Like death and taxes, there is no escaping DNA damage—it is an unpleasant fact of life. DNA breaks and other mishaps accumulate in neurons with age and in Alzheimer disease, and wreak havoc on the expression of many genes (see ARF related news story on Lu et al., 2005). Fortunately, cells have a way to fight back by turning on DNA repair pathways. A paper published this week in the online edition of PNAS suggests that the amyloid precursor protein (APP) and its partner, the neuronal adaptor protein Fe65, participate in one such pathway that allows cells to repair double-stranded DNA breaks. The work, from Tommaso Russo and colleagues at the Università di Napoli Federico II in Naples, Italy, makes a novel link between APP signaling and DNA repair. The increased DNA damage seen in AD is most often attributed to higher levels of oxidative damage, but the new data suggest defects in repair might also contribute.

Fe65 is a neuronal adaptor protein that binds the short cytoplasmic domain of APP. γ-secretase processing of APP liberates the intracellular domain (AICD), and AICD-Fe65 complexes move to the nucleus. Once there, the proteins meet up with additional partners, some of which are known to participate in transcriptional regulation or DNA repair. The Russo group previously implicated Fe65 in DNA repair when they reported that Fe65 knockout mice are sensitive to DNA damaging agents and radiation (Minopoli et al., 2007). The new paper shows that Fe65 promotes DNA repair via association with the histone acetyltransferase Tip60 at sites of DNA strand breaks. The data also suggest that APP association is required for the chromatin binding function of Fe65.

First author Maria Stante and colleagues build the case for Fe65 in Tip60-mediated repair using embryonic fibroblasts from Fe65 knockout mice, which show defects in DNA repair. The Tip60 protein, along with its partner TRRAP, makes up a major part of the NuA4 histone acetylase complex, which is recruited to strand breaks and participates in their mending. In the knockout cells, repair was rescued by expressing the wild-type Fe65, but not a deletion mutant that does not associate with Tip60.

To take a closer look, the authors used NIH-3T3 cells engineered to inducibly express a restriction enzyme that makes a single double strand DNA break in an introduced sequence, allowing analysis of repair activity at that site. Successful repair results in restoration of a complete coding sequence for green fluorescent protein, which allows the measure of DNA repair efficiency by flow cytometry. When Fe65 was knocked down in these cells using siRNA, the investigators found that Tip60/TRRAP was no longer recruited to the site of the double strand break, Tip60-dependent histone acetylation was decreased, and the efficiency of DNA repair declined by about a third.

Fe65 and Tip60 also associate with the APP cytoplasmic domain or the AICD, and the investigators next looked at whether that association was important for DNA repair. Two pieces of evidence suggest yes: DNA repair in Fe65-lacking cells could be reconstituted by expression of a wild-type Fe65 protein, but not a point mutant (C655F) that fails to associate with APP. The mutant not only failed to rescue DNA repair, but it also acted as a dominant negative in the engineered cells by blocking the actions of wild-type Fe65. In addition, siRNA knockdown of APP and its relative APLP2 in the same cells resulted in less recruitment of Tip60/TRRAP to the site of the break, and decreased histone H4 acetylation and less DNA repair.

From these results, the investigators conclude that Fe65 plays a significant role in the recruitment of Tip60/TRRAP to the DNA breaks and that this function depends on its interaction with APP. However, when they examined Fe65 interaction with chromatin, they found that although Fe65 was associated with chromatin at DNA breaks, that binding was not induced by DNA damage. Whole-chromatin immunoprecipitation experiments indicated that Fe65 was constitutively bound across the genome, and that its presence was only slightly increased after treatment of cells with a DNA damaging agent etoposide. In agreement with a role for APP in the chromatin localization, there was less Fe65 associated with chromatin in the APP/APLP knockdown cells. Interestingly, the C655F mutant did not associate with either intact or damaged chromatin, which could explain its inability to rescue the repair defect in Fe65 knockout cells.

The new data open up many questions. Fe65 is composed of three protein-protein interaction domains, and has several known binding partners, and likely more still to be discovered. The C655F mutant abolishes both APP binding and chromatin association, and it needs to be clarified if the two phenomena are related or independent, and if they occur in neuronal cells. The APP/APLP2 knockdown supports a role for APP in DNA repair, but more work will be needed to figure out how Fe65 gets from APP in the membrane to the nucleus, and whether this requires liberation of the AICD through processing by the γ-secretase, or dissociation of Fe65 from intact APP. Though these questions remain to be clarified, the authors write, “Our results suggest that the involvement the Fe65-APP complex in the response of cells to DNA damage and in the DNA repair machinery should be taken into account as a possible mechanism contributing to neuronal dysfunction observed in AD pathology.”—Pat McCaffrey


  1. This report from Tommaso Russo’s group represents a small but growing number of studies that are focused not on Aβ-ology but on the function of APP. Continuing from their previous observation that Fe65 protects cells from DNA damage, this study provides strong evidence that APP is required for the Fe65 mediated DNA repair. Knockdown of APP/APLP2 resulted in impaired recruitment of Fe65-Tip60-TRRAP complex to the DNA break sites in the nucleus and reduced the repair efficiency. The authors mention that this unexpected role of APP should be taken into account as a possible mechanism contributing to neuronal dysfunction in AD.

    These are important observations and would further stimulate experiments along the lines of APP function. One important issue that this study unfortunately did not address is whether or not the presence of AICD in the nucleus is essential for the observed effects of Fe65. Now that we have several excellent γ-secretase inhibitors available, this one should be an easy experiment to do. Although the Cao-Sudhof model suggests that the cleavage of APP is not required and that mere association with APP-CTF at the cell surface is enough to “activate” Fe65 (open conformation), there is really no solid data to support such a notion. This scenario might be plausible in non-neuronal cells where the distance from cell surface to the nucleus is not insurmountable but less so in neuronal cells, where both APP and Fe65 are known to accumulate in the synapses or growth cones. In any case, these are interesting findings on the novel aspect of APP function, and one can be sure the Russo lab is examining the effects of γ-secretase inhibitors and APPswe or D664A mutation in their assay.

Make a Comment

To make a comment you must login or register.


News Citations

  1. After 40, DNA Damage Accrues in Genes, Hampering Expression

Paper Citations

  1. . Essential roles for Fe65, Alzheimer amyloid precursor-binding protein, in the cellular response to DNA damage. J Biol Chem. 2007 Jan 12;282(2):831-5. PubMed.

Further Reading

Primary Papers

  1. . Fe65 is required for Tip60-directed histone H4 acetylation at DNA strand breaks. Proc Natl Acad Sci U S A. 2009 Mar 31;106(13):5093-8. PubMed.