Lethal (2) giant discs (Lgd)/CC2D1 is required for the full activity of the ESCRT machinery

Abstract

Background: A major task of the endosomal sorting complex required for transport (ESCRT) machinery is the pinching off of cargo-loaded intraluminal vesicles (ILVs) into the lumen of maturing endosomes (MEs), which is essential for the complete degradation of transmembrane proteins in the lysosome. The ESCRT machinery is also required for the termination of signalling through activated signalling receptors, as it separates their intracellular domains from the cytosol. At the heart of the machinery lies the ESCRT-III complex, which is required for an increasing number of processes where membrane regions are abscised away from the cytosol. The core of ESCRT-III, comprising four members of the CHMP protein family, organises the assembly of a homopolymer of CHMP4, Shrub in Drosophila, that is essential for abscission. We and others identified the tumour-suppressor lethal (2) giant discs (Lgd)/CC2D1 as a physical interactor of Shrub/CHMP4 in Drosophila and mammals, respectively.

Results: Here, we show that the loss of function of lgd constitutes a state of reduced activity of Shrub/CHMP4/ESCRT-III. This hypomorphic shrub mutant situation causes a slight decrease in the rate of ILV formation that appears to result in incomplete incorporation of Notch into ILVs. We found that the forced incorporation in ILVs of lgd mutant MEs suppresses the uncontrolled and ligand-independent activation of Notch. Moreover, the analysis of Su(dx) lgd double mutants clarifies their relationship and suggests that they are not operating in a linear pathway. We could show that, despite prolonged lifetime, the MEs of lgd mutants have a similar ILV density as wild-type but less than rab7 mutant MEs, suggesting the rate in lgd mutants is slightly reduced. The analysis of the MEs of wild-type and mutant cells in the electron microscope revealed that the ESCRT-containing electron-dense microdomains of ILV formation at the limiting membrane are elongated, indicating a change in ESCRT activity. Since lgd mutants can be rescued to normal adult flies if extra copies of shrub (or its mammalian ortholog CHMP4B) are added into the genome, we conclude that the net activity of Shrub is reduced upon loss of lgd function. Finally, we show that, in solution, CHMP4B/Shrub exists in two conformations. LGD1/Lgd binding does not affect the conformational state of Shrub, suggesting that Lgd is not a chaperone for Shrub/CHMP4B.

Conclusion: Our results suggest that Lgd is required for the full activity of Shrub/ESCRT-III. In its absence, the activity of the ESCRT machinery is reduced. This reduction causes the escape of a fraction of cargo, among it Notch, from incorporation into ILVs, which in turn leads to an activation of this fraction of Notch after fusion of the ME with the lysosome. Our results highlight the importance of the incorporation of Notch into ILV not only to assure complete degradation, but also to avoid uncontrolled activation of the pathway.

Keywords: CC2D1A; CC2D1B; CHMP4B; ESCRT; Endosomal pathway; Lethal (2) giant discs; Lgd; Notch pathway; Shrub.

Fig. 1

A fraction of Notch remains at the LM of lgd mutant MEs and is activated. (A, A’) Dx and Su(dx) act antagonistically to determine the fraction of Notch remaining at the LM of the ME. Over-expression of Su(dx) or loss of dx function favours incorporation of Notch into ILVs. (B–C’) Expression of the Notch target Wg and the Notch activity reporter Gbe+Su(H) in wild-type and lgd mutant discs. The expression domain of Wg along the D/V boundary of the wing anlage (arrow in B) is dramatically expanded in lgd mutant discs (arrow in C). In addition, the complex expression pattern of Gbe+Su(H) seen in wild-type is replaced by a uniform staining in the mutant discs (B’, C’). (D, D’) Over-expression of Su(dx) with hhGal4 in the posterior compartment normalises the expression of both read-outs of Notch activity (arrow). A similar effect is seen if the function of dx is removed in lgd mutant discs (E, E’, arrow). Note that the normal expression of the Notch targets, e.g. along the D/V boundary (arrow in D–E’), is not affected by the manipulations. Scale bars (B–E’) 200 μm. At least ten wing imaginal discs were analysed for each genotype

Fig. 2

lgd mutant MEs have a normal ILV content. (A–E) Ultra-structural analysis of lgd mutant and lgd cells rescued by additional genomic copies of shrub (lgd^d7; BAC^shrub). Representative TEM pictures of MVBs of wild-type (A), lgd^d7 mutant (B) and rescued lgd^d7 mutant (C) wing disc cells. (B”) As an example, colourized overlay of determined ILV area (46.9% of total MVB area) with the original electron microscopy image of the MVB is shown in (B’). Highlighted in bright red is the area which is not considered as ILV content/electron-dense material within the MVB. (D) Statistical analysis of MVBs of wild-type (blue), lgd mutant (red) and rescued lgd mutant (green) cells. The average area of lgd mutant MVBs is significantly increased in comparison to MVBs of wild-type cells. Furthermore, the enlargement of endosomes in lgd mutants is partially rescued in lgd mutant cells expressing two additional copies of BAC-shrub. (E) Pixelwise quantification of electron-dense material per area within the lumen of wt (blue) and lgd^d7 mutant (red) MVBs to measure ILV formation (as shown in B’, B”). A self-written macro for the image processing software “Fiji” was used (see the “Methods” section and [30]). Area quotients of several MVBs are collected and blotted in a box blot. The ILV content in lgd^d7 mutant is not changed in comparison to the wild-type [(D, E) wt n = 212 MVBs; lgd^d7 n = 230 MVBs; lgd^d7; BAC^shrub n = 277; (D) Kruskal-Wallis test, Dunn’s multiple comparison test p < 0.001 (***), p < 0.01 (**); (E) Mann-Whitney test, two-tailed; box plots: whiskers: 5–95 percentile; mean shown as “+”]. Scale bars (A–C) 250 nm

Fig. 3

The size of ESCRT-containing electron-dense microdomains is enlarged in lgd mutant MEs. Representative TEM pictures of MVBs of wild-type (A, A’) and lgd^d7 mutant (B, B’) wing disc cells. (A’, B’) Defined and measured ESCRT-containing electron-dense microdomains of the MVBs shown in A and B are highlighted in red. Magnification of ESCRT-containing electron-dense microdomains where the associated, nascent ILV is clearly observable (arrowhead). The overall morphology of the MVBs in lgd^d7 mutants appears to be normal (B, B’). (C) Statistical analysis of the ESCRT-containing electron-dense microdomains of wild-type and lgd^d7 mutant MVBs in wing imaginal disc cells. The average size of lgd mutant electron-dense microdomains is significantly increased in comparison to the wild-type [wt n = 72 MVBs (90 ESCRT-containing electron-dense microdomains); lgd^d7 n = 87 MVBs (111 ESCRT-containing electron-dense microdomains); Mann-Whitney test, two-tailed p < 0.0001 (****); scatter dot plot, mean shown as a “line”]. (D–H) Analysis of MEs of cells that are depleted of Rab7 function. (D, D’) A wing disc depleted of Rab7 function in the posterior compartment by the continuous expression of rab7-RNAi with hhGal4 (arrow). The magnification in (D’) highlights the efficiency of the depletion in the posterior compartment (arrow). (E, E’) Representative MEs/MVBs of anterior wild-type and posterior depleted cells. Analysis of the size (F), luminal content (G) and size (H) of the ESCRT-containing electron-dense microdomains of the wild-type and Rab7-depleted MEs/MVBs. While the average size and the luminal content is increased, there is no change in the size of the ESCRT-containing electron-dense microdomains in comparison to wild-type [(F, G) wt (ant.) n = 253 MVBs; (F) rab7-RNAi (post.) n = 282 MVBs; (G) rab7-RNAi (post.) n = 280 MVBs; (F, G) Mann-Whitney test, two-tailed p < 0.0001 (****); box plots: whiskers: 5–95 percentile; mean shown as “+”. (F) Values greater than 1 μm² are not shown (rab7-RNAi (post.): 8 out of 282 MVBs). (H) wt (ant.) n = 86 MVBs (114 ESCRT-containing electron-dense microdomains); rab7-RNAi (post.) n = 92 MVBs (110 ESCRT-containing electron-dense microdomains) Mann-Whitney test, two-tailed; scatter dot plot; mean shown as a “line”]. Scale bars (A–B’, E, E’) 250 nm; (D) 200 μm; (D’) 50 μm. (D, D’) At least ten wing imaginal discs were analysed for each genotype

Fig. 4

Reduction of shrub function causes a phenotype that resembles that of lgd mutants. (A) Expression of Gbe+Su(H) in a control wild-type disc. (B–C’) Depletion of shrub function by expression of UAS shrub-RNAi with hhGal4 tubGal80^ts for 32 h (in the posterior compartment). The expression of Gbe+Su(H) is significantly increased in the depleted posterior area (B, C). The depleted cells accumulate Notch in enlarged endosomes (B’), but the expression of E-cad is not affected (B”). (C, C’) Z-stack of the disc shown in (B–B”) to show the apical-basal axis of the disc cells. The comparison of the cells of the anterior with cells of the posterior compartment reveals that the subcellular localisation of E-cad is not affected by the depletion of shrub. (D–H) Rescue of shrub null mutants by BAC^shrub-CHMP4B. (D) Expression of Gbe+Su(H) in a control wild-type disc. (E) A shrub mutant disc completely rescued by one copy of the control BAC^shrub-cDNA construct. (F) A similar rescue is observed with two copies of BAC^shrub-CHMP4B. (G) One copy of BAC^shrub-CHMP4B leads only to a partial rescue. Gbe+Su(H) is slightly ectopically expressed. It resembles the expression of Gbe+Su(H) in a hypomorphic situation of lgd shown in (H). Scale bars (A–B”) 50 μm; (D–H) 200 μm. At least ten wing imaginal discs were analysed for each genotype

Fig. 5

Comparison of shrub clones in the absence or presence of one copy of BAC^shrub-CHMP4B in the genome with lgd^d7 mutant cell clones. Clones are labelled by the absence of GFP. (A–E) shrub mutant cell clones are rare and comprise only a few cells. (A, B) Overview of the notal region of a wing disc bearing a shrub mutant clone. (C–E) Magnification of the region boxed in (A, B). A small shrub mutant clone is boxed and shown at higher magnification in the insert. The few surviving mutant cells contain enlarged Notch-positive MEs (E). In contrast, large orphan wild-type twin-clones (GFP-positive (arrow in A, B and outlined in yellow D, E) are present in these discs. (F–H) shrub mutant cell clones partially rescued with one copy of BAC^shrub-CHMP4B. (F) Several large mutant clones can be observed, indicating that the lethality is suppressed. (G, H) Magnification of the region boxed in (F) which includes a large clone. The partially rescued cells of the clone display a similar endosomal defect as the cells of lgd mutant clones, shown in (I–K). In both cases, the enlarged MEs of the mutant cells accumulate Notch. (K, K’) Note that the levels of Shrub in the lgd mutant cells are similar to that of the neighbouring wild-type cells. Scale bars (A, B, F) 50 μm, (C–E, G–K) 10 μm. At least ten wing imaginal discs were analysed for each genotype

Fig. 6

Reduction of shrub activity causes moderate enlargement of MVBs. (A–D) Ultra-structural analysis of shrub mutant cells partially rescued by BAC^shrub-CHMP4B (shrub^4-1; BAC^shrub-CHMP4B/+). Representative TEM pictures of MVBs of wild-type (A) and partially rescued shrub mutant (B) wing disc cells. (B”) Colourized overlay of estimated ILV area (50.6% of total MVB area) with the original EM image of the MVB shown in (B’). The area that is not considered as ILV content/electron-dense material within the MVB is highlighted in bright red. (C) Statistical analysis of the MVBs of wild-type (blue) and partially rescued shrub mutant (red) cells. The average area of MVBs in shrub mutant cells is significantly increased. (D) Quantification of electron-dense material per area within the lumen of wt (blue) and partially rescued (red) MVBs to measure ILV formation (as shown in B’). There is no change in the ILV content in the partially rescued shrub mutant cells in comparison to wild-type [(C, D) wt n = 337 MVBs; (C) shrub^4-1; BAC^shrub-CHMP4B/+ n = 271; (D) shrub^4-1; BAC^shrub-CHMP4B/+ n = 263; (C) Mann-Whitney test, two-tailed p < 0.0001 (****). (C) Values greater than 0.5 μm² are not shown (wt, 1 out 337 MVBs; shrub^4-1; BAC^shrub-CHMP4B/+, 10 out of 271 MVBs). (D) Unpaired t test, p < 0.05 (*) (box plot: whiskers: 5–95 percentile; mean shown as “+”)]. Scale bar (A, B, B’) 250 nm

Fig. 7

Rescue of lgd mutants by extra copies of BAC^shrub or BAC^vps20. (A) Time of death of lgd mutants rescued with one or two copies of Bac^shrub, Bac^shrub-mut2 or Bac^vps20. (B–C’) The addition of one or two copies of BAC^shrub results in the normalisation of the expression patterns of Wg and Gbe+Su(H) (compare with Fig. 1B–C’). (D, D’) The addition of one copy of BAC^vps20 does not modify the lgd mutant phenotype (compare with Fig. 1C, C’), while the presence of two extra copies results in premature death during development (A, E, E’). At least ten wing imaginal discs were analysed for each genotype

Fig. 8

Lgd cycles between the cytosol and the LM of the ME. (A’–C”) Depletion of Rab7 in the posterior compartment of a wing disc using an UAS rab7-RNAi construct continuously driven by hhGal4. (A’–C’) Magnification of yellow outlined area, located in the wild-type anterior compartment and (A”–C”) of the white box, located in the posterior Rab7-depleted compartment of the same wing discs shown in (A–C). The arrows point to Notch-positive MEs that are associated with Lgd-RFP. (D–F’) RNAi mediated depletion of vps4 function in the posterior compartment of the wing imaginal disc. hhGal4 tubGal80^ts was used to drive expression of UAS vps4-RNAi for 24 h. The depletion results in an enlargement of Notch-positive MEs. Magnification of the region boxed in (D–F) is shown in (D’–F’). (G–I’) Depletion of shrub-RNAi for 38 h using enGal4 tubGal80ts leads to localisation of Lgd at enlarged ME. Magnification of the area boxed in (G–I) is shown in (G’–I’). (J–L’) Expression of the dominant-negative Shrub-GFP with hhGal4 for 15 h. Lgd-HA accumulates on the enlarged MEs together with Shrub-GFP. Magnification of the region boxed in (J–L) is shown in (J’–L’). Scale bars (A–L) 10 μm. At least ten wing imaginal discs were analysed for each genotype