The MYO6 interactome reveals adaptor complexes coordinating early endosome and cytoskeletal dynamics

Abstract

The intracellular functions of myosin motors requires a number of adaptor molecules, which control cargo attachment, but also fine-tune motor activity in time and space. These motor-adaptor-cargo interactions are often weak, transient or highly regulated. To overcome these problems, we use a proximity labelling-based proteomics strategy to map the interactome of the unique minus end-directed actin motor MYO6. Detailed biochemical and functional analysis identified several distinct MYO6-adaptor modules including two complexes containing RhoGEFs: the LIFT (LARG-Induced F-actin for Tethering) complex that controls endosome positioning and motility through RHO-driven actin polymerisation; and the DISP (DOCK7-Induced Septin disPlacement) complex, a novel regulator of the septin cytoskeleton. These complexes emphasise the role of MYO6 in coordinating endosome dynamics and cytoskeletal architecture. This study provides the first in vivo interactome of a myosin motor protein and highlights the power of this approach in uncovering dynamic and functionally diverse myosin motor complexes.

Keywords: BioID; MYO6; endosome; functional proteomics; interactome.

Figure 1. The MYO6 interactome reveals numerous novel binding partners

1. Schematic diagram of BirA*‐MYO6 CBD wild‐type and mutant constructs.

2. myc‐BirA*‐MYO6 CBD RPE cell lines were analysed by immunoblot using myc and GAPDH (loading control) antibodies.

3. Immunofluorescence microscope images of RPE cells stably expressing myc‐BirA*‐MYO6 CBD NI (top row) and myc‐BirA*‐MYO6 CBD LI (bottom row) treated with 50 μM biotin for 24 h. Cells were immunostained with antibodies to myc (blue), APPL1 (green, top row), DAB2 (green, bottom row) or streptavidin (red) to visualise biotinylated proteins. Scale bar, 20 μm.

4. Diagram of all direct and indirect interactions identified for MYO6 NI and LI using BioID. Edge length corresponds to FC‐A score (lower score = greater length) and node size to SAINT score (lower confidence = smaller node). Green and blue nodes indicate interactions specific to the NI and LI isoforms, respectively, and cyan indicates shared partners. Previously described interactions are highlighted by red outlines.

5. Graph depicting relative enrichment (the fold‐change ratio) of proteins in pull‐downs from MYO6 NI and LI expressing cells. Dot size corresponds to SAINT score (lower confidence = smaller dot).

6. Schematic diagram of SILAC workflow.

7. Plot depicting mean log₂ fold change for each protein in (H) across the different MYO6 mutants. Colours correspond to those in (H).

8. Table of proteins identified with a significant (significance A, FDR < 5%) loss in at least one experiment. Heavy/light ratios are provided and > twofold losses are highlighted in colour. Proteins are grouped by colour based on pattern of loss across the different MYO6 mutants. − = not identified in the experiment.

9. Plot of principal component analysis using the mean log₂ fold change from triplicate experiments. The figure shows the projections of the data on the first (x‐axis) and second (y‐axis) principal components. Principal components 1–3 account for 76.12, 17.6 and 6.274% of the variability in the data, respectively. CARD10 was excluded from the analysis due to its absence from the ΔRRL data set. Dashed boxes highlight the clusters likely to represent distinct MYO6‐associated protein complexes. Colours correspond to those in (H).

Figure EV1. Characterisation of BirA*‐MYO6 CBD NI mutant cell lines

Immunofluorescence microscope images of RPE cells stably expressing myc‐BirA*‐MYO6 CBD NI ΔWWY (top row), ΔRRL (middle row) and ΔPIP2 (bottom row) treated with 50 μM biotin for 24 h. Cells were immunostained with myc (blue) and APPL1 (green) antibodies. Scale bar, 20 μm.

Figure 2. The MYO6 interactome can be verified by secondary screens

Diagram of the MYO6 protein interaction network identified by BioID (solid lines) with MYO6 (white), GIPC1, LARG, LRCH3 and DOCK7 (yellow) baits and supplemented with interaction data available in public databases (dashed lines). Previously identified MYO6 binding partners are indicated in pink. Lower confidence interactions (> 3 FC‐A, < 0.8 SAINT) are indicated by red lines. All proteins < 2 interactions in the network were excluded for simplicity and further adjustments made manually. Proteins were clustered using a force‐directed layout function in Cytoscape 58. Putative complexes/subcellular locations are highlighted by the dashed boxes.

Figure 3. GIPC1 links MYO6 to multiple protein complexes

1. Dot plot of high and medium confidence interactions (> 3 FC‐A and > 0.8 SAINT or > 3 FC‐A and < 0.8 SAINT, respectively) identified in BirA*‐GIPC1 and BirA*‐LARG experiments and shared interactors from the BirA*‐MYO6 CBD interactome.

2. Network diagram of the LIFT complex.

3. Top: Schematic cartoon of LARG domain structure. Bottom: GFP nanobody immunoprecipitates from HEK293T cells transfected with GFP and GFP‐tagged LARG fragments encompassing amino acids 1–274 (PDZ‐CC), 274–721 (RGSL), 721–1172 (DH‐PH) and 1171–1544 (COOH) (left) or GFP and full‐length GFP‐SH3BP4 (right). Samples were analysed by Western blot with the indicated antibodies.

4. Confocal microscope images of SH3BP4 HeLa siRNA KD cells immunostained for SH3BP4 (red) and actin (green). Images are maximum intensity projections of confocal stacks. Views through the z‐stack (yellow dashed line) are shown. Scale bar, 20 μm. Graphs (i) and (ii) on the right: pixel intensity profiles of SH3BP4 and actin labelling along yellow line.

5. Confocal microscope images of HeLa cells transiently transfected with GFP‐MYO6⁺ (green) and immunostained for SH3BP4 (top row, red), LARG (bottom row, red) and actin (blue). Images are maximum intensity projections of confocal stacks. Scale bar, 20 μm.

6. Graph depicting the mean percentage of MYO6⁺ or MYO10‐positive cells treated with mock, LARG or SH3BP4 siRNA which generated filopodia. Counts were performed on cells from 10 fields of view (typically 10–30 cells/field) and n = 3 independent experiments. A two‐sample t‐test was used to determine statistical significance. ***P < 0.001. Error bars indicate SEM.

Figure EV2. GIPC1 links MYO6 to multiple protein complexes

1. A, B

   Top: The mammalian two‐hybrid assay was used to test binding of (A) full‐length GIPC1 against fragments of LARG encompassing amino acids 1–274 (PDZ‐CC), 274–721 (RGSL), 721–1172 (DH‐PH) and 1171–1544 (COOH) or LARG COOH against fragments GIPC1 encompassing amino acids 1–120 (NH2), 1–223 (NH2‐PDZ), 120–223 (PDZ), 120–333 (PDZ‐COOH), 223–333 (COOH) and (B) full‐length SH3BP4 against fragments GIPC1 encompassing amino acids 1–120 (NH2), 120–223 (PDZ), 223–333 (COOH). Graphs show the mean relative luciferase activity from single representative experiments. Bottom: Schematic cartoon highlighting domain structure, fragments and binding sites found in LARG (left) and GIPC1 (right).

2. C

   GFP nanobody immunoprecipitates from HEK293T cells transfected with GFP and GFP‐tagged CARD10 full length or missing the C‐terminal four amino acids (ΔSEA). Samples were analysed by Western blot with the indicated antibodies.

3. D

   The mammalian two‐hybrid assay was used to test binding of full‐length CARD10 and CARD10 ΔSEA to full‐length GIPC1. Graphs show the mean relative luciferase activity from single representative experiments.

4. E

   Widefield microscope images of HeLa cells transfected with GFP‐MYO6⁺ or GFP‐MYO10 and treated with mock, LARG or SH3BP4 siRNA. Cells were immunostained with a GFP antibody. Scale bar, 20 μm.

5. F

   HeLa cells treated with mock, LARG or SH3BP4 siRNA were analysed by Western blot using LARG, SH3BP4 and GAPDH (loading control) antibodies.

Figure 4. LPAR1‐LARG‐RHO‐dependent actin reorganisation controls endosome positioning

1. Schematic of the LPA‐LARG‐RHO signalling pathway.

2. Confocal microscope images of 0 min (upper panels) or 5 min (lower panels) LPAR1 uptake in HeLa cells expressing GFP‐MYO6 CBD and HA‐tagged LPAR1 in the presence of 10 μM LPA. Cells were immunostained for GFP (green), APPL1 (red) and LPAR1 (blue). Scale bar, 20 μm.

3. Confocal microscope images of HeLa cells serum starved (upper panels) or treated with 10 μM LPA for 5 mins (lower panels) and immunostained for APPL1 (red) and actin (green). Scale bar, 20 μm. Graph to the right depicts the mean Pearson's correlation coefficient calculated for actin and APPL1 in serum starved or LPA stimulated cells from ≥ 5 fields of view (2–6 cells/field) in n = 3 independent experiments (> 70 cells per condition). Paired t‐test P = 0.097. Error bars indicate SEM.

4. Confocal microscopy of HeLa cells expressing GFP‐LARG GEF immunostained for APPL1 (red) and actin (green). Scale bar, 20 μm. Graph to the right depicts the mean Pearson's correlation coefficient calculated for actin and APPL1 in GFP or GFP‐LARG GEF transfected cells from ≥ 7 fields of view (1–7 cells/field) in n = 4 independent experiments (> 100 cells per condition). Significance was calculated using a two‐sample t‐test. ***P < 0.001. Error bars indicate SEM.

Figure EV3. LPAR1‐LARG‐RHO‐dependent actin reorganisation controls endosome positioning and motility

1. Confocal microscope images of HeLa cells expressing GFP‐APPL1 and HA‐tagged LPAR1 and serum starved overnight. Cell surface LPAR1 was labelled with HA antibody (blue), and uptake was allowed to proceed in the presence of 10 μM LPA for 0 min (upper panels) or 5 min (lower panels). Cells were fixed and immunostained with GFP (green) and GIPC1 (red) antibodies. Scale bar, 20 μm.

2. Quantification of (i) phalloidin and (ii) APPL1 signal intensity in cells serum starved or treated with 10 μM LPA for 5 min. Data are the mean of n = 3 independent experiments. Paired t‐test P = 0.0974 (i) and P = 0.1481 (ii). Error bars indicate SEM.

3. Quantification of (i) phalloidin and (ii) APPL1 signal intensity in GFP or GFP‐LARG GEF transfected cells. Graph depicts mean from n = 4 independent experiments. Significance was calculated using a one‐sample t‐test. *P < 0.05, **P < 0.01. Error bars indicate SEM.

4. Image sequences from spinning disc confocal microscope showing mCherry‐APPL1 (red) and BFP‐LifeAct (green) in mock (top row) or GFP‐LARG GEF (bottom row) transfected HeLa cells. The motility of selected endosomes over time is highlighted by the arrowheads. Scale bar, 10 μm; 0.5 s/frame (see also Movie EV1).

5. Confocal microscope images of HeLa cells transfected with GFP‐RHOA Q63L (top row), GFP‐RHOB Q63L (middle row) and GFP‐RHOC Q63L (bottom row). Cells were immunostained with an APPL1 antibody (red) and labelled with phalloidin to visualise actin (green). Scale bar, 20 μm.

Figure 5. MYO6 is linked to the RhoGEF DOCK7 via LRCH3

1. Dot plot of high and medium confidence interactions (> 3 FC‐A and > 0.8 SAINT or > 3 FC‐A and < 0.8 SAINT) identified in BirA*‐LRCH3 and BirA*‐DOCK7 experiments and shared interactors from the BirA*‐MYO6 CBD interactome.

2. Network diagram of the DISP complex.

3. Top: Schematic cartoon highlighting domain structure, fragments and binding sites found in LRCH3 (left) and DOCK7 (right). Bottom: GFP nanobody immunoprecipitates from HEK293T cells transfected with GFP, full‐length GFP‐LRCH3 and GFP‐LRCH3 fragments corresponding to amino acids 1–382 (LRR), 383–648 (Unc) or 649–777 (CH) or GFP‐DOCK7 fragments corresponding to amino acids 1–500 (NH2), 501–1000 (DHR1), 1001–1500 (Int) or 1501–2140 (DHR2). Samples were analysed by Western blot with the indicated antibodies.

4. The mammalian two‐hybrid assay was used to test direct binding of full‐length LRCH3 and wild‐type, ΔWWY or ΔRRL MYO6 tail and full‐length LRCH3 or LRCH3 fragments and wild‐type MYO6 tail. Graph shows relative luciferase activity from a single representative experiment.

Source data are available online for this figure.

Figure EV4. LRCH1 is not a MYO6 binding partner

1. The mammalian two‐hybrid assay was used to test binding of full‐length LRCH3 and LRCH1 to wild‐type MYO6 tail. Graph shows the mean relative luciferase activity from a single representative experiment.

2. GFP nanobody immunoprecipitates from HEK293T cells transfected with GFP and full‐length GFP‐LRCH1 were analysed by Western blot with the indicated antibodies (GFP control IP same as Fig 5C).

3. Sequence alignment of the LRCH proteins. Boxes highlight the highly conserved leucine‐rich repeats and calponin homology domains (blue) and the unconserved MYO6 binding site in LRCH3 (red).

Figure EV5. The DISP complex regulates septin organisation

1. Confocal microscope images of RPE cells stably expressing GFP‐SEPT7 and untransfected and immunostained with a GFP (green) antibody and labelled with phalloidin to visualise actin (red). Scale bar, 20 μm.

2. GFP nanobody immunoprecipitates from HEK293T cells transfected with GFP, full‐length GFP‐LRCH3 and GFP‐LRCH3 fragments corresponding to amino acids 1–382 (LRR), 383–648 (Unc) or 649–777 (CH). Samples were analysed by Western blot with the indicated antibodies (same IP as Fig 5C).

3. Widefield microscope images of RPE cells transfected with HA‐MitoGBD and GFP‐LRCH3 fragments corresponding to amino acids 1–382 (LRR; top row) and 383–648 (Unc; middle row) or full‐length GFP‐LRCH1 (bottom row). Cells were immunostained with HA (blue), GFP (green) and SEPT7 (red) antibodies. Scale bar, 20 μm.

Figure 6. The DISP complex regulates septin organisation

1. Widefield microscope images of RPE cells expressing HA‐MitoGBD and GFP (top row), full‐length GFP‐LRCH3 (middle row) or GFP‐LRCH3 fragment 649–777 (CH; bottom row). Cells were immunostained with HA (blue), GFP (green) and SEPT7 (red) antibodies. Scale bar, 20 μm. Graph to the right depicts the Pearson's correlation coefficient calculated for GFP and SEPT7 in RPE cells transfected with GFP, full‐length GFP‐LRCH3, GFP‐LRCH3 fragments and GFP‐LRCH1. Graph displays the mean calculated from 10 fields of view (1 cell/field) from n = 3 independent experiments. Statistical significance was determined by repeated measures ANOVA and a Bonferroni post hoc test. ***P < 0.001. Error bars indicate SEM. Below, schematic cartoon highlighting domain structure of LRCH3 and the putative SEPT7 binding site.

2. Structured illumination microscope images of RPE cells transiently transfected with GFP‐tagged LRCH3, immunostained with GFP (green), SEPT7 (top row) or DOCK7 (bottom row, red) antibodies and labelled with phalloidin to visualise actin (blue). Scale bar, 5 μm or inset, 1 μm.

3. Confocal microscope images of RPE cells stably expressing myc‐LRCH3 and transiently transfected with HA‐MYO6 CBD NI (upper panels) or GFP‐DOCK7 DHR2 domain (lower panels, GFP‐DOCK7 GEF). Cells were immunostained with myc (blue), HA (top row, green) or GFP (bottom row, green) and SEPT7 (red) antibodies. Scale bar, 20 μm.

4. Graph depicting the mean percentage of GFP‐DOCK7 GEF or myc‐LRCH3‐positive cells which displayed septin oligomerisation. Counts were performed on > 100 cells per condition from n = 3 independent experiments. Statistical significance was determined by two‐sample t‐test. ***P < 0.001. Error bars indicate SEM.

Figure 7. Model of LIFT/DISP complex function

1. Upon activation of Gα12/13‐coupled receptor at the cell surface (e.g. LPAR1), LARG is activated and the receptor is internalised into MYO6‐GIPC1‐positive endosomes. LARG is able to catalyse the GDP‐GTP exchange of RHO GTPases which can then activate their downstream effectors to promote actin remodelling. The interaction between LARG and GIPC1 links this activity to the endosome to promote actin reorganisation in proximity to the trafficking receptor. This actin remodelling might affect endosome position and motility.

2. LRCH3 is able to displace septins (green) from actin filaments (blue) via its C‐terminal calponin homology domain. Once septins are displaced from the actin, DOCK7 can promote actin remodelling via its activity towards RAC1 and CDC42 GTPases.