In yesterday’s Nature, Kun Ping Lu at Beth Israel Deaconess Medical Center, Boston, together with colleagues in the U.S., Taiwan, and China, reports that isomerization of a proline in the C-terminal end of amyloid-β precursor protein (AβPP) reduces production of amyloid-β (Aβ). The conformational change is catalyzed by the prolyl isomerase Pin1, which may also prevent neurofibrillary tangles by isomerizing a proline residue in the microtubule binding protein tau (see ARF related news story). The actions of Pin1 on both AβPP and tau suggest that the prolyl isomerase might play a central role in AD pathology.

Proline, the sole α-imino acid in nature’s protein repertoire, is the only peptide residue that can flip-flop between two isomeric conformations, cis and trans. Such isomerizations have a profound effect on the shape of the protein backbone and in the late 1990s were predicted to be the basis for Pin1’s essential role in mitosis. Pin1 has since been linked to isomerization of a number of different proteins, including tau; transcription factors c-jun, NF-κB, and β-catenin; and the cell cycle protein cyclin D1. It has also been shown to bind to a plethora of other proteins, including, most recently, the neuronal harbinger of death BIM(EL), which triggers apoptosis. Whether Pin1 actually catalyzes proline isomerization in these binding partners remains to be determined.

In fact, proving that Pin1 causes proline isomerization has been a technical nightmare. Though miniscule amounts of the enzyme can elicit conformational changes in proteins such as Cdc25 (see Stukenberg and Kirschner, 2001), no one has conclusively demonstrated that these are caused by a cis/trans isomerization. But in collaboration with NMR spectroscopist Linda Nicholson at Cornell University, Lu and colleagues have now been able to visualize that flip-flop in full-length AβPP. Using ROESY (rotational frame Overhauser effect spectroscopy) NMR, they demonstrate the presence of both cis and trans isomers of proline 669, and transitions between the two isomers catalyzed by small amounts of Pin1.

Test tube experiments are one thing, but does this isomerization have physiological significance? Last year, Japanese researchers led by Takafumi Uchida at Tohoku University, showed that Pin1 binds to a threonine 668-proline 669 fragment in C99, the truncated form of AβPP that results from β-secretase cleavage of the full-length protein (see Akiyama et al., 2005). This was no surprise given that phosphorylated serine/threonine-proline is a well-known binding motif for the isomerase. Now, joint first authors Lucia Pastorino, Anyang Sun, Pei-Jung Lu, and their colleagues report that the isomerase also binds to full-length AβPP. They also tested how Pin1 affects processing of the full-length precursor protein in cell culture and in mice expressing the human AβPP carrying the Swedish double mutation (Tg2576).

The authors found that in Chinese hamster ovary (CHO) and H4 neuroglioma cells, AβPP and Pin1 colocalize and coimmunoprecipitate. They also found that overexpression of Pin1 reduces the amount of Aβ secreted from the CHO cells. In contrast, when the authors knocked out Pin1 in a breast cancer cell line, the cells produced about threefold less sAPPα, the soluble N-terminal fragment released by α-secretase. Taken together, the data seems to suggest that Pin1 reduces formation of Aβ by increasing the non-amyloidogenic processing of AβPP mediated by α-secretase.

In contrast, Uchida’s group had found that Pin1 increased production of Aβ from C99 when both were expressed in murine embryonic fibroblasts. Because C99 is poorly cleaved by α-secretase, this experiment might miss any effect Pin1 exerts on α-secretase-mediated AβPP processing. Nevertheless, the different effects seen on full-length versus β-secretase cleaved precursor suggest that the role of Pin1 in AβPP processing may be complex.

Pastorino and colleagues used Pin1-/- mice generated by Uchida’s lab to examine the relationship between the isomerase and AβPP processing in vivo. The Japanese group had found that soluble and insoluble Aβ40/42 were lower in Pin1-negative mice than in wild-type, though they did not report at what age they tested the animals. Now, Pastorino and colleagues report that while young (i.e., 2-6 months old), Pin1-negative mice had about the same amount of Aβ in the brain as wild-type littermates; in older animals (15 months), the absence of the isomerase caused about a 30 percent increase in the amount of insoluble Aβ42 (levels of soluble Aβ and insoluble Aβ40 stayed normal). It should be noted that this increase is relatively mild compared to some mouse models of AD. Tg2576 animals, for example, produce over 10-fold more Aβ42 than do wild-type mice and have abundant Aβ plaques. Tg2576 and other animal models also have well-documented learning and memory problems, so it will be interesting to see how the Pin1-/- animals fare in behavioral studies, too.

To test if the mild increase in insoluble Aβ in the Pin1-negative mice is due to some nonspecific, age-related phenomenon, the authors examined what happens to Tg2576 mice when Pin1 is absent. They found that by 6 months, Tg2576/Pin1-/- mice produced about 50 percent more insoluble Aβ42 than did age-matched transgenic littermates, suggesting that Pin1 may have a direct effect on Aβ production in vivo.

How could isomerization of proline 669 in the C-terminal, intracellular end of AβPP affect processing by α- and/or β-secretase on the extracellular side of the membrane or even by γ-secretase in the transmembrane domain? “That’s a very interesting question. We don’t know for sure, but we think it may be related to proteins that control trafficking of AβPP by binding to its C-terminal end,” said Lu.

Fe65, for example, which suppresses Aβ production, cannot bind to AβPP if threonine 668 is phosphorylated (see Ando et al., 2001), but dephosphorylation at that position requires that proline 669 be in the trans form. The natural, albeit slow, isomerization of the proline residue would forever protect a small portion of the total AβPP from dephosphorylation at threonine 668, increasing the chances for β- and subsequent γ-cleavage. “We believe that the role of Pin1 is to reset any cis-proline 669 to trans, so that the threonine can be dephosphorylated,” suggested Lu

This theory fits in with other observations. Li-Huei Tsai’s group at Harvard University has shown that phosphorylation of threonine 668 leads to increased production of Aβ in cell lines and that more of the phosphorylated form of the amino acid is found in tissue samples from AD brain (see Lee et al., 2003). Perhaps the big question is: How does threonine 668 get phosphorylated to begin with? It turns out that more of the amino acid is bound to phosphate in mitotic cells. This is curious, given Pin1’s role in mitosis and in light of a hypothesis suggesting that foiled attempts at cell cycle re-entry may be a trigger for neurodegeneration (see related ARF related Live Discussion). Stress may also lead to phosphorylation of threonine 668, suggested Lu.

The actions of Pin1 are likely to be multifaceted. It seems to function in the proliferation of breast cancer cells, and apoptosis mediated by mitochondria (see Becker and Bonni, 2006), yet without it, neurons are prone to neurodegeneration. As with most research, knowledge of Pin1 will benefit from distribution of the Pin1-/- mice and further exploration by other labs in the field.—Tom Fagan

Comments

Make a Comment

To make a comment you must login or register.

Comments on News and Primary Papers

  1. This paper elegantly demonstrates that the peptidyl prolyl isomerase Pin1 regulates amyloid-β production via binding to Thr668Pro of amyloid precursor protein (APP). Pin 1 binds to phosphorylated Ser/Thr-Pro motifs in proteins to regulate numerous functions via conversion of prolyl residues from cis to trans conformation and vice versa (Butterfield et al., 2006, submitted). For example, by regulating the conformation of key prolyl residues in the protein phosphatase, PP2A, Pin1 regulates, in part, the dephosphorylation of tau.

    Recently, our laboratory showed by redox proteomics that Pin1 was selectively oxidatively modified and dysfunctional in brain from subjects with Alzheimer disease and amnestic mild cognitive impairment, or MCI (Sultana et al., 2005; Butterfield et al., 2006). We also showed that purified Pin1, when subjected to oxidative damage, became dysfunctional, suggesting that in AD and MCI brain it is the oxidative modification of Pin1 that may lead to its loss of function. Lu and coworkers demonstrate that Pin1 knockout mice have greater deposition of Aβ42 in brain than do wild-type mice. Thus, it is interesting to speculate that the excess deposition of Aβ42 in AD brain may be due in part to oxidatively modified and dysfunctional Pin1. The diminished activity of Pin1 in AD brain also, via its effects on PP2A, could affect dephosphorylation of tau, contributing to the hyperphosphorylation of this key cytoskeletal protein with the downstream consequence of neurodegeneration. Consequently, I agree with Lu and colleagues that Pin1 may be an attractive therapeutic target for AD, and would add for MCI, as well.

    References:

    . Pin1 in Alzheimer's disease. J Neurochem. 2006 Sep;98(6):1697-706. PubMed.

    . Redox proteomics identification of oxidatively modified hippocampal proteins in mild cognitive impairment: insights into the development of Alzheimer's disease. Neurobiol Dis. 2006 May;22(2):223-32. PubMed.

    . Oxidative modification and down-regulation of Pin1 in Alzheimer's disease hippocampus: A redox proteomics analysis. Neurobiol Aging. 2006 Jul;27(7):918-25. PubMed.

  2. Kun Ping Lu’s group and his collaborators have been at the fore of elucidating Pin1’s cellular roles, including, since discovering that tau is a Pin1 target protein, its involvement in neurodegeneration. They accumulated data that depletion of Pin1 in HeLa cells causes apoptosis in HeLa cells, that patterns of AD pathology correlate with regions of lower Pin1 expression in normal human brain, that Pin1 knockout mice suffer neurodegeneration, and that Pin1 can ameliorate p-tau pathology. On the basis of that, they have suggested that a fuller elucidation of Pin1’s involvement in neurodegeneration (and cancer) might lead to new therapeutic strategies.

    Our group has acquired data confirmatory of, and complementary to, that of Lu and his coworkers. We have observed Pin1 deficits in a range of frontotemporal dementias and in aging normal brain neurons and have suggested that this might be a susceptibility factor both in neurodegenerative disease (Thorpe et al., 2004) and in aging-related neurodegeneration (Hashemzadeh-Bonehi et al., 2006).

    In this latest work, Lu and colleagues suggest that deficits of Pin1 would also be deleterious to neurons in respect of Aβ secretion; it binds to p-Thr668 of APP and its overexpression reduces Aβ secretion in cell cultures, whilst knockdown, both in cells and mice, selectively increases secretion of the toxic amyloid species, Aβ42. A concern is that this data is contradictory to the work of others (Akiyama et al., 2005), which is not referred to in this present work. Akiyama et al. also used knockdown mice and several cell lines (different from those used by Lu et al.). I can only presume that differences in genetic background might account for the discrepant data between these two studies; although the source of the Pin1 KO mouse is the same for both groups, it appears that Lu's group maintain their colony in an inbred C57/S129 line, whereas Akiyama’s group maintain a C57/B6 strain. Clear differences have been observed in these strains’ behavioral phenotypes. Additionally, the mouse brain gene expression database shows higher hippocampal GSK3β expression in an S129-derived strain than in a C57/B6 strain. Such strain differences, especially local concentrations of upstream APP kinases, could influence APP processing. Indeed, the elucidation of these differences might add important new insights into the neurodegenerative process.

    While research showing involvements of just one specific protein in molecular neuropathological pathways do not confirm their centrality to a disease, other recent research evidence supports such a view: Pin1 promoter polymorphisms, which result in lowered protein expression, correlate with AD (Segat et al., 2005), and Pin1 is one of a handful of proteins susceptible to oxidation in MCI hippocampus, with the authors suggesting that this may be involved in the progression from MCI to AD (Butterfield et al., 2006). Thus, if the above concern regarding conflicting data is addressed, this new data from Lu’s group could put Pin1 protein potentially at the heart of the ameliorative influences that might slow or halt the key twin molecular neuropathological pathways leading to plaque and tangle formation and thence neuronal cell death in AD.

    For more information, see our research website.

    View all comments by Julian Thorpe
  3. One lesion, two pathologies: Can Pin1 disturbance cause plaques and tangles?

    Nearly 20 years have passed since the first efforts to link neurofibrillary pathology and amyloid pathology via dysfunction of some common regulatory step involving protein phosphorylation. At one time or another, PKC, ERK, GSK3, Cdk5, and protein phosphatases 1 and 2A have all been proposed to be players in the story. Among these, GSK3 and Cdk5 have been two of the most tantalizing, since each can act as both tau kinases and APP kinases. Pastorino and colleagues report in the current issue of Nature the discovery of a possible missing link in the form of prolyl isomerization by the isomerase Pin 1.

    Part of the consensus sequence for GSK3/Cdk5 phosphorylation of Thr668 in the APP cytoplasmic tail is the presence of a prolyl residue at position 669. Pastorino et al. propose that the phosphorylation state of Thr668 regulates the susceptibility of Pro669 to isomerization by Pin1 by a factor of 1,000-fold, and that the isomerization state of that proline is a key mechanism that controls sorting of APP into amyloidogenic versus nonamyloidogenic processing pathways. The link between Pin1 and Aβ was strengthened by studying Pin1-/- mice, which have increased levels of brain Aβ. When the Tg2576 mouse is crossed with a Pin 1-deficient mouse, there is an apparent increase in Aβ immunoreactivity in multivesicular bodies. Pin 1-/- mice have previously been shown to exhibit tauopathy and spontaneous neurodegeneration. Hence, the conclusion is drawn that Pin 1 hypofunction might plausibly lead to both tauopathy and amyloidosis.
    The role of the Thr668 phosphorylation state, while an attractive molecular mechanism for Pin 1 action, is not an essential feature of the Pastorino Pin 1 model.

    Direct analysis of the role of phosphorylation of Thr668 on Aβ generation has provided little evidence for an important link between the two. In cultured cells, Thr668Ala APP, a non-phosphorylatable mutant, is metabolized normally, and the stoichiometry of Aβ species (both 40 and 42) generated during processing of Thr668Ala APP is not obviously different from that of wild-type APP. The question of whether cell culture data can be extrapolated to CNS neurons raises its head here once again. A crucial experiment aimed at testing the role of Thr668 phosphorylation involves knocking Thr668Ala into the genome of mice and then measuring brain Aβ levels. These experiments are in progress: The results should go a long way toward determining whether phosphorylation of Thr668 has a major impact on brain Aβ generation.

  4. In this very interesting paper, Pastorino et al. demonstrate that Pin1 catalyzes the conformational change of the phosphorylated APP cytoplasmic tail. The Pin1-mediated shift in the Thr668-Pro motif from a cis to trans conformation results in selectivity towards α-secretase processing of APP. Overexpression of Pin1 decreases amyloid-β production by 30-40 percent and conversely, the knockout of Pin1 favors β-secretase processing and results in an increase of Aβ42 by 50 percent (which rivals the change seen in FAD mutations in the presenilins).

    How does Pin1 alter Aβ production? It is interesting that a structural change in the intracellular tail of APP alters its extracellular (or endosomal) cleavage. One attractive mechanism by which Pin1 alters APP processing is by affecting APP’s cytoplasmic binding partners. There are several known binding partners of APP that interact with the TPEE motif including X11, Disabled, and Fe65. Of these candidate proteins, only Fe65 is shown to be sensitive to the phosphorylation state of Thr668 (Ando et al., 2001).

    How would Pin1 affect Fe65 binding to APP? Fe65 has been shown to bind less avidly to APP phosphorylated at Thr668 (Ando et al., 2001; Kimberly et al., 2005). Since Pin1 promotes formation of the trans isomer of phospho-Thr668, then it likely diminishes the interaction of Fe65 with APP (although this was not formally demonstrated in the current manuscript). Since Pin1 and Fe65 may represent competing pathways for APP processing, one would expect Fe65 to have the opposite effect of Pin1 on APP. In support of this speculation, Fe65 overexpression results in a fourfold increase in Aβ production (Sabo et al., 1999). Thus Pin1 and Fe65 may represent competing pathways for the proteolytic processing of APP. Pin1 interaction diverts APP away from β-secretase, whereas Fe65 interaction promotes β-secretase activity. However, it is unlikely that this represents the entire story, as overexpression of Fe65 and its family members also results in a less robust twofold increase in the α-secretase-generated α-APPs (Sabo et al., 1999; Guenette et al., 1999).

    While these potential pathways may represent a mechanism for APP proteolytic regulation, it must be remembered that only a fraction of APP is phosphorylated at Thr668 in neurons and brain lysate, and could therefore be subject to Pin1-mediated regulation. Nevertheless, this important work identifies a new chapter in APP biology and may even open new avenues for therapeutics for Alzheimer disease.

    View all comments by W. Taylor Kimberly
  5. Pastorino and colleagues demonstrate an interesting direct role for Pin1 in APP processing both in vivo and in vitro. The tie-in to the differential regulation of the interaction based on cell cycle phase in dividing cells is also intriguing. It leads one to wonder whether Pin1 might also have a potential function to protect neurons that may be pushed to entering the cell cycle by disease processes. Another exciting area to be followed up in subsequent studies, which was also mentioned in an above comment, is how the Pin1 interaction and conformational change in the APP intracellular domain which results from it may influence interactions with other C-terminal binding proteins. The functional consequences of the potentially altered interactions on signaling pathways in neurons may yield interesting information on APP’s normal role(s) and how these role(s) may be disrupted by the disease.

    View all comments by Donna McPhie

References

News Citations

  1. Stave off Your Tangles...with Pin1?

Webinar Citations

  1. Cell Cycle Hypothesis Pedaling into Mainstream Acceptance? Results in Fly, Mouse Models Warrant a Second Look

Paper Citations

  1. . Pin1 acts catalytically to promote a conformational change in Cdc25. Mol Cell. 2001 May;7(5):1071-83. PubMed.
  2. . Pin1 promotes production of Alzheimer's amyloid beta from beta-cleaved amyloid precursor protein. Biochem Biophys Res Commun. 2005 Oct 21;336(2):521-9. PubMed.
  3. . Phosphorylation-dependent regulation of the interaction of amyloid precursor protein with Fe65 affects the production of beta-amyloid. J Biol Chem. 2001 Oct 26;276(43):40353-61. PubMed.
  4. . APP processing is regulated by cytoplasmic phosphorylation. J Cell Biol. 2003 Oct 13;163(1):83-95. PubMed.
  5. . Pin1 mediates neural-specific activation of the mitochondrial apoptotic machinery. Neuron. 2006 Mar 2;49(5):655-62. PubMed.

Further Reading

No Available Further Reading

Primary Papers

  1. . The prolyl isomerase Pin1 regulates amyloid precursor protein processing and amyloid-beta production. Nature. 2006 Mar 23;440(7083):528-34. PubMed.