TRIM15 is a focal adhesion protein that regulates focal adhesion disassembly

Abstract

Focal adhesions are macromolecular complexes that connect the actin cytoskeleton to the extracellular matrix. Dynamic turnover of focal adhesions is crucial for cell migration. Paxillin is a multi-adaptor protein that plays an important role in regulating focal adhesion dynamics. Here, we identify TRIM15, a member of the tripartite motif protein family, as a paxillin-interacting factor and a component of focal adhesions. TRIM15 localizes to focal contacts in a myosin-II-independent manner by an interaction between its coiled-coil domain and the LD2 motif of paxillin. Unlike other focal adhesion proteins, TRIM15 is a stable focal adhesion component with restricted mobility due to its ability to form oligomers. TRIM15-depleted cells display impaired cell migration and reduced focal adhesion disassembly rates, in addition to enlarged focal adhesions. Thus, our studies demonstrate a cellular function for TRIM15 as a regulatory component of focal adhesion turnover and cell migration.

Keywords: Cell migration; Focal adhesion disassembly; Focal adhesions; Paxillin; TRIM E3 ligases; TRIM15.

Fig. 1.

TRIM15 interacts with paxillin and localizes to focal adhesions. (A) Western blot (WB) analyses of anti-FLAG immunoprecipitates (beads) and cell lysates using antibodies against the FLAG epitope or paxillin (PXN) from HeLa cell lysates expressing FLAG-tagged TRIM15 or control proteins TRIM62 and talin. (B) Western blot analyses of anti-paxillin (endogenous) or control IgG immunoprecipitates (IP) and cell lysates (CL) with antibodies against endogenous TRIM15 or paxillin from HepG2 cell lysates. Asterisk, non-specific background bands; arrowheads, TRIM15- or paxillin-specific band. The relative concentrations of immunoprecipitates (20×) with respect to cell lysates (1×) loaded on the gel are also indicated. (C) Individual and merged TIRFM images of the indicated cell lines showing endogenous TRIM15, paxillin and vinculin detected by immunofluorescence using respective antibodies. Arrows point to colocalized signals in individual focal adhesions. Areas outlined in white are shown at a higher magnification to the right. Scale bars: 10 µm.

Fig. 2.

TRIM15 recruitment to focal adhesions is myosin-II-independent and depends on paxillin. (A) HeLa cells expressing TRIM15–CFP (red) were immunostained and imaged for phosphotyrosine epitopes (anti-pY, green). Arrows, focal contacts where TRIM15–CFP and phosphotyrosine epitopes colocalize. Scale bars: 10 µm. (B) HeLa cells expressing TRIM15–YFP (green) and either paxillin–CFP or zyxin–CFP (red) were treated with DMSO (control) or 40 µM blebbistatin for 2 h. Arrows, TRIM15-containing focal contacts that are resistant to blebbistatin treatment. Scale bars: 2 µm. (C) Confocal microscopy images of HeLa cells treated with control (NTsi) or paxillin-specific (PXNsi) siRNA and immunostained for endogenous TRIM15. Zyxin–YFP was used as the focal adhesion marker. Areas outlined in white are shown at a higher magnification to the right. (D) A plot comparing the intensities of endogenous TRIM15 staining in individual focal adhesions (<100) normalized to the corresponding intensities of focal adhesion marker with or without paxillin siRNA for the experiment shown in C. FAs, focal adhesions; a.u., arbitrary units. (E) Confocal microscopy images of paxillin-null MEFs coexpressing the indicated fluorescent proteins. Areas outlined in white are shown at a higher magnification to the right. Scale bars: 10 µm; insets, 2 µm (C,E). (F) A plot comparing the intensities of TRIM15–YFP at individual focal adhesions (<200) normalized to the corresponding intensities of focal adhesion marker with or without paxillin coexpression in paxillin-null MEFs for the experiment shown in E. ****P<0.0001. (G) Western blot of cell lysates probed with the indicated antibodies to demonstrate similar expression levels of TRIM15 for an experiment as shown in E.

Fig. 3.

TRIM15 localizes to focal adhesions by binding to paxillin through its coiled-coil domain. The schematic in A shows the domain structure of TRIM15 as a guide. Asterisks, point mutations (see below); R, RING domain; B, B-box domain; CC, coiled-coil domain. (A,C) Lysates from HEK293 cells coexpressing either FLAG-tagged paxillin (PXN, lanes 1–6) or empty FLAG vector (lanes 7–12) with YFP-tagged wild-type (RBCPS) TRIM15 or versions of TRIM15 containing deletions were immunoprecipitated using antibodies against the FLAG epitope. The amino acid coordinates of TRIM15 deletions are indicated. The immunoprecipitates (IP, anti-FLAG beads) and cell lysates were analyzed by western blotting (WB) using antibodies against FLAG or GFP. TRIM15 Bm carries two point mutations in the B-box (C83A, H86A). TRIM15 Cm (full-length protein) or CCm (coiled-coil domain alone) carry two point mutations in the coiled-coil domain (V213G, L1216R), which contains the putative paxillin-binding subdomain. (B,D) TIRFM images of HeLa cells expressing YFP fusions of wild-type (RBCPS) TRIM15 or the indicated mutant TRIM15 versions displayed along with grayscale processed images showing focal adhesion (FA) outlines. See also supplementary material Fig. S1D; Fig. S3A. (E) TIRFM images of HeLa cells treated with either non-targeting control (NTsi) or paxillin-specific (PXNsi) siRNA and coexpressing the YFP-tagged coiled-coil domain of TRIM15 (green) along with zyxin–mCherry (red). Arrows, presence or absence of TRIM15-CC–YFP signal at focal adhesions. Scale bars: 10 µm.

Fig. 4.

TRIM15 interacts with the LD2 motif of paxillin. (A) TRIM15 interacts with the N-terminal LD-containing fragment of paxillin. Lysates from HEK293 cells coexpressing YFP-tagged wild-type or mutant paxillin (PXN) or empty vector together with either FLAG–TRIM15 (lanes 1–4) or empty FLAG vector control (lanes 5–8) were immunoprecipitated using antibodies against the FLAG epitope. The immunoprecipitates (IP, beads) and cell lysates were analyzed by western blotting (WB) using antibodies against GFP or FLAG as indicated. (B–D) Paxillin LD2 is required for an interaction with TRIM15. HEK293 cell lysates coexpressing the indicated wild-type or mutant FLAG-tagged paxillin variants or empty vector control together with full-length TRIM15–YFP, the coiled-coil domain (TRIM15CC–YFP) or the mutated coiled-coil domain (V213G, L1216R; TRIM15CCm–YFP) were immunoprecipitated and processed as described in A. The amino acid coordinates for paxillin deletions are also indicated. (E) Paxillin LD2 motif binds directly to TRIM15 coiled-coil domain in vitro. Paxillin LD2 [PXN(-LD2-)] protein (input) was incubated with glutathione-S-transferase (GST) fusion proteins of the TRIM15 coiled-coil domain (GST–TRIM15CC). GST was used as the specificity control. The co-precipitated LD2 protein was detected by western blotting using polyclonal anti-paxillin antibodies that can also recognize the LD2 region of paxillin (upper panel). A Coomassie-Blue-stained gel with the input LD2 and GST fusion proteins is shown below as a reference.

Fig. 5.

TRIM15-depleted cells are impaired in cell migration and chemotaxis. (A) Western blot analyses to determine the levels of the indicated proteins using specific antibodies in lysates from cells that were treated with the indicated siRNAs [against TRIM15 (T15si) or paxillin (Pxnsi)] or that expressed shRNAs [against TRIM15 (T15sh)], compared with those of cells treated with non-targeting controls (NTsi or NTsh). (B,C) Efficiency of TRIM15 and paxillin knockdown in HeLa cells for the experiments shown in A. The mRNA levels were measured by real-time PCR at 72 h post-siRNA treatment or shRNA expression. The data represent the mean±s.d. (two experiments performed in triplicate). (D) Western blot analysis using antibodies against GFP of cell lysates from HeLa cells treated with the indicated siRNAs and expressing either TRIM15–YFP or its siRNA-resistant version (T15–YFP si-res). (E) Phase images (triplicates) of a wounded HeLa cell monolayer treated with control siRNA (NTsi), TRIM15-specific siRNA (T15si#1) or paxillin-specific siRNA (PXNsi) taken at 0 and 36 h after scraping the confluent monolayer. Scale bar: 100 µm. (F) The percent open wound area was determined using Tscratch software for an experiment similar to that shown in E. (G) The percent open wound area at indicated time-points determined as in F during the course of a wound healing assay carried out in HeLa cells stably expressing the indicated shRNAs and transfected with plasmids expressing the indicated proteins. (H) A 72-h wound healing experiment similar to that shown in E, where TRIM15-siRNA-treated HeLa cells were transfected with plasmids expressing empty vector as control or 100 ng of plasmids expressing the indicated siRNA-resistant versions of TRIM15 to rescue the migration defect. NTsi-treated cells were used as an additional control. (I) A plot of the percent open wound area at 72 h determined during the course of a wound healing assay performed in HeLa cells treated with either non-targeting or paxillin-specific siRNA and transfected with 100 ng of plasmids expressing the indicated siRNA-resistant versions of paxillin for rescue. Data in F–I represent the mean±s.d. (three experiments performed in triplicate). (J–L) The indicated cell migration parameters were estimated from two-dimensional migration trajectories of >80 individual HeLa cells expressing non-targeting shRNA or TRIM15-specific shRNA with or without TRIM15–RFP for rescue after re-plating on fibronectin-coated coverslips. See supplementary material Movie 2. (M) The number of migrated cells in Boyden chambers was determined for HeLa cells stably expressing non-targeting shRNA or shRNA targeting two sequences in TRIM15 (T15sh1 and T15sh2). Data represent the mean±s.d. (two experiments performed in triplicate). *P<0.05; ****P<0.0001; ns, not significant.

Fig. 6.

TRIM15 regulates cell spreading and focal adhesion disassembly. (A) HeLa cells treated with control (NTsi), TRIM15-specific (T15si) and paxillin-specific (PXNsi) siRNAs for 72 h were plated onto fibronectin-coated plates for the indicated times (min). The number of cells attached at each time-point was normalized to the total number of cells attached after 3 h. Data represent the mean±s.d. (three experiments performed in triplicate). (B) The spreading efficiency (on fibronectin-coated plates) of HeLa cells stably expressing shRNA#1 targeting TRIM15 (T15sh) or non-targeting (NTsh) at 20, 45 and 90 min after plating. Data represent the mean±s.d. (experiments performed three times on separate days in triplicate). (C,D) A plot of >100 focal adhesion (FA) assembly and disassembly rates computed from 2–3 h of time-lapse microscopy of HeLa cells stably expressing zyxin–YFP to label focal adhesions and treated with non-targeting or TRIM15-specific siRNA for 48–72 h. See supplementary material Movie 3. (E) Raw and grayscale processed images with outlines of individual focal adhesions of HeLa cells stably expressing zyxin–YFP treated with non-targeting siRNA or TRIM15-specific siRNA for 48–72 h at steady state. (F,G) Mean (horizontal black lines) of average focal adhesion area and intensity computed per image from 17–22 images taken randomly encompassing >50 cells for an experiment as shown in E. a.u., arbitrary units. (H) HeLa cells stably expressing the focal adhesion marker zyxin–YFP (green) transfected with either non-targeting siRNA or TRIM15-specific siRNA were serum starved and treated with 10 µM nocodazole for 4 h. At the indicated time (min) after nocodazole washout, cells were fixed and visualized. Fixed and grayscale processed images with focal adhesion outlines are shown. (I) A plot of average focal adhesion area (mean±s.d.) determined from analyses of 6–8 images taken randomly per time-point per sample for three experiments as shown in H. (J) Split, merged and grayscale processed images of an experiment as shown in H for HeLa cells stably expressing zyxin–YFP, treated with TRIM15-specific siRNA and transfected with plasmid expressing siRNA-resistant FLAG–TRIM15 (red) for rescue. The plot shows individual focal adhesion intensities computed from 80–100 cells from the indicated conditions. Scale bars: 10 µm. (K) Time-lapse color-coded images (yellow, early time-points; light blue, late time-points) of control and TRIM15-depleted HeLa cells during the focal adhesion disassembly assay shown in supplementary material Movie 4. Scale bars: 10 µm. (L) A plot of the average relative intensities of zyxin–YFP-labeled focal adhesions as a function of time tracked post-nocodazole washout during the focal adhesion disassembly assay, shown for control and TRIM15-depleted HeLa cells. A total of 201 and 175 focal adhesions were tracked for non-targeting-siRNA-treated and TRIM15-siRNA-treated cells, respectively. The indicated average half-lives of focal adhesions (t_1/2) for each condition were calculated based on best-fit values using the exponential function y = exp(−k*t). The rate constant, k, for focal adhesions in non-targeting-siRNA-treated and TRIM15-siRNA-treated cells was 2.2×10⁻² and 3.3×10⁻³ per min, respectively, see supplementary material Movie 4. (M) A plot of >200 focal adhesion resident times computed from 12–14 h of time-lapse microscopy of HeLa cells stably expressing zyxin–YFP and treated with non-targeting siRNA or TRIM15-specific siRNA for 48–72 h. ***P<0.001; ****P<0.0001; ns, not significant.

Fig. 7.

Endocytosis of activated β1 integrin is delayed in TRIM15-depleted cells. (A) Subtracted images showing endocytosed β1 integrin (green) from an antibody feeding assay performed using Dylight-488-conjugated antibodies against activated β1 integrin (12G10) in HeLa cells treated with non-targeting or TRIM15-specific siRNA (NTsi or T15si, respectively) for the indicated times (min) are presented in the upper panel. Non-internalized antibodies corresponding to surface β1 integrin were detected using Alexa-Fluor-568-conjugated secondary antibodies (red). The merged images shown in the lower panel correspond to total β1 integrin. Scale bar: 10 µm. (B) The box and whisker plot shows the normalized percentage of endocytosed integrin as a function of time for an experiment as shown in A from analyses of 17–18 images taken per time point per sample. The box shows the median, 25th and 75th percentiles. The whiskers show maximum and minimum values. (C) A FACS histogram plot of control or TRIM15-depleted HeLa cells showing the uptake of transferrin conjugated to the pH-sensitive dye pHrodoRed at the end of 35 min. HeLa cells treated similarly but maintained at 4°C (0 min) were used as the negative control. 84.9% of control cells and 84% of TRIM15-siRNA-treated cells were positive for transferrin in the gated area. (D) The percentage open wound area at the indicated time-points (h) was determined using images analyzed by Tscratch software during the course of the wound healing assay that was performed using HeLa cells stably expressing either non-targeting or TRIM15-specific shRNA (NTsh or T15sh, respectively) and transfected with plasmids expressing the indicated proteins. Data show the mean±s.d.; *P<0.05; **P<0.01.

Fig. 8.

TRIM15 is a stable focal adhesion protein and its turnover depends on its ability to form oligomers. (A–F) The indicated YFP fusion proteins of zyxin, TRIM15, paxillin, vinculin, talin and TRIM15 carrying a B-box deletion (R_CPS) were expressed in HeLa cells and were subjected to FRAP analyses during which a micropoint laser bleached a small area within a single focal adhesion (arrows in A,B) and the recovery of fluorescence was recorded over time (in seconds). Selected images from FRAP time-lapse videos are shown for YFP-tagged zyxin (A) and TRIM15 (B). The plots in A–F show fluorescence recovery experiments performed on numerous focal adhesions for the indicated proteins in seconds following photobleaching (t = 0). A total of 47, 21, 24, 34, 11 and 49 focal adhesions were analyzed for YFP-tagged zyxin, TRIM15, paxillin, vinculin, talin and R_CPS, respectively. The indicated half times for recovery (t_1/2) were determined based on best-fit curves. Note that TRIM15 does not recover after photobleaching and therefore t_1/2 values could not be computed. Scale bars: 2 µm.