It has become part of the standard lore on Alzheimer's that a series of enzymes-dubbed α-, β-, and γ-secretase-successively chops away at the APP cell-surface protein, and that one product of γ-secretase is the infamous Aβ-peptide. Yet why does this orderly degradation of APP occur? The mystery only deepened last year, when combined knockouts of APP and its homologues proved lethal, suggesting the precursor protein must be doing something important (Heber et al., 2000). Just what that something could be is the subject of a July 6 paper in Science by Xinwei Cao and Thomas Sudhof of the Howard Hughes Medical Institute at the University of Texas Southwestern Medical Center in Dallas. The authors show that the less-studied product of γ-secretase cleavage-an intracellular fragment comprising the inner half of APP's transmembrane region plus its cytoplasmic tail-likely is involved in transcriptional activation.
"Our study provides a physiological reason for the processing of APP. That has implications for AD, because once you know why it is cut you can also look at why it is sometimes cut more and sometimes less," said Sudhof. Gene transcription is tightly regulated by extracellular signals, yet in sporadic AD there are no mutations that change any of the known APP processing components. Possibly, sporadic AD develops because the pathway that produces Aβ is dysregulated for many years, Sudhof argues.
The study also has implications for therapeutic attempts to block APP processing. "For example, γ-secretase inhibition clearly is going to have effects not only on Aβ production but also on nuclear signaling. That could be good or bad," says Sudhof.
In their search for a function for APP processing, the researchers pursued a trail laid down by notch, another cell-surface receptor that is a substrate for γ-secretase. Following cleavage, its cytoplasmic tail moves to the nucleus and regulates transcription there (Schroeter et al., 1998, and Struhl et al., 1998).
Cao created fusion proteins of APP with the DNA-binding domains of a yeast and a bacterial transcription factor. These chimeric proteins allowed Cao to measure if the APP cytoplasmic tail activates transcription of reporter genes in transfected cell lines. The scientists found that the APP fusion protein stimulated transcription only weakly on its own, but transcription jumped more than 2,000-fold when the tail was bound to the multi-domain adaptor protein Fe65. Fe65 is not specific to APP, but may have a role in AD, much like Ras has many physiological binding partners but plays a specific role in cancer, Sudhof says.
Moreover, Cao and Sudhof report that the APP cytoplasmic tail and Fe65 form a stable trimeric complex with Tip60, a histone acetyltransferase. (These enzymes control access of transcriptional enzymes to genes by modifying the packing density of the histone proteins wrapped around the DNA). Tip60 itself is part of a nuclear protein complex that acts as a general transcription factor.
Cao and Sudhof also report that all of Fe65's major domains are required for transcriptional activation and that any mutation disrupting Fe65 binding to the APP fragment or to Tip60 abolishes this function. Taken together, the authors suggest that the complex of APP's cytoplasmic tail with Fe65 and Tip60 directly acts in transcription. They add that this similarity between APP and Notch supports the idea that presenilin-dependent proteolysis functions as a general biological mechanism of transcriptional regulation (Brown et al., 2000).
Although the in vitro experiments in the present study need to be repeated with endogenous proteins, the work raises two important questions: Which proteins regulate APP proteolysis, and which genes does the cytoplasmic APP tail turn on or off? Stay tuned.—Gabrielle Strobel