An isoprenylation and palmitoylation motif promotes intraluminal vesicle delivery of proteins in cells from distant species

Abstract

The C-terminal ends of small GTPases contain hypervariable sequences which may be posttranslationally modified by defined lipid moieties. The diverse structural motifs generated direct proteins towards specific cellular membranes or organelles. However, knowledge on the factors that determine these selective associations is limited. Here we show, using advanced microscopy, that the isoprenylation and palmitoylation motif of human RhoB (-CINCCKVL) targets chimeric proteins to intraluminal vesicles of endolysosomes in human cells, displaying preferential co-localization with components of the late endocytic pathway. Moreover, this distribution is conserved in distant species, including cells from amphibians, insects and fungi. Blocking lipidic modifications results in accumulation of CINCCKVL chimeras in the cytosol, from where they can reach endolysosomes upon release of this block. Remarkably, CINCCKVL constructs are sorted to intraluminal vesicles in a cholesterol-dependent process. In the lower species, neither the C-terminal sequence of RhoB, nor the endosomal distribution of its homologs are conserved; in spite of this, CINCCKVL constructs also reach endolysosomes in Xenopus laevis and insect cells. Strikingly, this behavior is prominent in the filamentous ascomycete fungus Aspergillus nidulans, in which GFP-CINCCKVL is sorted into endosomes and vacuoles in a lipidation-dependent manner and allows monitoring endosomal movement in live fungi. In summary, the isoprenylated and palmitoylated CINCCKVL sequence constitutes a specific structure which delineates an endolysosomal sorting strategy operative in phylogenetically diverse organisms.

Figure 1. Localization of CINCCKVL chimeras in human cells.

(A) Human fibroblasts were transfected with the indicated constructs and observed live by confocal microscopy after 16 h in serum-depleted medium. To the right of every condition, zoom-ins of original pictures and fluorescence intensity profiles along a section (see dotted lines marked by asterisks) are shown. The magnification used for zoom-ins is seven-fold with respect to the corresponding full-sized images. F.I., fluorescence intensity. (B) HeLa cells were transiently transfected with the indicated constructs and treated with 25 nM LTR for 15 min, where indicated, prior to live confocal microscopy observation. Bottom panels show cells that were treated with 10 µM chloroquine 24 h post-transfection for a further 24 h. Insets show details of the pictured cells. (C) Cells transfected as in (B) were visualized live by super-resolution microscopy (SIM). Insets show individual MVB and their ILV decorated with CINCCKVL fluorescent proteins. Scale bar, 20 µm.

Figure 2. Endolysosomal localization of CINCCKVL-chimeric proteins depends on posttranslational processing.

(A) HeLa cells were transiently transfected with the indicated constructs and 24 h later they were cultured in serum-free medium for another 24 h. Cell lysis and fractionation were achieved as described in the Experimental section. Upper panels show the amount of the constructs in total cell lysates and lower panels depict the levels of the constructs in S100 (soluble) and P100 (particulate) fractions, assessed by western blot with an anti-GFP antibody. Hsp90 was used as a loading control, RhoGDI as a marker of the soluble fraction and vimentin as a particulate fraction marker. Results are representative from three experiments with virtually identical results. The positions of the 25 and 100 kDa markers are shown for reference. (B) HeLa cells were transfected with GFP-8, its palmitoylation defective mutant (GFP-8-C240, 243S), its isoprenylation defective mutant (GFP-8-C244S) or GFP, as indicated, and stained with LTR prior to live confocal microscopy imaging. (C) GFP-8-transfected HeLa cells were treated with 20 µM 2-bromopalmitate for 6 h or 10 µM simvastatin for 24 h in serum-free medium and stained with LTR before live observation by confocal microscopy. (D) BAEC were transfected with Dendra-8 and treated with simvastatin for 24 h (left panel) prior to eliciting a green-to-red photoswitch using UV light (middle panel). Immediately after photoswitching simvastatin was removed and the localization of the red, photoconverted protein was assessed 24 h later (right panel). Scale bar, 10 µm.

Figure 3. Effects of agents modulating cholesterol synthesis and traffic on the distribution of CINCCKVL constructs.

(A) HeLa cells were transfected with GFP-8 (upper panels) or GFP-CD63 (lower panels). 24 h later cells were treated with 100 µM ZGA or 10 µM U18666A for 24 h and stained with LTR as above. Insets show enlarged areas of interest. The single channels corresponding to the areas in insets are shown below each image. (B) BAEC were transfected with GFP-8 and treated with ZGA or U18666A as described in (A). Insets show enlarged areas of interest and lower panels depict fluorescence intensity profiles along a section (see dotted lines marked by asterisks).

Figure 4. C-terminal sequences of CINCCKVL-chimeric proteins, RhoB homologs and related proteins from diverse species.

The C-terminal sequences of GFP-8 and tRFP-T-8 are shown on top, together with a schematic view of the lipidic modifications; palmitates are shown in black and the geranylgeranyl moiety in blue. The Pubmed accession numbers of genes coding for proteins bearing C-terminal sequences similar to that of RhoB are shown in the lower panel. Potential sites for palmitoylation are shown in red and the isoprenylation cysteine in green.

Figure 5. Localization of RhoB-related proteins in amphibian cells.

(A) Xenopus laevis RhoB (GFP-RhoB X) or GFP-8 were expressed in Xenopus laevis A6 cells and acidic compartments were stained with LTR. The right panel depicts the extent of co-localization of GFP and LTR signals shown as Pearson coefficients (x100) or co-localization rates (in percentages). Results are average values of at least 30 cells per condition ± standard error of mean (SEM). *p<1×10⁻⁶ vs GFP-8 by Student’s t-test. (B) Effect of agents altering lysosomal function on the distribution of tRFP-T-8. Xenopus A6 cells were co-transfected with Lamp1-GFP to mark lysosomes and tRFP-T-8, and treated with 10 µM U18666A or 10 µM chloroquine, as indicated. (C) Dependence of tRFP-T-8 localization on posttranslational modification in Xenopus cells. A6 cells were transfected with tRFP-T or tRFP-T-8 and treated with 10 µM simvastatin or 20 µM 2-bromopalmitate and imaged live by confocal microscopy.

Figure 6. Localization of tRFP-T-8 in High Five insect cells.

High Five cells were transfected with Lamp1-GFP and tRFP-T-8 and live cells were visualized by confocal microscopy. The overlays of single fluorescent z-sections alone and with the Differential Interference Contrast image are shown.

Figure 7. Localization of GFP-8 in Aspergillus nidulans.

Strains of Aspergillus nidulans expressing GFP-8 or its palmitoylation deficient mutant (GFP-8-C240, 243S) were imaged as described in Methods. (A and B) show co-localization of the constructs with the vacuole marker, CMAC. Insets show grayscale images for better contrast. (C) Kymograph showing the movements of various GFP-8-positive compartments. Rapidly moving endosomes (likely corresponding to early endosomes) are marked by arrowheads (red), static vesicles (likely corresponding to vacuoles) are marked by arrows and potential points of contact between endosomes and vacuoles are depicted by asterisks. (D) Graph representing velocities of individual endosomes.