QUICK SEARCH:   [advanced] Author: Keyword(s):
Home     Help     Feedback     Subscriptions     Archive     Search     Table of Contents

First published online 6 October 2004
doi: 10.1242/dev.01413

This Article Summary Full Text Full Text (PDF) Supplementary Material Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Related articles in Development Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Wang, W. Articles by Struhl, G. Search for Related Content PubMed PubMed Citation Articles by Wang, W. Articles by Struhl, G.

Drosophila Epsin mediates a select endocytic pathway that DSL ligands must enter to activate Notch

Howard Hughes Medical Institute, Department of Genetics and Development, Columbia University College of Physicians and Surgeons, 701 West 168th Street, New York, NY 10032, USA

View larger version (65K): [in a new window]   Fig. 1. Lqf is required for Notch signaling in the developing wing disc. (A) Structure of Lqf/Epsin. The conserved ENTH (Epsin N-terminal homology) domain, Ubiquitin-interacting motifs (UIM) and `Clathrin box' motifs (CBM), as well as the DPW and NPF tripeptide repeat domains, are shown together with their interaction partners. The position of the lqf1227 mutation, a stop codon, is indicated. (B) Adult wing with clones of lqf- cells associated with wing notching (asterisk) and thickened veins (arrow). (C) Adult mesonotum containing a clone of lqf- cells (marked by the expression of a UAS-y+ transgene, which darkens bristles, and the mwh mutation, which splits each epidermal hair into a tuft; the clone is outlined by a red dotted line). lqf- cells next to the clone border can form normal bristles (arrows), as opposed to multi-shafted bristles or no bristles, indicating rescue of Notch signaling by adjacent wild-type cells. (D) Wing disc containing clones of lqf- cells, marked by nuclear GFP expression (green). Here, and in all subsequent figures, dorsal is up. Cells that abut the boundary between the dorsal (D) and ventral (V) compartments express Cut (red) in response to the activation of Notch. Cut is also expressed in isolated neural cells in more proximal portions of the wing disc. Ser produced by D cells activates Notch in V cells, and Dl produced by V cells activates Notch in D cells. Both Dl and Ser are upregulated in response to the activation of Notch, creating a positive-feedback loop responsible for Cut induction in D and V cells flanking the DV boundary. lqf- clones that abut the DV boundary block Cut expression on both sides, indicating a failure in Notch signaling (blockage of the signal coming from either direction interrupts the feedback loop necessary for Cut expression on both sides of the boundary). (D') Higher magnification of the DV boundary, showing that Cut expression is rescued in lqf- cells adjacent to wild-type cells along the clone border (rescued cells appear yellow).

View larger version (110K): [in a new window]   Fig. 2. Selective requirement for Lqf in sending DSL signals. (A) Clones of cells overexpressing Dl (marked by nuclear GFP, green) activate Notch in adjacent wild-type cells within the D compartment of the wing blade primordium, as indicated by the induction of Cut expression (red). The abnormally high levels of ectopic Dl expression generated under these conditions autonomously inhibit Notch transduction by cells within the clones. (B) Clones of lqf- cells that overexpress Dl (green) fail to activate Notch ectopically, and interrupt normal Cut expression when they abut the DV boundary. The same result was obtained using either the lqf1227 or lqfBTmutation; a lqf1227 clone is shown. (C,C') `Twin spots' comprising clones derived from the two daughters of single mother cells: one clone of each twin spot overexpresses Dl (marked by nuclear GFP, green), and induces Cut (red) in all the surrounding cells, including the cells of its sibling lqf- clone (marked by the absence of staining for ß-Gal, blue, C'). (D) Clones of cells overexpressing Ser (green) activate Notch in adjacent wild-type cells within the V compartment, as indicated by the induction of Cut expression (red); Notch tranduction within the clone is inhibited, as in A. (E) Clones of lqf- cells that overexpress Ser (green) fail to activate Notch ectopically, and interrupt normal Cut expression when they abut the DV boundary. (F,F') Twin spot, as in C,C', except that Ser, rather than Dl, is overexpressed, and the twin spot is located in the V compartment; Cut is expressed in all the surrounding cells, including cells of the sibling lqf- twin clone. (G,G') Clones of lqf- cells that ectopically express Dpp (marked by nuclear GFP, green) in a wing disc carrying the Dpp-responsive, omb-lacZ reporter gene. Dpp expressed by cells located just anterior to the AP compartment boundary functions as a gradient morphogen to control omb expression (red) in a broad, centrally located stripe flanking the boundary (see also H). Clones of lqf- cells that ectopically express Dpp, express omb and induce surrounding clones to do the same, indicating that they are competent both to send and to receive Dpp. (H,H') Clones of lqf- cells (marked by the absence of GFP, green) in an omb-lacZ wing disc, counterstained for both omb-lacZ (red) and Ci (blue) expression. Ci expression serves in this experiment to mark the A compartment (left); the AP boundary is shown in white. Note the presence of a large lqf- clone in the P compartment (arrow) that has no apparent effect on the broad domain of omb expression, indicating that Dpp has moved normally through the P compartment from its source in the A compartment. Note also that clones of lqf- cells are of similar size to their sibling twin clones (marked in this experiment by two copies of the GFP-producing transgene, and hence bright green), and have similarly wiggly borders.

View larger version (101K): [in a new window]   Fig. 4. Bulk endocytosis of DSL ligands is not impaired in the absence of Lqf. (A) Clones of lqf- cells (marked by cell surface CD8-GFP, green) in the eye disc stained for endogenous Dl (red) and expression of a Dl-lacZ transgene (blue). Undifferentiated cells (left) are recruited to form photoreceptors as they enter the morphogenetic furrow; the arrow marks a lqf- clone in the vicinity of the furrow. Expression of Dl on the apical surface is strongly enhanced in the clone, as is expression of the Dl-lacZ reporter gene (shown at a deeper plane of section to visualize ß-Gal, which is nuclear). (B) As in A, except both Dl and ß-Gal staining are shown at deeper planes of section. Note that Dl staining within the clone is punctate, consistent with localization in endocytic compartments. (C) Clone of lqf- cells (marked by CD8-GFP, green) in the wing blade primordium stained for endogenous Dl (red). The focal plane is at the apical surface in C, and approximately 6 µm beneath the apical surface in C' and C''. Dl appears to be generally unaffected by the lqf- clone at both planes of focus; note the presence of cytosolic puncta at the deeper plane both inside and outside of the clone. The clone is located just dorsal to the DV boundary, and interrupts the normal, Notch-dependent upregulation of Dl in cells flanking the boundary. The same result was obtained with clones of either lqf1227 or lqfARI cells; a lqfARI clone is shown. (D) Clone of hrs- wg- wing cells overexpressing an HRP-tagged form of Dl (HRPDl), stained for HRP (green) and Wg (red). The clone (marked by HRPDl expression) is located close to the DV boundary, the source of Wg (at the top of the image). HRPDl and Wg co-localize in large puncta. Because these cells are wg-, the Wg protein that co-localizes with HRPDl serves as an in vivo marker for an endocytic compartment, presumably the abnormal endosomal structures that result from the loss of Hrs. (E) As in D, except the clone is triply mutant for hrs- wg- and lqf-. Note that HRPDl and Wg still co-localize, indicating that HRPDl has been internalized, like Wg, into the abnormal Hrs-deficient endosome.

View larger version (99K): [in a new window]   Fig. 5. Evidence for a Lqf-dependent pathway of Dl endocytosis. (A) Clone of lqf- cells (marked by the absence of ß-Gal, green) in a wing disc overexpressing Myc-tagged Dl, stained for extracellular (red) and total (blue) Dl accumulation. Plane of focus is at the apical cell surface. Surface accumulation of Dl is not impaired by the absence of Lqf. (A') Same clone as in A. Four adjacent planes of focus located around 10-15 µm below the cell surface; elements in focus in each plane are shown, summed, by Auto-montage software (Syncroscopy). Punctate accumulation is detected only by staining for total Dl protein (blue); no accumulation is detected using the extracellular staining protocol (red), as expected. The punctate accumulation of Dl is not impaired by the absence of Lqf. (B) Clone of lqf- cells (marked by the absence of ß-Gal, green) in a wing disc overexpressing both Myc-tagged Dl and Neur, stained for Myc (red) and Hrs (blue). Plane of focus is at the apical cell surface. Surface depletion of Dl is impaired in the lqf- clone. (B') Same clone as in B, shown at a deeper plane of section (around 10 µm beneath the surface). Note that the accumulation of Dl in cytosolic puncta is not affected by the absence of Lqf. Although the number of Dl staining puncta appears similar to that seen in in A', where Neur was not overexpressed, only one focal plane has been sampled, as opposed to the four planes in A'. (B'') High magnification view of the boxed region in B', showing Hrs and Dl staining. Note that some Dl staining puncta co-stain with Hrs (arrows). (C) Clone of lqf- cells overexpressing Myc-tagged Dl (red) and Neur (marked by CD8-GFP, green) fail to induce Cut (blue), indicating that the absence of Lqf blocks the signaling activity of Dl, even when co-overexpressed with Neur. (C',C'') High magnification of the boxed region in C, showing Dl staining (red) at apical and sub-apical planes of focus. As in B', Dl accumulates at both the surface and in cytosolic puncta; nevertheless, signaling is blocked, owing to the absence of Lqf.

View larger version (79K): [in a new window]   Fig. 3. DSL signaling depends on endocytosis, ubiquitination and Lqf. (A) Wing disc with clones of cells overexpressing Myc-tagged DlLDL+ (green) stained for Cut (red). Clones in the D compartment located close to the DV boundary activate Cut, indicating that DlLDL+ has signaling activity. The white box marks a clone that is shown at higher magnification in apical and sub-apical planes of focus in A'' and A''. Note the presence of cytosolic puncta in A''. In this, and all other images in this figure, the various Dl chimeric proteins were detected with guinea pig -Dl antisera to allow us to visualize Cut expression (which requires a mouse -Cut antisera). Hence, both endogenous Dl and the exogenous Dl chimera are detected. However, the exogenous Myc-tagged Dl chimeras are expressed at several-fold higher levels than the endogenous protein, so the contribution of the endogenous protein to the Dl stain is negligible. All of these experiments were also performed using mouse -Myc antisera, and identical results were obtained for the subcellular distribution of the various Myc-tagged Dl chimeras. (B) lqf- clones overexpressing Myc-tagged DlLDL+ located close to the DV boundary can induce Cut, indicating that the presence of the LDL internalization signal allows the chimeric Dl protein to bypass the requirement for Lqf. The subcellular distribution of Myc-tagged DlLDL+ appears to be unaffected by the absence of Lqf. (C) Clones of cells overexpressing Myc-tagged DlLDLm fail to activate Cut ectopically, and also block endogenous Cut expression when they abut the DV boundary. Myc-tagged DlLDLm accumulates on the apical surface, but not in cytosolic puncta, indicating that mutation of the LDL internalization signal blocks endocytosis as well as signaling activity. (D) Clones of cells overexpressing Myc-tagged DlR+ induce Cut; Myc-tagged DlR+ accumulates both apically and in cytosolic punta. (E) Clones of lqf- cells overexpressing Myc-tagged DlR+ do not induce ectopic Cut, and block normal Cut expression when they abut the DV boundary; nevertheless, the subcellular distribution of Myc-tagged DlR+ is not detectably altered by the absence of Lqf. (F) Clones of cells overexpressing Myc-tagged DlRm fail to activate Cut, and block normal Cut expression at the DV boundary. Myc-tagged DlRm accumulates on the apical surface, but not in cytosolic puncta. (G) Clones of cells overexpressing Myc-tagged DlUbi+ upregulate expression of the vg-boundary enhancer-lacZ reporter gene (red), although they do not activate Cut (not shown), indicating detectable, but weak, signaling activity. Myc-tagged DlUbi+ accumulates apically, as well as in cytosolic puncta. (H) Clones of cells overexpressing Myc-tagged DlUbim fail to ectopically upregulate the vg-boundary enhancer-lacZ reporter gene and block its normal expression when they abut the DV boundary. Myc-tagged DlUbim accumulates apically; however, accumulation in cytosolic puncta is greatly reduced when compared with Myc-tagged DlUbi+.

View larger version (62K): [in a new window]   Fig. 6. Lqf-dependent processing of Dl. Western blot analysis of Myc-tagged Dl co-overexpressed with Neur in clones of wild-type (lane 1) or lqf- (lane 2) cells; only cells within the clones overexpress MycDl and Neur (see Materials and methods; the same result was obtained in all of three independent experiments). In wild-type cells, two bands are found, which correspond in size, respectively, to full-length MYCDl (105 kDa) and a truncated form (50 kDa) which is smaller than the size expected for the Myc-tagged Dl extracellular domain (75 kDa), but larger than that expected for the complementary Myc-tagged portion of Dl consisting of the transmembrane and cytosolic domains (40 kDa). In lqf- cells, only the upper, apparently unprocessed, band is observed.

View larger version (22K): [in a new window]   Fig. 7. Role of endocytosis and Lqf in sending DSL signals. (A,B) Examples of two general models are shown, distinguished by whether the activation of Notch is triggered by the early events of DSL endocytosis leading up to pinching off of the coated vesicle (A), or is dependent on the recycling of DSL ligands (B). To accommodate our results, we envisage that the first model (A) would require Lqf (red) to be present, or active, in only a subset of coated pits or other structures that provide a specialized micro-environment (dark gray) necessary for productive interactions (pink scissors) between DSL ligands and Notch. Cargo proteins, including DSL ligands (colored gold), that carry only mono-Ubiquitin internalization signals would depend on Lqf to be recruited to these specialized structures. Other adapters (blue) would internalize mono-ubiquitinated DSL ligands via other structures that lack the necessary environment for productive interactions to occur. In the second model (B), Lqf and other adapters could co-exist in coated pits, with Lqf allowing mono-ubiquitinated cargo to gain access, subsequently, to a recycling pathway. Entry into this pathway would be essential for the conversion of nascent DSL proteins into active ligands, for example by proteolytic processing (not shown), which can then interact productively with Notch. In both models, introduction of other internalization signals, such as the LDL receptor signal, would allow DSL ligands to bypass the requirement for Lqf to enter the required surface structures or recycling pathways. EE, early endosome; RE, recycling endosome; LE/MVB, late endosome/multi-vesicular body; TGN, trans-Golgi network.