The findings by Li et al. on the regulation of APP processing by sumoylation indeed bring a new perspective to the field of AD research. Although the authors documented a lack of direct sumoylation of APP and BACE, they have demonstrated a clear increase of α-NTF (derived from α-secretase cleavage of APP) and a reduction of β-NTF (derived from β-secretase cleavage of APP) when SUMO-3 is overexpressed. When mutant SUMO-3 is overexpressed, all of α-NTF, β-NTF and Aβ are increased. Importantly, overexpressing either wildtype or mutant SUMO-3 leads to increased steady-state levels of APP.

Based on these observations, it will be interesting to search for any effect of sumoylation on α-secretase, which is not discussed in the paper. I would predict an increase of α-secretase activity by sumoylation. Overexpression of wildtype SUMO-3 could lead to enhanced α-secretase, resulting in increased α-NTF levels, reduced β-NTF levels (and subsequently reduced Aβ levels.) On the contrary, overexpression of mutant forms of SUMO-3 may not affect α-secretase activity, and increased levels of...  Read more

The findings by Li et al. on the regulation of APP processing by sumoylation indeed bring a new perspective to the field of AD research. Although the authors documented a lack of direct sumoylation of APP and BACE, they have demonstrated a clear increase of α-NTF (derived from α-secretase cleavage of APP) and a reduction of β-NTF (derived from β-secretase cleavage of APP) when SUMO-3 is overexpressed. When mutant SUMO-3 is overexpressed, all of α-NTF, β-NTF and Aβ are increased. Importantly, overexpressing either wildtype or mutant SUMO-3 leads to increased steady-state levels of APP.

Based on these observations, it will be interesting to search for any effect of sumoylation on α-secretase, which is not discussed in the paper. I would predict an increase of α-secretase activity by sumoylation. Overexpression of wildtype SUMO-3 could lead to enhanced α-secretase, resulting in increased α-NTF levels, reduced β-NTF levels (and subsequently reduced Aβ levels.) On the contrary, overexpression of mutant forms of SUMO-3 may not affect α-secretase activity, and increased levels of α-NTF, β-NTF and Aβ could result simply from increased steady-state levels of APP in these cells.

If the above prediction is correct, several conclusions in the paper might need to be reevaluated, e.g., a positive effect of monosumoylation on Aβ production, an indirect effect of sumoylation on APP processing. Nevertheless, the authors may have already examined the α-secretase and failed to find any effect.

Regarding the physiological relevance of sumoylation to AD pathology, this paper has mentioned a SUMO-3 distribution to the neuronal soma from AD and Down’s syndrome brains, in contrast to both soma and nuclear from nondemented brains. Sumoylation of transcription factors plays an important role in regulating gene expression, and the observations described in this paper can be pursued to explore differential gene expressions during the development of AD pathology.

View all comments by Weiming Xia

This article presents a novel regulatory mechanism for APP processing. Li et al. observe that increased expression of SUMO-3 increases production of Aβ and reduces production of APPsα. SUMO (small ubiquitin-like modifier) proteins are small peptides that are added to proteins in a manner similar to ubiquitin, unlike ubiquitin however, they do not target proteins for proteasomal degradation. The function of SUMO is not well understood, but in cases where it has been examined SUMO appears to control localization of proteins within the cell, such as targeting to particular organelles or the nucleus. For instance, herpes virus particles appear to form inclusions by manipulating the SUMO system. Consistent with this model, Li and colleagues do not observe a significant effect of SUMO on the turnover of APP. Since little is known about SUMO, Li’s observation opens up a new chapter in the regulation of APP processing and will likely yield new and interesting findings in the upcoming months and years.

View all comments by Benjamin Wolozin

Posttranslational modification of the cytoplasmic tail of APP has emerged as an important mechanism for controlling APP metabolism and Aβ generation. These phenomena can be divided into two basic cellular phenotypes: modifications that regulate the α- and β-secretase pathways in parallel, and others that regulate the two pathways in a reciprocal fashion. In both cases, the molecular basis appears to be alteration of APP trafficking. Some APP-tail-binding proteins appear to promote retention of APP in the ER, slowing APP flux down both to the α and β pathways. The best known means of regulating the α and β pathways in a reciprocal manner is accomplished by activation of signal transduction pathways, such as those mediated by protein kinase C (PKC) and extracellular signal regulated protein kinase (ERK). To date, no "cellular machinery" has been discovered that is capable of transducing changes in cytoplasmic protein phosphorylation into activation or inhibition of intralumenal proteolysis. Accelerated vesicle biogenesis at the trans-Golgi network (TGN) appears to explain at least...  Read more

Posttranslational modification of the cytoplasmic tail of APP has emerged as an important mechanism for controlling APP metabolism and Aβ generation. These phenomena can be divided into two basic cellular phenotypes: modifications that regulate the α- and β-secretase pathways in parallel, and others that regulate the two pathways in a reciprocal fashion. In both cases, the molecular basis appears to be alteration of APP trafficking. Some APP-tail-binding proteins appear to promote retention of APP in the ER, slowing APP flux down both to the α and β pathways. The best known means of regulating the α and β pathways in a reciprocal manner is accomplished by activation of signal transduction pathways, such as those mediated by protein kinase C (PKC) and extracellular signal regulated protein kinase (ERK). To date, no "cellular machinery" has been discovered that is capable of transducing changes in cytoplasmic protein phosphorylation into activation or inhibition of intralumenal proteolysis. Accelerated vesicle biogenesis at the trans-Golgi network (TGN) appears to explain at least part of the response to activation of PKC or ERK.

Now, Barbara Cordell and her colleagues have demonstrated that the "regulated cleavage" phenotype can be induced by addition of small ubiquitin-like modifiers to the APP cytoplasmic tail. α cleavage is activated upon "SUMOylation" of APP, and β cleavage is reciprocally reduced. As with PKC and ERK activation, one subcellular mechanism might include the redistribution of "SUMOylated" APP toward the plasma membrane, where α-secretase is encountered. This mechanism does not completely explain the phenomenon unless the sorting machinery at the TGN is operating under limiting substrate conditions. Otherwise, there must also be a separate explanation for how the β pathway is inhibited.

"SUMOylation" joins the growing list of signals that can modulate APP metabolism: acetylcholine, interleukin-1, estrogen, testosterone, insulin, IGF-1. The life cycle of APP is coming to light, revealing an exquisitely fine-tuned balance of competing proteolytic pathways. The relative activity of these competing pathways is governed by the activation state or the hormone responsiveness of the cell. These regulatory factors may contribute to the association of Alzheimer's with disturbances in the levels of gonadal and peptide hormones.

View all comments by Samuel Gandy

This paper is groundbreaking in more than one respect. First of all, it presents a novel strategy for identifying modulators of APP processing. The authors screened 100,000 cDNAs from a human brain cDNA expression library. They began with 1,001 pools, each with about 100 plasmids, transfected the DNA pool together with an AβPP expression plasmid into 293 T cells, and assayed the growth medium by ELISA for the level of Aβ, and for the levels of secreted AβPP generated by α-secretase (α-NTF) and β-secretase (β-NTF). They identified a cDNA whose protein product decreased production of Aβ and β-NTF, and increased production of α-NTF. This strategy is not trivial to execute, and represents quite a tour de force.

Remarkably, the authors showed that the cDNA they retrieved encoded the small ubiquitin-like protein SUMO-3. I say "remarkably." because the ubiquitin-like proteins have not yet been implicated in AD, except insofar as our laboratory has reported that APP-BP1, which interacts with the C-terminus of APP, is part of the activating enzyme for the related ubiquitin-like...  Read more

This paper is groundbreaking in more than one respect. First of all, it presents a novel strategy for identifying modulators of APP processing. The authors screened 100,000 cDNAs from a human brain cDNA expression library. They began with 1,001 pools, each with about 100 plasmids, transfected the DNA pool together with an AβPP expression plasmid into 293 T cells, and assayed the growth medium by ELISA for the level of Aβ, and for the levels of secreted AβPP generated by α-secretase (α-NTF) and β-secretase (β-NTF). They identified a cDNA whose protein product decreased production of Aβ and β-NTF, and increased production of α-NTF. This strategy is not trivial to execute, and represents quite a tour de force.

Remarkably, the authors showed that the cDNA they retrieved encoded the small ubiquitin-like protein SUMO-3. I say "remarkably." because the ubiquitin-like proteins have not yet been implicated in AD, except insofar as our laboratory has reported that APP-BP1, which interacts with the C-terminus of APP, is part of the activating enzyme for the related ubiquitin-like protein NEDD8 (which, the authors show, does not affect APP processing). The cellular reactions involving SUMO and NEDD8 ubiquitin-like proteins appear to be quite similar to those involving ubiquitin, but the ubiquitin-like proteins have novel regulatory functions not necessarily linked to proteolysis. For example, protein neddylation is emerging as a major means of regulating the cell cycle. While the functions of the SUMO ubiquitin-like proteins are poorly understood, they appear to affect protein localization in addition to protein stability. Thus, this paper is also groundbreaking in revealing the involvement of a completely new signaling pathway in AD. In fact, the authors show that SUMO-3 immunoreactivity is present in neurons in the brains of individuals with AD and Down’s syndrome, and of unaffected individuals. They make the interesting comment that SUMO-3 appears segregated in the neuronal soma in a larger percentage of neurons in AD and Down’s syndrome brains than in unaffected brains, where it is largely nuclear. We have made a similar observation (unpublished) about NEDD8, which shows a pronounced change in intraneuronal localization in affected regions of AD brain. In summary, this paper uses a novel strategy to identify the existence of a new pathway that modulates APP α- versus β-secretase cleavage. It also presents a new set of targets for potential therapeutic intervention in the disease. It has been difficult to generate β-secretase inhibitors; perhaps stimulation of sumoylation with SUMO-3, apparently a natural inhibitor of β-secretase, will be a viable alternative.

View all comments by Rachael Neve