Research Article

De novo design of a biologically active amyloid

See allHide authors and affiliations

Science  11 Nov 2016:
Vol. 354, Issue 6313, aah4949
DOI: 10.1126/science.aah4949

You are currently viewing the abstract.

View Full Text

Aggregation by design

Amyloid aggregation is driven by short sequences within proteins that self-assemble into characteristic amyloid structures. About 30 human proteins are implicated in amyloid-associated diseases, but many more contain short sequences that are potentially amyloidogenic. Gallardo et al. designed a peptide based on an amyloidogenic sequence in the vascular endothelial growth factor receptor VEGFR2. The peptide induced VEGFR2 to form aggregates with features characteristic of amyloids. Amyloids were toxic only in cells that required VEGFR2 activity, suggesting that the toxicity was due to loss of function of VEGFR2, rather than to inherent toxicity of the aggregates. The peptide inhibited VEGFR2-dependent tumor growth in a mouse tumor model.

Science, this issue p. 10.1126/science.aah4949

Structured Abstract

INTRODUCTION

It has been shown that most proteins possess amyloidogenic segments. However, only about 30 human proteins are known to be involved in amyloid-associated pathologies, and it is still not clear what determines amyloid toxicity in these diseases. We investigated whether an endogenously expressed protein that contains sequences with known amyloidogenic segments, but is not known to aggregate either under normal or pathological conditions, can be induced to do so by seeding it with a peptide comprising the protein’s own amyloidogenic fragment. We chose to target the protein vascular endothelial growth factor receptor 2 (VEGFR2) because it has well-characterized biological function and so could provide a model system with which to investigate the relationship between protein loss of function and amyloid toxicity in different cellular contexts.

RATIONALE

The capacity of the amyloid conformation of disease proteins to catalyze their own amyloid conversion demonstrates the sequence specificity of amyloid assembly. Because the core of amyloids consists of short amyloidogenic sequence fragments, we hypothesized that a short amyloidogenic protein sequence of VEGFR2, a protein normally not associated with protein aggregation, should be able to interact with and specifically induce the aggregation of VEGFR2, resulting in its functional knockdown. We used TANGO, an algorithm that predicts aggregation-prone sequences, to identify potential amyloidogenic fragments in VEGFR2. We synthesized these fragments as a tandem repeat in a peptide framework in which each unit is flanked by charged residues and coupled by a short peptide linker. The thinking behind this design was that the tandem repeats would promote the formation of diffusable soluble oligomeric aggregates, whereas the charged residues would kinetically stabilize these oligomers and reduce the rate of insoluble fibril formation.

RESULTS

By screening for loss of function, we identified one peptide, termed “vascin,” that was highly potent at inhibiting VEGFR2. This sequence was derived from the translocation signaling sequence of VEGFR2. We found that vascin is an amyloidogenic peptide that readily forms small β-structured oligomers, ranging from dimers to nonamers, that slowly convert to amyloid fibrils. When added to cell culture medium, these oligomers are efficiently absorbed by the cell, where they interact with and promote the aggregation and partial degradation of nascent VEGFR2. Vascin aggregation does not induce the aggregation of known disease amyloids. Neither do vascin oligomers affect the function of the related EGF receptor or the surface translocation of other receptors. We found vascin only to be toxic to cells that are dependent on VEGFR2 function, suggesting that toxicity is due to loss of VEGFR2 function and not to vascin aggregation or vascin-induced VEGFR2 aggregation. Consistent with this, we found that vascin is active in vivo and could reduce tumor growth in a VEGFR2-sensitive subcutaneous B16 melanoma syngenic tumor model in mice but is not intrinsically toxic to other tissues.

CONCLUSION

We found that a short amyloidogenic protein fragment can induce the aggregation of a protein normally not associated with amyloidosis in a manner that recapitulates key biophysical and biochemical characteristics of natural amyloids. In addition, we found that amyloid toxicity is observed only in cells that both express VEGFR2 and are dependent on VEGFR2 activity for survival. Thus, rather than being generic, amyloid toxicity here appears to be both protein-specific and conditional on a requirement for VEGFR2 protein function.

A synthetic amyloid peptide induces aggregation.

We designed vascin, a synthetic amyloid peptide based on an amyloidogenic fragment of the signal peptide of VEGFR2. Vascin forms prefibrillar oligomers that penetrate mammalian cells and interacts with the nascent VEGFR2 protein, resulting in its aggregation and functional knockdown. [Composition includes parts of an image from iStock.com/luismmolina.]

Abstract

Most human proteins possess amyloidogenic segments, but only about 30 are associated with amyloid-associated pathologies, and it remains unclear what determines amyloid toxicity. We designed vascin, a synthetic amyloid peptide, based on an amyloidogenic fragment of vascular endothelial growth factor receptor 2 (VEGFR2), a protein that is not associated to amyloidosis. Vascin recapitulates key biophysical and biochemical characteristics of natural amyloids, penetrates cells, and seeds the aggregation of VEGFR2 through direct interaction. We found that amyloid toxicity is observed only in cells that both express VEGFR2 and are dependent on VEGFR2 activity for survival. Thus, amyloid toxicity here appears to be both protein-specific and conditional—determined by VEGFR2 loss of function in a biological context in which target protein function is essential.

View Full Text