Loss of association of REEP2 with membranes leads to hereditary spastic paraplegia

Abstract

Hereditary spastic paraplegias (HSPs) are clinically and genetically heterogeneous neurological conditions. Their main pathogenic mechanisms are thought to involve alterations in endomembrane trafficking, mitochondrial function, and lipid metabolism. With a combination of whole-genome mapping and exome sequencing, we identified three mutations in REEP2 in two families with HSP: a missense variant (c.107T>A [p.Val36Glu]) that segregated in the heterozygous state in a family with autosomal-dominant inheritance and a missense change (c.215T>A [p.Phe72Tyr]) that segregated in trans with a splice site mutation (c.105+3G>T) in a family with autosomal-recessive transmission. REEP2 belongs to a family of proteins that shape the endoplasmic reticulum, an organelle that was altered in fibroblasts from an affected subject. In vitro, the p.Val36Glu variant in the autosomal-dominant family had a dominant-negative effect; it inhibited the normal binding of wild-type REEP2 to membranes. The missense substitution p.Phe72Tyr, in the recessive family, decreased the affinity of the mutant protein for membranes that, together with the splice site mutation, is expected to cause complete loss of REEP2 function. Our findings illustrate how dominant and recessive inheritance can be explained by the effects and nature of mutations in the same gene. They have also important implications for genetic diagnosis and counseling in clinical practice because of the association of various modes of inheritance to this new clinico-genetic entity.

Figure 1

Pedigrees and Segregation of the REEP2 Mutations (A) Pedigrees of two families with hereditary spastic paraplegia. Segregation of the mutations is indicated (plus sign indicates wild-type, M indicates mutation). Affected individuals are designated by black squares (men) or circles (women). Sampled individuals are designated by an asterisk (^∗). Subjects with unknown status are designated with a question mark. The arrow indicates the index subject. (B) Schematic representation of the gene showing the sites of the mutations. (C) Electrophoregrams showing the mutations. (D) Conservation of amino acids in the N-terminal domain of REEP2 according to the phylogenic evolutionary tree (top) and in comparison with other members of the REEP family in H. sapiens. The positions of the missense mutations are shown by arrows. Conserved amino acids are in yellow (among species), red (among all REEP proteins), or blue (among members of the REEP1–REEP4 subfamily). Horizontal bars represent the hydrophobic segments of the proteins.

Figure 2

REEP2 Is Implicated in ER Morphogenesis (A) Immunoblot showing the amount of endogenous REEP2, actin, and CLIMP-63 in fibroblasts from a control and family member FSP200-IV.4. (B) Control and FSP200-IV.4 fibroblasts were transfected with vectors directing expression of GFP-Sec61β (red, false color) and immunostained with antibodies against CLIMP-63 (green). (C) Quantification of the area occupied by CLIMP-63 staining in cells (control n = 50, FSP200-IV.4 n = 57). Data represent mean ± SEM. ^∗∗∗p < 0.0001, t test. (D) Transmission electron micrographs of control and FSP200-IV.4 fibroblasts. Asterisk indicates swollen ER; arrowhead points to a very thin ER sheet. (E) Quantification of the thickness of ER sheets in fibroblasts. Control n = 210, FSP200-IV.4 n = 237. ^∗∗∗p = 4 × 10⁻⁸, chi-square test.

Figure 3

REEP2 Downregulation Leads to Expansion of ER Sheets (A) COS7 cells were transfected with vectors directing the expression of shRNAs obtained from Sigma (shRNA REEP2_A hairpin sequence: CCCAGCCTATTCTTCCTACAACTCGAGTTGTAGGAAGAATAGGCTGGGT, shRNA REEP2_B hairpin sequence: ACTGGCTTCCAAGACACTGAACTCGAGTTCAGTGTCTTGGAAGCCAGTT). Cell lysates were immunoblotted with antibodies against endogenous REEP2 and actin. (B) COS7 cells in which endogenous REEP2 was downregulated by shRNA were immunostained with antibodies against the endoplasmic reticulum marker calreticulin and CLIMP-63 that labels ER sheets. (C) Quantification of the area occupied by CLIMP-63 staining in cells. n > 50 cells in three independent experiments. Data represent mean ± SEM. ^∗∗∗p < 0.001, one-way ANOVA. (D) COS7 cells were transfected with vectors directing the expression of wild-type or V5-tagged p.Val36Glu REEP2. Cells were immunostained with anti-V5 and anti-CLIMP-63 antibodies. (E) Quantification of the cellular area occupied by CLIMP-63 staining. n > 50 cells per condition. Data represent mean ± SEM. ^∗p < 0.05, one-way ANOVA.

Figure 4

Interaction of REEP2 with REEP1 and Membranes (A) Homogenates (Homog.) from V5-tagged REEP2-expressing COS7 cells were separated into soluble (Sol.) and membrane (Memb.) fractions and then immunoblotted to detect REEP2 (anti-V5 antibody), the membrane marker calnexin, and the cytosolic marker tubulin. (B) Lysed membranes from V5-tagged REEP2-expressing COS7 cells were subjected to alkaline extraction, and the soluble and pellet fractions were analyzed by immunoblot to detect REEP2 (anti-V5 antibody), the soluble protein calreticulin, and the integral membrane protein calnexin. (C) Schema representing the liposome flotation experiment. Liposomes were prepared by mixing chloroform solutions of brain phosphatidylcholine and phosphatidylserine (Avanti Polar Lipids) in 75:25 ratio. After drying the lipids under nitrogen, they were resuspended in 100 mM NaCl, 10 mM Tris (pH 7.4) and sonicated. Liposomes were incubated with recombinant protein for 30 min at 22°C, and then mixed with equal amount of 50% sucrose in 100 mM NaCl, 10 mM Tris (pH 7.4). The mixture was overlaid with 40 μl of 15% sucrose solution. After centrifugation, the liposomes were collected at the top of the upper phase (top) and unbound proteins in the lower phase (bottom). (D) Liposome flotation assay with wild-type REEP2, p.Val36Glu REEP2, p.Phe72Tyr REEP2, and an equimolar amount of wild-type and p.Val36Glu REEP2 analyzed by immunoblot with an antibody raised against REEP2. (E) Quantification of the association of recombinant REEP2 with membranes. Data represent mean ± SEM (n = 4). ^∗∗∗p < 0.001, one-way ANOVA. (F) V5-tagged REEP2 and HA-tagged REEP1 were coexpressed in COS-7 cells and REEP2 immunoprecipitated with an anti-V5 antibody. The eluates were analyzed by immunoblot. (G) Quantification coimmunoprecipitation of HA-REEP1 with REEP2. The p.Phe72Tyr variant in REEP2 decreased the interaction of REEP2 with REEP1. Data represent mean ± SEM (n = 3). ^∗p < 0.05.