Clarin-1, encoded by the Usher Syndrome III causative gene, forms a membranous microdomain: possible role of clarin-1 in organizing the actin cytoskeleton

Abstract

Clarin-1 is the protein product encoded by the gene mutated in Usher syndrome III. Although the molecular function of clarin-1 is unknown, its primary structure predicts four transmembrane domains similar to a large family of membrane proteins that include tetraspanins. Here we investigated the role of clarin-1 by using heterologous expression and in vivo model systems. When expressed in HEK293 cells, clarin-1 localized to the plasma membrane and concentrated in low density compartments distinct from lipid rafts. Clarin-1 reorganized actin filament structures and induced lamellipodia. This actin-reorganizing function was absent in the modified protein encoded by the most prevalent North American Usher syndrome III mutation, the N48K form of clarin-1 deficient in N-linked glycosylation. Proteomics analyses revealed a number of clarin-1-interacting proteins involved in cell-cell adhesion, focal adhesions, cell migration, tight junctions, and regulation of the actin cytoskeleton. Consistent with the hypothesized role of clarin-1 in actin organization, F-actin-enriched stereocilia of auditory hair cells evidenced structural disorganization in Clrn1(-/-) mice. These observations suggest a possible role for clarin-1 in the regulation and homeostasis of actin filaments, and link clarin-1 to the interactive network of Usher syndrome gene products.

FIGURE 1.

CLRN1 is a plasma membrane protein localized at F-actin-enriched protrusions. A, the topology and transmembrane domains shown were predicted with the HMMTOP transmembrane topology prediction server (55). The possible N-linked glycosylation site is indicated. Also shown (red circle) is the previously predicted motif near the CLRN1 C-terminal tail that may serve as a PDZ-binding site (4). B, immunolocalization of Human WT CLRN1. C, immunolocalization of Na/K ATPase in HEK293 cells stably expressing CLRN1. D, merged image of B and C indicates that CLRN1 and Na/K ATPase co-localize. Images B–D are single optical sections of HEK293 cells. E, cell surface biotinylation was performed to separate cell surface proteins (avidin-bound) (AB) from intracellular proteins (flow-through) (FT). Immunoblots of both fractions reveal that most of the CLRN1 protein localized to the plasma membrane. HEK293 cells alone and HEK293 cells expressing CLRN1 were preincubated for 30 min with Sulfo-NHS-SS-Biotin to label cell surface proteins. After cells were harvested, biotin-labeled CLRN1 protein levels were measured by immunoblotting. F, localization of human WT CLRN1 in HEK293 cells stably expressing CLRN1. G, F-actin in HEK293 cells stably expressing CLRN1. F-actins were labeled with phalloidin-Alexa 488. H, merged image of F and G. CLRN1 localized at both microvilli (arrows) and lamellipodia (arrowheads). I–K, CLRN1 localization studied by immunofluorescence confocal microscopy after disruption of F-actin by cytochalasin D treatment. I, CLRN1 localized diffusely on the plasma membrane. J, F-actin localization is shown. K, merged image of I and J. After disruption of F-actin, CLRN1 and F-actin no longer co-localize. Images F–K were generated from multiple optical sections by a maximum intensity projection. Scale bars, 50 μm.

FIGURE 2.

CLRN1 induces reorganization of actin filaments and changes cellular morphology. Here HEK293 cells stably expressing wild-type CLRN1 are indicated as HEK-CLRN C1, D1, and D2, and HEK 293 cells stably expressing the CLRN1^N48K are indicated as HEK-CLRN^N48K. A, HEK-CLRN cell lines express different levels of CLRN1 protein. HEK293 and HEK-CLRN C1, D1, and D2 cells were lysed in 1% Igpal and 0.5% sodium deoxycholate, and CLRN1 was detected by immunoblotting with anti-Clarin-1 antibody 10B5. HEK293 cells did not show any detectable protein signals, whereas HEK-CLRN D2 cells expressed the highest levels of CLRN1 protein. B–F, HEK293 and stable cells expressing CLRN1 and CLRN1^N48K were imaged by bright field microscopy. The cell line used for each experiment is indicated below each image. Columns G–K, HEK293 cells, HEK-CLRN cell lines, and HEK-CLRN^N48K cells were imaged by confocal microscopy. Cells were labeled with Phalloidin-Alexa488 (top row, green) and anti-clarin-1 antibody 10B5 (middle row, red). The bottom row shows merged images from the top and middle rows. The cell line used for each experiment is indicated above each column. Images G–K were generated from multiple optical sections by a maximum intensity projection. Note: CLRN1 staining is weak in I, but the staining pattern is similar to H and J.

FIGURE 3.

Identification of clarin-1-enriched microdomains. Human CLRN1 forms specific microdomains on plasma membranes. A, HEK-CLRN cells were lysed in buffer containing different detergents and analyzed in a discontinuous sucrose gradient. Ten fractions (0.4 ml) were collected from the top to the bottom of the gradient and analyzed by Western blotting for CLRN1 and transferrin receptor. Detergents shown on the right of the CLRN1 panel are listed from the top to bottom in order of increasing hydrophobicity. CLRN1 was enriched in low density fractions and shifted to higher density fractions with increasing detergent hydrophobicity. Localization of transferrin receptor, a transmembrane protein, was not affected by the hydrophobicity of detergents. Flotillin, a lipid raft protein, partitioned into low density fractions in the presence of either Triton X-100 or Brij-98. B, cholesterol depletion causes CLRN1 to move into higher density fractions. HEK-CLRN cells were treated without or with MeβCD for 30 min at 37 °C before being lysed in 0.5% Brij-98. The cell lysate was centrifuged in a discontinuous sucrose gradient, and CLRN1 was detected by Western blotting. Cholesterol depletion partially disrupted the membrane microdomains containing CLRN1. C, affinity purification of CLRN1 from crude membrane preparations of HEK-CLRN C1 cells. Proteins were resolved by 8% SDS-PAGE, and stained with silver nitrate. Lane 1, crude membrane preparation from HEK-CLRN-C1 cells; lane 2, CLRN1-protein complex purified from a crude membrane preparation of HEK-CLRN-C1 cells by HA affinity gel purification; lane 3, after blocking by HA epitope peptide, only a small amount of proteins from the crude membrane preparation bound to HA affinity gel; lane 4, In the presence of 1% Igpal and 0.5% sodium deoxycholate, CLRN1 was purified by the HA affinity gel. The location of purified CLRN1 is indicated by an arrowhead.

FIGURE 4.

CLRN1 induces compartmentalization of integrin, N-cadherin, and α-catenin. HEK-CLRN C1 cells were co-stained with 10B5-Cy3 (mAb anti-CLRN1, red, images in the left column) and one of following mAbs (green, images in the middle column): anti-integrin-β1 subunit (A), anti-N-cadherin (C), and anti-α-catenin (E). Arrows indicate some locations where CLRN1 and integrin/N-cadherin/α-catinin co-localize (yellow or orange, images in the right column). HEK293 cells were co-stained with 10B5-Cy3 (mAb anti-CLRN1, red, images in the left column) and one of the following mAbs (green, images in the middle column): anti-integrin-β1 subunit (B), anti-N-cadherin (D), or anti-α-catenin (F). The distribution of integrin, N-cadherin, or α-catinin is more uniform on the plasma membrane in the absence (B, D, and F) than in the presence (A, C, and E) of CLRN1. Images are single optical sections. Scale bar, 50 μm.

FIGURE 5.

N-Linked glycosylation is required for the stability and plasma membrane localization of CLRN1. A, Western blotting analysis of HEK293 cells stably expressing human CLRN1 and CLRN1^N48K (top gel image). Five samples were treated with either Endo H_f or PNGase F to study glycosylation patterns. Three samples were treated with MG132 to inhibit degradation of CLRN1^N48K by proteasomes. Combinations of those treatments are indicated above the top gel image. **, position corresponding to glycosylated wild-type CLRN1; *, position of non-glycosylated CLRN1. The cell type used for each sample is shown below the bottom gel image; WT represents stable cells expressing wild-type human CLRN1, N48K represents stable cells expressing CLRN1^N48K, and MOCK represents HEK 293 cells. Anti-β-tubulin antibody was used to study the loading of each sample (bottom gel image). B, localization of human CLRN1^N48K mutant in HEK293 cells. Top row: i, expression of CLRN1^N48K (green immunofluorescence) is barely detectable under normal conditions; ii, cells were stained with anti-calreticulin (red), a marker for the endoplasmic reticulum; iii, merged image of i and ii. Bottom row: iv, after MG132 treatment, expression of CLRN1^N48K is detectable as shown by green immunofluorescence; v, cells were stained with anti-calreticulin; vi, merged image of iv and v. CLRN1^N48K localized to the endoplasmic reticulum, as shown by co-localization of calreticulin and CLRN1^N48K; CLRN1^N48K is also observed in the cytoplasm. C, tunicamycin promotes proteasome-mediated degradation of wild-type CLRN1. In this experiment, CLRN1 expression was regulated by a doxycycline-inducible promoter in HEK293-CLRN-ind cells. i, in the absence of doxycycline, CLRN1 was not detected; ii, 24 h after doxycycline induction, most of the CLRN1 (red) was observed on plasma membranes; iii, combination treatment with doxycycline and tunicamycin abolished the expression of CLRN1; iv, after treatment of the cells with doxycycline, tunicamycin, and MG132, CLRN1 (red) was noted as intracellular inclusions. Nuclei were stained with Hoechst 33342 (blue). D, Western blotting analysis of HEK293-CLRN-ind cells. Two samples were treated with tunicamycin (TM) to inhibit N-linked glycosylation, and three samples were treated with doxycycline (DC) to induce expression of CLRN1. Cells used for lane 3 were also treated with MG132. Combinations of these treatments are indicated over the gel images. The cell type used for each lane is indicated below the gel images: WT are from HEK-CLRN-ind, N48K are from HEK-CLRN^N48K, and MOCK are from HEK 293 cells. β-Tubulin antibody was used to study the loading of each sample (bottom gel images). Images in B and C are single optical sections. Scale bar, 50 μm.

FIGURE 6.

F-actin-enriched stereocilia are poorly developed in cochlear hair cells of Clrn1 knock-out mice. A, Clrn1 knock-out (Clrn1^−/−) mice did not express full-length Clrn1 mRNA in the cochlea (RT panel). Total RNA was isolated from the cochlea and used for RT-PCR. Total RNA without reverse transcription reaction (no RT panel) was used as a negative control to confirm that cDNAs, instead of genomic DNAs, were amplified by PCR. As an internal control, actin cDNA was amplified to confirm that a similar amount of mRNA was used for each experiment. Based on sequencing analysis, band a was composed of two splicing variants: one with all four exons and the other with exons 1, 3, and 4. Band b was a transcript with exons 1 and 4. Band c was a nonspecific product. B, organs of Corti from Clrn1^−/− (left panel) and Clrn1^+/+ (right panel) mice were fluorescently labeled with Alexa 488-labeled phalloidin to reveal F-actin structures in green. F-actin-enriched stereocilia were poorly developed in Clrn1^−/− outer hair cells, whereas V-shaped bundles of stereocilia were observed in Clrn1^+/+ outer hair cells. Images were generated from multiple optical sections by a maximum intensity projection. In C: left panel, Clrn1^−/− mice show disorganization of OHC stereocilia (white arrowheads) and the inner hair cells (IHCs) evidence subtle changes in stereocilial organization. Right panel, wild-type mouse illustrating normally configured V-shaped stereocilia in outer hair cells (OHCs, rows 1–3) and crescent-shaped bundles in IHCs. Scale bars, 10 μm.