Soluble and multivalent Jag1 DNA origami nanopatterns activate Notch without pulling force

Abstract

The Notch signaling pathway has fundamental roles in embryonic development and in the nervous system. The current model of receptor activation involves initiation via a force-induced conformational change. Here, we define conditions that reveal pulling force-independent Notch activation using soluble multivalent constructs. We treat neuroepithelial stem-like cells with molecularly precise ligand nanopatterns displayed from solution using DNA origami. Notch signaling follows with clusters of Jag1, and with chimeric structures where most Jag1 proteins are replaced by other binders not targeting Notch. Our data rule out several confounding factors and suggest a model where Jag1 activates Notch upon prolonged binding without appearing to need a pulling force. These findings reveal a distinct mode of activation of Notch and lay the foundation for the development of soluble agonists.

Fig. 1. Schematic view of the Notch pathway and experimental principle.

a The extracellular domain of the Notch receptor comprises a negatively regulatory region (NRR, dashed box) containing a pre-cleaved S1 site held together non-covalently and 29-36 (depending on Notch subtype) EGF repeats. Whether these are stretched out (left) or form tertiary structures (right), is unclear. b The canonical Notch pathway. Notch heterodimers are formed via Furin cleavage of the S1 in the Golgi, followed by membrane insertion. Binding to ligands induces cleavage of S2 by ADAM proteases, which enables cleavage of S3 via γ-Secretase to release the Notch intracellular domain (NICD). The NICD translocates to the nucleus forming a transcription activator complex with CSL and Mastermind. c DNA origami nanopatterns of Jag1 were designed where Jag1-Fc molecules covalently conjugated to DNA oligonucleotides could be attached to the nanostructures to form many types of precisely controlled multivalent Jag1 binder patterns.

Fig. 2. Characterization of Jag1Fc DNA nano-patterns.

a A rod like DNA origami was used to create 1-8x Jag1Fc nanopatterns (JNPs). b Gel shift assay. 1Kb ladder (L), origami scaffold alone (Sc), empty JNP(0x), a repeat of(0x) but with added Jag1Fc-DNA conjugates (control for non-specific binding), structures loaded with Jag1Fc patterns: 1-8x JNP (2% agarose, EthBr stained). Agarose gel repeated independently for n = 20 c Schematic representation of the DNA origami used for the DNA PAINT experiments. Here, the Jag1-DNA conjugates also contain an extension for DNA-PAINT docking sites. d DNA PAINT on Jag1Fcs on DNA origami Average (cyan, thick) and individual cropped DNA PAINT superresolution images (white, thin) of JNPs (scale bars = 50 nm). Bar graphs show the Jag1 site occupancy distributions of the different nanorods. e Zoom-ins of negative stain TEM of (unpurified) 4x JNPs and empty DNA origami rods (0x JNP) and zoom-out of 4x JNPs (bottom). Scale bars are 100 nm. Micrographs were repeated independently for n = 5 f Receptor immobilized on SPR chip surface and increasing concentrations of 1-8x JNP as analyte. The bar graph shows the mean apparent kD of different Jag1Fc nano-patterns, dots represent two individual repeats.

Fig. 3. Activation of the Notch pathway by Jag1Fc-nanopatterns.

a Origami decorated with 1-4 (28 nm separations) or 8 (14 nm separations) Jag1Fc are used to stimulate iPSc-derived neuroepithelial stem-like (lt-NES) cells. b Microscopy of lt-NES cells shown in i) brightfield and ii) fluorescence channel immunostained for Notch1 (magenta), F-actin (green), and nucleus (blue). Scale bar, 50 μm. c Time course of lt-NES cells stimulated with 3x JNP. Relative expression of HES1 from qPCR normalized to parallel samples stimulated with empty(0x) structures. Points represent individual data for n = 2 biological repeats d Effect of different number of Jag1Fcs. Relative expression of HES1 from qPCR normalized to sample with empty structures for 1-4x JNPs. Bar graphs represent mean expression levels ± SD and black dots indicate individual data points for n  =  4 biological replicates. Statistical analysis of the data was performed using one-way analysis of variance (ANOVA) followed by Dunnett multiple-comparison test (**P = 0.0026, ****P < 0.0001). e Dosage effect. lt-NES cells stimulated with 1-8x JNP at increasing concentrations. Relative expression of HES1 from qPCR normalized as in (c). Points represent individual data for n = 2 biological repeats f Proximity Ligation Assay (PLA) performed using antibodies against cleaved NICD on stimulated lt-NES cells. Representative images of cells for each condition: PLA dots (magenta), F-actin (green) and nucleus (blue). Scale bars, 10 μm. g Violin plot of the PLA experiment from image analysis of 50 cells for each condition. Statistical analysis of the data was performed using ANOVA followed by Tukey multiple-comparison test (**P = 0.0011). h Heat map diagram of mRNA sequencing experiment performed on lt-NES cells after stimulation with JNPs. Three biological repeats for each condition were shown for genes with FDR < 0.05 and absolute log2FC > 0.5. Data is shown automatically clustered using hierarchical complete-linkage clustering of Euclidean distances. i Selection of Notch pathway-related genes from RNA seq, transcripts per million (TPM) plotted for each condition. Box plots shown as median, first and third quartiles with whiskers extending up to 1.5 x inter-quartile range (IQR) and individual data points represent data for n = 3 biological repeats j Volcano plots of genes upregulated by 1x JNP and 8x JNP relative to 0x JNP. Source data are provided as a Source Data file.

Fig. 4. Multivalency drive Notch activation despite potential force sources and ADAM10 is the driving metalloprotease.

a Electrostatic repulsion between negative DNA origami and negative cell membrane could cause pull. qPCR on lt-NES cells stimulated with normal JNPs, or oligolysine (K10) coated JNPs (charge changed towards neutral). Bar graphs show the mean expression levels and dots represent 2 biological repeats. b Non-specific surface attachment of JNPs could drive force generation in motile cells. qPCR of part of the DNA origami scaffold (M13mp18 ssDNA) to detect concentrations of structures in solution and, retrieved by harsh washing of cell culture surfaces exposed to JNP solution, any remaining non-specifically bound DNA origami. *Denotes measurements below the linear range of the assay. Bar graphs show the mean value of 2 technical repeats shown with dots c Endocytosis could drive force generation. qPCR of HES1 with, or without, Pitstop2 (inhibitor for clathrin-mediated endocytosis) on JNP-stimulated lt-NES cells. Bar graphs show the mean value of 3 technical repeats shown with dots. d RNA sequencing data of gene ontology genesets that are implied in different levels of endocytosis. Data is shown automatically clustered using hierarchical complete-linkage clustering of Euclidean distances. e Increasing concentrations of ADAM inhibitor GI 254023X added on iPS cells and stimulated with 4x JNPs. qPCR of HES1. Square symbols represent the mean value of 3 technical repeats. Source data are provided as a source data file.

Fig. 5. Chimeric structures indicate that multivalency effect is primarily caused by increased binding and suggests hypothesis of tertiary structure occlusion.

a Are the effects observed reliant on receptor clustering and following conformational change (top), or are they due to a reduced off-rate of bound structures leading to more time for activation of single receptors (bottom)? b Cholesterol chimeras. Activation data from qPCR of stimulated lt-NES cells (left) and schematic representation (right) of 1x JNP with protein placed at the side as before, in the middle, and in the middle with cholesterol modifications. Bar graphs show the mean value and dots represent n = 2 biological repeats. c Micrographs of cholesterol-modified JNPs Negative stain TEM. The scale bar is 100 nm. Micrographs repeated independently n = 3 d Confocal microscopy of stimulated cells after stimulation with fluorophore labeled 0x JNP, 1x JNP, 1x JNP(m), and 1x JNP(chol). Images repeated independently n = 3. Scale bars are 10 μm. e BAI1 chimeras – where Jag1Fc is swapped at certain positions with the protein BAI1. Fold change of HES1 from qPCR in lt-NES cells stimulated with BAI1/Jag1Fc chimeric JNPs. Bar graphs show the mean value and dots represent n = 2 biological repeats. f, g Different conceptual models of Notch activation. f Where EGF repeats of Notch (green) extend out like a rod from the cell membrane (lines at the bottom), a force would be able to transport the signal of binding down to the cell membrane to expose the NRR. g Where EGF repeats of Notch are curled up in a tertiary structure, a binding of ligand would be able to induce unraveling of the tertiary structure over time to allow exposure of the NRR. Source data are provided as a Source Data file.