Skip to main page content
U.S. flag

An official website of the United States government

Dot gov

The .gov means it’s official.
Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

Https

The site is secure.
The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2023 Sep 4;222(9):e202204020.
doi: 10.1083/jcb.202204020. Epub 2023 Aug 1.

Early proteostasis of caveolins synchronizes trafficking, degradation, and oligomerization to prevent toxic aggregation

Frederic Morales-Paytuví  1 , Alba Fajardo  1 , Carles Ruiz-Mirapeix  1 , James Rae  2 , Francesc Tebar  3 , Marta Bosch  1   3 , Carlos Enrich  3 , Brett M Collins  2 , Robert G Parton  2   4 , Albert Pol  1   3   5
Affiliations

Early proteostasis of caveolins synchronizes trafficking, degradation, and oligomerization to prevent toxic aggregation

Frederic Morales-Paytuví et al. J Cell Biol. .

Abstract

Caveolin-1 (CAV1) and CAV3 are membrane-sculpting proteins driving the formation of the plasma membrane (PM) caveolae. Within the PM mosaic environment, caveola assembly is unique as it requires progressive oligomerization of newly synthesized caveolins while trafficking through the biosynthetic-secretory pathway. Here, we have investigated these early events by combining structural, biochemical, and microscopy studies. We uncover striking trafficking differences between caveolins, with CAV1 rapidly exported to the Golgi and PM while CAV3 is initially retained in the endoplasmic reticulum and laterally moves into lipid droplets. The levels of caveolins in the endoplasmic reticulum are controlled by proteasomal degradation, and only monomeric/low oligomeric caveolins are exported into the cis-Golgi with higher-order oligomers assembling beyond this compartment. When any of those early proteostatic mechanisms are compromised, chemically or genetically, caveolins tend to accumulate along the secretory pathway forming non-functional aggregates, causing organelle damage and triggering cellular stress. Accordingly, we propose a model in which disrupted proteostasis of newly synthesized caveolins contributes to pathogenesis.

PubMed Disclaimer

Conflict of interest statement

Disclosures: The authors declare no competing interests exist.

Figures

None
Graphical abstract
Figure 1.
Figure 1.
Structure of CAV1 and modeling of CAV2 and CAV3. (A) Sequence alignment of human CAV1, CAV2, and CAV3. Regions of the structure are indicated above and colored: SM, signature motif (red); SD, scaffolding domain (green); and IMD, intermembrane domain (purple); PM, pin motif (yellow); SR, spoke region (gray); and beta-strand (cyan). The “DFE” sequence of CAV1 and CAV3 mediating COPII binding and ER export is highlighted by asterisks (*). The site of pathogenic Pro to Leu mutation (P132L in CAV1 and P104L in CAV3) is indicated by a red asterisk. (B) Structural alignment of an individual monomer of CAV1 from AlphaFold2 predictions. The structure is colored using the same scheme as Porta et al. (2022) with the positions of regions labeled as in A. The previously proposed oligomerization domain (OD) contains the SM and SD. (C) Structure of the CAV1 11 subunit oligomer determined by cryoEM (Porta et al., 2022; see AlphaFold prediction in Fig. S1). Each CAV1 protomer is shown in cartoon representation in a different color. Note that only residues from 49–178 were resolved in the cryoEM maps. (D) CAV1 cryoEM structure shown in surface representation colored according to hydrophobicity. The CAV1 coordinates were docked into a model lipid bilayer using PyMOL to show the scale of the protein complex compared to the membrane, and to visualize the proposed mechanism of membrane docking and insertion into the cytoplasmic leaflet. (E and F) Five subunits of full-length CAV3 modeled with AlphaFold2. In E, protomers are colored according to the pLDDT scores, and in F, according to the domains depicted in A and B. The CAV3 (40DFE42) inter-subunit interactions are detailed. The central Phe (F) side chain is stacked within a hydrophobic pocket while the flanking Asp (D) and Glu (E) side chains are shielded by the PM motif of the adjacent protomer.
Figure S1.
Figure S1.
AlphaFold2 predictions of pentameric homooligomers of human CAV1, CAV2, and CAV3. Supplemental for Fig. 1. (A) The top-ranking structures of homo-pentameric models of human CAV1, CAV2, and CAV3 predicted by AlphaFold2 are shown in the ribbon diagram and colored according to the pLDDT scores. The right-hand panels show the plots of the Predicted Alignment Error (PAE) for each top-ranking model. There is a strong degree of correlation between the five chains in each structure indicating their physical association with each other.
Figure 2.
Figure 2.
Differential proteostasis of newly synthesized caveolins. (A) Experimental design for studying the early proteostasis of caveolins. This protocol has been used throughout the work for the transfection (1 μg of DNA), expression (for 3 h), and analysis of caveolins and mutants, unless otherwise specified. (B and C) N-terminally GFP-tagged CAV1, CAV2, or CAV3 were transfected in parallel and after 3 h protein steady-state levels were determined in cell lysates (equal protein concentration) by immunoblotting (IB) with anti-GFP antibodies (B). Band intensity was quantified by densitometry, corrected with respect to the GAPDH levels determined by IB in each sample, and expressed relative to the GFP levels in CAV1 transfected cells (C; n ≥ 5 independent experiments). (D–H) The intracellular distribution of GFP-caveolins after 3 h was analyzed by confocal microscopy using specific antibodies against organellar markers; Cavin-1 (PM caveolae), TGN46 (GC), Calreticulin (ER), and Plin-3 (LDs). Representative images are included (D–G and Fig. S2). GFP-caveolin colocalization with each organelle was quantified using the Manders M2 overlapping coefficient (n ≥ 3 independent experiments with at least seven cells per condition; H). (I and J) Cells transfected with the indicated proteins were homogenized after 3 h with 1% Triton X-100 (4°C) and soluble and insoluble fractions separated by centrifugation. Protein distribution was analyzed in equal volumes of each fraction by IB with anti-GFP antibodies. (I and J) A representative result (I) and the relative distribution (J) of the protein in both fractions (left panel), and the relative insoluble fraction of CAV2 and CAV3 when compared with CAV1 (right panel; n = 3 independent experiments). (K–P) Transfected cells were solubilized after 3 h with either 0.5% Triton X-100 (TX) or 0.2% TX-100 + 0.4% SDS (TX+SDS), loaded at the top of sucrose density gradients, and fractionated by centrifugation according to their molecular weight. Protein distribution was quantified in equal volumes of each fraction by IB with anti-GFP antibodies. (K and M) show a representative result and (L–P) the relative distribution of the proteins in each fraction of the gradients quantified by densitometry (n ≥ 3 independent experiments). (Q and R) Transfected cells were homogenized after 3 h by nitrogen cavitation in the absence of detergents, loaded at the bottom of sucrose gradients, and fractionated according to their buoyant density by centrifugation for 1 h. Protein distribution was calculated in equal volumes of each fraction by IB with anti-GFP antibodies (n = 3 independent experiments). Upon these conditions, LDs float into the top fraction, as corroborated by enrichment in endogenous Plin-3. (Q and R) A representative result (Q) and the relative enrichment (R) of each caveolin on LDs. All graphs show means ± SD; ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 calculated in a one-way ANOVA (C, J, and R) or two-way ANOVA tests (H, L, N, O, P, and Q). Scale bars are 20 μm. Source data are available for this figure: SourceData F2.
Figure S2.
Figure S2.
Differential early transport of newly synthesized caveolins. Supplemental for Fig. 2. (A–H) Colocalization of GFP-CAV1, CAV2, and CAV3 with antibodies recognizing Cavin-1 (PM caveolae), TGN46 (GC), Calreticulin (ER), and Plin-3 (LDs) and analyzed by confocal microscopy. (I) Distribution of GFP-caveolins in C2C12 myoblasts after 3 h. The location of caveolin in the PM (open arrowhead), GC (arrowhead), ER (open arrow), and LDs (arrow) is indicated. (J) Gain of the insolubility of caveolins in C2C12 cells after a 3 h expression as explained in Fig. 2 I. Source data are available for this figure: SourceData FS2.
Figure 3.
Figure 3.
Proteasomal degradation of newly synthesized caveolins. (A–C) The indicated proteins were expressed (as in Fig. 2 A) in control (CTL) or in cells additionally treated with MG132 (MG). Protein steady-state levels determined after 3 h by IB with anti-GFP antibodies (A; as in Fig. 2 B). Proteins levels were referred to the expression of each protein in the absence of MG132 (red line; B; n ≥ 4 independent experiments). CAV1 K*R and CAV3 K*R steady-state levels were compared to CTL and MG132-treated cells and referred to the wt caveolin (C; n ≥ 3 independent experiments). (D–G) CAV1 K*R and CAV3 K*R were expressed for 3 h, and their intracellular distribution was analyzed by confocal microscopy. D and E show representative images of the mutants and their co-localization with Calreticulin (ER marker). Arrows indicate punctate structures formed by the mutants and are especially apparent along the ER. Additional images are included in Fig. S3. F and G show the colocalization of the mutants with markers of each compartment in confocal microscopy images quantified using the Manders M2 overlapping coefficient (n ≥ 3 independent experiments and at least nine cells per condition). (H–O) The indicated proteins were expressed in CTL cells or, in the case of CAV1 and CAV3, in cells additionally treated with MG. After 3 h, cells were solubilized with either TX or TX+SDS and fractionated in sucrose sedimentation gradients (as in Fig. 2, K and M). Protein distribution was determined in equal volumes of each fraction by IB with anti-GFP antibodies. H and K show a representative result and (I–O) the relative distribution of the proteins in each fraction of the gradients (n = 3 independent experiments). All graphs show means ± SD; ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 calculated in a one-way ANOVA test (B and C) and two-way ANOVA test (F, G, I, J, L, and O). Scale bars are 20 μm. Source data are available for this figure: SourceData F3.
Figure S3.
Figure S3.
Proteasomal degradation determines trafficking of newly synthesized caveolins. Supplemental for Fig. 3. (A–D) CAV1 and CAV3 (in CTL or in cells additionally treated with MG132) and CAV1 K*R and CAV3 K*R were expressed for 3 h. Proteins were colocalized with antibodies recognizing Cavin-1 (PM caveolae), TGN46 (GC), Calreticulin (ER), and Plin-3 (LDs) and analyzed by confocal microscopy. Scale bars are 20 μm.
Figure 4.
Figure 4.
Molecular determinants for the retention of CAV3 in the ER. (A–C) GFP-tagged alpha-CAV1 and beta-CAV1 (lacking the initial 32 amino acids of alpha) were expressed for 3 h, and steady-state protein levels were determined by IB with anti-GFP antibodies as in Fig. 2 B (A; n = 3 independent experiments). Protein distribution was assessed by confocal microscopy as in Fig. 2 H (B), and the gain of insolubility determined as in Fig. 2, I and J (C). (D) The N-terminal domains of beta-CAV1 and CAV3 are aligned, relevant differences in amino acids are indicated in red, and equivalent changes in green. The DFE motif binding COPII is underlined. The hydrophobicity (H), the hydrophobic moment (μH), and the tendency to fold as an alpha helix of the hypervariable domains of both proteins are depicted (hydrophobic residues in yellow and charged amino acids in red and blue). (E–H) The initial 14 amino acids of beta-CAV1 were exchanged by the first 16 residues of CAV3 to generate CAV1 with the N-terminus of CAV3 (CAV1 N-CAV3) and CAV3 N-CAV1. The mutants were expressed for 3 h, and their distribution was analyzed by confocal microscopy as in Fig. 2 D. (E and F) show representative images of the mutants and their co-localization with Calreticulin (ER marker) or Cavin-1 (PM marker). Additional images are included in Fig. S4. (G and H) show the colocalization of the mutants with organellar markers as in Fig. 2 H and the quantification using the Manders M2 overlapping coefficient (n ≥ 3 independent experiments and at least eight cells per condition). (I and J) The indicated proteins were transfected for 3 h and the steady state levels of wt and mutants determined by IB with anti GFP antibodies as in Fig. 2 B. The insolubility to 1% TX of the mutants was determined as in Fig. 2, I and J. Arrows in J indicate the bands showing reduction or gain of insolubility of the mutants. The graph below shows the quantification of the protein levels with GFP antibodies and densitometry with respect to the wt proteins calculated as in Fig. 2, I and J (n ≥ 3 independent experiments). All graphs show means ± SD; ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 in a two-tailed unpaired t test (A and C), one-way ANOVA (I and J), or two-way ANOVA test (G and H). Scale bars are 20 μm. Source data are available for this figure: SourceData F4.
Figure S4.
Figure S4.
Caveolin N-terminus determines the localization of newly synthesized caveolins. Supplemental for Fig. 4. (A–F) CAV1 N-CAV3 and CAV3 N-CAV1 were expressed for 3 h. Proteins were colocalized with antibodies recognizing Cavin-1 (PM caveolae), TGN46 (GC), Calreticulin (ER), and Plin-3 (LDs) and analyzed by confocal microscopy. Scale bars are 20 μm.
Figure 5.
Figure 5.
Impaired arrival of CAV1 to late-GC avoids oligomerization. (A) Amino acid sequence of the C-terminal domains of wt CAV1 and the CAV1- P158PfsX22 mutant (CAV1-P158). The newly generated ER-retrieval signal is indicated with red letters. (B) GFP-tagged CAV1-P158 was expressed for 3 h in CTL cells or in cells additionally treated with MG132. Protein levels were determined by densitometry as in Fig. 2 B (n = 3 independent experiments). (C and D) GFP-CAV1 and GFP-CAV1-P158 were expressed for 3 h, and their intracellular distribution analyzed by confocal microscopy. C shows representative images of the mutants and their co-localization with Plin-3 (LD marker). Additional images are included in Fig. S5. D shows the colocalization of CAV1-P158 with organellar markers as in Fig. 2 H (n ≥ 3 independent experiments and at least eight cells per condition). (E–G) CAV1-P158 was expressed for 3 h in CTL cells or in cells additionally treated with MG. Cells were solubilized with either TX or TX+SDS and fractionated in sucrose sedimentation gradients and protein distribution was analyzed as in Fig. 2 K with anti-GFP antibodies. E shows a representative result, and F and G show the relative distribution of the protein in each fraction of the gradient (n = 3 independent experiments). (H) CAV1 wt and CAV1-P158 were transfected in parallel and expressed for 16 h. The ER stress of cells was evaluated by measuring expression of XBP1 and DDIT3 by qPCR. Results are referred to the expression of these stress markers in cells transfected with CAV1 (n = 4 independent experiments). All graphs show means ± SD; ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 in a two-tailed paired t test (B and H) or a two-way ANOVA test (D, F, and G). Scale bars are 20 μm. Source data are available for this figure: SourceData F5.
Figure S5.
Figure S5.
Intracellular distribution of caveolin pathogenic mutants. Supplemental for Figs. 5 and 6. (A–C) CAV1-P158, CAV1-P132L, and CAV3-P104L were expressed for 3 h. Proteins were colocalized with antibodies recognizing Cavin-1 (PM caveolae), TGN46 (GC), Calreticulin (ER), and Plin-3 (LDs) and analyzed by confocal microscopy. (D–I) Mutants were expressed for 24 h. Cells were solubilized in TX and fractionated in sedimentation gradients as in Fig. 2 K. The IB shows representative experiments and the graphs the average distribution when compared to the 3 h expression (n ≥ 3 independent experiments). All graphs show means ± SD; ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 in a two-way ANOVA test (E, G, and I). Scale bars are 20 μm. Source data are available for this figure: SourceData FS5.
Figure 6.
Figure 6.
Oligomerization defective caveolin mutants aggregate and cause ER stress. (A) The scheme shows a detail of the position and interactions of the Pro 104 of human CAV3 within the oligomer predicted by AlphaFold2 (shown in ribbon diagram and colored according to the pLDDT score; Fig. 1 for additional details). (B and C) GFP-tagged CAV1-P132L (B) and CAV3-P104L (C) were expressed for 3 h in untreated cells or in cells additionally treated with MG132 and protein steady levels determined as in Fig. 2 B (n = 3 independent experiments). (D–G) The distribution of the mutants after 3 h was analyzed by confocal microscopy. (D and F) show representative images of the mutants and their colocalization with TGN36 (GC marker). Additional images are included in Fig. S5. (E and G) show the colocalization of the mutants with organellar markers calculated as in Fig. 2 H (n ≥ 3 independent experiments and at eight cells per condition). (H–O) Cells were transfected with GFP-tagged CAV1-P132L or CAV3-P104L and expressed in CTL cells or cells additionally treated with MG132. After 3 h cells were solubilized with either TX or TX+SDS and fractionated in sucrose sedimentation gradients as in Fig. 2, K and M. Protein distribution was quantified in equal volumes of each fraction by IB with anti-GFP antibodies. (H and L) shows a representative result and (I, J, M, and N) the relative distribution of each protein in each fraction of the gradient calculated by densitometry (n = 3 independent experiments). (K and O) CAV1 and CAV1-P132L or CAV3 and CAV3-P104L were transfected in parallel for 16 h and the ER stress evaluated by measuring expression of XBP1 and DDIT3 by qPCR. Results are referred to the expression of the stress markers in cells transfected with the wt caveolins (n ≥ 4 independent experiments). All graphs show means ± SD; ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 in a two-tailed paired t test (B and C, and stress in K and O) two-way ANOVA test (E and G) and sedimentation gradient in I, J, O, and P. Scale bars are 20 μm. Source data are available for this figure: SourceData F6.
Figure 7.
Figure 7.
Impaired proteostasis of newly-synthesized caveolins causes organellar damage. (A) CAV1 and CAV1 K*R or CAV3 and CAV3 K*R were transfected in parallel for 16 h and the ER stress evaluated by measuring expression of XBP1 and DDIT3 by qPCR. Results are referred to the expression of the stress markers in cells transfected with the wt caveolins (n ≥ 3 independent experiments). The graphs show means ± SD; ns, not significant; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 in a two-way ANOVA test. (B–I) Cells transfected with GFP-CAV1 (B and D–F) or GFP-CAV1 K*R (C and G–I) were incubated with Mitotracker (red) and fixed for CLEM. The figure includes single confocal slices (B for CAV1 and C for K*R) and the overlays of the fluorescently labeled images over the corresponding EM sections (D and E for CAV1 and G and H for K*R). The squares indicate the areas selected for high magnification panels shown next. GFP-CAV1 shows labeling of puncta throughout the cell (D) whereas GFP-CAV1 K*R is very concentrated in the perinuclear area (G). GFP-CAV1 labels typical Golgi cisternae (E and F). In contrast, high labeling for GFP-CAV1 K*R is associated with abnormal small vesicles/tubules clustered in the perinuclear area of the cell, demonstrating the disruption of Golgi complex structure (H and I). Bars for low- and high-magnification panels are indicated.
Figure 8.
Figure 8.
Early proteostasis of caveolins and disruption by pathogenic mutants. Caveolin protomers (black circles) are co-translationally inserted into the ER through the central hydrophobic domain in a signal recognition particle (SRP)-dependent manner. CAV1 is rapidly transported to the GC in COPII vesicles. CAV3 is retained for longer times in the ER and LDs than CAV1, with fewer CAV3 protomers moving to the GC. Because the COPII binding residues of caveolins are completely hidden within large oligomeric complexes, only monomeric or small oligomers of caveolins can be transported from the ER to the cis-GC. Caveolins are submitted to a robust degradation in the ER (dashed circles), a process that prevents the tendency of caveolins to aggregate (see below). Once in the GC, probably in medial- or late-GC, and likely by finding a specific lipid environment, caveolins oligomerize, initially into 8S oligomers (11 protomers) and next into 70S oligomers (∼144 protomers), which are rapidly transported to the PM to assemble mature caveolae. Active proteasomal degradation is needed for efficient caveolin trafficking, likely by avoiding the aggregation of caveolins, a process also observed in artificial membranes and probably masking key sorting motifs. Degradation-insensitive mutants (K*R, red circles) have a marked tendency to form aggregates and display trafficking defects; slow transport to the PM for CAV1 K*R and to LDs in the case of CAV3 K*R. Similarly, wt caveolins display trafficking defects when the proteasome is chemically inhibited. The K*R mutants trigger ER stress and disrupt secretory pathway organelles. Caveolin pathogenic mutants unable to oligomerize, such as CAV1-P132L (red squares), accumulate in the GC. Although subjected to degradation, these mutants also progressively form aggregates and trigger ER stress. The aggregates formed by mutants have the capacity to retain wt caveolins (black to red circles) and reduce caveola formation. Equivalent mutants of CAV3, such as CAV3-P104L, also accumulate and aggregate in the ER and trigger stress. In support of this model, caveolin mutants unable to reach the late-GC, such as CAV1-P158, fail to form functional oligomers, aggregate, and cause stress. When early proteostatic mechanisms fail, for example in the environment of unhealthy or senescent cells, instead of forming PM caveolae, wt caveolins could be toxic proteins and additional stressors.
See this image and copyright information in PMC

Similar articles

See all similar articles

References

    1. Aslanyan, M.G., Doornbos C., Diwan G.D., Anvarian Z., Beyer T., Junger K., van Beersum S.E.C., Russell R.B., Ueffing M., Ludwig A., et al. . 2023. A targeted multi-proteomics approach generates a blueprint of the ciliary ubiquitinome. Front. Cell Dev. Biol. 11:1113656. 10.3389/fcell.2023.1113656 - DOI - PMC - PubMed
    1. Balchin, D., Hayer-Hartl M., and Hartl F.U.. 2016. In vivo aspects of protein folding and quality control. Science. 353:aac4354. 10.1126/science.aac4354 - DOI - PubMed
    1. Bersuker, K., Peterson C.W.H., To M., Sahl S.J., Savikhin V., Grossman E.A., Nomura D.K., and Olzmann J.A.. 2018. A proximity labeling strategy provides insights into the composition and dynamics of lipid droplet proteomes. Dev. Cell. 44:97–112 e7. 10.1016/j.devcel.2017.11.020 - DOI - PMC - PubMed
    1. Bolte, S., and Cordelières F.P.. 2006. A guided tour into subcellular colocalization analysis in light microscopy. J. Microsc. 224:213–232. 10.1111/j.1365-2818.2006.01706.x - DOI - PubMed
    1. Bosch, M., Parton R.G., and Pol A.. 2020a. Lipid droplets, bioenergetic fluxes, and metabolic flexibility. Semin. Cell Dev. Biol. 108:33–46. 10.1016/j.semcdb.2020.02.010 - DOI - PubMed

Publication types

-