Ubiquitylation-independent activation of Notch signalling by Delta
- Institute of Genetics, Heinrich-Heine-Universitaet Duesseldorf, Germany
- Universität zu Köln, Germany
- Heinrich-Heine-Universitaet Duesseldorf, Germany
- Institute of Molecular Biology and Biotechnology, Foundation for Research and Technology - Hellas, Greece
- University of Crete, Greece
Figures
Figure 1 with 1 supplement
Dl-induced activity of Notch in mib1 mutant wing imaginal discs.
(A) Expression of Wg in a normal wing disc. (B) Establishment of the Wg expression domain along the DV boundary by Notch signalling. For further explanation, see text. (C–E) Expression of Wg in mutants of genes involved in Notch signalling. (C) Psn-, (D) Ser- and (E) mib1-mutant wing imaginal disc. The arrow in (A) points to the expression along the D/V boundary that is dependent on Notch signalling. It is lost in the mutants. The arrowhead in (A, C–E) points to the inner ring-like expression domain of Wg. Its diameter is a measure of the loss of Notch pathway activity. (F) Expression of Gbe + Su(H) in a wildtype disc. (G–I) Expression of Wg (G, I), Gbe + Su(H) (H, I) in a mib1 mutant disc of the early third instar stage. The arrow highlights the expression along the D/V boundary. (J, K) Expression of vgBE (J, K) and Wg (K) in a mib1 mutant disc of the early third instar stage. The arrow points to the expression of vgBE and Wg along the D/V boundary. (L–O) Expression of Gbe + Su(H) in the notum of a wildtype (L), mib1 (M), and Notch depleted disc (N, O). The arrow highlights the S3 expression domain of Gbe + Su(H). S3 is present in mib1 mutants (M, arrow), but absent in Notch depleted nota (O, the arrow points to the expected location of S3).
Figure 1—figure supplement 1
Residual Dl induced Notch activity in mib1 mutant wing imaginal discs.
(A–F) The phenotype of mib1 H double mutant late third instar wing discs. Expression of Wg (A, C) and Gbe + Su(H) (B, C). Both Notch targets are expressed along the D/V boundary, while they are lost in a mib1 mutant disc at this stage (see Figure 1D). In addition, the diameter of the inner ring of Wg expression is dramatically increased (A, arrowhead). (D–F) Expression of Gbe + Su(H) and Ap. The arrow in (D) highlights the region shown at higher magnification in (E, F). The expression of Gbe + Su(H) along the D/V boundary is restricted to dorsal Ap expressing boundary cells.
Figure 2 with 1 supplement
Dl induced activity in mib1 mutant wing imaginal discs.
(A, B) Normal expression patterns of Gbe + Su(H) (A), Wg and ptc Gal4 (B). The arrow in (A) points to the S3 expression of Gbe + Su(H) and in (B) to the expression domain of ptcGal4. (C) The expression of ptcGal4 occurs in a gradient that increases from anterior until the A/P compartment boundary. It is not expressed in posterior cells and occurs perpendicular to the expression of Wg along the D/V boundary. A: anterior; P: posterior; D: dorsal; V ventral. (D, E) Expression of Dl-HA in a wildtype disc results in ectopic activation of Wg (E, white arrows) and Gbe + Su(H)-lacZ (D, white arrows). The yellow arrow in (D, E) points to the intersection of ptcGal4 with the D/V boundary where the expression of Wg and Gbe + Su(H) is interrupted due to the cis-inhibitory effect of high Dl expression. Note that Gbe + Su(H) is ectopically induced throughout the whole ptcGal4 domain, which also runs through the notal region (arrowhead in D). In contrast the ectopic expression domain of Wg is restricted to the wing area and stops at the inner ring of Wg expression (arrowhead in E). (F) One stripe of the ectopic expression domains of Wg is located in regions of low expression of the anterior compartment. The second stripe is induced in posterior boundary cells that do not express Dl-HA by trans-signalling from anterior boundary cells. (G, H) Expression of ptcGal4, Wg and Gbe + Su(H) in mib1 mutant wing imaginal discs. The arrowhead in (H) points to the inner ring of Wg expression. (I, J) Expression of Dl-HA in a mib1 mutant disc. Weak ectopic expression of Wg (I, arrow) is observed . The ring-like domains of Wg expression are expanded in comparison to mib1 mutants (arrowhead, compare with H). (J) Ectopic expression of Gbe + Su(H) is induced throughout the ptcGal4 domain (arrow and arrowhead). (K–L) Expression of UAS Dli2 ala-HA in mib1 mutant discs results in a similar phenotype as expression of Dl-HA (compare with I, J). (M, N) Expression of Ser-HA is not able to induce expression of Gbe + Su(H) (M) or Wg (N) in mib1 mutant wing discs.
Figure 2—figure supplement 1
Dl-HA cannot induce activation of the Notch pathway in Psn mutant wing imaginal dioscs.
(A–C) Expression of Dl-HA fails to induce Gbe + Su(H) (A, C) or Wg (B, C) in Psn mutant wing discs. The arrow and arrowhead in (B) point to the strongly reduced ring-like domains of Wg expression.
Figure 3 with 4 supplements
Expression of DlK2R-HA (inserted into the landing site 51C) with ptcGal4.
(A) Schematic representation of Dl-HA and DlK2R-HA and the localisation of the i1 and i2 binding boxes for Neur and Mib1 respectively in the ICD of Dl. (B–D). Expression in wildtype discs interrupts the expression of Wg and of Gbe + Su(H) at the D/V boundary (B, C, yellow arrow). Expression Gbe + Su(H) is ectopically induced (C, D, arrowhead). (E) Ectopic expression of Wg (arrows) is observed in a fraction of the discs. (F) Expression of DlK2R-HA in Psn mutant discs does not induce Gbe + Su(H). (G–J) Expression of DlK2R-HA in mib1 mutant discs. (H–J) Magnification of the wing area of the disc shown in (G). The expression of DlK2R-HA causes the induction of two parallel stripes of Gbe + Su(H) (arrowheads in H, J) along the ptc domain. Wg expression is only weakly induced (I, arrow). (K, L) Expression of DlK2Ri1 ala-HA results in a similar induction of Gbe + Su(H) expression as DlK2R-HA (arrowheads). (M) Co-expression of Dl-HA with Notch results in the induction of a broad stripe of expression of Wg in the wing area. Moreover, no gap in the expression domain of Wg along the D/V boundary is observed (yellow arrow, compare with Figure 2E). (N–P) A similar suppression of the increased cis-inhibitory effect of DlK2R-HA was observed upon co-expression with Notch (yellow arrow in N). (Q, R) Co-expression of DlK2R-HA with Notch in a mib1 mutant disc, reliably induces a broad stripe of Gbe + Su(H) expression (arrowhead). (S, T) Co-expression of Notch and Dl-HA in mib1 mutants. A broad band of expression of Gbe + Su(H) is induced throughout the ptcGal4 domain (arrowhead). (U, V) Dl Ser double mutant MARCM clones co-expressing DlK2R-HA and Notch. Wg is induced throughout the clone in the absence of endogenous ligands (arrows). (W, X) A mib1 mutant disc that bears Dl Ser double mutant clones (loss of GFP) and expresses DlK2R-HA with ptcGal4. DlK2R-HA is able to induce expression of Gbe + Su(H) in the Ser Dl double mutant cells (arrow).
Figure 3—figure supplement 1
Ubiquitylation of Dl-HA or DlK2R-HA in S2 cells by Neur and Mib1.
Dl-HA or DlK2R-HA was detected by anti-HA from S2 cell extracts. Inputs and neutravidin (NA) pulldowns are shown. Arrows point to the Dl protein. All samples had been cotransfected with ac5-Gal4 (under the actin 5 promoter) and the following combinations of expression plasmids: 1: pIZ-Dl-HA, UAS-mib1HM, ac5-bioUb6-2A-BirA; 2: pIZ-Dl-HA, UAS-GFP-neur, ac5-bioUb6-2A-BirA; 3: pIZ-Dl-HA, ac5-bioUb6-2A-BirA; 4: pIZ-Dl-HA, ac5-BirA; 5: pIZ-DlK2R, ac5-bioUb6-2A-BirA; 6: pIZ-DlK2R, UAS-GFP-neur, ac5-bioUb6-2A-BirA; 7: pIZ-DlK2R, UAS-mib1HM, ac5-bioUb6-2A-BirA ac5-bioUb6-2A-BirA expresses a polyprotein, which gets processed by endogenous Ubiquitin proteases to bioUb, a BirA-recognition-motif-tagged ubiquitin, and the BirA biotin ligase itself (ref). ac5-BirA expresses only BirA as a control for spurious (non-Ubi) biotinylation. In this way, neutravidin pulls down ubiquitylated adducts of proteins. Note that the NA- pulled down Dl is of higher MW than the Dl protein, due to the addition of ubiquitin or poly-ubiquitin moieties. Also shorter Ubi-modified proteolytic products are detected. No bioUb-adducts are detected for DlK2R (conditions 5–7).
Figure 3—figure supplement 2
Quantification of the gap in the endogenous expression of Wg along the D/V boundary induced by cis-inhibition of the Dl constructs.
(A-A` Dl-HA). (B-B` DlK2R-HA). (A`and B`) magnification of the region boxed in (A, B). In each case the Hoechst channel (A`, B`) is used to count the nuclei. Arrows in (A`, B`) point to the boundary of Wg expression. (C) Quantification of the gap size.
-
Figure 3—figure supplement 2—source data 1
Quantification of cis-inhibition of Dl variants.
- https://doi.org/10.7554/eLife.27346.012
Figure 3—figure supplement 3
The consequences of expression of a randomly inserted DlK2R-HA construct with ptcGal4.
(A, B) Expression of DlK2R-HA in wildtype discs results in the ectopic activation of Wg in posterior boundary cells (arrowheads). The yellow arrow points to the intersection of the ptcGal4 domain with the D/V boundary where the wing primordium is divided into parts due to strong increase in cis-inhibition. (C) Higher magnification of the wing area of the disc shown in (B). The arrow points to the ectopic expression domain of Wg in posterior boundary cells. (D) The expression of DlK2R-HA causes the formation of a scar (yellow arrow). (E–H) Expression of the dominant-negative Dlstu (E, F) or an N-RNAi construct (G, H). result in similar phenotypes as in the case of DK2R-HA. Here a scar and division of the wing area is observed as well (yellow arrow in E, G). (I, J) Co-expression of Dl with N-RNAi phenocopies DlK2R-HA expression (compare with A, B). The arrowheads point to the ectopic expression of Wg and the yellow arrow highlights the division of the wing area.
Figure 3—figure supplement 4
Induction of Dl expression by the Dl variants.
The Dl-MIMIC-GFP line was used for measurement of the activity of the Dl promoter. (A) Expression of Dl in normal discs results in the induction of ectopic expression of Dl (arrows). Expression of Dl-HA (B–D) or DlK2R-HA (E–G) in mib1 mutants fail to induce significant ectopic expression of Dl, although it induces ectopic expression of Gbe +Su(H) (C–F).
Figure 4 with 1 supplement
Endocytosis of Dl-HA and DlK2R-HA.
(A–C) Subcellular localisation of Dl-HA and DlK2R-HA (D–F) in wing imaginal disc cells. The variants are detected at the apical plasma membrane (arrowhead in A, (D) and in Rab7 positive MEs (arrows). (G) Quantification of the percentage of Rab7 positive endosomes that are also positive for the Dl variants. (H) Summed fluorescence HA intensity of Rab7 positive vesicles in discs that either express UAS Dl-HA or DlK2R-HA expressed with ciGal4. For further information of the method, see M and M. The intensity is significantly higher for DlK2R-HA expressing discs. (I) Representative Western blot of the expression of Dl-HA and DlK2R-HA with tubGal4 Gal80ts for 48 hr, revealing stronger expression of DlK2R-HA. (J) Quantification of the expression of the Dl variants. The average value of ratio between the band of the Dl variant and the Tubulin loading control from eight independent Western blots was determined. DlK2R-HA is expressed approximately 2.5x more than Dl-HA. The P-value was calculated with a Student’s t-test.
-
Figure 4—source data 1
Raw median data of cargo intensity measurements and raw percentage data of colocalization analysis.
Raw data of vesicle cargo intensity measurements (medians of HA-intensity sum in Rab7 positive vesicles; biological replicates Dl-HA/ DlK2R-HA) and vesicle colocalization analysis (percentages of Rab7positive vesicles with minimum HA-signal; biological replicates Dl-HA/ DlK2R-HA). See M and M for further information.
- https://doi.org/10.7554/eLife.27346.015
-
Figure 4—source data 2
Quantifications of Western blots.
See M and M for further information.
- https://doi.org/10.7554/eLife.27346.016
Figure 4—figure supplement 1
Degradation of Dl-HA and DlK2R-HA in the wing imaginal disc.
Pulse chase experiment with Dl-HA (A–D) and DlK2R-HA (E–J). The constructs were expressed for 16 hr with ciGal4 tub.Gal80ts at 29°C. Expressing was stopped through the transfer to 18°C and the discs prepared and stained at various time points. Dl-HA was not detected after a case of 24 hr (C, D). At this time point DLK2R-HA was still strongly expressed (G, H). A significant fraction was still located in the apical membrane (H, arrows). (I, J) Even after a chase of 36 hr a weak DlK2R-HA signal was still detectable.
Figure 5
Expression of Dli1/2 with ptcGal4.
(A–C) Expression in wildtype discs causes an interruption of the expression of Wg along the D/V boundary at the point of intersection with the ptc domain (A, arrow) and induces ectopic expression of Gbe + Su(H) (B, arrowhead). (D–F) Expression of Dli1/2 in mib1 mutant discs fails to induce Wg expression (D), but induces strong expression of Gbe + Su(H) throughout the ptcGal4 domain (E, arrow and arrowhead). (G–I) Co-expression of Dli1/2 with Notch in the wildtype results in the induction of a broad stripe of ectopic expression of Wg (G, arrow) and of Gbe + Su(H) (H, arrow and arrowhead) in a similar manner as co-expression of Notch with DlK2R-HA (Compare with Figure 3N-P, Figure 2J–L, ). (J–L) Co-expression of Dli1/2 with Notch in mib1 mutant results in the expansion of the ring-like expression domains of Wg (J, arrow) and ectopic expression of Gbe +Su(H) throughout the ptcGal4 domain (K, arrow and arrowhead). (M, N) Dl Ser double mutant MARCM clones (arrow, green in N) that co-express Dli1/2 with Notch. The mutant cells express Wg.
Figure 6 with 1 supplement
Analysis of cis-inhibition in mib1 null mutant wing discs in the early third instar stage.
(A) Relief of cis-inhibition. In several areas of the notum of the wing disc cis-inhibition prevents significant Notch signalling (a). (b) Cells that are located at both sides of the boundary of a clone double mutant for Ser and Dl. Mutant cells have lost their ligands and therefore the Notch receptor is no longer engaged in cis-inhibitory interactions and can be activated by the ligands of the adjacent Dl and Ser positive cells on the other side of the boundary. Consequently the Notch pathway is activated in the mutant cells, which consequently activate the expression of the Notch target genes. (B–D) Cis-inhibition occurs in the notum. Dl Ser double mutant clones in the notum of a wildtype wing imaginal disc. (C, D) Magnification of the area with a clone highlighted by the arrow in (B). The clone is labelled by the absence of GFP. Expression of Gbe + Su(H) is induced in the mutant cells at the clone boundary because of relief of cis-inhibition ( arrowhead). (E–J) Expression of Gbe + Su(H) is also induced in Dl Ser double mutant cells at the boundary of Dl Ser double mutant clones in mib1 mutant cells in the notal area (arrows). Thus, ligand dependent Notch activation caused by relief of cis-inhibition also occurs in the absence of mib1 function. (K, L) A large Dl Ser double mutant clone that covers most of the notal area of a mib1 mutant disc. The S3 expression of Gbe + Su(H) normally present in mib1 mutant discs is lost. The arrows point to the expected region of S3 expression (compare with Figure 1M, arrow). (M–O) Depletion of Notch function in the anterior compartment of a wing disc by expression of Notch-RNAi with ciGal4. It causes the induction of Gbe + Su(H) expression in the posterior cells adjacent to the ciGal4 expression domain (O, arrows). The arrowhead in (O) points to the unrelated expression of Gbe + Su(H) in the trachea, which is attached to the imaginal disc, but is no part of it. (P, Q) Loss of Neur function does not affect the S3 expression of Gbe + Su(H) in mib1 mutant discs. neur mutant clones are labelled by the absence of GFP. The arrow points to S3 expression in a clone.
Figure 6—figure supplement 1
Activation of E(spl)mß expression in mib1 discs through relief of cis-inhibition.
(A–D) Expression of Dl (antibody staining) and Ser (revealed by a MIMIC-GFP gene trap).
(B–D) Magnification of the notal area of the disc shown in (A). It reveals that the patterns of Ser and Dl combined cover most of the notum. (E–G) Expression of E(spl)mß-lacZ (F, G) and Dl (E, G) in a wildtype wing imaginal disc. (H) Expression of E(spl)mß-lacZ in a mib1 mutant disc. The arrow points to the S3-like expression domain. (I–L) Relief of cis-inhibition results in the induction of expression of E(spl)mß-lacZ in Dl Ser double mutant cells at the boundary of Dl Ser clones (arrowhead) in mib1 mutant discs. Clones are labelled by the absence of GFP. The asterisk labels a leg disc adjacent to the wing disc. The arrow points to the S3-like domain of E(spl)mß-lacZ. (M–O) Depletion of Notch function in the anterior compartment by ciGal4 driven UAS N-RNAi results in the loss of all expression of E(spl)mß-lacZ. The arrow in (N) points to the expected location of the S3-like domain. The asterisk labels a leg disc adjacent to the wing disc.
Figure 7 with 1 supplement
Activation of Dl-HA and DlK2R-HA by Neur.
With exception of (H, I) ptcGal4 was used for expression of the constructs. (A) Expression of Neur has no effect on expression of Wg along the D/V boundary. (B) Co-expression of Neur and Dl-HA causes a strong ectopic expression of Wg (arrows). The arrowhead points to the dorsal half where the anterior stripe of ectopic Wg expression is suppressed. (C) Expression of Neur in mib1 mutants restores expression of Wg along the D/V boundary (arrow). (D) Co-expression of Dl-HA with Neur in mib1 mutants induces ectopic Wg expression in a manner comparable to wildtype (arrows, compare with B). (E, F) Co-expression of Neur with DlK2R-HA in wildtype (E) or mib1 mutants (F) results in a strong ectopic activation of Wg expression (arrows). (G) Co-expression of an independently generated Neur construct with DlK2R-HA induces a similar ectopic activation of Wg expression (arrows). (H, I) Dl Ser mutant MARCM clones (green in I) co-expressing Neur and DlK2R-HA. Wg is ectopically induced in the mutant cells despite the lack of the function of Dl and Ser (arrow). (J) Co-expression of Mib1 and DlK2R-HA does not induce ectopic Wg expression. The yellow arrow highlights the gap in the expression of Wg along the D/V boundary, which is typical for expression of DlK2R-HA. (K) Expression of NeurΔNHR1 cannot restore the expression of Wg along the D/V boundary. (L) Its co-expression with DlK2R-HA cannot induce ectopic expression of Wg. The phenotype resembles that induced by expression of DlK2R-HA alone (compare with Figure 2A). Likewise Co-expression of NeurΔNHR1 with DlK2R-HA (M) or Dl-HA (N) in mib1 mutants fails to induce ectopic expression of Wg. (O) Co-expression of Neur with Dli1/2, cannot induce ectopic expression of Wg. As in case of expression of Dli1/2 alone, the expression of Wg along the D/V boundary is interrupted (arrow). (P, Q) Expression of NeurΔRF does not affect expression of Wg along the D/V boundary in wildtype discs and co-expression with Dl-HA in the wildtype results in ectopic expression of Wg comparable to expression of Dl-HA alone (Q, arrows, compare with Figure 2E and Figure 1F). (R) NeurΔRF cannot restore the D/V expression of Wg in mib1 mutants. (S) Co-expression of NeurΔRF with Dl-HA in mib1 mutants also induces strong expression of Wg (arrows). Note, that the expression is weaker than in the wildtype disc shown in (Q). (T) Likewise, co-expression of NeurΔRF with DlK2R-HA can induce strong ectopic expression of Wg (arrows). (U–W) Co-expression in mib1 mutants induces weak ectopic expression of Wg and Gbe + Su(H). (V, W) Magnification of the wing area of the disc shown in (U, arrow). The comparison with (T) reveals that the ectopic expression of Wg is weaker than in wildtype discs. The arrows point to the ectopic expression of Wg (V) and Gbe + Su(H) (W).
Figure 7—figure supplement 1
Subcellular localisation of Neur and NeurΔRF.
Optical cross-sections of wing disks expressing EGFP-Neur (A) or Neur-ΔR-GFP (B). E-cadherin (red) marks the apical domain of the epithelium (up in both panels) and DAPI (cyan) marks nuclei. (A', B') show the GFP channel alone. Note, the cortical localization of both Neur and NeurΔR. Scale bars are 10 μm.
Figure 8
(A) Graphic of the larval CNS and a neuroblast lineage.
Arrows depict events of Notch signalling. (B) A brain hemisphere (area boxed in A) with clonal lineages in green. One marked lineage is highlighted by the arrow. The large cell nearest to the arrow is the neuroblast. Green: GFP, Red: Hey, Blue: Delta. (C–H) Enlarged areas of brain lobes containing marked clones of various genotypes, highlighted by arrows. In all panels GFP (green) marks the mutant lineages. Red: Hey. C' –H' show the Hey pattern alone; selected clone borders are drawn in yellow. (C) Several GFP marked lineages homozygous for the null allele neur (Glittenberg et al., 2006). (D) A GFP-marked clone double mutant for Dl and Ser. None of the marked cells express Hey. (E) Dl Ser; Dl-HA. Wt Dl restores Hey expression in the mutant lineages. (F) Dl Ser; Dli1/2. No rescue of Hey expression is observed. (G) Dl Ser; DlK2R-HA. Hey is rescued as efficiently as by wt Dl. (H): neur (Glittenberg et al., 2006); NeurΔR. The catalytically inactive form of Neur rescues Hey expression.
Figure 9
Specification of SOPs in an ubi-independent manner.
SOPs are revealed by anti Hnt staining. (A) Schematic drawing of the distribution of the SOPs of the large bristles in a wildtype disc (red cells). In the wildtype the nascent SOP sends an inhibitory signal via the Notch pathway to prevent SOP development in its neighbours. The loss of the activity of the Notch pathway, e.g. in clones mutant for genes involved in Notch signalling, results in the formation of clusters of SOPs (neurogenic phenotype). (B–I) Dl Ser double mutant MARCM clones. The SOPs are revealed by Hnt expression. Clones are labelled by GFP (green). (B, C) A Dl Ser double mutant MARCM clone. The clone includes a proneural cluster (arrow). Many cells have adopted the SOP fate in the absence of Dl and Ser function (neurogenic phenotype). (D, E) The neurogenic phenotype is suppressed and a single SOP is observed in a Dl Ser double mutant MARCM clone that expresses Dl-HA (arrows), indicating that selection of the SOP is restored. (F, G) A similar rescue of the selection process is observed when DlK2R-HA is expressed in Dl Ser double mutant MARCM clones (arrows). (H, I) In contrast no rescue is observed when DlK2Ri1 ala is expressed in the clones (arrows). (J) Notum of a wildtype fly showing the regular bristle pattern. (K, L) Notum of a lqf mutant fly rescued by Lqf UIM13E/3A-ΔUIM2-GFP. (L) A magnification of the notal area of the fly shown in (K). An occasional duplication of a large bristle is highlighted by the arrowhead. The density of the smaller bristles is increased compared to the wildtype. However, most bristles are well separated from each other. (M, O) Comparative analysis of SOP formation in wing discs of lqf mutant flies rescued by Lqf UIM13E/3A-ΔUIM2-GFP and mib1 mutant discs. The arrow highlights a SOP at the posterior dorso-central position. A nearly wildtype pattern of SOPs is observed in both genotypes. The arrowhead in (M) points to two adjacent developing SOPs that probably give rise to the bristle duplication observed in the imago (compare with K, L, arrowhead). However, no neurogenic phenotype is observed and the pattern of SOPs of both genotypes is similar. (P) lqf mutant clones in the notum of a mib1 mutant wing disc revealed by absence of GFP (green). The arrow highlights the region where the clone includes a proneural cluster. The loss of the function of lqf causes a neurogenic phenotype.
Additional files
-
Transparent reporting form
- https://doi.org/10.7554/eLife.27346.024
Download links
A two-part list of links to download the article, or parts of the article, in various formats.
Downloads (link to download the article as PDF)
Open citations (links to open the citations from this article in various online reference manager services)
Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)
Ubiquitylation-independent activation of Notch signalling by Delta
eLife 6:e27346.
https://doi.org/10.7554/eLife.27346
