The type V myosin-containing complex HUM is a RAB11 effector powering movement of secretory vesicles

Abstract

In the apex-directed RAB11 exocytic pathway of Aspergillus nidulans, kinesin-1/KinA conveys secretory vesicles (SVs) to the hyphal tip, where they are transferred to the type V myosin MyoE. MyoE concentrates SVs at an apical store located underneath the PM resembling the presynaptic active zone. A rod-shaped RAB11 effector, UDS1, and the intrinsically disordered and coiled-coil HMSV associate with MyoE in a stable HUM (HMSV-UDS1-MyoE) complex recruited by RAB11 to SVs through an interaction network involving RAB11 and HUM components, with the MyoE globular tail domain (GTD) binding both HMSV and RAB11-GTP and RAB11-GTP binding both the MyoE-GTD and UDS1. UDS1 bridges RAB11-GTP to HMSV, an avid interactor of the MyoE-GTD. The interaction between the UDS1-HMSV sub-complex and RAB11-GTP can be reconstituted in vitro. Ablating UDS1 or HMSV impairs actomyosin-mediated transport of SVs to the apex, resulting in spreading of RAB11 SVs across the apical dome as KinA/microtubule-dependent transport gains prominence.

Keywords: Biochemistry; Biological sciences; Molecular biology.

Graphical abstract

Figure 1

The type V myosin MyoE is a RAB11 effector (A) Schemes depicting cooperation of kinesin-1 (KinA) and MyoE to deliver RAB11 SVs to the SPK in the wild type and in a myoEΔ mutant in which SVs distribute across the apical dome, unable to be focused at the SPK due to the absence of F-actin-dependent transport. (B) (Top) Localization of GFP-RAB11 vesicles in a myoEΔ mutant. The picture is displayed with the same contrast adjustments as the wild-type. (Bottom) The image of myoEΔ has been magnified and contrasted to reveal the spreading of RAB11 SVs across the apical dome better. (C) Growth of indicated strains on solid complete medium at 37°C. (D) (Left) Localization of GFP-RAB11 vesicles in sec4Δ hyphae. The picture is displayed with the same contrast adjustments as the wild type. Right, quantitation of the area occupied by the apical cluster of RAB11 vesicles. The datasets were significantly different in a Mann-Whitney test (p < 0.0001). Bars indicate SD. (E) Localization of endogenously tagged MyoE-GFP in wild type and sec4Δ tips. The bottom graph shows that the MyoE-GFP signal at the SPK was significantly lower in sec4Δ cells compared with the wild type. The datasets were significantly different in a t student test with Welch’s correction (p < 0.0001). Bars indicate SD. (F) Localization of GFP-tagged MyoE GTD in different genetic backgrounds. Unlike myoEΔ hyphae, wild type, and sec4Δ cells contain the resident copy of myoE intact. (G) GFP-GTD colocalizes with mCh-RAB11 in the tip cluster of SVs (top) and in cytosolic puncta (bottom). (H) Kymograph derived from a time-lapse movie built with near-cortical planes of a hypha co-expressing mCh-RAB11 and GFP-GTD. Tip-directed trajectories of SVs are indicated with arrows. (I) Dual-channel imaging of MyoE-mCh and GFP-MyoE GTD illustrating the larger area occupied by the latter. Graph, average results of linescans traced across the apical dome (scheme) (n = 10). Note the high cytosolic background of GFP-GTD. (J) GST pull-downs with GST-MyoE GTD as bait and the indicated bacterially expressed purified RABs, loaded with GDP or GTPγS, as preys. GST-GFP was used as bait for a negative control. Pulled-down material was analyzed by α-His western blotting. The Coomassie-stained gel shows levels of GST fusion protein baits. See also Figure S1A and Videos 1 through 4.

Figure 2

UDS1, a novel, direct effector of RAB11 (A) Proteins retained by GTPγS- and GDP-loaded RAB11-GST columns were identified by shotgun proteomics. Spectral counts obtained for each protein and condition and the relative enrichment detected in one sample versus the other are listed. Note that markedly abundant GdiA (GDP dissociation factor) interacts preferentially with GDP-RAB11. AP-2α was used as negative control. (B) Features of UDS1. The probability of forming coiled-coils (red graph) and disordered regions (gray area) is indicated, as are the positions of the UDS1 and SCOP superfamily domains. (C) GST pull-down assays with the indicated baits, using a prey extract of A. nidulans-expressing UDS1-HA3 from the endogenously tagged gene. Pull-downs were analyzed by western blotting with α-HA3 antibody. GST-GFP was used as negative bait control. (D) UDS1 is a dimer in vitro. Equilibrium ultracentrifugation of purified UDS1 at a concentration of 4 μM; (top) the concentration gradient obtained (empty circles) is shown together with the best-fit analysis assuming that the protein is a dimer. (Bottom plot) Differences between experimental data and estimated values for the dimer model (residuals). See also Figure S2E Negative-stain electron microscopy of purified UDS1. The proteins were stained with uranyl acetate and examined in a JEOL-1230 electron microscope. Four examples selected showed the extended screw-like form of UDS1. The lengths of N = 71 molecules were measured (plot; average 496 Å +/− 73 SD). (F) AlphaFold prediction of the UDS1 dimer. See Figures S3A and S3B. (G) Enlarged view of the predicted fold of the UDS1 domain dimer. (H) GST pull-down assays with the indicated RABs and purified UDS1-His as prey. UDS1 was detected by α-His western blotting.

Figure 3

Subcellular localization of UDS1 (A) Hyphal tip cell expressing mCh-RAB11 and MyoE-GFP; images are MIPs of deconvolved Z-stacks, all at the same magnification. (B) A hyphal tip cell expressing mCh-RAB11 and UDS1-GFP. Images are MIPs of deconvolved Z-stacks. See also Video S5. (C) Left, a hyphal tip cell expressing endogenously tagged MyoE-GFP and UDS1-tdT. Images are MIPs of deconvolved Z-stacks. (D) MyoE-GFP and UDS1-tdT strictly colocalize across time: A 4D movie made with MIPs of Z-stacks acquired every 30 s (Video S6) was used to draw a kymograph across the long axis of a hypha growing at 0.9 μm/min. The diagonal lines traced by apical spots reflect apical extension growth. (E) UDS1-GFP strictly colocalizes with mCh-RAB11 SVs transported by MTs to the apical dome in a cell lacking MyoE. Images are MIPs of deconvolved Z-stacks. See also Video S7.

Figure 4

HMSV, an uncharacterized interactor of the MyoE GTD (A) Cell extracts expressing the indicated GFP-tagged baits by allelic replacement were immunoprecipitated with GFP-Trap. Pulled-down proteins were analyzed by LC-MS/MS. The table lists the spectral counts obtained for each of the indicated co-precipitating proteins. A Uso1-GFP-expressing strain was used as negative control. (B) Prediction of coiled-coils (red graph) and disordered regions (gray area) along the primary sequence of HMSV. (C) AlphaFold prediction of HMSV. Roman numerals indicate α-helical regions (color-coded). Two large disordered regions are indicated as loop 1 and loop 2. See also Figures S3C and S3D. (D) Growth tests at 37°C of indicated strains. See also Figure S1B. (E) Endogenously tagged MyoE-mCh and HMSV-GFP strictly colocalize. MIPs of deconvolved Z-stacks. (F) Kymograph derived from the 4D sequence shown in Video S8, mounted with MIPs of Z-stacks acquired with a beam splitter every 15 s for 15 min. The hypha was growing at 0.62 μm/min. (G) Widths of wild-type and mutant hyphae stained with calcofluor to label the cell walls. Top, middle planes of representative tips. (Bottom) x,y images of septa alongside with the corresponding orthogonal views. (Right) Quantitation of cell width. Bars are the average value for ∼20 hyphae per condition; error bars are SD. Significance was assessed with an ANOVA Kruskal-Wallis test with Dunn’s multiple comparison correction. n.s., not significant. (H) MyoE and HMSV interact directly: MyoE and HMSV-HA3 obtained from TNT reactions were mixed with Protein A beads preloaded with polyclonal α-MyoE antiserum or with antiserum against the unrelated protein Uso1. Immunoprecipitates were analyzed by α-HA western blotting. Equal IgG loading was confirmed with Coomassie-stained heavy chains. (I) Spectral counts of HMSV detected in GFP-trap immunoprecipitates of cell extracts expressing the MyoE GTD or the analogue cargo-binding, C-terminal domain (CTD) of KinA/kinesin-1. (J) HMSV associates with MyoE through the GTD of the motor. Extracts of myoEΔ cells co-expressing HMSV-HA3 with GFP-tagged MyoE GTD or MyoE ΔGTD (MyoE lacking the GTD) were immunoprecipitated with GFP-Trap nanobody. A Uso1-GFP HMSV-HA3 strain was used as negative control. (Left) Pulled-down material was analyzed by α-HA western blotting. (Right) Relative levels of the preys by α-GFP western blotting. (K) HMSV interacts directly with the MyoE GTD: pull-down assays with indicated GST baits. Preys were in vitro expressed (with TNT) HMVS-HA3 or, as control, UDS1-HA3. Blots were revealed with α-HA antibody.

Figure 5

UDS1 bridges its direct interactor HMSV to the active form of RAB11. HMSV does not interact directly with RAB11 (A) Control showing that a Uds1-HA3 is efficiently pulled down from extracts by GTPγS RAB11 but not by GDP-RAB11. (B) As in (A), but using HMSV-HA3 extracts as preys. Uds1-HA3 and HMSV-HA3 were expressed from allelic replacements. Pull-downs analyzed by α-HA western blotting. (C) HMSV and UDS1 interact directly: pull-down assays with GST-UDS1 as bait and HMSV-HA3 or, as negative control, Uso1-HA3 as preys, which were obtained by TNT expression. Pull-downs analyzed by α-HA3 western blotting. (D) HMSV is recruited to RAB11 only when UDS1 is present and the GTPase is in the active conformation. GST pull-down assays with RAB11 and, as negative control, RAB5b baits. The preys, combined as indicated, were purified UDS1-His and TNT-synthesized HMSV-HA3. Samples were analyzed by α-His and α-HA western blotting.

Figure 6

UDS1 and MyoE are components of the HMSV-scaffolded HUM complex (A) MyoE associates with UDS1 and HMSV. Extracts of cells expressing endogenously tagged GFP proteins were immunoprecipitated with α-GFP nanobody. (Left) α-MyoE western blot analysis of immunoprecipitates. The band indicated with an asterisk is unspecific (α-MyoE is a polyclonal antiserum). (Right top) Silver staining of immunoprecipitates. MyoE is remarkably abundant in the HMSV sample, detectable in the UDS1 sample, and absent in the Uso1 sample. (Right bottom) Relative levels of the preys revealed by α-GFP western blotting. (B) UDS1 and MyoE associate only if HMSV is present. (C) HMSV and MyoE associate efficiently when UDS1 is absent. (D) UDS1 associates with HMSV regardless of whether MyoE is absent or present. GFP-Trap immunoprecipitates were analyzed by α-HA3 western blotting to detect HMSV and by α-GFP western blotting to reveal the relative levels of the baits.

Figure 7

HUM complex components cooperate with RAB11 to recruit MyoE to SVs. A model (A) (Top) Localization of the HUM complex components in different genetic backgrounds. Images are MIPs of deconvolved Z-stacks. As GFP reporters were endogenously tagged, the corresponding null background images are empty. (Bottom) Quantitation (arbitrary units, A.U.) of the MyoE-GFP signal in the SPK of uds1Δ and hmsVΔ cells compared with the wild type. Means (±SD) were: left, wild-type 2,842 ± 227 versus uds1Δ 558 ± 159 (p < 0.0001 in unpaired t test). (Right) Wild type, 3496 ± 245 versus 654 ± 149 in the hmsVΔ mutant (p < 0.0001 in unpaired t test). See also Videos S9 and S10. (B) A prototypic cargo of the RAB11 recycling pathway is delocalized from the SPK by uds1Δ and hmsVΔ. The scheme depicts endocytic recycling followed by the chitin-synthase ChsB. ChsB and RAB11 are similarly delocalized from the SPK, indicated by red arrowheads. (C) Model for the engagement of HUM with RAB11 SVs. In the wild type, RAB11 is recruited to SVs during the Golgi-to-post-Golgi transition. RAB11 interacts both with the GTD of MyoE and with UDS1 in the HUM complex. UDS1 bridges active RAB11 to the HMSV scaffold. HMSV bridges RAB11/UDS1 to MyoE by direct interaction with the motor’s GTD. MyoE transport is most efficient in the context of the whole complex. However, in the absence of UDS1 or HMSV, MyoE-mediated SV transport remains partially operative due to the direct interaction between RAB11 and the MyoE GTD, albeit this transport is less efficient, accumulation of SVs in the SPK is impaired and MT-dependent transport becomes more prominent, leading to the characteristic apical dome distribution of SVs in these mutants.