretase or members of the TACE and ADAM protease families. The function of pen-2 is unknown. We have shown that during the assembly of presenilin complexes, aph-1 forms an initial kDa scaffolding complex of about 123 kDa, together with nicastrin. This initial scaffolding complex is then supplemented by the addition of PS1/PS2 and pen-2 to generate an abundant, but inactive, kDa complex of about 440 kDa. The final, functionally active complex has a molecular weight of about 650 kDa. This maturation process is marked by glycosylation of nicastrin and by the autocatalytic endoproteolytic cleavage of PS1/PS2 within the TM6-TM7 hydrophilic loop to generate N- and C-terminal fragments (NTF and CTF). We have shown that about half of PS1 complexes also contain the p23 cargo protein TMP21. TMP21 acts as a regulatory subunit that specifically modulates the activity of the γ-site cleavage (but not ζ- and ε-site cleavage).
There are no experimentally proven models of the three-dimensional structure of presenilin complexes. Two conflicting, low-resolution structures have been generated by negative stain electron microscopy for PS1 complexes isolated in digitonin (which does not support enzymatic function). The currently published hypothetical TM domain topologies of PS1/PS2 and aph-1 are largely based on hydrophobicity plots and canonical, vertical α-helical TM domains that completely span the membrane. However, the C-termini of PS1/PS2 are particularly hydrophobic, and it is possible that some of the putative TM domains in the C-terminus might create long, kinked, imperfect α-helices that run oblique to the normal axis of the membrane, as have been observed in aquaporin 1 and RyR1. Nevertheless, most models of PS1/PS2 favor nine TM domains, although some favor seven or eight TM domains.
Our studies will generate an initial topological model depicting close interactions between TM domains of the presenilins, nicastrin, aph-1, and pen-2 in the presenilin complex. When combined with other biophysical studies, this model will facilitate the discovery of the three-dimensional structure of the presenilin complexes. Knowledge of these interactions and how they are affected by disease-causing mutations (PS1 mutations in Alzheimer's disease, for example) will be key to the rational design of specific inhibitors of the presenilin enzymes as therapeutics for human diseases. Such inhibitors will need to be designed to selectively block only the deleterious generation of Aβ but permit cleavage of substrates such as Notch, p75, and so forth. This objective is difficult to achieve by high throughput screening of compound libraries but would be dramatically facilitated by knowledge of presenilin complex structure.
Last updated June 2010
