The hereditary spastic paraplegia protein strumpellin: characterisation in neurons and of the effect of disease mutations on WASH complex assembly and function

Abstract

Mutations in the gene encoding strumpellin cause autosomal dominant hereditary spastic paraplegia (HSP), in which there is degeneration of corticospinal tract axons. Strumpellin is a component of the WASH complex, an actin-regulating complex that is recruited to endosomes by interactions with the retromer complex. The WASH complex and its relationship to retromer have not been fully characterised in neurons, and the molecular pathological mechanism of strumpellin mutation is unclear. Here we demonstrate that the WASH complex assembles in the brain, where it interacts with retromer. Members of both complexes co-localise with each other and with endosomes in primary cortical neurons, and are present in somato-dendritic and axonal compartments. We show that strumpellin is not required for normal transferrin receptor traffic, but is required for the correct subcellular distribution of the β-2-adrenergic receptor. However, strumpellin disease mutations do not affect its incorporation into the WASH complex or its subcellular localisation, nor do they have a dominant effect on functions of the WASH complex, including regulation of endosomal tubulation, transferrin receptor traffic or β-2-adrenergic receptor localisation. Models of the WASH complex indicate that it contains a single strumpellin molecule, so in patients with strumpellin mutations, complexes containing wild-type and mutant strumpellin should be present in equal numbers. In most cell types this would provide sufficient functional WASH to allow normal cellular physiology. However, owing to the demands on membrane traffic imposed by their exceptionally long axons, we suggest that corticospinal neurons are especially vulnerable to reductions in functional WASH.

Supplementary Fig. 1

WASH complex member FAM21 co-localises with Wash1 and retromer in neuronal cells. GFP-tagged FAM21 was transiently transfected into rat primary cortical neurons and the cells were labelled with the markers indicated.

Supplementary Fig. 2

Wild-type and disease mutant strumpellin co-localise with WASH and retromer markers in HeLa cells. HeLa cells were transiently transfected with myc-tagged wild-type strumpellin (A, C, E) or disease mutant strumpellin-myc V626F (B, D, F) and labelled with the markers indicated.

Supplementary Fig. 2

Wild-type and disease mutant strumpellin co-localise with WASH and retromer markers in HeLa cells. HeLa cells were transiently transfected with myc-tagged wild-type strumpellin (A, C, E) or disease mutant strumpellin-myc V626F (B, D, F) and labelled with the markers indicated.

Supplementary Fig. 3

Wild-type and disease mutant strumpellin co-localise with EEA1 in HeLa cells. HeLa cells were transiently transfected with myc-tagged wild-type strumpellin (A) or disease mutant strumpellin-myc V626F (B) and labelled with the early endosomal marker EEA1.

Supplementary Fig. 4

SNX27 co-localises with retromer markers. HeLa cells were transfected with HA-tagged SNX27 and labelled with VPS26 (A) or Wash1 (B).

Supplementary Fig. 5

Strumpellin siRNA transfection depletes strumpellin from HeLa cells. Cell lysates prepared from cells transfected with strumpellin siRNA (n = 3 experiments) or from mock-transfected cells were immunoblotted with anti-strumpellin. Blotting versus tubulin serves as a loading control. This experiment also serves as validation that the signal detected by the anti-strumpellin antibody is specific. This strumpellin band runs at approximately 120 kD.

Fig. 1

The WASH and retromer complexes. The WASH complex comprises the core proteins KIAA1033, strumpellin, Wash1 and FAM21. Wash1, FAM21 and strumpellin interact directly with KIAA1033, while the association between the WASH and retromer complexes occurs via the tail domain of FAM21. FKBP15, a protein of unknown function, interacts with both retromer and the FAM21 tail. Retromer consists of a core complex of VPS26, 29 and 35, with an associated sorting nexin dimer comprising a combination of SNXs 1, 2, 5 and 6. SNX27 has been reported as an additional interactor of the WASH complex. Interactions between the sorting nexins and the WASH and retromer complexes are indicated by arrows.

Fig. 2

WASH complex is present in the brain and interacts with retromer. (A) and (B). Rat brain lysates were immunoprecipitated (IP) with the antibodies indicated, then immunoblotted (WB) with the antibodies shown. Spurious lanes on the blots have been spliced out, as indicated by the vertical lines. The strumpellin antibody appears relatively inefficient for IP, but sensitive for WB, whereas the WASH1 antibody is efficient for IP but relatively insensitive for WB.

Fig. 3

Wash1 and VPS26 localise to the dendrites and axons in neuronal cells. Rat primary cortical neurons were co-labelled with antibodies against MAP2, a marker of the cell body and dendrites and Wash1 (A) or VPS26 (B). Wash1 and VPS26 present in the MAP2 compartment, in both the cell body and dendrites (examples indicated with arrowheads). They were also present in non-MAP2 labelled regions (inset magnified images in A and B), indicating that the WASH complex and retromer are in axons. In these and subsequent images, the colour of the lettering in each greyscale panel indicates the colour of that image in the corresponding merged colour panel.

Fig. 4

Wash1 and VPS26 localise to the early endosomes in neurons. Rat primary cortical neurons were labelled with the markers indicated. Examples of puncta showing co-localisation are indicated by arrowheads. In the case of cells labelled with EEA1 and Wash1 (A) or EEA1 and VPS26 (B), the arrowheads indicate puncta in which the WASH or retromer complex marker appeared to label a subdomain of the EEA1 compartment. The magnified images (inset or below the labelled panels) correspond to the boxes indicated on the labelled figure panels.

Fig. 4

Wash1 and VPS26 localise to the early endosomes in neurons. Rat primary cortical neurons were labelled with the markers indicated. Examples of puncta showing co-localisation are indicated by arrowheads. In the case of cells labelled with EEA1 and Wash1 (A) or EEA1 and VPS26 (B), the arrowheads indicate puncta in which the WASH or retromer complex marker appeared to label a subdomain of the EEA1 compartment. The magnified images (inset or below the labelled panels) correspond to the boxes indicated on the labelled figure panels.

Fig. 5

Both wild type and disease mutant strumpellin co-immunoprecipitate with WASH and retromer complex components. Lysates prepared from HeLa cells transfected with either wild type or disease mutant strumpellin were used in co-immunoprecipitation experiments with either anti-Wash1 (A) or anti-SNX1 (B) as the IP antibody. Spurious lanes on the blots have been spliced out, as indicated by the vertical lines.

Fig. 6

Wild-type and disease mutant strumpellin co-localises with WASH, retromer and EEA1 in neurons. HeLa cells (A and B) or rat primary cortical neurons (C–E) were transiently transfected with myc-tagged wild-type strumpellin (A) or disease mutant strumpellin-myc V626F (C–E), and labelled with the markers indicated. The arrowheads indicate examples of co-localised puncta.

Fig. 7

Strumpellin mutation does not induce endosomal tubulation. HeLa cells were transfected with strumpellin-myc (A) or strumpellin-myc-L619F (B), then labelled with SNX1. In (B) the arrowheads indicate a tubular structure. (C) The mean number of SNX1 tubules per cell was counted in mock transfected cells (Mock) and in cells transfected with wild type strumpellin-myc (Wild type) or with the disease mutants strumpellin-myc V626F (V626F), strumpellin-myc L619F (L619F) and strumpellin-myc N471D (N471D). Error bars represent S.E.M., n = 3 experiments.

Fig. 8

Strumpellin depletion, but not mutation, alters the distribution of SNX27 and β2AR. (A–D) HeLa cells were subjected to mock siRNA transfection (A and C) or were transfected with siRNA versus strumpellin (B and D), and in addition were transfected with HA-SNX27 (A and B) or HA-β2AR (C and D). (E–H) HeLa cells were transfected with wild-type strumpellin-myc (E and G) or with strumpellin-myc V626F (F and H), as well as HA-SNX27 (E and F) or HA-β2AR (G and H). Strumpellin knock-down was confirmed by immunoblotting (Supplementary Fig. 5).

Fig. 8

Strumpellin depletion, but not mutation, alters the distribution of SNX27 and β2AR. (A–D) HeLa cells were subjected to mock siRNA transfection (A and C) or were transfected with siRNA versus strumpellin (B and D), and in addition were transfected with HA-SNX27 (A and B) or HA-β2AR (C and D). (E–H) HeLa cells were transfected with wild-type strumpellin-myc (E and G) or with strumpellin-myc V626F (F and H), as well as HA-SNX27 (E and F) or HA-β2AR (G and H). Strumpellin knock-down was confirmed by immunoblotting (Supplementary Fig. 5).

Fig. 9

Strumpellin depletion or mutation does not affect the uptake or recycling of the transferrin receptor. HeLa cells depleted of either strumpellin or clathrin heavy chain (as a positive control) were incubated with fluorescent transferrin and used in either uptake (A) or recycling (B) assays. The amount of internalised transferrin present after 20 min in each assay was measured by flow cytometry. To examine the steady-state distribution of the transferrin receptor, mock-transfected cells or cells transfected with siRNA against strumpellin were labelled for TfnR (C). Cells transfected with wild-type or disease mutant strumpellin, or with an empty vector, were also used in uptake (D) and recycling (E) assays, with depletion of the clathrin heavy chain used as a positive control. Error bars = S.E.M, n = 3 experiments in (A), (B), (D) and (E).