Figure 1.
Half desmosome internalisation is cPKC and actin dependent.
(A) MDCK cells cultured in NCM had desmosomes at cell-cell contacts as indicated by DP staining. (B) Internalised rings of DP were present in cells treated for 60 minutes with LCM (arrows). (All internalisation controls in LCM were carried out in the presence of the appropriate drug vehicle.) (C) Internalisation of desmosomes was prevented by co-treatment with LCM and Gö6976 (0.8 µM), as cell contact was lost but half desmosomes remained at the cell surface giving rise to the appearance of intercellular gaps (arrows). (D) In NCM, DP (red) was localised to the cell surface in association with the surface marker con-A (green). (E) LCM treatment caused internalisation of DP and separation from con-A, which remained at the surface. (F) Co-treatment with LCM and latrunculin A (5 µM) inhibited desmosome internalisation, as indicated by persistent association of DP with con-A. Fluorescence profiles depict the intensity of staining along the white line in the images.
Figure 2.
Half desmosome intracellular transport is regulated by microtubules.
(A–C) Co-treatment with LCM and nocodazole (33 µM) (A), AMP-PNP (500 µM) (B) or ATA (50 µM) (C) caused DP to be internalised but to remain just beneath the cell surface (arrows). Confocal images of MDCK cultured in NCM and stained for KIFC3 (green) and DP (red) (D) which co-localise following 30minutes LCM treatment (E). This co-localisation was inhibited by LCM co-treatment with nocodazole (33 µM) (F). Yellow arrows indicate XZ axis. Bar, 5 µm. Fluorescence profiles depict the intensity of staining along the white line.
Figure 3.
Intermediate filaments are not involved in internalisation of half desmosomes.
(A) Western blot showing partial knockdown of DP in HaCaT cells transfected with 50 nM DP siRNA or compared to scrambled DP siRNA and mock transfected controls. (B) Single confocal slices of HaCaT cells or transfected with 50 nM scrambled or DP siRNA showing that the latter cells had disrupted intermediate filament organisation as indicated by keratin 8 & 18 staining. (C) HaCaT cells transfected with DP siRNA or scrambled siRNA were treated with either NCM or LCM for 1 hour and then stained for Dsg2. Half desmosomes were internalised in DP knockdown cells as in the controls. Bar, 5 µm.
Figure 4.
Internalised desmoplakin co-localises with the centrosome.
(A–C) Co-localisation of DP with three centrosome markers. Single confocal slices of MDCK cells stained for DP (red) and the centrosome markers aurora A (A) ninein (B) and γ-tubulin (C) (all green) following 2-3 hours of LCM treatment. Fluorescence profiles depict the intensity of staining along the white line in the merged images. (D–F) The time course of co-localisation. Representative images of desmoplakin and γ-tubulin localisation demonstrate the 3 categories of staining pattern used for quantification with DP surrounding the centrosome (D), beside the centrosome (E) or co-localised with the centrosome (F) during LCM treatment, data are mean values ± s.e.m. Yellow arrows indicate XZ axis. Bar, 5 µm. Asterisk indicates statistical significance (p<0.0286, Mann-Whitney test).
Figure 5.
Desmosomal proteins remain co-localised following internalisation.
(A–D) Co-localisation of the desmosomal proteins DP, PG, Dsc2a and Dsg2 in NCM persists following LCM-induced internalisation for 1 hour (A,B) and 24 hours (C,D). Yellow arrows indicate XZ axis. Bar, 5 µm. Fluorescence profiles depict the intensity of staining along the white line in the merged images. (E) Cells were treated with either NCM or LCM for 90 minutes and then separated into their insoluble (INS) and soluble (SOL) fractions. Western blots for desmosomal proteins (E) were quantified by densitometry (F) (dashed lines indicate lanes which have been re-ordered from the same western blot). Asterisk indicates statistical significance (p<0.0286, Mann-Whitney test). For further details see text.
Figure 6.
Internalised Dsg2 is not recycled to the cell surface.
(A) LCM-induced internalisation protects Dsg2 from trypsin. MDCK cells were treated with trypsin/EDTA after incubation LCM for 0, 30 or 60 minutes. Western blots show that cell surface Dsg2 (150 KDa) was degraded by trypsin/EDTA generating a membrane-protected cytoplasmic fragment (arrow) (LCM (mins) 0). (The band above the trypsin fragment and present in each lane is believed to be a natural degradation product of Dsg2.) However, after LCM treatment a substantial amount of full length Dsg2 remained after trypsin/EDTA treatment (LCM (mins) 30 and 60) showing that it had become membrane-protected. (B, C) The amount of membrane-protected Dsg2 (arrow) remains the same after induction of new desmosome formation following LCM treatment. Cells were treated either with LCM for 2 hours, or with LCM for 1 hour followed by NCM for 1 hour (total time 2 hours) to induce new desmosome formation (verified by immunofluorescence but not shown). The latter treatment is referred to as reverse calcium switching (REV). Western blots (B, quantified in C) show that the amount of membrane protected Dsg2 was identical after both treatments (2 hr LCM, 2 hr REV). Bar in C indicates s.e.m. (D) Internalised DP (arrows) does not colocalise with Rab11 following a reverse calcium switch (REV). Bar, 5 µm. Fluorescence profiles depict the intensity of staining along the white line in the images.
Figure 7.
Desmosomal proteins are degraded by lysosomes and proteasomes, and co-localise with the lysosomal marker lamp1.
(A) Internalised desmosomal proteins are gradually degraded. Western blots of whole MDCK cell lysates following LCM treatment for 0, 8, 16 or 24 hours show that Dsg2, DP and PG were gradually degraded. (B–D) Lysosomal and proteasomal degradation. Cells were treated with LCM for 16 or 24 hours in the presence of the noted inhibitors or vehicle alone (DMS0 or water). Western blots of whole cell lysates show that the lysosomal inhibitors bafilomycin A1 (250 nM), chloroquine (100 µM) and leupeptin (100 µM) and the proteasomal inhibitor MG132 (10 µM) inhibited degradation of Dsg2 (B) and chloroquine and bafilomycin A1 inhibited LCM-induced Dsg3 and PG degradation (C, D). (E) Internalised Dsg2 co-localises with the lysosomal marker lamp1. Cells cultured in NCM or treated with LCM for 16 hours stained for Dsg2 (red) and Lamp1 (green). Co-localisation in the latter cells is indicated by white arrows.Yellow arrows indicate XZ axis. Bar, 5 µm. Fluorescence profile depicts the intensity of staining along the white line in the image. (F) DP degradation was not inhibited by lysosomal inihibitors, but instead by the proteasomal inhibitors bortezomib (20–200 nM) and MG132 (10 µM) (dashed lines indicate lanes which have been re-ordered from the same western blot). Bortizomib had no effect on PG or Dsg2 degradation (F). (G) Western blots of whole cell lysates co-treated with LCM and nocodazole (33 µM) for 16 or 24 hrs shows degradation of DP is unaltered, whilst degradation of PG and Dsg2 is partially inhibited. Bar, 5 µm.
Figure 8.
Model of the internalisation, transport and degradation of desmosomal halves.
Treatment of calcium dependent desmosomes (A) with LCM causes loss of intercellular adhesion and formation of half desmosomes (B). These are internalised by a mechanism dependent on cPKC and actin filaments (C). Once internalised, desmosomal halves are transported in a microtubule/kinesin dependent manner to the centrosome and degraded by lysosomes and proteasomes (D).