Filamin FLN-2 promotes MVB biogenesis by mediating vesicle docking on the actin cytoskeleton

Abstract

Multivesicular bodies (MVBs) contain intralumenal vesicles that are delivered to lysosomes for degradation or released extracellularly for intercellular signaling. Here, we identified Caenorhabditis elegans filamin FLN-2 as a novel regulator of MVB biogenesis. FLN-2 co-localizes with V-ATPase subunits on MVBs, and the loss of FLN-2 affects MVB biogenesis, reducing the number of MVBs in C. elegans hypodermis. FLN-2 associates with actin filaments and is required for F-actin organization. Like fln-2(lf) mutation, inactivation of the V0 or V1 sector of V-ATPase or inhibition of actin polymerization impairs MVB biogenesis. Super-resolution imaging shows that FLN-2 docks V-ATPase-decorated MVBs onto actin filaments. FLN-2 interacts via its calponin-homology domains with F-actin and the V1-E subunit, VHA-8. Our data suggest that FLN-2 mediates the docking of MVBs on the actin cytoskeleton, which is required for MVB biogenesis.

Figure 1.

FLN-2 is required for MVB formation. (A–D) Confocal fluorescence images of the epidermis in adult day 2 WT (A–A″) and fln-2 (B–C″) co-expressing VHA-5::RFP and VHA-8::GFP. (D) Mean fluorescence intensity; n = 17 worms per strain. (E–H) TEM images of epidermal MVBs in WT (E), fln-2 (F), cup-5 (G), and cup-5;fln-2 (H) at adult day 2. Blue arrows, light MVBs; red arrows, dark MVBs. (I–L) MVB number/µm² epidermal surface (I and L), ILV number (J), and light MVB diameter (K) in the indicated strains. Around 4–6 cross-sections were scored in each of the 5 worms per group in (I and L) and n ≥ 30 light MVBs were scored per group in J and K. (M–N″) DIC (M) and confocal fluorescence images of epidermis in WT adults expressing FLN-2::GFP (M′) or FLN-2::GFP and VHA-5::RFP (N–N″). Yellow arrowheads, strip-like structures; white arrowheads, vesicular-tubular structures. D and I–L show mean ± SD. **, P < 0.01; ***, P < 0.001; N.S, not significant, by unpaired two-tailed Student’s t test. Scale bars, 5 μm (A–C″ and M–N″), 500 nm (E–H).

Figure S1.

Loss of FLN-2 disrupts VHA-5–positive vesicles. (A–B″) Confocal fluorescence images of the apical epidermis in WT adults co-expressing VHA-5::RFP and SCAV-3::GFP (A–A″) or expressing VHA-5::GFP and stained by lysotracker red (B–B″). (C) Schematic diagram showing the fln-2 gene structure and the FLN-2A domain structure. Black boxes indicate exons and lines designate introns. The red, green, and blue arrows indicate the mutation sites identified in qx416, qx439, and qx463, respectively. (D–K) Confocal fluorescence images of VHA-5::RFP in the hypodermis of WT (D), qx439 (E and F), qx463 (G and H), qx416 (I) and qx416 (J) expressing FLN-2 at adult day 2. The average fluorescence intensity of VHA-5 in the indicated strains is quantified in K. 15 animals were scored in each strain. Data are shown as mean ± SD. Unpaired two-tailed Student’s t test was performed to compare mutant datasets with WT or datasets that are linked by lines. ***, P < 0.001. (L–M′) DIC and confocal fluorescence images of WT adults expressing FLN-2::GFP in the head and tail regions. (N–O″) Confocal fluorescence images of the epidermis (N–N″) or the excretory canal (O–O″) in WT adults co-expressing FLN-2::RFP and VAB-10A::GFP (N–N″) or VHA-5::RFP (O–O″). FLN-2 overlaps with VAB-10A on the stripe-like fibrous organelles (N–N″). In the excretory canal cell, FLN-2 is aligned at the apical side of VHA-5 as indicated by arrowheads (N–N″). Scale bars, 5 μm.

Figure S2.

FLN-2 and actin cytoskeleton are required for MVB formation. (A–D) Confocal fluorescence images of the epidermis in WT (A–A″) and fln-2 (B–C″) co-expressing VHA-5::RFP and VHA-13::GFP at adult day 2. Average fluorescence intensity of VHA-5 and VHA-13 is quantified in D. At least 17 animals were scored in each strain. (E–G) The protein levels of VHA-5::RFP and VHA-8::GFP are unaltered in fln-2. Three independent immunoblot experiments were performed; E shows a representative result. (H–N) Confocal fluorescence images of the epidermis in WT animals expressing HGRS-1::GFP with the indicated RNAi treatments at adult day 2. The number of HGRS-1–positive vesicles per unit area (238 μm²) was quantified and shown in N. At least 14 animals were scored in each strain. (O–R) Merged confocal fluorescence images of the epidermis in WT day 2 adults co-expressing VHA-5::RFP and VHA-13::GFP (O and P) or VHA-8::GFP (Q and R) treated with either DMSO or LatA. (S and T) Mean fluorescence intensity and number of VHA-5–, VHA-13–, or VHA-8–positive vesicles per unit area (134 μm²) under the indicated treatments. 15 animals were scored in each treatment condition. (U–X) Representative TEM images of MVBs in the epidermis of WT day 2 adults treated with DMSO (U) or LatA (V and W). The number of MVBs per µm² of epidermis in the indicated treatments is shown in X. 23 and 32 cross sections were quantified under DMSO and LatA treatment, respectively. In D, F, G, N, S, T, and X, data are shown as mean ± SD. Unpaired two-tailed Student’s t test was performed to compare mutant datasets with WT or control treatment or datasets that are linked by lines. *, P < 0.05; ***, P < 0.001. Scale bars, 5 μm in A–C″, H–M, and O–R; 500 nm in U–W. Source data are available for this figure: SourceData FS2.

Figure 2.

V-ATPase is important for MVB formation. (A–F) Confocal fluorescence images of epidermis in control RNAi (A–A″ and D–D″), vha-8 RNAi (B–B″), or vha-13 RNAi worms (E–E″) co-expressing VHA-5::RFP and VHA-13::GFP (A–B″) or VHA-8::GFP (D–E″) at adult day 1. (C and F) Average fluorescence; n = 16 worms per group. (G–J) Confocal images of epidermis in control RNAi and vha-5 RNAi day 1 adults expressing VHA-8::GFP (G and H) or VHA-13::GFP (I and J). (K–R) TEM images and quantification of MVBs in day 2 adult epidermis. Blue arrows, light MVBs; red arrows, dark MVBs. For each strain, four to six cross-sections were scored in each of three to five worms. (C, F, and R) show mean ± SD. **, P < 0.01; ***, P < 0.001; N.S, not significant, by unpaired two-tailed Student’s t test. Scale bars, 5 μm (A–B″, D–E″, and G–J), 500 nm (K–Q).

Figure 3.

ESCRT complex is required for MVB biogenesis. (A–E″) Confocal fluorescence images of epidermis in WT day 2 adults co-expressing VHA-5::RFP and FLN-2::GFP with the indicated RNAi treatments. (F and G) Line scan analyses of VHA-5::RFP and FLN-2::GFP along the dotted lines in (A″) and (B″). (H and I) Mean fluorescence intensity of VHA-5 (H) and FLN-2 (I) in the indicated RNAi treatments. n ≥ 15 worms per group. (J–L) TEM images of MVBs in the epidermis of control RNAi (J and K) and hgrs-1 RNAi (L) day 2 adults. Blue arrows, light MVBs; red arrows; dark MVBs. (M–O) MVB number/µm² epidermal surface (M), ILV number (N) and light MVB diameter (O) in control RNAi and hgrs-1 RNAi worms. 4–5 cross sections were scored in each of 5 worms per group (M). n ≥ 59 light MVBs were scored per group (N and O). (P–R) Merged fluorescence images of epidermis in WT adults co-expressing HGRS-1::GFP and VHA-5::RFP (P) or FLN-2::RFP (Q). (R and S) Line scan analyses along the white lines in (P and Q). H, I, and M–O show mean ± SD. ***, P < 0.001, by unpaired two-tailed Student’s t test. Scale bars, 5 μm (A–E″, P, and Q), 500 nm (J–L).

Figure 4.

FLN-2 docks VHA-5 vesicles on the actin cytoskeleton. (A and B) Confocal fluorescence images (A) and line scan analysis (B) of the epidermis in WT adults co-expressing FLN-2::RFP and the actin probe ABD::GFP. (C–F) Confocal images of the epidermis in WT (C) and fln-2 (D) day 2 adults expressing ABD::GFP. (E) Line scan analyses. (F) Mean fluorescence intensity of ABD::GFP; n ≥ 17 worms per group. (G–J) Merged images of the epidermis in WT day 2 adults co-expressing VHA-5::RFP and ABD::GFP treated with DMSO (G) or LatA (H). (I) Line scan analyses. (J) Mean fluorescence intensity of VHA-5::RFP and ABD::GFP; n = 16 worms per group. (K–S) SIM fluorescence images of epidermis WT adults co-expressing VHA-5::RFP, ABD::GFP, and FLN-2::BFP (K–R). Boxed area in K is magnified in L–R. (S) Three different co-localization patterns of FLN-2, VHA-5, and actin as indicated in R. Yellow, red, and white arrows indicate VHA-5-, FLN-2-, and F-actin–positive structures, respectively. (T and U) Quantification of FLN-2-actin attachment (T) and position of VHA-5 vesicles relative to FLN-2 and actin (U). 10 (T) and 20 (U) different regions were scored in 8 and 15 worms, respectively. F, J, T, and U show mean ± SD. **, P < 0.01; ***, P < 0.001, by unpaired two-tailed Student’s t test. Scale bars, 5 μm (A–A″, C, D, G, and H), 1 μm (K–R).

Figure S3.

Inactivation of vha-5 and vha-8 does not affect the association of FLN-2 with actin. (A–F) Confocal fluorescence images of the epidermis in WT co-expressing FLN-2::RFP and ABD::GFP with the indicated RNAi treatments. (D–F) Line scan analyses of FLN-2::RFP and ABD::GFP along the dotted lines (A″, B″, and C″) in the indicated RNAi treatments. Scale bars, 5 μm. (G and H) In vitro F-actin co-sedimentation assay showing that GST-tagged FLN-2A(3xCH), but not GST, co-precipitates with F-actin. Three independent experiments were performed. Quantification is shown in H. Data are presented as mean ± SD. Unpaired two-tailed Student’s t test was performed to compare the two datasets (with actin versus without actin). *, P < 0.05; N.S not significant. (I) Schematic diagram showing domains of FLN-2A and human filamin A (hFLNA) and the truncations that were used in yeast-2 hybrid analyses. (J) Interactions between FLN-2A or hFLNA and VHA-8 were examined by Y2H analyses. FLN-2A(1-330, 3xCH) and hFLNA(1-264, 2xCH), but not FLN-2A(331-3611, Δ3xCH), interacted with VHA-8. Source data are available for this figure: SourceData FS3.

Figure 5.

FLN-2 interacts with VHA-8 and actin. (A) Domains of FLN-2A and FLN-2D. (B–D) Y2H (B and C) and GST (D) pulldown analyses of interactions between FLN-2A(3xCH) or FLN-2D and VHA-5/8/13 subunits. FLN-2A(3xCH) interacts with VHA-8 but not VHA-13 or VHA-5. n ≥ 3 independent GST pull-down experiments; D is a representative result. IB, immunoblot. (E) Model of FLN-2 and actin in MVB biogenesis. FLN-2 facilitates MVB biogenesis by acting as a cross-linking protein to organize actin filaments and as a bridging molecule to dock MVBs on the actin cytoskeleton. ESCRT components may function upstream of FLN-2 to regulate MVB formation. Source data are available for this figure: SourceData F5.