Legius syndrome mutations in the Ras-regulator SPRED1 abolish its membrane localization and potentially cause neurodegeneration

Abstract

The SPRED family proteins act as negative regulators of the Ras-ERK pathway: the N-terminal EVH1 domain interacts with the Ras-GAP domain (GRD) of the NF1 protein, while the C-terminal Sprouty-related (SPR) domain promotes membrane localization of SPRED, thereby recruiting NF-1 to Ras. Loss-of-function mutations in the hSPRED1 cause Legius syndrome in an autosomal dominant manner. In this study, we investigated the effects of missense mutations in the SPR domain identified in patients with Legius syndrome. Among the 18 mutations we examined, six (C368S, M369L, V408E, P415A, P415L, and P422R) have defects in the palmitoylation of the SPRED1 protein, losing plasma membrane localization and forming cytoplasmic granular aggregates. To evaluate the in vivo effects of SPR mutations, knock-in (KI) mice with P415A and P415V substitutions or M417Afs∗4, a C-terminal 28 amino acid deletion, were generated. All these KI mice exhibited cranial malformations, a characteristic feature of Legius syndrome. However, both P415A and P415V mutants formed granular aggregates, whereas M417Afs∗4 showed a diffuse cytoplasmic distribution, and Spred1^P415A and Spred1^P415V mice, but not Spred1^M417Afs∗4 mice, developed cerebellar ataxia and Purkinje cell loss with age. These data suggest that in addition to loss of palmitoylation, the C-terminal region is required for the granular aggregate formation and Purkinje cell loss. The autophagy inducer spermidine rescued the ataxia phenotypes and Purkinje cell loss in Spred1^P415A mice. These results suggest that some, but not all, SPR mutations that lose lipid modification induce abnormal cytoplasmic aggregation, which could be a target for autophagic clearance, and potentially cause neurodegenerative diseases.

Keywords: Legius syndrome; SPRED1; autophagy; membrane localization; neurodegeneration; palmitoylation; spermidine.

Figure 1

Mutations of Legius syndrome examined in this study.A, comparison of the amino acid sequences of human SPRED1 and mouse Spred1. Domain assignment was based on the categorization of mSpred1 domains. The identity between human SPRED1 and mouse Spred1 is 93.0%, with a similarity of 94.6%, indicating a high degree of conservation. The asterisks indicate the positions of 21 mutations in hSPRED1 found in Legius syndrome that we investigated in this study, all of which were found to be conserved between humans and mice. B, suppression of EGF-induced Erk reporter activity by SPRED1 variants. As a loss-of-function control, the T102R mutation in the EVH1 domain was used. After 24 h transfection of 10 ng SPRED1 cDNAs and reporter genes, HEK293T cells were stimulated with 50 ng/ml EGF for 6 h, followed by the luciferase assay. The transfection efficiency was monitored by β-galactosidase activity. Elk reporter activity was normalized to an empty vector (vector) as 100%. The number of each group was six (left panel) and three (right panel) and the error bars represent the mean ± standard error of the mean (SD). The graphs show representative data from three independent experiments. C, detection of ERK phosphorylation. WT or P415A SPRED1 plasmids together with GFP-ERK2 cDNA were introduced into HEK293T cells, followed by EGF stimulation at a final concentration of 50 ng/ml for the indicated time periods. The ratio of the quantified density of the bands (pERK/total ERK) is shown at the bottom. Data are representative of three independent experiments. D, mutations in hSPRED1 that profoundly affect its ability to inhibit ERK upon EGF stimulation are highlighted in red, those that cause a slight decrease are highlighted in blue, and mutations that have no effect on ERK inhibition are highlighted in black.

Figure 2

Effect of SPRED1 mutations on membrane localization.A, fluorescence images of HEK293T cells expressing mutant SPRED1-EGFP. Representative images from at least three independent experiments are shown. Bar; 20 μm. B, membrane raft localization of WT, R325∗, P415A and C416R EGFP-tagged SPRED1. HEK293T cells were transfected with the indicated cDNA plasmids and the lipid raft fraction was obtained by sucrose density gradient centrifugation. Total cell lysates (TCL) before centrifugation and lipid raft fraction (Raft) were analyzed by Western blotting. α-Tubulin and flotillin-1 were used as non-raft and raft markers, respectively. The data are representative of three independent experiments.

Figure 3

Effect of SPRED1 mutations on palmitoylation.A, detection of palmitoylation of SPRED1 by metabolic labeling with an alkyne-containing palmitate analogue, 17-ODYA. HEK293T cells were transfected with plasmids carrying EGFP- hSPRED1 WT or SPR mutants were metabolically labeled with 17-ODYA for 14h. After EGFP- hSPRED1 protein was immunoprecipitated with anti-GFP agarose, followed by click chemistry was performed to introduce biotin into ODYA. Palmitoylated EGFP-SPRED1 was detected using streptavidin-HRP. The R325∗ mutant, which lacks the C-terminal SPR domain, was used as a negative control. B, inhibition of palmitoylation by 2-BP (2-bromohexadecanoic acid). WT EGFP-hSPRED1 was expressed in 293T cells on 4-well chamber slides and incubated with 2-BP (200μM or 500 μM) for 20 h after transduction. The cells were then fixed in 4% paraformaldehyde and the localization of WT EGFP-hSPRED1 was examined by fluorescence microscopy. Bar; 20 μm. C and D, effect of cysteine to serine substitution at the C-terminal region of the SPR domain on subcellular localization, ERK inhibitory activity, and palmitoylation. The indicated cysteine (C) to serine (S) substitution mutations and C-terminal deletion mutants, C435∗ (C-terminal 10 aa deletion), C419∗ (C-terminal 26 aa deletion), S388∗ (C-terminal 56 aa deletion) of EGFP-hSPRED1 were examined for their subcellular localization by fluorescence microscopy. Bar; 20 μm. ERK suppression activity was accessed by transfecting WT and mutant hSPRED1 plasmids together with Elk reporter plasmids. EGF-induced Elk reporter activity without SPRED1 vector was normalized to 100%. N = 3 for each transfection. In (D), palmitoylation of SPRED1 was detected by metabolic labeling with 17-ODYA as shown in (A). Representative data from at least two independent experiments are shown.

Figure 4

Effect of zDHHC enzymes on SPRED1 localization and activity.A, HEK293T cells were transfected with 50 ng of each Elk-1 reporter, β-galactosidase gene, 10 ng zDHHC expression vectors, and 5 ng FLAG-tagged SPRED1 expression vectors. After 24 h, cells were stimulated with 50 ng/ml EGF for 6 h and then the activity of luciferase and β-galactosidase was analyzed. ∗∗p < 0.01 Dunnet’s test. The number of each group was three (N = 3), and the error bar represents mean ± SEM. B, effect of zDHHC24 overexpression on SPRED1 palmitoylation. Palmitoylation of SPRED1 was detected by metabolic labeling with 17-ODYA as shown in Figure 3A. C, effect of zDHHC1 or zDHHC24 on WT and SPR mutant SPRED1 activity. EGF-induced Elk reporter activity was measured in the presence or absence of SPRED1 and/or zDHHC plasmids. D, subcellular localization of WT and mutant EGFP-SPRED1 when co-expressed with zDHHC1 or zDHHC24. EGFP-hSPRED1 was examined for its subcellular localization by fluorescence microscopy. Bar; 20 μm.

Figure 5

Generation of Spred1-P415 knock-in (KI) mice.A, the P415A substitution was designed by homologous recombination using oligo-DNA and CRISPR/Cas9. As by-products, a P415V substitution and a 4 bp deletion (M417Afs∗4) resulting in a frameshift that lacked C-terminal 28 amino acids were also obtained. B, subcellular localization of EGFP-labeled P415A, P415V, and M417Afs∗4 mutants of murine Spred1 expressed in HEK293 cells. Bar; 20 μm. C, brain deformities of 14-week-old homozygous Spred1^P415A/P415A (A/A), Spred1^P415V/P415V (V/V), and Spred1^{M417Afs∗4/M417Afs∗4} (del/del) mice. D, immunostaining of the brain from 8-week-old BL6 male mice with SPRED1 antibody (green) and anti-MAP2 antibody (red). The arrows indicate membrane localization of WT SPRED1 protein. Bars; 100 μm and 20 μm, respectively. E, enlarged view of anti-Spred1 antibody staining (red) of WT and A/A mutant the Purkinje cells of the cerebellum. Arrows indicate membrane localization of Spred1 protein. F, the number of cerebellar Purkinje cells in WT and A/A mutant mice (19-month-old male mice; N = 6). Unpaired t test ∗∗∗p < 0.0001.

Figure 6

Cerebellar Purkinje cells in WT and heterozygous Spred1^P415A/+(A/+), Spred1^P415V/+(V/+), and Spred1^{M417Afs∗4/+}(del/+) mice. (Upper panels) Purkinje cells in the cerebellum were stained with anti-calbindin antibody (green) and anti-Spred1 antibody (red) in 5-month-old (young) and 19-month-old (aged) KI heterozygous mice and their littermates (+/+). The number of cerebellar vermis lobules is expressed as a Roman number. Bars; 100 μm (green) and 20 μm (red), respectively. The arrow indicates membrane localization of WT SPRED1 protein and arrow heads indicate intracellular aggregates of mutant SPRED1. (Bottom panel) Quantitative counting of cerebellar Purkinje cells. Three sections from three mice per group were used. Data are expressed as the mean ± SEM ∗p < 0.05 (Tukey's test).

Figure 7

Cerebellar ataxia symptoms of aged Spred1^P415A/+(A/+), Spred1^P415V/+(V/+) mice, and Spred1^{M417Afs∗4/+}(del/+) mice.A, the hind-limb clasping test. The number of mice tested is shown. B, footprint test. For the footprint assay, the step width of the hindlimb (mm) was measured (left panels). N = 4 to 7. Representative footprint patterns of WT (+/+) and heterozygous (A/+) or (del/+) mice are shown (right). Data are expressed as the mean ± SEM ∗p < 0.05 (Tukey's test).

Figure 8

Effect of spermidine (SPD) treatment on ataxia symptoms (A) and cerebellar Purkinje cell number (B) in WT (+/+) and heterozygous Spred1^P415A/+(A/+) mice.A, the hind-limb clasping test of young (5–6 months old), aged (19–20 months old) (+/+) or (A/+) mice. SPD (3 mM) in water was administered ad libitum. Control animals received normal autoclaved drinking water. Eight to nine 11- to 12-month-old mice were treated with SPD daily for 3 months, then every 2 weeks for a total of 8 months (A) The hind-limb clasping test was performed. B, Purkinje cells in the cerebellum were stained with an anti-calbindin antibody and quantified as shown in Figure 6. p < 0.05 (Tukey's test).