reep1 null mice reveal a converging role for hereditary...

O R I G I N A L A R T I C L E

Reep1 null mice reveal a converging role for hereditary

spastic paraplegia proteins in lipid droplet regulationBenoıt Renvoise1 Brianna Malone1 Melanie Falgairolle2Jeeva Munasinghe3 Julia Stadler1 Caroline Sibilla1dagger Seong H Park 1 andCraig Blackstone11Cell Biology Section Neurogenetics Branch 2Developmental Neurobiology Section and 3Laboratory ofFunctional and Molecular Imaging National Institute of Neurological Disorders and Stroke National Institutesof Health Bethesda Maryland USA

To whom correspondence should be addressed at Craig Blackstone Neurogenetics Branch NINDS NIH Building 35 Room 2A-201 9000 Rockville PikeBethesda MD 20892-3738 USA Tel thorn1 3014519680 Fax thorn1 3014804888 E-mail blackstcnindsnihgov

AbstractHereditary spastic paraplegias (HSPs SPG1-76 plus others) are length-dependent disorders affecting long corticospinal axonsand the most common autosomal dominant forms are caused by mutations in genes that encode the spastin (SPG4) atlastin-1 (SPG3A) and REEP1 (SPG31) proteins These proteins bind one another and shape the tubular endoplasmic reticulum (ER)network throughout cells They also are involved in lipid droplet formation enlargement or both in cells though mecha-nisms remain unclear Here we have identified evidence of partial lipoatrophy in Reep1 null mice in addition to prominentspastic paraparesis Furthermore Reep1-- embryonic fibroblasts and neurons in the cerebral cortex both show lipid dropletabnormalities The apparent partial lipodystrophy in Reep1 null mice although less severe is reminiscent of the lipoatrophyphenotype observed in the most common form of autosomal recessive lipodystrophy Berardinelli-Seip congenital lipodystro-phy Berardinelli-Seip lipodystrophy is caused by autosomal recessive mutations in the BSCL2 gene that encodes an ERprotein seipin that is also mutated in the autosomal dominant HSP SPG17 (Silver syndrome) Furthermore REEP1co-immunoprecipitates with seipin in cells This strengthens the link between alterations in ER morphogenesis and lipidabnormalities with important pathogenic implications for the most common forms of HSP

Introduction

Hereditary spastic paraplegias (HSPs) are characterized byprominent lower extremity spasticity and typically more mildweakness resulting from a length-dependent axonopathy ofcorticospinal motor neurons (1ndash5) Historically HSPs have beenclassified as lsquopurersquo or lsquocomplicatedrsquo based on the absence (pure)or presence (complicated or complex) of additional clinicalfeatures including distal amyotrophy ataxia neuropathyparkinsonism cognitive changes or visual changes amongothers More recently a genetic classification scheme has

predominated and HSPs are most often referred to by their ge-netic loci (numbered in order of identification SPG1-76) Stillothers have not been assigned an SPG locus (1ndash46) To dateover 60 HSP gene products have been identified and functionalanalyses support pathogenic convergence within a relativelysmall number of common cellular themes (2)

SPG4 SPG3A and SPG31 are the most common autosomaldominant HSPs (listed in order of frequency) together compris-ing 50 of all patients with HSP Indeed the three disease

daggerPresent address MRC Laboratory of Molecular Biology Cambridge UKReceived May 25 2016 Revised August 4 2016 Accepted September 12 2016

Published by Oxford University Press 2016 This work is written by US Government employees and is in the public domain in the US

1

Human Molecular Genetics 2016 Vol 0 No 0 1ndash15

doi 101093hmgddw315Advance Access Publication Date 16 September 2016Original Article

gene products atlastin-1 spastin and receptor expression-enhancing protein 1 (REEP1) respectively together with theSPG72 protein REEP2 SPG12 protein reticulon 2 SPG61 proteinARL6IP1 and SPG69 protein RAB3GAP2 play roles in shapingand distributing the tubular endoplasmic reticulum (ER)network in cells (26ndash8) REEP1 atlastin-1 and the largermembrane-bound M1 isoform of spastin interact through hy-drophobic hairpin domains within the tubular ER to coordinateER shaping and microtubule interactions (79ndash12) Interestinglyatlastin-1 spastin and REEP1 localize prominently to the corti-cospinal neurons whose axons are impaired in HSPs (13ndash17)

SPG31 results from autosomal dominant REEP1 mutations(18) It is mostly a pure form only rarely associated withneuropathy and in just one family was an axonal neuropathypresent without signs of hyperreflexia Some mutations disrupt apredicted miRNA binding site in the 3rsquo untranslated region sup-porting haploinsufficiency in at least a subset of patients andheterozygous missense and truncation mutations have also beendescribed throughout the gene (1819) A young boy with a homo-zygous loss-of-function mutation presented with a spinal muscu-lar atrophy with respiratory distress type 1 (SMARD1)-likephenotype with additional features of congenital axonal neurop-athy hyperreflexia and diaphragmatic palsy (20)

REEP proteins were first described based on their ability toenhance surface expression of some G-protein coupled recep-tors including olfactory and taste receptors (2122) REEP1 is amember of a family of ER shaping proteins comprising sixmembers in humans divided structurally and functionally intotwo main subgroups REEP1-4 and REEP5-6 (72324) withREEP1-4 proteins involved in microtubule interactions with ERtubules (725) REEP1 has also been reported to facilitateER-mitochondrial interactions (26) Finally REEP proteins havealso been implicated in enhancing ER cargo capacity regu-lating ER-Golgi processing and interacting selectively withsome cargo proteins (23) Given these numerous functionalroles the in vivo consequences of REEP1 mutation remainlargely unknown

The ER is a continuous membrane organelle comprising thenuclear envelope and a peripheral network of membrane tu-bules and sheets (2728) Tubular ER presents high membranecurvature in cross-section stabilized by members of the reticu-lon and REEP protein families (1011) likely by forming wedge-like structures in the lipid bilayer as well as arc-like scaffoldsaround tubules (10112428) Identifying effects of changes in ERmorphology thus appears critical for understanding HSP patho-logic mechanisms Of course this is complicated by the factthat the ER has many different functions and interactions withnumerous organelles and cytoskeletal elements (2728)

Recent attention has focused on the role of the ER in the for-mation of lipid droplets (LDs) organelles involved in fat storagein eukaryotes (29) LDs contain a core of neutral lipids (triglycer-ides and sterol esters) surrounded by a phospholipid monolayercontaining proteins such as perilipins Interestingly atlastin-1regulates LD size with REEP1 also playing a role albeit less clear(3031) Furthermore the HSP proteins seipin (SPG17) spartin(SPG20) and spastin (SPG4) also function in LD maintenance in-dicating a possible convergent pathogenic theme (232)Atlastin-1 and REEP1 can be co-immunoprecipitated from cells(7) and SPG17 protein seipin (Fld1p in yeast) is a conserved inte-gral membrane ER protein thought to act at the interface of theER and LDs (33) Recently overexpressed atlastin-1 has been co-precipitated with the SPG73 gene product CPT1C an ER-localized carnitine palmitoyltransferase (34)

Here we report that Reep1-- mice exhibit a gait phenotypeand prominent spasticity along with significant lipodystrophyThese symptoms are associated with prominent changes in cel-lular LDs Reep1 localizes to the ER and interacts with atlastin-1as well as with the SPG17 protein seipin that is also mutated inthe most common form of autosomal recessive lipodystrophyBerardinelli-Seip congenital lipodystrophy suggesting a modelwhereby REEP1 may function with other key HSP gene productsto regulate LD formation and size

ResultsReep1-- mice exhibit spasticity and motor dysfunction

Reep1-- mice were generated as shown in SupplementaryMaterial Figure S1A and disruption of the Reep1 gene was con-firmed using PCR (Supplementary Material Fig S1B)Immunoblotting of multiple mouse tissues established the ab-sence of the Reep1 protein (Supplementary Material Fig S1C)Reep1-- mice bred normally and their offspring appeared nor-mal at birth Early in life they moved throughout the cage muchlike littermate controls and appeared to feed normallyHowever they soon developed age-dependent impairments inmotor function Compared to wild-type animals Reep1-- miceexhibited spasticity with lower extremity clonus at postnatalday 3 (compare Supplementary Material Videos S1 and S2)prominent impairment at 2 months in the Rotarod test (Fig 1A)and significant differences in gait parameters such as stridetime stance time and stance length upon formal TreadScantesting (Fig 1A and Supplementary Materials Fig S1D andMovies S3ndashS7) However grip test data showed no differences(Supplementary Material Fig S1E) At 4 months 50 ofReep1-- mice were unable to walk at the same speed as Reep1thornthornmice Thus Reep1-- mice have a progressive defect in ambu-lation reminiscent of that in SPG31 patients as reported previ-ously (17) In preliminary studies Reep1thorn- animals hadessentially no phenotype even at more advanced ages (data notshown) also consistent with a previous report (17) and theywere not investigated further

Reep1-- motoneurons exhibit increased activity

Gait dysfunction affecting the hind limbs in mice can resultfrom different types of impairments in the pyramidal motorpathway To understand the basis for the Reep1-- phenotypeneurophysiologically we analysed motoneuron activity (35) In3-day-old Reep1-- mice locomotor-like activity frequency washigher (Fig 1B Supplementary Material Fig S1FndashH) Also la-tency of the suprathreshold monosynaptic reflex (L5 segment)was shortened in Reep1-- mice (Fig 1CndashE SupplementaryMaterial Fig S1F and H) independent of the site of stimulation(Supplementary Material Fig S1H) Together these data sup-port spasticity as a principal cause of the impaired gait in theReep1-- mice Motoneuron staining using fluorescent dyes alsoshowed that the Reep1-- lateral motor pool is slightly smaller(Fig 1F and G) Intracellular input resistances recorded fromReep1-- and Reep1thornthornmotoneurons were not different how-ever suggesting that Reep1-- motoneuron input resistancechanges are either due to a decrease in the extent of the den-dritic tree or else increased membrane resistivityMonosynaptic reflexes did not show differences intracellularly(Supplementary Material Fig S1F and H)

2 | Human Molecular Genetics 2016 Vol xx No xx

Disruption of the Reep1 gene causes reduction inneuronal axon length and decreased branching

Four month-old male Reep1 Reep1thornthornand Reep1thorn- micewere phenotyped pathologically (NIH Veterinary ResearchProgram) with no gross differences in particular within theCNS Similarly more detailed evaluations of sections throughcerebral cortex and spinal cord revealed no obvious differencesin Reep1-- mice (Supplementary Material Fig S2) To assess the

Reep1 developmental expression in the CNS we measuredmRNA levels of Reep1 and its closest orthologue Reep2 in sev-eral regions at multiple time points (Supplementary MaterialFig S3) Reep1 mRNA expression is particularly high in the cere-bral cortex at post-natal day 7 (P7) while Reep2 expression ismore constant throughout the CNS and during development

The overall structure of the spinal cord and numbers of ante-rior horn cells are not changed in Reep1-- mice (SupplementaryMaterial Fig S2C and E) However myelin staining of thoracic

Figure 1 Reep1-- mice exhibit motor dysfunction and axon abnormalities (A) Rotarod duration testing of Reep1thornthornand Reep1mice at 2 months (nfrac1420) Stride time

and print area were obtained using a TreadScan Gait device (speedfrac1414 cmsec nfrac1420) means 6 SEM Plt0001 (B) Frequencies of locomotor-like activity evoked by

10 threshold train of stimuli to dorsal root ganglion Plt0001 (C) Measurement of monosynaptic reflex latency from L5 ventral root Left Experimental set-up mo-

toneurons are blue and afferents are red Right Superimposed traces showing averaged monosynaptic reflexes recorded in one Reep1thornthornand one Reep1-- animal

(D) Average monosynaptic latencies for 2and 10 thresholds Means 6 SEM Plt005 Plt0001 ns not significant (E) Antidromic latencies of intracellularly-

recorded L5 motoneurons (F) Transverse L5 spinal cord sections showing dorsal root fibres filled with fluorescent dextran (red) with motoneurons in blue Scale bar

100 lm (G) Mean sizes 6 SEM of L5 motoneurons calculated from F Plt0001 (H) Microslice 100 lm sections of cervical spinal cord from 4-month old mice were

stained with toluidine blue Scale bar 100 lm (I) Axon diameters in the ventral tract (means 6 SEM nfrac143) Plt 0001 (J) Immunoblot of Reep1 protein in lysates from

cortical neurons with actin as a loading control (K) Representative DIV6 cortical neurons were stained for b-tubulin (black) Scale bar 20 mm (L) Left Quantifications of

primary axon length in DIV3 neurons (means 6 SEM nfrac143 with 30 neurons per trial) with overexpression of REEP1 where indicated Plt005 ns not significant Right

Quantifications of primary axon branches in neurons (means 6 SEM nfrac14 3 with 30 neurons per trial) Plt001

3Human Molecular Genetics 2016 Vol xx No xx |

spinal cord sections from 4-month-old Reep1-- mice revealedvisible degeneration and quantifications showed a significantreduction in axon diameter in both ventral and dorsal tracts ofReep1-- spinal cord (Fig 1H and I and Supplementary MaterialFig S1G) A number of mammalian cell culture models for HSPshave exhibited alterations in axon elongation and branchingtherefore we analysed effects of Reep1 loss on axon outgrowthin cultured cortical neurons at 6 days in vitro (DIV) (Fig 1JndashL)Axon length and branch numbers were significantly reduced inReep1-- neurons an effect rescued upon expression of recombi-nant human REEP1 (Fig 1L)

Reep1 disruption leads to lipoatrophy

While there were no differences in body weight (SupplementaryMaterial Fig S4A) visually Reep1-- mice appeared thinner (Fig2A) and torso dissection showed a significant reduction of thefat tissue (Fig 2B) Though gross structures of white adipose(WAT) and liver tissues seemed similar (Fig 2C SupplementaryMaterial Fig S4B) the proportion of total adipose tissue mea-sured by body CT scan was significantly decreased in Reep1male mice (Fig 2D and E) To establish whether there weredirect effects of Reep1 loss in fat tissue we performed immuno-blotting Reep1 was absent in Reep1 fat but present inReep1thornthorn and Reep2 levels were also markedly decreased inReep1 fat (Fig 2F) This is likely not due to widespreadchanges in ER shaping proteins since Reep1-- mice had no dif-ferences in levels of the ER protein Reep5 (SupplementaryMaterial Fig S4C) Finally fasting triglyceride cholesterol andinsulin levels were also significantly reduced in Reep1-- serum(Fig 3A Supplementary Material Fig S2A) consistent with dys-lipidemia but not the commonly observed pattern of hypercho-lesterolemia and hypertriglyceridemia typically associated withhuman lipodystrophies

We next examined whether expression of Reep1 was associ-ated with adipogenesis regulation in WAT from Reep1-- andReep1thornthornmice using qPCR The expression of many pro-adipogenesis markers such as Cebpd Dkk1 Fgf2 and Jun wassignificantly reduced (Fig 3B) In contrast expressions of anti-adipogenic such as Foxo1 and Taz1 were significantly up-regulated in Reep1-- mice compared to Reep1thornthornmice (Fig 3C)consistent with our earlier observation that Reep1-- mice ex-hibit a significant reduction of adipose tissue accumulationTogether these findings suggest that Reep1 acts during adipo-cyte differentiation to regulate adipogenesis possibly via itsbinding partners in the ER

Lipid changes in Reep1-- brain

To assess any roles of Reep1 in CNS lipid metabolism we exam-ined lipid profiles in brain of Reep1and Reep1thornthorn animalsin vivo Importantly SPG54 patient brains have an abnormallipid peak on MR spectroscopy (36) While Reep1thornthorn and Reep1--mice had a similar overall lipid spectrum our analysis of MRI T2

values from several brain regions revealed significant alter-ations in Reep1-- cerebral cortex suggesting that waterlipidcompositions are altered This reduction appeared selectivesince MRI T2 values in other regions were not different (Fig 4ASupplementary Material Fig S4DndashE) In addition in the Reep1--mice numerous metabolite levels were higher in particularglutamate (Supplementary Material Fig S4EndashG) Together theseresults support a role for Reep1 in CNS lipid metabolism partic-ularly in cerebral cortex

Reep1-- cells have LD defects and altered ERmorphology

Several groups have shown that overexpressed REEP1 is in-volved in shaping ER tubules and LD formation though in somecases the LD-like structures do not stain for neutral lipid(7303137) Analysis of Reep1-- mouse embryonic fibroblasts(MEFs) showed a dramatic alteration of ER organization(Supplementary Material Fig S5A and B) with a larger propor-tion of apparent sheet ER in Reep1-- cells (SupplementaryMaterial Fig S5B and C) in agreement with a previous study(17) Cell fractionation studies using Reep1thornthorn and Reep1-- MEFsshowed that Reep1 specifically localizes to ER and there are nodifferences in mitochondrial complex I activity (SupplementaryMaterial Fig S5DndashF)

To investigate further any role for Reep1 in LD formation westained cerebral cortical sections for LDs using the ValaSciences Lipid Droplet Screen-Certified Kit and significant LDabnormalities were noted (Fig 4B and C) We also investigatedMEFs which express low levels of Reep1 protein (Fig 5A)Quantification of the fluorescence intensities of BODIPY 493 andLD540 staining were similar in Reep1and Reepthornthorn cells witha trend toward modest reduction in Reep1-- cells (Fig 5B and C)However after cells were stimulated with oleic acid a monoun-saturated fatty acid that stimulates LD formation staining in-tensities in Reep1thornthorn cells increased significantly while theyremained essentially unchanged in Reep1 cells (Fig 5B and Cand Supplementary Material Fig S5G) Upon differentiation ofMEFs into adipocytes Reep1thornthorn cells form large LDs whileReep1-- cells do not (Fig 5D) Together these results establish aspecific role for Reep1 in LD formation

Interestingly we observed that the number of LDs (Fig 5E) issimilar in Reep1-- MEFs before and after treatment with oleicacid (as assessed using LD540) while the size of LDs is signifi-cantly reduced compared to the wild-type cells (Fig 5F) In MEFstransfected with a Reep1 expression construct the number ofLDs increased in both Reep1thornthornand Reep1-- cells after treat-ment with oleic acid (Fig 5E) recombinant Reep1 expression in-creased LD size in the Reep1-- cells These results suggest thatReep1 is involved in determining the size of LDs in cells as pre-viously reported for its binding partner atlastin-1 (30)

Reep1-- cells have perilipin defects in LDs

There is an abnormal accumulation of TIP47 a member of theperilipin family on the surface of LDs in Reep1-- MEFs (Fig 5G-I)Along these lines we observed that levels of perilipin A whichcoats LDs and is involved in regulating lipid stores are decreasedin Reep1-- WAT (Fig 6A) Furthermore phosphorylation of perili-pin was significantly increased in Reep1--tissue (Fig 6B) whichmay trigger fragmentation and dispersion of LDs (38) This hy-pothesis was confirmed by treating MEFs with a specific inhibitorof protein kinase A (PKA)-dependent phosphorylation at perilipinSer492 Rp-8-PIP-cAMPs which restored oleic acid-mediated in-creases in LD size in Reep1-- MEFs (Fig 6C and D)

Reep1 and atlastin-1 in LD formation

To understand how absence of Reep1 can modify ER shapingand LD formation we analysed its binding partner atlastin-1a membrane-bound ER GTPase mutated in SPG3A (7)Recombinant atlastin-1 and Reep1 expressed in MEFs localize toER tubules and for Reep1 in particular along bundled ERtubule-like structures (Supplementary Material Fig S6A) as


described previously (7) In NIH-3T3 and COS7 cells Reep1 andatlastin-1 co-overexpression can induce formation of large LDsand this requires atlastin-1 GTPase activity (SupplementaryMaterial Fig S6B) In addition there is a massive recruitment ofadipose differentiation-related protein (ADRP)perilipin-2 to LDsin cells overexpressing both REEP1 and ADRP (SupplementaryMaterial Fig S6C) Unexpectedly in atlastin-1REEP1 overex-pressing COS7 cells these proteins localize not only to rims sur-rounding LDs but also co-localize with the ER luminal proteincalregulin (Supplementary Material Fig S6D) indicating that ERtubules may surround LDs thus complicating clear interpreta-tion of LD versus ER membrane protein localizations

Interestingly atlastin-1 protein levels are significantly de-creased in Reep1-- cells (Supplementary Material Fig S6E) consis-tent with a role for atlastin-1 in regulating LD size (30) Moreoverwhile expression of recombinant human REEP1 in MEFs rescuedthe Reep1-- LD phenotype before and after treatment with oleicacid expression of atlastin-1 or GTPase-defective atlastin-1 K80Adid not restore normal LD formation (Fig 7) These observationsindicate that Reep1 is important for the formation of LDs andcould mechanistically link atlastin-1 and Reep1

Reep1 binds the SPG17BSCL2 protein seipin

Finally the SPG17BSCL2 protein seipin that also localizes to ERis known to be involved in regulating LD formation and

morphology In fact recombinant Myc-REEP1 and HA-seipin in-teract robustly as assessed by co-immunoprecipitation studiesin NIH-3T3 cells (Fig 8A) Endogenous co-immunoprecipitationof CHAPS-solubilized mouse brain extracts did not reveal spe-cific co-precipitation though Reep1 robustly interacted withatlastin-1 under these conditions (Fig 8B) Importantly mousebrain extracts chemically cross-linked using dithiobis(succini-midyl propionate) (DSP) exhibited a clear interaction betweenendogenous Reep1 and atlastin-1 proteins suggesting that theinteraction might be transient (Fig 8C) Additional supportiveexperiments in COS7 cells show that recombinant Reep1 co-localizes closely with wild-type seipin and two well-describedpathological seipin SPG17 missense mutant proteins (Fig 8D)

DiscussionMost cases of pure HSP result from autosomal dominant muta-tions in genes for SPG4 SPG3A or SPG31 (2) Recently Reep1ndashndashmice were generated (17) and these animals developed hindlimb dysfunction with weakness spasticity and degeneration ofaxons of the corticospinal tract Reep1ndashndash cells also had defectsin ER structure (17) The Reep1-- mice described here similarlyrecapitulate many aspects of human SPG31 with neurophysio-logic studies showing the physiologic basis for the spasticityThus our results suggest that Reep1-- mouse spasticity resultsfrom increased motoneuron responsiveness possibly in the

Figure 2 Reep1 disruption causes lipodystrophy (A) Representative photograph of Reep1thornthornand the thinner Reep1mice (B) Images showing a decrease in the ab-

dominal white fat deposits (black arrows) in Reep1mice (C) Photos of sections of adipose tissue stained with BODIPY 594 Scale bar 500 mm (D) Representative im-

ages of CT scan analysis of total fat tissue in Reep1thornthornand Reep1mice (E) Volumes of fat tissue in Reep1thornthornand Reep1mice quantified from CT scans as shown

in D (nfrac14 10 mice means 6 SEM Plt0001) (F) Immunoblot analysis of Reep1 and Reep2 Actin was monitored as a control for protein loading WAT white adipose tis-

sue non-specific bands


Figure 3 Altered serum lipid parameters in Reep1-- mice (A) Serum triglycerides cholesterol and insulin are decreased in Reep1mice Chol cholesterol HDL high-

density lipoprotein LDL low-density lipoprotein (BndashC) Quantitative qRT-PCR analysis of pro- (B) and anti- (C) adipogenesis mRNAs using the Qiagen Mouse

Adipogenesis PCR Array Graphs show means 6 SEM (nfrac143 mice each in triplicate) Plt005 Plt001 Plt0001

Figure 4 LD alterations in Reep1-- mouse cerebral cortex (A) Representative T2 relaxation time map (additional details in Supplementary Material Fig S4D and E) for

Reep1thornthornand Reep1brain The pixel intensity of the T2 maps represents the quantitative T2 values (in ms scale bar to the right) (B) Cortical sections of Reep1thornthornand

Reep1mice stained for tubulin (gray) LDs (Vala green) and DAPI (blue) Boxed regions are enlarged below Scale bar 100 mm (C) Quantification of number and size

of LDs in cell bodies of neurons from sections as in B (nfrac143 mice means 6 SEM) Plt005


setting of higher levels of the neurotransmitter glutamate Infact several groups have already shown that glutamate re-lease can maintain and amplify the ongoing activity of moto-neurons contributing to clinical signs of spasticity and rigidity(3940)

Another key finding here is that Reep1 is also expressed inWAT albeit at low levels and that Reep1 depletion causes lip-oatrophy with significantly decreased visceral fat Furthermore

Reep1-- MEFs exhibit a clear and significant impairment of dif-ferentiation into adipocytes and LD size Our data further showthat the Reep1 deletion modifies perilipin phosphorylation inMEFs Earlier studies have shown that phosphorylation of perili-pin is critical for triglycerol storage and hydrolysis by regulatingLD formation and degradation by autophagy (4142) More re-cently it has been demonstrated that PKA mediates Ser492phosphorylation of mouse perilipin (equivalent to Ser497 in

Figure 5 LD abnormalities in Reep1-- mice (A) Immunoblot analysis of endogenous Reep1 protein expression in Reep1thornthornneurons and MEFs in total homogenates

(Homog) and membrane fraction (Memb) (B) LD540 staining of MEFs from Reep1thornthornand Reep1mice before and after treatment with oleic acid Boxed regions are en-

larged in the upper right hand-corner insets Scale bars 10 mm (C) Graphical representation of total LD540 and BODIPY 594 staining intensities in Reep1thornthornand Reep1

MEFs before and after treatment with oleic acid (means 6 SEM nfrac1460) Plt0001 (D) Impairment of MEF differentiation MEFs from Reep1thornthorn and Reep1-- mice were

cultured and differentiated into adipocytes Oil red O staining of LDs with DAPI (blue) staining nuclear is shown Scale bar 10 mm (E-F) Quantification of number

(E) and size (F) of LDs in MEF cells transfected (thornREEP1) or not with a REEP1 expression construct and treated (red) or not (blue) with oleic acid Plt001 Plt0001

(G-H) Quantification of cells (as in C) with TIP47thornLDs in Reep1thornthornand Reep1MEFs after treatment with oleic acid for the indicated times Boxed regions are enlarged

directly below Scale bar 20 mm (I) Quantification of cells with TIP47thornLDs at several time points after oleic acid treatment Plt0001


human perilipin) driving fragmentation and dispersion of LDs(38) In our experiments incubation of Reep1-- MEFs with a spe-cific PKA inhibitor increased the formation of large LDs in wild-type cells and importantly rescued the LD phenotypein Reep1-- cells Along these lines overexpression of REEP1 inCOS7 cells caused a significant enrichment of ADRPperilipin 2at LDs

Though Reep1 appears to play an important role in LD for-mation Reep1-- mice still accumulate some adipose tissue andReep1-- MEFs still produce LDs This is not surprising sincethere are at least three other closely-related Reep proteinsReep2-4 in mice However LD quantity (and size in some celltypes) is significantly decreased suggesting that any redun-dancy is not complete

Interestingly the Fsp27 protein localizes to LDs and pro-motes lipid storage in adipocytes and while the WAT volume ofFsp27 null mice is unchanged from wild-type (43) the visceralratio of total volume WAT to mass in Fsp27mice is lower(44) Histological analysis of WAT indicated that adipocytes

in Fsp27mice have multiple small LDs but adipocyte sizeis not decreased Another related study showed that FIT2(Fat storage-Inducing Transmembrane protein 2) regulates LDformation (45) FIT2 is a 6-transmembrane domain-containingprotein whose conformation likely regulates its activity in mod-elling the ER and mediating LD formation The authors sug-gested that FIT2 is regulated by conformational changesinduced by ER membrane lipids or interactions with other ERproteins It is tempting to speculate that atlastin-1 activity maybe similarly regulated by its interaction with Reep1 and in factthis type of regulation has been well characterized for other pro-teins such as HMG-CoA reductase (46)

Recent studies have revealed important roles of the SPG17BSCL2 protein seipin which localizes to the ER in lipid storageat both the cellular (LDs) and whole-body (adipose tissue devel-opment) levels (4748) We show that REEP1 can interact withseipin which is highly expressed in adipose tissue andstrongly induced during adipocyte differentiation (49ndash51)Modelling loss-of-function seipin mutations in Berardinelli-

Figure 6 Increased phosphorylation of perilipin in Reep1-- cells (A) Immunoblots of perilipin in total MEF lysates from Reep1thornthornand Reep1mice (25 and 08 rela-

tive expression levels respectively) PLCc levels were monitored as a control for protein loading (B) Phosphorylation defect of perilipin in the Reep1-- MEFs (C) Reep1thorn

thornand Reep1--MEFs were incubated with a site-selective lipophilic PKA inhibitor (Rp-8-PIP-cAMPS) in presence or absence of oleic acid Reep1thornthornand Reep1-cells show

a high accumulation of LDs after inhibitor treatment suggesting a perilipin phosphorylation defect in Reep1-- mice (D) Quantification of LD perimeters from C

(means 6 SEM nfrac1460) Plt0001 Scale bars 10 mm


Seip congenital lipodystrophy Bscl2-- mice have prominentlipodystrophy (52ndash54) However though lipoatrophic thesemice exhibit residual fat pads for which histological analysisshows immature adipocytes with a defect in the LD develop-ment (5253) Furthermore Bscl2-- MEFs show impairment ofadipocyte differentiation (49) Studies performed in yeast andhuman non-adipose cells have shown that Fld1 and Bscl2 null

mutations respectively alter LD morphology with formationof giant LDs or clustering of numerous tiny droplets (4855)Reep1-- mice exhibit a similar phenotype with reduction ofWAT volume and disruptions in LD morphology The similari-ties among these mutant mice combined with the interactionbetween Reep1 and seipin indicate that these proteins couldfunction together in LD formation

Figure 7 LD540 staining of Reep1thornthornand Reep1MEFs transfected with Reep1 atlastin-1 or atlastin-1 K80A constructs as indicated Graphs represent the quantifica-

tion of cells before and after treatment with oleic acid (means 6 SEM nfrac1450) Plt005 Plt001 Plt0001 Scale bars 10 mm

Figure 8 SPG17 protein seipin interacts with REEP1 (A) NIH-3T3 cells were co-transfected with Myc-REEP1 and HA-seipin constructs Cell lysates were immunoprecipi-

tated with anti-Myc and anti-HA antibodies and immunoblotted for HA (for seipin) and Myc (for REEP1) epitopes as indicated (B) Mouse brain extracts from the indi-

cated genotypes were immunoprecipitated (IP) as indicated and then immunoblotted for atlasitn-1 seipin and Reep1 Panels on the left used Reep1thornthornextracts (C)

Imunoprecipitations (IP) were carried out as in B except that extracts were first subjected to cross-linking with DSP which is then cleaved in SDS-PAGE sample buffer

(D) Distribution of HA-seipin (green) and Reep1 (red) constructs after co-transfection in COS7 cells Scale bar 10 lm


LD formation and size are highly regulated and atlastinsfunction in regulating LD size (30) Atlastin GTPase activity isknown to mediate the formation of the tubular ER networkthrough interactions with proteins of the reticulon and Yop1pDP1REEP superfamilies (7956) Co-transfection of Reep1 andatlastin-1 causes the formation of large and numerous LDs inCOS7 cells Several studies have already shown that ER proteinsincluding reticulons and REEPs have curvature-promoting prop-erties through hydrophobic hairpin domains and scaffolding(10ndash1224) We suggest that REEP1 and atlastin-1 cooperate toregulate LD size though mechanisms still remain unclear

LD biogenesis is also facilitated by the special structures ofsome ER proteins One model for LD formation emphasizes thephysical properties of lipid emulsion as well as the impact of cur-vature induced by ER proteins (57) The authors suggested thattwo cytosolic populations of LDs exist smaller and expanding LDsSmaller LDs bud from ER and are characterized by the presence ofdiacylglycerol O-acyltransferase 1 (DGAT1) The expanding LDs re-quire the accumulation and the activity of glycerol-3-phosphateacyltransferase 4 (GPAT4)-DGAT2 which enhances the synthesisof triacylglycerol Finally they suggested that proteins such as sei-pin and FIT2 influence the surface tension and the budding pro-cess during LD formation from the bilayer ER-membranes (57)

Surface tension is critical for the formation and release of LDs(57) and we speculate that recruitment of Reep1 and atlastin-1to sites of formation of the expanding LD induces negative cur-vature of the ER membrane atlastin-1 then can form trans inter-actions necessary for membrane fusion This could explain theobservation of smaller LDs in Reep1-- cells Reep1-- cells couldalso have defective DGAT2 accumulation in LDs which inhibitsaccumulation of triglycerides and influences LD size

The decrease in LD number but not size in neurons of Reep1--cerebral cortex is seemingly at odds with the alterations in LDsize observed in MEFs This could be for a number of reasons in-cluding technical factors (eg inability to assess accuratelysmaller LDs in these brain sections) or else differential interac-tions of Reep1 across different cell types Finally Reep1-4 proteinsmay function similarly and the relative contribution of Reep1 inthe brain (which has high Reep1 levels) to Reep1-4 dosage likelydiffers from the contribution of Reep1 to Reep1-4 dosage in pe-ripheral tissues (where Reep1 levels are much lower even ac-counting for effects of Reep1 depletion on Reep2 stability that weobserved in WAT) A better mechanistic understanding of theroles of Reep1 in LD biology should clarify this in the future

Our study thus provides a new model for HSP and exposesa link between regulation of LD formation and morphologyby Reep1 via interactions with atlastin-1 and seipin with dysre-gulation possibly contributing to axonal dysfunction in longcorticospinal neurons In addition Reep1 has been shown topromote ER stress resistance in Drosophila and it also preventsTau-mediated degeneration in vivo by inhibiting the formationor accumulation of thioflavin-S-positive Tau clusters (58) Theseobservations combined with our results supports the idea thatneuronal degeneration in HSP patients results from functionalER changes with consequences for LD formation metabolism ofTau ER stress and ER shaping ndash all of which are vital in thestructural organization of the CNS

Materials and MethodsGeneration and breeding of Reep1-- mice

All animal experiments were approved by the NINDSNIDCDAnimal Care and Use Committee in Bethesda Maryland

Reep1-- mice on a 129SvEv x C57BL6 background producedby homologous recombination-based gene targeting from thegene trap clone OST398247 (Supplementary Material Fig S1A)were purchased from the Texas AampM Institute for GenomicMedicine For construction of the targeting vector the mousechromosome 6 sequence (nucleotide 71623408-71745067)was used as a reference The heterozygous mice producedwere then backcrossed for at least 10 generations into theC57BL6J background For subsequent breeding one male andone female at sexual maturity were housed in the same cageAfter gestation and delivery P1 pups were sacrificed to generateneuronal cell cultures or else the young mice were weaned21ndash28 days after birth for other studies For genotyping genomicDNA was isolated from tail snips using standard proceduresGene-specific PCR was carried out with Taq DNA polymerase(Invitrogen) and primers specific for Reep1 (forward 50-ACGCTACAGACCCATGGGTAGC-30) Neo cassette (reverse 50-ATAAACCCTCTTGCAGTTGCATC-30) and genomic reverse (50-CCAAGGCATTTTCTTCCAAAGG-30) (Supplementary Material Fig S1B)

Anatomic and histologic analyses

Mice were transcardially perfused with 4 paraformaldehyde(PFA) or 2 PFA2 glutaraldehyde Brain and spinal cordwere then excised and post-fixed in 4 PFA or 2 PFA2 glu-taraldehyde respectively for 16 h Sections (50 lm thick) werecut with a Leica CM3050S cryostat and stained with H amp ENissl Luxol fast blue or Cresyl violet or else immunostainedas described (59) For measurement of axon size in the spinalcord sections (1 lm thick) were cut with a PELCO easiSlicer(Ted Pella) and then transferred to a drop of distilled water ona glass slide Sections were dried on a slide warmer thencovered with a drop of toluidine blue for 1 min Next sec-tions were gently rinsed with distilled water and dried Forfat tissue preparation abdominal white fat was dissectedfrom the mice and stored at -80 C Thick sections (50 lm)were prepared using a cryostat at -50 C then fixed with 4formaldehyde and stained with H amp E or BODIPY 493Images were acquired using a LSM710 laser scanning micro-scope (Carl Zeiss Microscopy) and analysed using ImageJ soft-ware (NIH)

Behavioural studies

Mice (n10) of ages 2 and 4 months were compared in two dif-ferent trials Observers were blind to their genetic status duringtesting To assess motor function mice were tested using aRotarod apparatus (MedAssociates ENV-575M 5 Station RotarodTreadmill) as previously published (59) Grip strength was moni-tored quantitatively in these same mice using a grip strengthmeter (Bioseb) (39) Analysis of gait was performed with theTreadScan apparatus and software (CleverSys) Videos of gaitduring treadmill locomotion were captured and frame-by-frame measures of mouse walking patterns were acquired at aconstant running speed of 14 cms According to the manufac-turer TreadScan generates measures of stride length and widthas well as fore and rear foot-placement rotation-angle duringmovement (with degrees relative to the body midline) It alsocomputes stance time (including stationary time) propel time(latter portion of stance when the paw is propelling the body)swing and break time (early part of stance until the paw reachesmaximum contact)


Tissue preparation and immunoblotting

Preparation of cell extracts gel electrophoresis and immuno-blotting were performed as described previously (1356) Proteinconcentrations were determined using Pierce BCA ProteinAssay Reagent (Life Technologies) For immunoblotting pro-teins were resolved by SDS-PAGE and membranes were incu-bated with antibodies (1ndash5 lgml) overnight at 4˚C

For whole tissue lysate preparation mice were dissected andtissues were weighed Briefly fresh cold lysis buffer (150 mMNaCl 50 mM Tris HCl pH 74 1 mM EDTA 1 NP40 05 deoxy-cholate 01 SDS 1Roche Complete Protease Inhibitor 1 mMPMSF) was added to each tissue sample then homogenized andsonicated Samples were spun down twice in a 4˚C refrigeratedcentrifuge at 13000 g for 30 min The supernatant was retainedand proteins quantified 20ndash30 lg of the sample was evaluatedby immunoblotting following SDS-PAGE For immunoprecipita-tion experiments brains from 2-month old mice were lysed inPBS with 1 CHAPS and 1Roche Complete Protease InhibitorWhere indicated DSP was added to a final concentration of15 mM for 60 min on ice before the reaction was quenched with1 M Tris (pH 75) NIH-3T3 and COS7 cells were maintained un-der standard conditions and DNA vector transfections wereperformed with Avalanche-Omni Transfection Reagent (EZBiosystems) for 24 h Immunoprecipitations were conducted asdescribed previously using protein AG PLUS-agarose beads(Santa Cruz Biotechnology) (7)

For adipose tissue isolation mice were dissected at 2 monthsof age and tissue was weighed After rinsing once with cold1PBS immediately 15ndash110 (wv) tissue to buffer of fresh coldlysis buffer (150 mM NaCl 50 mM Tris HCl pH 74 1 mM EDTA1 NP-40 05 deoxycholate 08 SDS 1Roche CompleteProtease Inhibitor 1 mM PMSF) was added Signal quantifica-tions were carried out with Image Lab Software (BioRad) andnormalized on the actin loading control signal

Cortical neuron cultures

Primary cultures of mouse cerebral cortical neurons were pre-pared from P1 mice plated at a density of 10 x 104cm2 on cov-erslips and maintained and immunostained as describedpreviously (59) Briefly brain tissue was subjected to papain di-gestion (5 Uml Worthington) for 45 min at 37 C The digestedtissue was carefully triturated into single cells then centrifugedat 1000 rpm for 5 min and resuspended in warm NeurobasalMedium supplemented with 10 heat-inactivated FBS 1B27(Gibco) 1GlutaMAX 045 D-glucose (Sigma-Aldrich) and pen-icillinstreptomycin (Gibco 15140-122) Dissociated cells wereplated onto dishes or coverslips pre-coated with poly-D-lysine(BD Biosciences) and maintained at 37 C in a 5 CO2 humidifiedincubator

Preparation of MEFs

Pregnant mice at day 14 post coitum were sacrificed and theuterine horns dissected Each embryo was separated from itsplacenta and brain and dark red organs were excised Afterwashing with fresh PBS embryonic tissues were minced andsuspended in trypsin-EDTA then incubated with gentle shakingat 37 ˚C for 15 min After addition of 2 volumes of fresh MEF me-dium [DMEM (high glucose Gibco 41966-052) 10 (vv) FBS1100 (vv) L-glutamine (200 mM Gibco 25030-024) and 1100(vv) penicillinstreptomycin (Gibco 15140-122)] remaining tis-sue pieces were removed and the supernatant was spun down

(5 min 1000 g) The cell pellet was resuspended in MEF mediumand plated (59)

MEF and neuronal transfections

MEFs and neurons were cultured in DMEM supplemented with10 FBS and Neurobasal Medium supplemented with 10 heat-inactivated FBS B27 (Gibco) GlutaMAX 045 D-glucose (Sigma-Aldrich) and penicillinstreptomycin (Gibco 15140-122) at 37 Cin a 5 CO2 humidified incubator respectively DNA transfec-tions were performed using Avalanche-Omni TransfectionReagent (EZ Biosystems) for 1 (MEFs) or 6 (neurons) days follow-ing the instructions of the manufacturer Eukaryotic expressionconstructs for HA-atlastin-1 HA-atlastin-1 K80A HA-Reep1 anduntagged Reep1 have been described (713) Quantifications offluorescence intensity and measure of axon size and branchingwere carried out with ImageJ and NeuronJ (NIH) respectively

Mouse CT imaging

In vivo CT studies were conducted under an animal study proto-col approved by the NIH NINDSNIDCD Animal Care and UseCommittee Animals were maintained under general anaesthe-sia (1ndash2 isoflurane delivered by a gas mixture of oxygen nitro-gen medical air) administered via nosecone and they wereplaced in a secure anatomic position that restricted motion andallowed for constant respiratory monitoring Mouse anatomicalCT was performed on a Bruker SkyScan1176 micro-CT scanner(Bruker) with the X-ray source (energy range 20ndash90 kV) biasedwith 50 kV500 mA using an AI 05 mm filter to reduce beamhardening The integrated scanner heater was used to maintainbody temperature Images were acquired with a pixel size of3613 mm camera-to-source distance was 170 mm and object-to-source distance of 123 mm Projections (360) were acquiredwith an angular resolution of 05 over a 180 rotation Twoframes were averaged for each projection with an exposuretime of 55 ms per frame Scan duration was approximately20 min Tomographic images were reconstructed using vendor-supplied software based on the Feldkamp cone beam algorithmAfter scanning images were reconstructed and processedReconstruction was performed using NRecon software (BrukerMicroCT) with an isotropic pixel size of 3613 mm To removenoise and clarify areas of fat from surrounding soft tissuesoftware-specific artefact reductions were utilized (smoothingfactor of 3 ring artefact correction of 8 and beam hardening cor-rection of 30) The contrast range provided an adequate win-dow for fat tissue with saturation of bone (allowing the air andsoft tissue window to encompass the majority of the greyscalerange) For post-processing Bruker CT-Analyser was used to se-lect the volume of interest (VOI) which began at the first tho-racic vertebrae and terminated at the midsection of femoralheads in the hip joint Upon visual inspection of cross-sectionswithin the VOI grayscale values of each animalrsquos fat windowwere determined a threshold was applied to create a mask offat within the VOI A larger threshold was used to create a maskof the entire body volume These masks were then compared toyield the percentage of fat versus body volume over the VOI

Antibodies and dyes

Mouse monoclonal antibodies were used against b-tubulin(IgG1 clone D66 Sigma-Aldrich) neuronal b-tubulin (14944302Covance) actin (clone AC40 Sigma-Aldrich) PLCc1 (2822 Cell


Signalling) HA-epitope (ab9110 Abcam) and Myc-epitope (IgG1clone 9E10 Santa Cruz Biotechnology) Rabbit polyclonal anti-bodies were used against perilipin (34705 Cell Signalling)anti-P-perilipin 1-serine 497 antibody (4855 Vala Sciences) my-elin (ab30490 Abcam) BSCL2seipin (106793 Abcam) Reep1(17988-1-AP Proteintech) Reep2 (15684-1-AP Proteintech)Reep5 (14643-1-AP Proteintech) calreticulin (ab2907 Abcam)calregulin (Clone T19 Santa Cruz Biotechnology) AIF (CloneE20 Epitomics) VDAC (ab15895 Abcam) HA-probe (Y-11Santa Cruz Biotechnology) and an affinity-purified atlastin-1 an-tibody (56) BODIPY 493 and BODIPY 594 dyes were obtainedfrom Invitrogen and LD540 was a gift from Prof ChristophThiele (60)

Confocal immunofluorescence microscopy

To observe the ER and the cytoskeleton cells plated on cover-slips were fixed for 20 min with 4 formaldehyde or 5 min withmethanol respectively Then cells were blocked for 30 min with10 goat serum 01 Triton X-100 in PBS (pH 74) at room tem-perature After three washes coverslips were incubated withprimary antibodies diluted in 1 goat serum Alexa Fluor anti-rabbit and anti-mouse secondary antibodies (Invitrogen) wereused at 1400 Cells were counterstained with 4rsquo6-diamidino-2-phenylindole (DAPI 01 mgml) where indicated and mountedusing Fluoromount-G (SouthernBiotech) The cells were imagedusing a Zeiss LSM710 confocal microscope with a 63 14 NAPlan-Apochromat oil differential interference contrast objectiveand image acquisition was performed using Zen software (CarlZeiss Microscopy) Images were processed with ImageJ AdobePhotoshop 70 and Adobe Illustrator CC software

Oleic acid treatment and staining of MEFs

To analyse LDs MEFs were cultured as described above and in-cubated with 200 lM oleic acid (Sigma-Aldrich) as describedpreviously (56) After overnight incubation cells were fixed per-meabilized and immunostained with antibodies then incu-bated with a solution of 01 lgml LD540 for 5 min or elseBODIPY 493 or 594 (Invitrogen) in PBS overnight Then cells werewashed and mounted in Fluoromount-G (SouthernBiotech)Fluorescence images were acquired with a 63objective as a se-ries of the optical z-stacks spanning the entire cell using thetransmitted light of a Zeiss LSM710 laser scanning confocal mi-croscope (Carl Zeiss Microimaging) Recording conditions wereset to obtain a fluorescence signal below saturation levels Cellboundaries were defined by applying a threshold to each of thez-sections The proportion of fluorescence was calculated usingImageJ Typically 300 cells in three different experiments wererandomly selected and examined for each condition

Electrophysiology

Experiments were performed on 2-to-4 day-old mice Mice weredecapitated and eviscerated then placed in a dissecting cham-ber and continuously perfused with artificial cerebrospinal fluid(concentrations in mM) 12835 NaCl 4 KCl 15 CaCl2H2O 1MgSO47H2O 058 NaH2PO4H2O 21 NaHCO3 30 D-glucose) bub-bled with 95 O2 and 5 CO2 After a ventral laminectomy thespinal cord was isolated together with attached roots and gan-glia and maintained at room temperature Motoneuron extra-cellular activity was recorded with plastic suction electrodesinto which individual ventral roots were drawn Recordings

were obtained from the L1 (left and right) and L5 (left or right)ventral roots Locomotor-like activity was elicited by stimulat-ing a sacral (S3 or S4) dorsal root ganglion (4 Hz duration 10 sthe duration of each stimulus 250 ls) The lowest intensity atwhich the train elicited locomotor-like activity was defined asthe threshold Frequencies were measured from two trials at 25 and 10 thresholds

Monosynaptic reflex responses recorded from L5 ventral rootwere evoked by a single electrical stimulus to the ipsilateral L5lumbar dorsal root or ganglion (stimulus duration 250 ls)We measured the latency of the response as the time betweenthe beginning of the stimulus artefact and the start of the re-sponse (ie when the inflection in the signal reaches 10 timesthe size of the pre-stimulus noise) We measured the responsesfor 2and 10 thresholds The stimulus threshold was definedas the current at which the minimal evoked response wasrecorded

We used whole-cell current clamp to record motoneurons invitro Whole-cell recordings were obtained with patch elec-trodes (resistance 6-10 MX) lowered blindly into the ventralhorn of the spinal cord Patch electrodes were filled with intra-cellular solution containing (in mM) 10 NaCl 130 K-Gluconate10 HEPES 11 EGTA 1 MgCl2 01 CaCl2 and 1 Na2ATP with pH ad-justed to 72ndash74 with KOH Motoneurons were only consideredfor subsequent analysis if they exhibited a stable resting mem-brane potential of -50 mV or more and an overshootingantidromically-evoked action potential Input resistance wascalculated from the slope of the currentvoltage plot within thelinear range The liquid junction potential was not correctedWe recorded monosynaptic reflexes intracellularly in motoneu-rons The threshold was determined independently for eachneuron as the intensity that triggers an action potential in therecorded neuron

Tracings

L5 ventral and dorsal roots were placed inside suction elec-trodes and backfilled with Cascade Blue dextran or Texas Reddextran for 12 h at room temperature (30ndash40 mM 10000 MWInvitrogen) (61) Segments were then fixed for 24 h in 4 PFA atroom temperature and then washed They were then embeddedin warm 5 Agar and transverse sections (100 lm) were cut on avibratome (Leica V1000) Sections were mounted on slides andcover-slipped with a solution made of glycerol and PBS (37)Images were acquired using a Carl Zeiss LSM710 confocal micro-scope with a 10objective

Polymerase chain reaction

Gene expression of Reep1 and Reep2 were analysed using theRapidScan kit (OriGene) following the instructions of the manu-facturer Primers were Reep1 forward 50-CAAGGCTGTGAAGTCCAAGG-30 reverse 50-GGCACTCTCAGAAGCACTCC-30 Reep2forward 50-GGATCATCTCTCGCCTGGTG-30 reverse 5rsquo-TTAGGGCAGGCTCATCTCCT-3rsquo For real-time PCR total RNA was ex-tracted from white adipocyte tissue with TRIzol reagent(Invitrogen) according to the manufacturerrsquos protocol Reversetranscription of total RNA with Oligo(dT)20 (Invitrogen) was per-formed using ThermoScript RT-PCR System (Invitrogen)Quantitative real-time PCR was carried out using an AppliedBiosystems SYBR green PCR Master Mix 7900HT The followingthermal cycling program was applied 10 min at 95 C 40 cyclesof 15 s at 95 C 30 s at 60 C and 1 min at 72 C Data were


normalized to GAPDH expression levels using the comparativethreshold cycle method

MRI analysis

Animal study protocols to conduct MRI and MR spectroscopystudies were approved by the NINDSNIDCD Animal Care andUse Committee Age- and sex-matched Reep1-- mice and theirwild-type littermates (nfrac14 4ndash5) were anaesthetized with 15 iso-fluorane and then positioned in a stereotactic holder Core bodytemperature was maintained at 37 C using circulating waterpad and monitored throughout the procedure MRI was per-formed on a 7T 21 cm horizontal Bruker Avance scanner withthe brain centered in a 72 vol (transmit)25 mm surface (receive)radio frequency coil ensemble to excite and detect the MR signalA pilot scan depicting gross brain anatomy in three mutuallyperpendicular directions was acquired which in turn was usedto obtain MR images and MR spectra Quantitative spin-spin re-laxation (T2) time weighted (echo time [TE]frac14 10 ms repetitiontime [TR]frac14 3000 ms 16 echos in-plane resolutionfrac14 150 mm 15slices slice thicknessfrac14 1 mm) images whose intensity wasweighted by one of the inherent timing parameters spin-spin(T2) relaxation were acquired The region for MR spectroscopywas positioned on a localized voxel (06 2 2 mm3) in thefrontal cortex positioned approximately 1 mm lateral and15 mm posterior to the Bregma encompassing major parts ofthe frontal cortex and minimizing overlap of the hippocampusand CSF (06 mm thickness) After uniformity of the magneticfield experienced within that chosen voxel was maximizedusing localized optimization techniques (gt 5 min) watersuppressed spectra of the chosen voxel were acquired usingPoint Resolved Spectroscopy (PRESS) sequence (NAfrac14 512 to-tal scan timefrac14 25 min TRTEfrac14 150016 ms 4K data pointsbandwidthfrac14 150 Hz) The water suppressed MR data were pro-cessed and spectral chemical shifts were assigned with respectto the creatine peak (305 ppm) Analysis of the complex spectrawas confined to those peaks that corresponded to creatineglutamate (375 ppm) choline (322 ppm) N-acetyl aspartate(NAA 206 ppm) myo-inositol (36 ppm) and GABA (228 ppm) ndashevaluated by deconvoluting those peaks to a Lorenzian lineshape and calculating areas under those peaks proportional tometabolite concentrations within the chosen voxel Spin-spinrelaxation (T2) maps were calculated using routines written inMATLAB (Mathworks) Areas from left and right hemispheresand each region (frontal parietal) were averaged for each sliceand subsequently averaged over all slices

Mitochondrial fractionation

Cells harvested from 80 10-cm plates of nearly confluent cellswere spun down at 1000 g at 4 C for 10 min Then cells wereresuspended in 20 ml ice-cold mitochondrial homogenizationbuffer (MHB 20 mM HEPES 1 mM EGTA 75 mM sucrose 225 mMmannitol 2 mgml BSA) and left on ice for 5 min to swell Cellswere homogenized with a motorized Potter-Elvehjem homoge-nizer with 40-45 updown strokes To pellet nuclei and unbro-ken cells homogenates were centrifuged (5 min 600 g 4 C)Supernatants were centrifuged again for 5 min at 600 g 4 C toeliminate nuclei and debris To pellet mitochondria lysosomesand ER supernatants were centrifuged for 10 min at 7000 g20000 g and 100000 g respectively Mitochondria pellets werewashed 3 times with 20 ml of MHB buffer and centrifuged for10 min at 10000 g 4 C Finally the crude mitochondrial pellet

was resuspended in 22 ml of mitochondrial resuspension buffer(MRB 5 mM HEPES 05 mM EGTA 250 mM mannitol 2 mgmlBSA) and transferred into 8 ml of 30 Percoll (in 2Percoll me-dium 50 mM HEPES 2 mM EGTA 450 mM mannitol 4 mgmlBSA) and ultracentrifuged for 30 min at 95000 g 4 C in a swing-ing bucket rotor (SW41Ti Beckman) The mitochondrial bandwas collected in and diluted with 14 ml of MRB Mitochondriawere washed twice with 14 ml MRB and centrifuged for 10 minat 6300 g 4 C Purified mitochondria were finally resuspendedin 50ndash200 ml of MRB Samples were quantified with Pierce BCAProtein Assay Reagent (Thermo Fisher Scientific) and analysedby SDS-PAGE

Complex I enzyme activity assay

To measure complex I activity in MEFs we used the Complex IEnzyme Activity Dipstick Assay Kit (MS130 Abcam) followingthe manufacturerrsquos instructions Briefly cells were cultured andthen lysed using a specific Extraction Buffer Dipsticks wereadded and wicked for 30 min then transferred to fresh activitybuffer for 30 min to reveal the signal Quantifications and imag-ing were carried out with Image Lab Software (BioRad)

Statistical analysis

Results are expressed as means 6 SEM Statistical analysis wasperformed using Studentrsquos t-test assuming unequal variancewith Plt 005 considered significant Mann-Whitney U-statisticsor one-way ANOVA test were used for comparisons among dif-ferent data sets Asterisks indicate significant differences(Plt 005 Plt 001 Plt 0005)

Supplementary MaterialSupplementary Material is available at HMG online

Acknowledgements

We thank V Diaz and D Donahue for assistance with the MRIand CT scans respectively We also thank M Lussier AKoretsky and M OrsquoDonovan for scientific advice and stimulatingdiscussions

Conflict of Interest statement None declared

FundingThis work was supported by the Intramural Research Programof the National Institute of Neurological Disorders and StrokeNational Institutes of Health

References1 Blackstone C OrsquoKane CJ and Reid E (2011) Hereditary

spastic paraplegias membrane traffic and the motor path-way Nat Rev Neurosci 12 31ndash42

2 Blackstone C (2012) Cellular pathways of hereditary spasticparaplegia Annu Rev Neurosci 35 25ndash47

3 Tesson C Koht J and Stevanin G (2015) Delving into thecomplexity of hereditary spastic paraplegias phenotypesand inheritance modes are revolutionizing their nosologyHum Genet 134 511ndash538


4 Fink JK (2013) Hereditary spastic paraplegia clinico-pathologic features and emerging molecular mechanismsActa Neuropathol 126 307ndash328

5 Harding AE (1993) Hereditary spastic paraplegia SeminNeurol 13 333ndash336

6 Novarino G Fenstermaker AG Zaki MS Hofree MSilhavy JL Heiberg AD Abdellateef M Rosti B Scott EMansour L et al (2014) Exome sequencing links corticospi-nal motor neuron disease to common neurodegenerativedisorders Science 343 506ndash511

7 Park SH Zhu PP Parker RL and Blackstone C (2010)Hereditary spastic paraplegia proteins REEP1 spastin andatlastin-1 coordinate microtubule interactions with the tu-bular ER network J Clin Invest 120 1097ndash1110

8 Gerondopoulos A Bastos RN Yoshimura S AndersonR Carpanini S Aligianis I Handley MT and Barr FA(2014) Rab18 and a Rab18 GEF complex are required for nor-mal ER structure J Cell Biol 205 707ndash720

9 Hu J Shibata Y Zhu PP Voss C Rismanchi N PrinzWA Rapoport TA and Blackstone C (2009) A class ofdynamin-like GTPases involved in the generation of the tu-bular ER network Cell 138 549ndash561

10 Voeltz GK Prinz WA Shibata Y Rist JM and RapoportTA (2006) A class of membrane proteins shaping the tubu-lar endoplasmic reticulum Cell 124 573ndash586

11 Hu J Shibata Y Voss C Shemesh T Li Z Coughlin MKozlov MM Rapoport TA and Prinz WA (2008)Membrane proteins of the endoplasmic reticulum inducehigh-curvature tubules Science 319 1247ndash1250

12 Tolley N Sparkes IA Hunter PR Craddock CP NuttallJ Roberts LM Hawes C Pedrazzini E and Frigerio L(2008) Overexpression of a plant reticulon remodels the lu-men of the cortical endoplasmic reticulum but does not per-turb protein transport Traffic 9 94ndash102

13 Zhu PP Patterson A Lavoie B Stadler J Shoeb M Patel Rand Blackstone C (2003) Cellular localization oligomerizationand membrane association of the hereditary spastic paraple-gia 3A (SPG3A) protein atlastin J Biol Chem 278 49063ndash49071

14 Zhu PP Soderblom C Tao-Cheng JH Stadler J andBlackstone C (2006) SPG3A protein atlastin-1 is enriched ingrowth cones and promotes axon elongation during neuro-nal development Hum Mol Genet 15 1343ndash1353

15 Solowska JM Morfini G Falnikar A Himes BT BradyST Huang D and Baas PW (2008) Quantitative and func-tional analyses of spastin in the nervous system implicationsfor hereditary spastic paraplegia J Neurosci 28 2147ndash2157

16 Riano E Martignoni M Mancuso G Cartelli D Crippa FToldo I Siciliano G Di Bella D Taroni F Bassi MT et al(2009) Pleiotropic effects of spastin on neurite growth de-pending on expression levels J Neurochem 108 1277ndash1288

17 Beetz C Koch N Khundadze M Zimmer G Nietzsche SHertel N Huebner AK Mumtaz R Schweizer M DirrenE et al (2013) A spastic paraplegia mouse model revealsREEP1-dependent ER shaping J Clin Invest 123 4273ndash4282

18 Zuchner S Wang G Tran-Viet KN Nance MA GaskellPC Vance JM Ashley-Koch AE and Pericak-Vance MA(2006) Mutations in the novel mitochondrial protein REEP1cause hereditary spastic paraplegia type 31 Am J HumGenet 79 365ndash369

19 Hewamadduma C McDermott C Kirby J Grierson APanayi M Dalton A Rajabally Y and Shaw P (2009) Newpedigrees and novel mutation expand the phenotype ofREEP1-associated hereditary spastic paraplegia (HSP)Neurogenetics 10 105ndash110

20 Schottmann G Seelow D Seifert F Morales-Gonzalez SGill E von Au K von Moers A Stenzel W and SchuelkeM (2015) Recessive REEP1 mutation is associated with con-genital axonal neuropathy and diaphragmatic palsy NeurolGenet 1 e32

21 Saito H Kubota M Roberts RW Chi Q and MatsunamiH (2004) RTP family members induce functional expressionof mammalian odorant receptors Cell 119 679ndash691

22 Behrens M Bartelt J Reichling C Winnig M Kuhn Cand Meyerhof W Members of RTP and REEP gene familiesinfluence bitter taste receptor expression J Biol Chem 28120650ndash20659

23 Bjork S Hurt CM Ho VK and Angelotti T REEPs aremembrane shaping adapter proteins that modulate specificG protein-coupled receptor trafficking by affecting ER cargocapacity PLoS One 8 e76366

24 Brady JP Claridge JK Smith PG and Schnell JR (2015) Aconserved amphipathic helix is required for membrane tu-bule formation by Yop1p Proc Natl Acad Sci USA 112E639ndashE648

25 Schlaitz AL Thompson J Wong CCL Yates JR 3rdand Heald R (2013) REEP34 ensure endoplasmic reticulumclearance from metaphase chromatin and proper nuclearenvelope architecture Dev Cell 26 315ndash323

26 Lim Y Cho IT Schoel LJ and Golden JA (2015)Hereditary spastic paraplegia-linked REEP1 modulates endo-plasmic reticulummitochondria contacts Ann Neurol 78679ndash696

27 Goyal U and Blackstone C (2013) Untangling the webmechanisms underlying ER network formation BiochimBiophys Acta 1833 2492ndash2498

28 Westrate LM Lee JE Prinz WA and Voeltz GK (2015)Form follows function the importance of endoplasmic retic-ulum shape Annu Rev Biochem 84 791ndash811

29 Walther TC and Farese RV Jr (2012) Lipid droplets andcellular lipid metabolism Annu Rev Biochem 81 687ndash714

30 Klemm RW Norton JP Cole RA Li CS Park SHCrane MM Li L Jin D Boye-Doe A Liu TY et al (2013)A conserved role for atlastin GTPases in regulating lipiddroplet size Cell Rep 3 1465ndash1475

31 Falk J Rohde M Bekhite MM Neugebauer SHemmerich P Kiehntopf M Deufel T Hubner CA andBeetz C (2014) Functional mutation analysis provides evi-dence for a role of REEP1 in lipid droplet biology HumMutat 35 497ndash504

32 Papadopoulos C Orso G Mancuso G Herholz MGumeni S Tadepalle N Jungst C Tzschichholz ASchauss A Honing S et al (2015) Spastin binds to lipiddroplets and affects lipid metabolism PLoS Genet 11e1005149

33 Szymanski KM Binns D Bartz R Grishin NV Li WPAgarwal AK Garg A Anderson RGW and Goodman JM(2007) The lipodystrophy protein seipin is found at endo-plasmic reticulum lipid droplet junctions and is importantfor droplet morphology Proc Natl Acad Sci USA 10420890ndash20895

34 Rinaldi C Schmidt T Situ AJ Johnson JO Lee PRChen K Bott LC Fado R Harmison GH Parodi S et al(2015) Mutation in CPT1C associated with pure auto-somal dominant spastic paraplegia JAMA Neurol 72561ndash570

35 Fulton BP and Walton K (1986) Electrophysiological prop-erties of neonatal rat motoneurones studied in vitroJ Physiol 370 651ndash678


36 Gonzalez M Nampoothiri S Kornblum C Oteyza ACWalter J Konidari I Hulme W Speziani F Schols LZuchner S and Schule R (2013) Mutations in phospholipaseDDHD2 cause autosomal recessive hereditary spastic para-plegia (SPG54) Eur J Hum Genet 21 1214ndash1218

37 Ulengin I Park JJ and Lee TH (2015) ER network forma-tion and membrane fusion by atlastin1SPG3A disease vari-ants Mol Biol Cell 26 1616ndash1628

38 Marcinkiewicz A Gauthier D Garcia A and BrasaemleDL (2006) The phosphorylation of serine 492 of perilipin adirects lipid droplet fragmentation and dispersion J BiolChem 281 11901ndash11909

39 Hefferan MP Kucharova K Kinjo K Kakinohana OSekerkova G Nakamura S Fuchigami T Tomori ZYaksh TL Kurtz N and Marsala M (2007) Spinal astrocyteglutamate receptor 1 overexpression after ischemic in-sult facilitates behavioral signs of spasticity and rigidityJ Neurosci 27 11179ndash11191

40 Tian GF Azmi H Takano T Xu Q Peng W Lin JOberheim N Lou N Wang X Zielke HR et al (2005) Anastrocytic basis of epilepsy Nat Med 11 973ndash981

41 Holm C (2003) Molecular mechanisms regulating hormone-sensitive lipase and lipolysis Biochem Soc Trans 31 1120ndash1124

42 Yeaman SJ (2004) Hormone-sensitive lipasendashnew roles foran old enzyme Biochem J 379 11ndash22

43 Toh SY Gong J Du G Li JZ Yang S Ye J Yao HZhang Y Xue B Li Q et al (2008) Up-regulation of mito-chondrial activity and acquirement of brown adipose tissue-like property in the white adipose tissue of Fsp27 deficientmice PLoS One 3 e2890

44 Peng XG Ju S Fang F Wang Y Fang K Cui X Liu G LiP Mao H and Teng GJ (2013) Comparison of brown andwhite adipose tissue fat fractions in ob seipin and Fsp27 geneknockout mice by chemical shift-selective imaging and 1H-MRspectroscopy Am J Physiol Endocrinol Metab 304 160ndash167

45 Gross DA Snapp EL and Silver DL (2010) Structural in-sights into triglyceride storage mediated by fat storage-inducing transmembrane (FIT) protein 2 PLoS One 5 e10796

46 DeBose-Boyd RA (2008) Feedback regulation of cholesterolsynthesis sterol-accelerated ubiquitination and degrada-tion of HMG CoA reductase Cell Res 18 609ndash621

47 Agarwal AK and Garg A (2006) Genetic disorders of adi-pose tissue development differentiation and death AnnuRev Genomics Hum Genet 7 175ndash199

48 Boutet E El Mourabit H Prot M Nemani M Khallouf EColard O Maurice M Durand-Schneider AM ChretienY Gres S et al (2009) Seipin deficiency alters fatty acid D9desaturation and lipid droplet formation in Berardinelli-Seipcongenital lipodystrophy Biochimie 91 796ndash803

49 Chen W Yechoor VK Chang BHJ Li MV March KLand Chan L The human lipodystrophy gene product

Berardinelli-Seip congenital lipodystrophy 2seipin plays akey role in adipocyte differentiation Endocrinology 1504552ndash4561

50 Fei W Du X and Yang H (2011) Seipin adipogenesis andlipid droplets Trends Endocrinol Metab 22 204ndash210

51 Payne AV Grimsey N Tuthill A Virtue S Gray SLDalla Nora E Semple RK OrsquoRahilly S and Rochford JJ(2008) The human lipodystrophy gene BSCL2Seipin may beessential for normal adipocyte differentiation Diabetes 572055ndash2060

52 Cui X Wang Y Meng L Fei W Deng J Xu G Peng XJu S Zhang L Liu G et al (2012) Overexpression of a shorthuman seipinBSCL2 isoform in mouse adipose tissue re-sults in mild lipodystrophy Am J Physiol Endocrinol Metab302 705ndash713

53 Chen W Chang B Saha P Hartig SM Li L Reddy VTYang Y Yechoor V Mancini MA and Chan L (2012)Berardinelli-Seip congenital lipodystrophy 2seipin is a cell-autonomous regulator of lipolysis essential for adipocytedifferentiation Mol Cell Biol 32 1099ndash1111

54 Prieur X Dollet L Takahashi M Nemani M Pillot B LeMay C Mounier C Takigawa-Imamura H Zelenika DMatsuda F et al (2013) Thiazolidinediones partially reversethe metabolic disturbances observed in Bscl2seipin-deficient mice Diabetologia 56 1813ndash1825

55 Fei W Shui G Gaeta B Du X Kuerschner L Li PBrown AJ Wenk MR Parton RG and Yang H (2008)Fld1p a functional homologue of human seipin regulatesthe size of lipid droplets in yeast J Cell Biol 180 473ndash482

56 Rismanchi N Soderblom S Stadler J Zhu PP andBlackstone C (2008) Atlastin GTPases are required for Golgiapparatus and ER morphogenesis Hum Mol Genet 171591ndash1604

57 Thiam AR Farese RV Jr and Walther TC (2013) The bio-physics and cell biology of lipid droplets Nat Rev Mol CellBiol 14 775ndash786

58 Appocher C Klima R and Feiguin F (2014) Functionalscreening in Drosophila reveals the conserved role of REEP1in promoting stress resistance and preventing the formationof tau aggregates Hum Mol Genet 23 6762ndash6772

59 Renvoise B Stadler J Singh R Bakowska JC andBlackstone C (2012) Spg20-- mice reveal multimodal func-tions for Troyer syndrome protein spartin in lipid dropletmaintenance cytokinesis and BMP signaling Hum MolGenet 21 3604ndash3618

60 Spandl J White DJ Peychl J and Thiele C (2009) Live cellmulticolor imaging of lipid droplets with a new dye LD540Traffic 10 1579ndash1584

61 Blivis D and OrsquoDonovan MJ (2012) Retrograde loading ofnerves tracts and spinal roots with fluorescent dyes J VisExp pii 4008


gene products atlastin-1 spastin and receptor expression-enhancing protein 1 (REEP1) respectively together with theSPG72 protein REEP2 SPG12 protein reticulon 2 SPG61 proteinARL6IP1 and SPG69 protein RAB3GAP2 play roles in shapingand distributing the tubular endoplasmic reticulum (ER)network in cells (26ndash8) REEP1 atlastin-1 and the largermembrane-bound M1 isoform of spastin interact through hy-drophobic hairpin domains within the tubular ER to coordinateER shaping and microtubule interactions (79ndash12) Interestinglyatlastin-1 spastin and REEP1 localize prominently to the corti-cospinal neurons whose axons are impaired in HSPs (13ndash17)

SPG31 results from autosomal dominant REEP1 mutations(18) It is mostly a pure form only rarely associated withneuropathy and in just one family was an axonal neuropathypresent without signs of hyperreflexia Some mutations disrupt apredicted miRNA binding site in the 3rsquo untranslated region sup-porting haploinsufficiency in at least a subset of patients andheterozygous missense and truncation mutations have also beendescribed throughout the gene (1819) A young boy with a homo-zygous loss-of-function mutation presented with a spinal muscu-lar atrophy with respiratory distress type 1 (SMARD1)-likephenotype with additional features of congenital axonal neurop-athy hyperreflexia and diaphragmatic palsy (20)

REEP proteins were first described based on their ability toenhance surface expression of some G-protein coupled recep-tors including olfactory and taste receptors (2122) REEP1 is amember of a family of ER shaping proteins comprising sixmembers in humans divided structurally and functionally intotwo main subgroups REEP1-4 and REEP5-6 (72324) withREEP1-4 proteins involved in microtubule interactions with ERtubules (725) REEP1 has also been reported to facilitateER-mitochondrial interactions (26) Finally REEP proteins havealso been implicated in enhancing ER cargo capacity regu-lating ER-Golgi processing and interacting selectively withsome cargo proteins (23) Given these numerous functionalroles the in vivo consequences of REEP1 mutation remainlargely unknown

The ER is a continuous membrane organelle comprising thenuclear envelope and a peripheral network of membrane tu-bules and sheets (2728) Tubular ER presents high membranecurvature in cross-section stabilized by members of the reticu-lon and REEP protein families (1011) likely by forming wedge-like structures in the lipid bilayer as well as arc-like scaffoldsaround tubules (10112428) Identifying effects of changes in ERmorphology thus appears critical for understanding HSP patho-logic mechanisms Of course this is complicated by the factthat the ER has many different functions and interactions withnumerous organelles and cytoskeletal elements (2728)

Recent attention has focused on the role of the ER in the for-mation of lipid droplets (LDs) organelles involved in fat storagein eukaryotes (29) LDs contain a core of neutral lipids (triglycer-ides and sterol esters) surrounded by a phospholipid monolayercontaining proteins such as perilipins Interestingly atlastin-1regulates LD size with REEP1 also playing a role albeit less clear(3031) Furthermore the HSP proteins seipin (SPG17) spartin(SPG20) and spastin (SPG4) also function in LD maintenance in-dicating a possible convergent pathogenic theme (232)Atlastin-1 and REEP1 can be co-immunoprecipitated from cells(7) and SPG17 protein seipin (Fld1p in yeast) is a conserved inte-gral membrane ER protein thought to act at the interface of theER and LDs (33) Recently overexpressed atlastin-1 has been co-precipitated with the SPG73 gene product CPT1C an ER-localized carnitine palmitoyltransferase (34)

Here we report that Reep1-- mice exhibit a gait phenotypeand prominent spasticity along with significant lipodystrophyThese symptoms are associated with prominent changes in cel-lular LDs Reep1 localizes to the ER and interacts with atlastin-1as well as with the SPG17 protein seipin that is also mutated inthe most common form of autosomal recessive lipodystrophyBerardinelli-Seip congenital lipodystrophy suggesting a modelwhereby REEP1 may function with other key HSP gene productsto regulate LD formation and size

ResultsReep1-- mice exhibit spasticity and motor dysfunction

Reep1-- mice were generated as shown in SupplementaryMaterial Figure S1A and disruption of the Reep1 gene was con-firmed using PCR (Supplementary Material Fig S1B)Immunoblotting of multiple mouse tissues established the ab-sence of the Reep1 protein (Supplementary Material Fig S1C)Reep1-- mice bred normally and their offspring appeared nor-mal at birth Early in life they moved throughout the cage muchlike littermate controls and appeared to feed normallyHowever they soon developed age-dependent impairments inmotor function Compared to wild-type animals Reep1-- miceexhibited spasticity with lower extremity clonus at postnatalday 3 (compare Supplementary Material Videos S1 and S2)prominent impairment at 2 months in the Rotarod test (Fig 1A)and significant differences in gait parameters such as stridetime stance time and stance length upon formal TreadScantesting (Fig 1A and Supplementary Materials Fig S1D andMovies S3ndashS7) However grip test data showed no differences(Supplementary Material Fig S1E) At 4 months 50 ofReep1-- mice were unable to walk at the same speed as Reep1thornthornmice Thus Reep1-- mice have a progressive defect in ambu-lation reminiscent of that in SPG31 patients as reported previ-ously (17) In preliminary studies Reep1thorn- animals hadessentially no phenotype even at more advanced ages (data notshown) also consistent with a previous report (17) and theywere not investigated further

Reep1-- motoneurons exhibit increased activity

Gait dysfunction affecting the hind limbs in mice can resultfrom different types of impairments in the pyramidal motorpathway To understand the basis for the Reep1-- phenotypeneurophysiologically we analysed motoneuron activity (35) In3-day-old Reep1-- mice locomotor-like activity frequency washigher (Fig 1B Supplementary Material Fig S1FndashH) Also la-tency of the suprathreshold monosynaptic reflex (L5 segment)was shortened in Reep1-- mice (Fig 1CndashE SupplementaryMaterial Fig S1F and H) independent of the site of stimulation(Supplementary Material Fig S1H) Together these data sup-port spasticity as a principal cause of the impaired gait in theReep1-- mice Motoneuron staining using fluorescent dyes alsoshowed that the Reep1-- lateral motor pool is slightly smaller(Fig 1F and G) Intracellular input resistances recorded fromReep1-- and Reep1thornthornmotoneurons were not different how-ever suggesting that Reep1-- motoneuron input resistancechanges are either due to a decrease in the extent of the den-dritic tree or else increased membrane resistivityMonosynaptic reflexes did not show differences intracellularly(Supplementary Material Fig S1F and H)

































































































Behavioural studies









Preparation of MEFs





Mouse CT imaging


Antibodies and dyes








Electrophysiology





Tracings






MRI analysis










Acknowledgements





































































































































































Behavioural studies









Preparation of MEFs





Mouse CT imaging


Antibodies and dyes








Electrophysiology





Tracings






MRI analysis










Acknowledgements












































































Preparation of MEFs





Mouse CT imaging


Antibodies and dyes








Electrophysiology





Tracings






MRI analysis










Acknowledgements











































































Electrophysiology





Tracings






MRI analysis










Acknowledgements







































































MRI analysis










Acknowledgements






































































reep1 null mice reveal a converging role for hereditary...

Documents