Human Molecular Genetics

Human Molecular Genetics Advance Access originally published online on November 10, 2004
Human Molecular Genetics 2005 14(1):19-38; doi:10.1093/hmg/ddi003

This Article Abstract FREE Full Text (PDF) Supplementary Material All Versions of this Article: 14/1/19    most recent ddi003v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (58) Request Permissions Google Scholar Articles by Reid, E. Articles by Sanderson, C. M. Search for Related Content PubMed PubMed Citation Articles by Reid, E. Articles by Sanderson, C. M. Social Bookmarking          What's this?

Human Molecular Genetics, Vol. 14, No. 1 © Oxford University Press 2005; all rights reserved

The hereditary spastic paraplegia protein spastin interacts with the ESCRT-III complex-associated endosomal protein CHMP1B

Evan Reid^1,*,, James Connell^2,, Thomas L. Edwards¹, Simon Duley², Stephanie E. Brown² and Christopher M. Sanderson²

¹Cambridge Institute for Medical Research and Department of Medical Genetics, University of Cambridge, Cambridge CB2 2XY, UK and ²Medical Research Council Rosalind Franklin Centre for Genomics Research, Hinxton, Cambridge CB10 1SB, UK

^* To whom correspondence should be addressed at: Cambridge Institute for Medical Research, University of Cambridge, Wellcome Trust/MRC Building, Addenbrooke's Hospital, Cambridge, CB2 2XY, UK. Tel: +44 1223 762632; Fax: +44 1223 217054; Email: ereid{at}hgmp.mrc.ac.uk

Received August 6, 2004; Revised October 6, 2004; Accepted October 21, 2004

	ABSTRACT

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
Pure hereditary spastic paraplegia is characterized by length-dependent degeneration of the distal ends of long axons. Mutations in spastin are the most common cause of the condition. We set out to investigate the function of spastin using a yeast two-hybrid approach to identify interacting proteins. Using full-length spastin as bait, we identified CHMP1B, a protein associated with the ESCRT (endosomal sorting complex required for transport)–III complex, as a binding partner. Several different approaches confirmed the physiological relevance of the interaction in mammalian cells. Epitope-tagged CHMP1B and spastin showed clear cytoplasmic co-localization in Cos-7 and PC12 cells. CHMP1B and spastin interacted specifically in vitro and in vivo in ß-lactamase protein fragment complementation assays, and spastin co-immunoprecipitated with CHMP1B. The interaction was mediated by a region of spastin lying between residues 80 and 196 and containing a microtubule interacting and trafficking domain. Expression of epitope-tagged CHMP1B in mammalian cells prevented the development of the abnormal microtubule phenotype associated with expression of ATPase-defective spastin. These data point to a role for spastin in intracellular membrane traffic events and provide further evidence to support the emerging recognition that defects in intracellular membrane traffic are a significant cause of motor neuron pathology.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
The main voluntary motor pathway can be divided into two stages. Firstly, upper motor neurons (UMNs) in the motor cortex of the brain connect via the corticospinal tracts to cells in the grey matter of the spinal cord. Secondly, lower motor neurons (LMNs) in the anterior horn of the spinal cord connect to the skeletal muscles at the neuromuscular junction (1). The axons of UMNs and LMNs are exceptionally long, with cytoplasmic volumes much greater than that of the cell body. They therefore provide an extreme environment for many cellular processes, such as intracellular trafficking, transport and energy metabolism.

In addition to being anatomically connected, UMNs and LMNs are linked by disease pathology. The most widely recognized disease of motor neurons is amyotrophic lateral sclerosis (motor neuron disease), a usually sporadic condition in which both UMNs and LMNs are affected. In addition, several groups of genetically determined conditions are associated with motor neuron dysfunction. Although individually rare, these conditions offer an opportunity to dissect the cellular processes involved in maintenance and degeneration of motor neurons. The hereditary spastic paraplegias (HSPs) are one such group and their defining clinical characteristic is a progressive spastic paralysis affecting the legs, usually caused by a length-dependent, distal degeneration of the corticospinal tract axons (2–4).

Clinically, the HSPs are divided into pure and complex forms, with pure HSPs, where spasticity occurs in relative isolation, forming the single largest subgroup. More than 20 HSP loci have been mapped and 10 of the associated genes have been identified (reviewed in 2). The SPG4 gene, spastin, is the most frequent HSP gene, with mutations found in 40% of autosomal dominant pure HSP families (5,6). The full-length spastin transcript is widely expressed within the nervous system and in non-neuronal tissues, including a range of leukocyte-related cells. It is most strongly expressed in subclasses of neurons (5,7,8). Spastin transcripts lacking either exons 4, 8 or 15 have been found, although only the exon 4-deleted variant is expressed in significant levels (7,9).

Data on the subcellular localization of endogenous spastin have been conflicting. Although early immunofluorescence studies using anti-spastin polyclonal antibodies on mouse tissue and interphase HeLa cells suggested that endogenous spastin is localized exclusively in the nucleus (7), subsequent studies using other polyclonal antibodies have found mixed nuclear and cytoplasmic distributions in HeLa cells (10,11), motor neuron cell lines (11) and human nervous system tissues (8). In human nervous tissues, spastin was located in the cytoplasm and synaptic terminals of several neuronal cell types, including motor neurons, although in other neuronal cell types it was present in the nucleus (8).

The different subcellular localizations identified for spastin imply that the protein might have several different functional roles, the full range of which is not yet clear. Early sequence analysis placed spastin in the AAA (ATPases associated with diverse cellular activities) family, a large multi-group family of proteins that contain a characteristic ATPase cassette and participate in many different cellular processes (5,12). Spastin is a member of AAA subgroup-7 which contains p60 katanin, a microtubule severing protein, and cell-culture and in vivo data suggest that at least one function of spastin involves regulation of microtubules. In the early stages of expression in cultured Cos-7 cells, epitope-tagged spastin localized in the cytoplasm, close to the centre of the microtubule asters. With later expression, it localized to discrete punctate cytoplasmic structures not corresponding to known organelles (13). Expression of wild-type spastin was associated with broken microtubule bundles and when ATPase-defective epitope-tagged spastin was expressed, it co-localized with abnormally thick and long perinuclear bundles of microtubules (13). Similar findings were made in cultured rat cortical neurons (14). These results have recently been extended by an immunolocalization study which identified high levels of endogenous spastin in subcellular regions where microtubules are dynamically regulated. In actively dividing Cos-7 and HeLa cells, spastin was found on the centrosome and spindle microtubules during metaphase, and on the central spindle and midbody during cytokinesis. In a motor neuron cell line, spastin was highly expressed at the distal ends of neurite extensions and at branch points. This study also identified the centrosomal protein NA14 as a binding partner for spastin (11).

Data from Drosophila models also support a role for spastin in microtubule regulation (15). Drosophila spastin (D-spastin) is particularly enriched at synaptic boutons, co-localizing with the synaptic vesicle marker synaptotagmin at neuronal regions where stable microtubules are normally absent. Targeted RNA interference knockdown of D-spastin in the nervous system caused synaptic undergrowth and a reduction of total synaptic area, with a corresponding enhancement of presynaptic neurotransmission strength that was corrected pharmacologically by nocodazole, a microtubule destabilizing agent. The animals had an increase in structurally stable tubulin at the neuromuscular junction presynaptic terminal, with encroachment of microtubules into the presynaptic boutons. Conversely, overexpression of D-spastin led to a decrease in the amount of structurally stable tubulin at the neuromuscular junction presynaptic terminal and a reduction in neurotransmission strength that was corrected by the microtubule stabilizing agent taxol. These data are consistent with the hypothesis that D-spastin may directly or indirectly regulate microtubule stability and may spatially restrict stabilized microtubules at presynaptic boutons (15).

Despite all of the data pointing towards a microtubule regulating function for spastin, the fact that it is found in punctate or tubular cytoplasmic structures when expressed in mammalian cells and that it is spatially distinct from stable microtubules in Drosophila axons suggests that it may have other roles. One proposal is that spastin may be involved in intracellular membrane traffic events. Spastin has an N-terminal microtubule interacting and trafficking (MIT) domain. This domain is shared with several proteins that have defined roles in membrane traffic, including VPS4, an AAA (ATPases associated with diverse cellular activities) protein from the same subgroup as spastin, sorting nexin 15 and calpain-7/PalB (16,17). Interestingly, the MIT domain is also found in spartin, a protein mutated in Troyer syndrome, an autosomal recessive HSP complicated by dysarthria, distal amyotrophy and mild developmental delay (16).

Intracellular membrane traffic pathways are divided into secretory and endocytic components (18). In the endocytic pathway, internalized receptors and transmembrane proteins from the plasma membrane are delivered to early endosomes, from where they may be recycled to the plasma membrane, targeted to the Golgi, delivered to late endosomes and lysosomes for degradation or, in polarized cells, transcytosed (18). As endosomes connect the secretory, internalization and degradative processes, they function as important sorting sites. Key molecular components of the endocytic pathway have been identified by analysis of yeast VPS mutants, grouped into classes A–F depending on the site of effect (19). The class E VPS mutants are of particular relevance to this study; at least 17 yeast class E VPS mutants are known and each has at least one human homologue. They are essential for the formation of the multivesicular body, a late endosomal structure formed by invagination and budding of the limiting membrane into the vesicle lumen (19). The multivesicular body fuses with the lysosome, with consequent exposure of the internalized membranes to the lumenal degradative compartment (19). Recent studies have identified three related membrane-associated protein complexes, termed ESCRT (endosomal sorting complexes required for transport)-I, -II and -III, which are made up of class E VPS proteins (20–22). These complexes may be sequentially recruited from cytoplasmic precursors and mediate sorting of cargoes to the multivesicular body. The ESCRT-III complex is thought to be responsible for the final sorting and concentration of cargo to a subset of multivesicular body membranes (22). This complex is made up of at least four class E VPS proteins that are all members of the same protein family, Chm2p, Chm3p, Chm4p and Chm6p (also called Vps2p, Vps24p, Vps32p/Snf7 and Vps20p, respectively in yeast, corresponding to CHMP2A and 2B, CHMP3, CHMP4A, 4B and 4C and CHMP6 in humans) (22–24). Two other members of the Chm family have been identified, Chm1p/Did2p (corresponding to CHMP1A and CHMP1B in humans) and Chm5p/Vps60p (corresponding to CHMP5 in humans) (22–24), although it is not clear whether they are part of the ESCRT-III complex or whether they regulate ESCRT-III function (22). The membrane association of the Chm proteins is regulated by the class E Vps protein Vps4p (discussed earlier), which may catalyse disassembly of the ESCRT-III complex (19,22–25). VPS4 is present in two forms in humans, VPS4A and VPS4B (23,24,26). In addition to its role in membrane traffic, an isoform of human CHMP1A is localized to the nucleus (27).

In this study, we set out to gain new insights into the function of spastin, by using a yeast two-hybrid approach to identify binding partners for the protein. Using full-length spastin as bait we identified two proteins involved in membrane traffic, CHMP1B and gp25L2, as specific binding partners. Given the similarity between spastin and VPS4 and taking into account the relative strength of the yeast two-hybrid interactions obtained with the two putative binding partners, we chose to prioritize the interaction with CHMP1B for study. We have confirmed the physiological importance of the interaction in mammalian cells by subcellular co-localization studies, split ß-lactamase protein fragment complementation assays and co-immunoprecipitation studies. Given the role of CHMP1B in endosomal membrane trafficking, these data suggest that spastin, like other HSP genes, may play a role in membrane traffic events.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
Yeast two-hybrid experiments identify CHMP1B and gp25L2 as potential binding partners of spastin
We used full-length spastin protein as a ‘bait’ to screen a high-complexity human yeast two-hybrid erythroleukaemia cDNA library, as described previously (28). This screen generated 16 positive diploid clones that grew on selective media lacking adenine and were positive for ß-galactosidase activity. When these primary interactions were retested in fresh yeast only two were found to be both reproducible and bait specific. The two positive ‘prey’ clones encoded the gp25L2 protein (NM_017510 [GenBank] ), which exhibited a weak yet discernable ß-galactosidase activity and restricted growth on selective media lacking tryptophan, leucine and adenine (–W/L/A), and CHMP1B (NM_020412 [GenBank] , also known as CHMP1.5), which exhibited strong ß-galactosidase activity and strong growth on selective (–W/L/A) media (Fig. 1A and B). As spastin appeared to interact more strongly with the CHMP1B prey clone, we chose to analyse the specificity and physiological relevance of this interaction in mammalian cells.

View larger version (37K): [in this window] [in a new window]   Figure 1. Spastin interacts with gp25L and CHMP1B in the yeast two-hybrid system. (A) Positive diploid yeast colonies isolated from a yeast two-hybrid screen in which full-length spastin was used as the bait protein were transferred to filters and tested for activation of the ß-galactosidase reporter gene. Diploid cells containing spastin and one of two different non-specific prey proteins (NSP1 and NSP2) show no blue colouration after the ß-galactosidase assay, whereas diploid cells containing spastin and gp25L2 or spastin and CHMP1B show weak or strong ß-galactosidase activity, respectively. (B) Reconfirmation of the observed interaction between spastin and gp25L2 or CHMP1B in fresh yeast. Column 1 shows the growth of diploid colonies on media lacking uracil and leucine (–U/L) to select for the presence of both bait and prey vectors, while column 2 shows the growth of diploid colonies on media lacking uracil, leucine and histidine, which requires the activation of the His reporter gene. In each case spastin was used as the ‘bait’, while gp25L2 or CHMP1B were used as ‘prey’ proteins.

 
Intracellular distribution of epitope-tagged spastin and CHMP1B
Existing data on the subcellular localization of spastin are conflicting and the subcellular localization of CHMP1B has not been reported previously, so we first performed immunofluorescence studies to examine the subcellular distribution of epitope-tagged human spastin and CHMP1B proteins. To avoid misinterpretation and to assess the potential for tag-induced changes in protein distribution, we compared the relative subcellular localization of three different tagged versions of each protein. Proteins were tagged at either the N or the C-terminus with GFP (GFP–Spastin, Spastin–GFP, GFP–CHMP1B and CHMP1B–GFP) or at the N-terminus with the c-myc epitope (myc–Spastin and myc–CHMP1B). To help assess whether the discrepancy in the subcellular localization of spastin in cell-culture transfection experiments versus studies using antibodies against the native protein is caused by differential localization of exon 4-deleted and full-length isoforms of spastin, we made a C-terminal GFP-tagged construct containing the exon 4-deleted form of spastin (Spastinex4–GFP, Fig. 2). The constructs were transiently expressed in Cos-7 and PC12 cells, and the relative sub-cellular distribution of each protein was assessed by fluorescence microscopy.

View larger version (37K): [in this window] [in a new window]   Figure 2. Schematic diagram of full-length spastin and four other mutated forms of spastin used in this study. MIT, microtubule interacting and trafficking domain; AAA, ATPase associated with diverse cellular activities domain.

 
Spastin.
Expression of GFP–Spastin, Spastin–GFP or myc–Spastin gave similar results. All three were expressed within the cytoplasm of Cos-7 cells (data not shown) and undifferentiated (Fig. 3A) or differentiated PC12 cells (Fig. 3B and C), consistent with previous observations (13,14). Spastin was predominantly localized to either punctate or tubular structures which were initially observed within the perinuclear region. At later time points (48 h or longer), punctate and more extensive tubular structures were also observed in more peripheral regions of the cell. In PC12 cells differentiated to have neurite extensions, punctate spastin expression was seen in the soma and also within neurites. Within neurites, expression was often most prominent at the distal ends (Fig. 3B and C). Expression of the Spastinex4–GFP construct in Cos-7 cells gave results identical to the expression of full-length spastin, with no nuclear localization (data not shown).

View larger version (114K): [in this window] [in a new window]   Figure 3. Subcellular distribution of spastin and CHMP1B epitope-tagged proteins in Cos-7 and PC12 cells. (A–C and G–I) PC12 cells, (D–F) Cos-7 cells. (A–C) show the distribution of spastin fused to GFP at its N-terminus (A), C-terminus (B), or to the c-myc epitope at its N-terminus (C). (D–I) show the distribution of CHMP1B fused to GFP at its N-terminus (D and G) or C-terminus (E and H), or to the c-myc epitope at its N-terminus (F and I). In Cos-7 cells, punctate nuclear structures were more prominent and perinuclear staining less prominent with the GFP–CHMP1B fusion protein (D) when compared with the other two CHMP1B fusion proteins. Occasional linear arrays of labelled puncta were seen (inset in F). Cos-7 cells were fixed and processed for microscopy 24 h after transfection and PC12 cells 48 h after transfection.

 
CHMP1B.
The subcellular distribution of the myc–CHMP1B and the CHMP1B–GFP fusion proteins in transiently transfected Cos-7 was broadly similar. Both were observed in the nucleus and the cytoplasm (Fig. 3E and F). Within the nucleus CHMP1B predominantly exhibited a diffuse/granular distribution, although occasional cells showed additional punctate nuclear labelling. Within the cytoplasm the fusion proteins were predominantly localized to very numerous, small, punctate structures that were present throughout the cytoplasm. In addition, clumps of larger circular structures, very reminiscent of membrane-bound vesicles, were present, predominantly in a focal perinuclear cap. In a proportion of the cells transfected with CHMP1B–GFP, but not myc–CHMP1B, the perinuclear cap of vesicles contained abnormally enlarged, labelled vesicular structures. Overall, this cellular distribution is similar to that previously described for the closely related CHMP1A protein (24). Although numerous punctate structures were also labelled throughout the cytoplasm in some cells expressing GFP–CHMP1B, there were differences versus the other fusion proteins. Many cells had diffuse cytoplasmic labelling, and where labelling was not diffuse, the perinuclear cap of vesicular structures seen with the other fusion proteins was seldom observed. Instead, labelled punctate and tubular structures tended to be present in the cytoplasm all around the nucleus, particularly in higher expressing cells. In addition, punctate nuclear staining was much more common than with the other constructs (Fig. 3D).

In PC12 cells, each of the three epitope-tagged CHMP1B constructs had a similar subcellular distribution (Fig. 3G–I). Most transfected cells had numerous cytoplasmic circular/vesicular labelled structures on a background of diffuse cytoplasmic staining. In cells showing a more neuronal differentiation pattern, CHMP1B was seen in neurite processes and in some cells the distal ends of neurites were particularly strongly stained. In addition to the cytoplasmic staining, many cells showed diffuse nuclear staining, with a small number of cells having punctate nuclear staining.

Epitope-tagged CHMP1B co-localizes with endosomal markers
All members of the CHMP family of proteins studied to date have had an endosomal subcellular localization. We examined whether this was true of CHMP1B by transiently co-expressing myc–CHMP1B with an endosomal marker, BD Living Colors^TM GFP-tagged RhoB. We saw striking co-localization of this marker and CHMP1B in PC12 cells (Fig. 4A–C) and Cos-7 cells (data not shown). We confirmed the endosomal location of CHMP1B using antibody markers for early endosomes (EEA1), late endosomes (mannose 6-phosphate receptor, M6PR) and lysosomes (Lamp-1), in Cos-7 cells. Transiently transfected CHMP1B–GFP partially co-localized with M6PR (Fig. 4D–F) and EEA1 (Fig. 4G–I), but did not show any clear co-localization with Lamp1 (data not shown), indicating that CHMP1B is present on early and late endosomes, but not lysosomes.

View larger version (92K): [in this window] [in a new window]   Figure 4. Epitope-tagged CHMP1B co-localizes with endosomal markers. (A–C) show co-localization in a partially differentiated PC12 cell co-transfected with the endosomal marker RhoB–GFP (A) and myc–CHMP1B (B). Note strong co-localization in the distal ends of neurite extensions and in a perinuclear region (C). (D–L) show Cos-7 cells. (D–F) show co-localization between CHMP1B–GFP (D) and the mannose 6-phosphate receptor (M6PR) (E), a marker of late endosomes. The co-localization is partial (F). (G–I) show co-localization between CHMP1B–GFP (G) and early endosome antigen-1 (EEA1) (H), a marker of early endosomes. Again, the co-localization is partial (I). (J–L) show a Cos-7 cell transfected with CHMP1B–GFP (J) and stained with the free cholesterol marker filipin (K). Filipin staining is seen inside CHMP1B–GFP coated vesicles (L). Coloured squares shown in the bottom right hand corner of each grey-scale picture indicate the colour of that image in the corresponding merged picture (C, F, I and L). Cos-7 cells were fixed and processed for microscopy 24 h after transfection and PC12 cells 48 h after transfection.

 
Esterified cholesterol taken up via LDL receptors is transported to the late endocytic compartment, where it is hydrolysed to release free cholesterol. Overexpression of some members of the CHMP family in mammalian cells can give rise to abnormal accumulation of free cholesterol in endosomal compartments (29). We examined whether this was the case when CHMP1B–GFP was expressed in Cos-7 cells, using the cholesterol-specific marker filipin. There was filipin staining of a subset of the CHMP1B–GFP labelled vesicles, although the strength of the filipin labelling in these vesicles did not appear increased when compared with filipin positive vesicles in untransfected cells (Fig. 4J–L).

Epitope-tagged spastin co-localizes with the endosomal protein RhoB
The potential interaction between spastin and CHMP1B led us to investigate whether spastin could also be present on endosomes. We examined cells co-expressing the BD Living Colors GFP-tagged RhoB with myc–Spastin and again saw striking co-localization in Cos-7 (Fig. 5A–C) and PC12 cells (Fig. 5D–F). In PC12 cells, strong co-localization was seen in the distal ends of neurite extensions as well as the cell body (Fig. 5D–F). In both cell types, examination of large numbers of cells indicated that the co-localization was independent of cell cycle stage. Antibody staining showed no co-localization between singly transfected, epitope-tagged spastin and EEA1, M6PR or Lamp-1 in Cos-7 cells (data not shown).

View larger version (55K): [in this window] [in a new window]   Figure 5. Epitope-tagged spastin co-localizes with RhoB–GFP in Cos-7 and PC12 cells. In (A–C), Cos-7 cells were co-transfected with RhoB–GFP (A) and myc–Spastin (B). Strong co-localization between RhoB and myc–Spastin is seen (C). (D–F) show a differentiated PC12 cell co-transfected with RhoB–GFP (D) and myc–Spastin (E). Again, strong co-localization is apparent, both in the cell body and in the foci along neurite extensions (F, arrowheads). The distal ends of neurites were often strongly co-stained (long arrow). Coloured squares shown in the bottom right hand corner of each grey-scale picture indicate the colour of that image in the corresponding merged picture (C and F). Cos-7 cells were fixed and processed for microscopy 24 h after transfection and PC12 cells 48 h after transfection.

 
Epitope-tagged spastin co-localizes with epitope-tagged CHMP1B in Cos-7 and PC12 cells
We next analysed the extent to which spastin and CHMP1B co-localized within mammalian cells. We examined, with immunofluorescence experiments in transiently transfected Cos-7 and PC12 cells, whether there was any co-localization between epitope-tagged spastin and epitope-tagged CHMP1B. We compared myc–Spastin with GFP–CHMP1B and CHMP1B–GFP, and myc–CHMP1B with GFP–Spastin and Spastin–GFP. In each case the spastin and CHMP1B proteins were clearly co-localized to small punctate or tubular cytoplasmic structures (Fig. 6). Although striking, the observed co-localization was clearly not complete, as spastin was observed to associate with only a subset of the CHMP1B positive structures. Interestingly, punctate structures exhibiting spastin–CHMP1B co-localization were occasionally arranged in linear arrays, reminiscent of transport vesicles lining up on cytoskeletal elements (Supplementary Material, Fig. S1). There were some differences in the pattern of co-localization seen with different combinations of fusion proteins, particularly in Cos-7 cells. In co-localization studies that used myc–CHMP1B versus spastin fusion proteins, most co-localization was in punctate and tubular structures that were distributed in the perinuclear and more peripheral cytoplasm (Fig. 6A–F). The perinuclear cap of vesicular structures typically seen with myc–CHMP1B expression alone was usually absent. In co-localization studies using CHMP1B–GFP, although peripheral punctate co-localization was found, the strongest co-localization was typically present within the perinuclear cap of vesicular structures (Fig. 6G–I). In co-localization studies that used GFP–CHMP1B, most co-localization was again in punctate and tubular structures that were distributed in the perinuclear and more peripheral cytoplasm (Fig. 6J–L).

View larger version (74K): [in this window] [in a new window]   Figure 6. Spastin and CHMP1B proteins co-localize in mammalian cells. Cos-7 cells were transiently transfected with Spastin–GFP and myc–CHMP1B (A–C), GFP–Spastin and myc–CHMP1B (D–F), CHMP1B–GFP and myc–Spastin (G–I) or GFP–CHMP1B and myc–Spastin (J–L). PC12 cells were transiently transfected with GFP–CHMP1B and myc–Spastin (M–O). The arrow points to a region of co-localization in a neurite extension. In all transfections there is co-localization of the two proteins. Coloured squares shown in the bottom right hand corner of each grey-scale picture indicate the colour of that image in the corresponding merged picture (C, F, I, L and O). Cos-7 cells were fixed and processed for microscopy 24 h after transfection and PC12 cells 48 h after transfection.

 
In PC12 cells, epitope-tagged spastin and CHMP1B were co-localized in punctate structures within neurite processes, as well as being co-localized in punctate and tubular structures in the cell body cytoplasm (Fig. 6M–O).

In vitro and in vivo ß-lactamase protein fragment complementation assays support a physical interaction between spastin and CHMP1B in mammalian cells
We further examined the physiological relevance of the observed interaction between spastin and CHMP1B using a series of ß-lactamase protein fragment complementation assays (PCAs). This approach provides a specific and sensitive independent method of assessing the validity of primary yeast two-hybrid interactions in a mammalian cell system (30–32). To perform the PCAs, the ß-lactamase protein was fragmented into two complementary domains termed BLF1 (aminoacids 24–197) and BLF2 (aminoacids 198–288), as previously described (30,31). When expressed in isolation these two domains do not efficiently reconstitute ß-lactamase activity. However, when the complementary fragments are expressed as separate in-frame fusions with two interacting proteins the two component parts of the ß-lactamase enzyme are brought into close proximity, promoting correct domain folding and the reconstitution of enzyme activity. Unlike in the yeast two-hybrid approach, protein fragments are likely to interact at their normal subcellular location.

The CHMP1B protein was fused upstream of the BLF1 fragment (CHMP1B–BLF1), whereas spastin was fused upstream of the complementary BLF2 fragment (Spastin–BLF2). As a positive control we generated two further constructs in which two subunits (ATP6E and ATP6G) of the vacuolar ATPase H⁺ pump (V-ATPase) protein complex were fused upstream of the BLF1 and BLF2 fragments, respectively (Fig. 7A). These proteins were chosen because there is evidence that they physically interact in vivo (33) and they exhibit highly specific interaction profiles in yeast two-hybrid assays (28).

View larger version (109K): [in this window] [in a new window]   Figure 7. Protein complementation assays. (A) Schematic representation of the six constructs used to perform ß-lactamase protein fragment complementation assays. BLF1 is an N-terminal fragment of the ß-lactamase enzyme encoding amino acids 24–197, whereas BFL2 corresponds to the C-terminal domain of ß-lactamase encoding amino acids 198–288. (B) Data from a PCA assay comparing the relative reconstituted ß-lactamase activity observed when HEK293T cells were transiently transfected with either BLF1+BLF2, CHMP1B–BLF1+ATP6G–BLF2, ATP6E–BLF1+ATP6G–BLF2 or CHMP1B–BLF1+Spastin–BLF2 fusion proteins. Data bars show the average of three independent measurements from each of two separate co-transfection experiments. (C) Data from a PCA assay comparing the relative reconstituted ß-lactamase activity observed when HEK293T cells were transiently transfected with either ATP6E–BLF1+ATP6G–BLF2, ATP6E–BLF1+Spastin–BLF2, or CHMP1B–BFL1+Spastin–BLF2 fusion proteins. In each case ß-lactamase activity was determined 48 h after transfection. Data bars show the average of three independent measurements from each of two separate co-transfection experiments. (D–H) show the results of an in vivo ß-lactamase protein fragment complementation assay. HEK293T cells were transiently co-transfected with corresponding BLF1 and BLF2 fusions shown in (A) and protein–protein interactions were detected using the Gene BLAzerTM In Vivo Detection Kit (Invitrogen). Forty-eight hours after transfection cells were loaded with the cell permeable fluorescent substrate CCF2/AM containing two fluorophores, coumarin and fluorescein. In the absence of reconstituted ß-lactamase activity the substrate remains intact and excitation of the coumarin at 409 nm leads to FRET and the emission of green light from the fluorescein at 520 nm. In the presence of reconstituted ß-lactamase activity CCF2/AM is cleaved and FRET is disrupted leading to emission of blue light from the coumarin at 447 nm. Cells containing a positive protein–protein interaction fluoresce blue, and those that do not fluoresce green. (I) ß-lactamase activity was quantified in vivo by counting the total number of green and blue cells in 10 different fields of view per transfection. (J) Graphical representation of the percentage of total cells positive for ß-lactamase activity from 10 fields of view.

 
In eight independent experiments we consistently observed a significant increase in ß-lactamase activity when the CHMP1B–BLF1 and the Spastin–BLF2 fusion proteins were co-expressed, as compared to the levels of activity observed when the unfused BLF1 and BLF2 fragments were expressed together (Fig. 7B). The level of ß-lactamase activity observed when the CHMP1B-BLF1 and the Spastin–BLF2 fusion proteins were co-expressed was broadly comparable to that observed when the positive control ATP6E–BLF1 and ATP6G–BLF2 fragments were used together (Fig. 7B and C). To test the specificity of the CHMP1B–spastin interaction in this system, the relative ß-lactamase activity observed when CHMP1B–BLF1 and Spastin–BLF2 were co-expressed was compared with that obtained when either the CHMP1B–BLF1 or the Spastin–BLF2 construct was replaced by the corresponding ATP6E–BLF1 or ATP6G-BLF2 construct (Fig. 7B and C). In each case the level of ß-lactamase activity observed was significantly reduced when the specific partner was replaced. The same pattern of specificity was observed in multiple replicates of these experiments.

We next used an in vivo ß-lactamase protein fragment complementation assay to verify the results obtained from the in vitro total cell lysate ß-lactamase assay. The in vivo assay is carried out in live cells and so provides a more physiological approach to explore protein–protein interactions (30,31). HEK293T cells were transfected on glass chamber slides and ß-lactamase activity was visualized in vivo using an inverted epifluorescence microscope. ß-lactamase activity was detected by emission of blue fluorescence, which occurs on hydrolysis of the fluorogenic substrate CCF2/AM, and was assessed by counting the total number of blue and green cells in 10 fields of view. Results were very similar to those obtained with the in vitro assay. Again, we consistently observed an increase in ß-Lactamase activity when the CHMP1B-BLF1 and the Spastin–BLF2 fusion proteins were co-expressed, as compared to the levels of activity observed when the unfused BLF1 and BLF2 fragments were expressed together (Fig. 7D, H–J). The level of ß-lactamase activity observed when the CHMP1B–BLF1 and the Spastin–BLF2 fusion proteins were co-expressed was broadly comparable to that observed when the positive control ATP6E–BLF1 and ATP6G–BLF2 fragments were used together (Fig. 7E, I and J). ß-Lactamase activity was reduced when either the CHMP1B–BLF1 or the Spastin–BLF2 construct was replaced by the corresponding ATP6E–BLF1 or ATP6G–BLF2 construct (Fig. 7F, G, I and J). These results were replicated in three independent experiments. Together, the in vitro and in vivo ß-lactamase protein fragment complementation assays strongly suggest that the interaction between spastin and CHMP1B is physiologically relevant.

Co-immunoprecipitation assays support an interaction between spastin and CHMP1B in mammalian cells
We carried out co-immunoprecipitation experiments to verify further the physiological significance of the interaction between spastin and CHMP1B. We first carried out subcellular fractionation studies to determine whether it would be feasible to co-immunoprecipitate spastin and CHMP1B from soluble cellular fractions. Subcellular fractionation of myc–Spastin transfected Cos-7 cells indicated that spastin is present in cytosolic, membrane, nuclear and, predominantly, in the cytoskeletal/insoluble subcellular fractions (Fig. 8A). CHMP1B–GFP had a similar subcellular distribution (Fig. 8B), and the subcellular distribution of either protein was not altered significantly by co-expression with the other (data not shown). Although the expression of each protein was strongest in the insoluble fraction, because of the technical difficulties of co-immunoprecipitation from an insoluble pellet, we attempted co-immunoprecipitation of spastin and CHMP1B from the soluble fraction. We transiently co-expressed myc-Spastin and CHMP1B–GFP or myc–Spastin and GFP empty vector in HEK293T cells. We separated the myc–Spastin/CHMP1B–GFP or myc–Spastin/empty-GFP whole-cell lysates into soluble and insoluble fractions by centrifugation. We confirmed that the proteins of interest were present in each of the two soluble fractions by immunoblotting with anti-myc and anti-GFP antibodies (Fig. 8C and D), then immunoprecipitated from the soluble fraction using an anti-GFP antibody. Subsequent western blotting with anti-myc antibody showed that myc–Spastin had been co-immunoprecipitated with CHMP1B–GFP (Fig. 8E), but that no myc–Spastin was co-immunoprecipitated from the myc–Spastin/empty-GFP vector co-transfection (Fig. 8F). myc–Spastin was not detected in mock co-immunoprecipitations which were carried out without GFP antibody (Fig. 8E and F).

View larger version (66K): [in this window] [in a new window]   Figure 8. myc–Spastin is predominantly present in an insoluble subcellular fraction and co-immunoprecipitates with CHMP1B. (A and B) Immunoblotting of subcellular fractions from Cos-7 cells transfected with myc–Spastin (A) or CHMP1B–GFP (B), using anti-myc and anti-GFP antibody, respectively. Lane order in (B) is as shown in (A). Bands corresponding to the expected sizes of both proteins are located in all four cellular fractions, with highest expression in a subcellular fraction that contains cytoskeletal elements and other insoluble components. Subcellular fractions were obtained using a subcellular proteome extraction kit (Calbiochem), 48 hours after cell transfection. (C–F) myc–Spastin co-immunoprecipitates with CHMP1B–GFP. HEK293T cells were co-transfected with myc–spastin and CHMP1B–GFP or, as a control, myc–Spastin and empty vector-EGFP. Forty-eight hours after transfection, cells were harvested into lysis buffer and separated into soluble and insoluble fractions by centrifugation. Expression of proteins of expected size in the soluble fraction from each co-transfection is shown in (C) (myc–Spastin/CHMP1B–GFP co-transfection) and (D) (myc–Spastin/empty vector-EGFP co-transfection). Both soluble fractions were immunoprecipitated using anti-GFP antibody, the immunoprecipitated proteins were separated by SDS–PAGE and then immunoblotted with anti-myc antibody. For the myc–Spastin/CHMP1B–GFP co-transfection, a band corresponding to myc–Spastin is seen in the immunoprecipitate (E). This band was not present when anti-GFP antibody was omitted from the immunoprecipitation (E). In the control myc–Spastin/empty vector-EGFP immunoprecipitation, no myc–Spastin bands were present (F).

 
The interaction between spastin and CHMP1B is mediated by the N-terminal region of spastin
We examined which sequences of the spastin molecule were responsible for the interaction with CHMP1B, using the yeast two-hybrid assay. We made yeast two-hybrid ‘baits’ containing full-length spastin, spastin with a C-terminal deletion removing the whole AAA domain (spastinAAA), spastin with an N-terminal deletion removing the first 194 amino acids of the protein, including the MIT domain (spastinN1) and spastin with an N-terminal deletion removing the first 80 amino acids and leaving the MIT domain intact (spastinN2) (Fig. 2). Although the CHMP1B prey interacted strongly with wild-type spastin, spastinAAA and spastinN2 baits, the spastinN1 bait failed to interact (Fig. 9).

View larger version (51K): [in this window] [in a new window]   Figure 9. The interaction between spastin and CHMP1B is mediated by a region of spastin that contains the MIT domain. Investigation of domains responsible for the interaction between spastin and CHMP1B was carried out using yeast two-hybrid. Column 1 shows the growth of diploid colonies on media lacking tryptophan and leucine (–U/L) to select for the presence of both bait and prey vectors. Column 2 shows the growth of diploid colonies on media lacking tryptophan, leucine and histidine, which requires the activation of the His reporter gene, while column 3 shows the growth of diploid colonies on media lacking tryptophan, leucine and adenine, which requires the activation of the Ade reporter gene. In each case spastin or a spastin mutant was used as the ‘bait’ protein, whereas CHMP1B was used as ‘prey’. Full-length spastin, spastinAAA and spastinN2 all interact with CHMP1B, but spastinN1 does not.

 
Expression of epitope-tagged CHMP1B prevents development of the abnormal microtubular cellular phenotype associated with ATPase-defective spastin
We next examined the effect of expression of CHMP1B on the abnormal microtubular cellular phenotype that has been described with expression of ATPase-defective spastin (13). We generated an N-terminal myc-tagged construct in which the entire AAA-domain-containing C-terminal end of spastin was deleted (myc–SpastinAAA, Fig. 2). We also made N-terminal myc-tagged constructs containing three different missense mutations, S362C, K388R and T615I. The S362C and K388R mutations are in the ATPase cassette and the K388R mutation is known to prevent normal ATPase function. The T615I mutation is at the extreme C-terminal end of the protein, outside the AAA domain. When we transiently expressed these constructs in Cos-7 cells, we found labelling localized to punctate, perinuclear structures in cells with low expression levels (Fig. 10A). However, with moderate (Fig. 10B) to high (Fig. 10C) expression levels, we found prominent labelled filamentous structures in nearly all cells. These findings are very similar to previous descriptions of the cellular phenotype caused by expression of spastin with defective ATPase function, and the filaments have been characterized as abnormally elongated and thickened bundles of microtubules (13). We also made a myc-tagged construct containing the S44L sequence change that has been found in the homozygous state in one patient with spastic paraplegia, although it is not clear whether the mutation is pathogenic in this situation. As previously reported, the expression pattern of this construct did not differ from wild-type spastin (13).

View larger version (115K): [in this window] [in a new window]   Figure 10. Co-expression with CHMP1B reverses the cellular phenotype caused by ATPase-defective spastin. (A–C) show three separate cells showing different appearances of myc–SpastinAAA transfected Cos-7 cells. With low expression levels, myc–Spastin is concentrated in a perinuclear region (A). With increased expression levels, cells have a mixed filamentous/punctate distribution of label (B). With higher expression, there are numerous filamentous structures in the cytoplasm (C). (D–F) show the typical appearance of a Cos-7 cell co-transfected with CHMP1B–GFP (D) and myc–SpastinAAA (E). The filamentous appearance is not present and SpastinAAA and CHMP1B are strongly co-localized in circular structures. Occasionally, the larger cytoplasmic vesicles sometimes seen with CHMP1B–GFP expression were co-stained (arrow). Similar results were found when the same spastin mutant was expressed with GFP–CHMP1B and when myc–SpastinK338R was expressed with GFP–CHMP1B or CHMP1B–GFP. A small percentage of co-transfected cells retained the filamentous phenotype, and the typical appearance of one of these cells is shown in (G–I). The CHMP1B–GFP labelling pattern (G) is completely altered from that normally found, and instead strongly co-localizes with myc–SpastinAAA labelled filaments (H and I). CHMP1B–GFP expression alone had no gross effect on polymerized microtubules (J–L). Coloured squares shown in the bottom right hand corner of each grey-scale picture indicate the colour of that image in the corresponding merged picture (C, F, I and L). Cos-7 cells were fixed and processed for microscopy 24 h after transfection.

 
To assess whether CHMP1B expression modified the abnormal cellular phenotype associated with the spastin mutants, we co-expressed GFP–CHMP1B or CHMP1B–GFP with each mutant construct. In the case of myc–SpastinAAA, myc–SpastinK388R, myc–SpastinS362C or myc–SpastinT615I constructs, co-expression with either CHMP1B construct resulted in a complete loss of the filamentous phenotype in virtually all cells that were co-transfected. The mutant spastin was strongly co-localized with CHMP1B in the cytoplasm. In the case of CHMP1B–GFP, this co-localization was on circular structures of varying sizes within the cytoplasm (Fig. 10D–F). The strongest co-localization was in small punctate structures, although a proportion of the larger vesicles sometimes seen with CHMP1B–GFP expression were also double-labelled. These larger vesicles stained positive with filipin, indicating that they were cholesterol-rich and implying that they were of endosomal origin (data not shown). In the small minority of co-transfected cells which did have a filamentous appearance (5% versus >60% of cells when the spastin mutant was expressed independently), CHMP1B was redistributed and co-localized with the mutant spastin on filaments (Fig. 10G–I). The appearances with co-expression of myc–SpastinS44L versus GFP–CHMP1B or CHMP1B–GFP did not differ from co-expression of the CHMP1B constructs with wild-type epitope-tagged spastin (data not shown). Numerous microtubule bundles of apparently normal length and diameter were seen when CHMP1B constructs were expressed alone (Fig. 10J–L).

CHMP1B is expressed in fetal and adult brain
CHMP1B would be expected to be present in central nervous system tissue, if the interactions between it and spastin have a significant relationship to the neurodegeneration seen in HSP. We identified expressed sequence tags (ESTs) corresponding to CHMP1B by carrying out a BLAST search against human EST databases (34). Several of the ESTs, including full-length or near full-length cDNAs (e.g. BX418704 [GenBank] , BI596792 [GenBank] ) were derived from adult or fetal brain cDNA libraries, indicating that CHMP1B is expressed in the central nervous system.

Mutation screening of CHMP1B in HSP families
As we found strong evidence that CHMP1B is a likely binding partner for spastin, we considered CHMP1B an HSP disease gene candidate. We therefore undertook mutation screening of the gene in a group of HSP families. The CHMP1B cDNA has a coding sequence of 600 bp which is oriented onto genomic DNA in a single exon. We designed PCR primers to amplify the genomic coding sequence of the gene, along with flanking non-coding DNA, in two overlapping fragments. These fragments were amplified from genomic DNA samples of an affected member from 25 families with HSP (in 23 of which the inheritance pattern was compatible with autosomal dominant inheritance, in two compatible with autosomal recessive inheritance). Nineteen of these families had already been screened for spastin mutations and 24 had been screened for atlastin mutations, with negative results. Sequencing of the CHMP1B amplified fragments revealed no coding sequence changes.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
We set out to gain insight into the function of spastin, a protein mutated in HSP, by identifying potential binding partners. Using a yeast two-hybrid approach, we identified a specific interaction between spastin and CHMP1B, an endosomal protein involved in membrane traffic events. The physiological relevance of this interaction in mammalian cells was strongly supported by striking co-localization of epitope-tagged spastin with CHMP1B, by positive and specific results with in vitro and in vivo ß-lactamase protein fragment complementation assays and co-immunoprecipitation studies. In addition, we saw effects on the subcellular distribution of epitope-tagged wild-type and mutant spastin with expression of CHMP1B. Together, these data give strong support for a physiological interaction between spastin and CHMP1B in mammalian cells, with the caveat that all of the confirmatory experiments used overexpression systems. This means that it is possible that some of the positive results may be over-expression artefacts and studies aimed at finding evidence for the endogenous relevance of the spastin–CHMP1B interaction would be helpful.

Our study supports previous observations on the distribution of epitope-tagged wild-type spastin in Cos-7 cells, with spastin present in punctate or tubular structures within the cell cytoplasm (13). In addition, we provide data on the subcellular localization of epitope-tagged spastin in the rat PC12 cell line, a cell line derived from pheochromocytoma cells that can be differentiated into a neuronal phenotype using nerve growth factor (NGF). In undifferentiated PC12 cells, spastin was present as punctate structures within the cytoplasm. In more differentiated cells, spastin was again present as puncta in the soma, but was also present in neurite extensions, with strongest expression often at the distal end of the neurite. This distribution is similar to that described for endogenous spastin in a motor neuron cell line, where expression was enriched in the distal axon and in branching regions (11). However, it differs from the expression pattern of FLAG-tagged wild-type spastin in rat cortical neurons, where expression was confined to the soma and did not extend into the axon or dendrites (14). The reasons for this discrepancy are not clear, but could include effects of cell-type or of the particular epitope tag used.

This study also characterizes for the first time the subcellular distribution of CHMP1B and we show that it has a mixed nuclear and cytoplasmic expression pattern in Cos-7 and PC12 cells. In PC12 cells, cytoplasmic CHMP1B expression was found in punctate structures in the cell soma, and also in neurite extensions, with the distal ends of labelled neurites often showing strong expression. Within the cytoplasm of Cos-7 cells, CHMP1B was typically present in a focal perinuclear cap of vesicles and in more peripheral, smaller punctate structures. At least a proportion of the CHMP1B positive perinuclear vesicles were endosomes, because they co-stained with markers of early and late endosomes and contained cholesterol. Some construct-dependent differences were noted in the subcellular distribution of CHMP1B in Cos-7 cells. Perinuclear caps of vesicular structures were commonly seen with myc–CHMP1B and CHMP1B–GFP, but not GFP–CHMP1B, whereas abnormally enlarged cytoplasmic vesicular structures were only seen with CHMP1B–GFP. These effects are likely to be secondary to effects of the site and nature of the epitope tag. One possible interpretation of these observations is that (i) the N-terminal GFP tag impedes binding to vesicular membranes, so GFP–CHMP1B does not label the perinuclear vesicular structures seen with the other fusion proteins, (ii) the C-terminal GFP tag allows binding to vesicular membranes but causes abnormal enlargement of endosomes by preventing removal of CHMP1B, (iii) the N-terminal c-myc tag allows both normal binding and removal from membranes, resulting in labelled perinuclear vesicles that are not enlarged. This interpretation could also explain some of the construct-dependent differences we found in our spastin–CHMP1B co-localization studies. In co-localization studies that used myc–CHMP1B versus spastin fusion proteins, most co-localization was in punctate and tubular structures that were distributed in the perinuclear and more peripheral cytoplasm. The perinuclear caps of vesicular structures typically seen with myc–CHMP1B were usually absent, suggesting that spastin expression caused CHMP1B to be mobilized from these vesicular foci. However, in co-localization studies using CHMP1B–GFP, although peripheral punctate co-localization was found, strongest co-localization was typically present in the perinuclear cap of vesicular structures, suggesting that in this situation it is spastin which is redistributed to the immobile CHMP1B. Effects of epitope tags on the function of CHMP proteins have been described (23). These effects can depend on the nature and position of the tag, although no clear rule as to which tags have a dominant negative effect on CHMP function has emerged so far.

When we expressed a variety of ATPase-defective spastin constructs in Cos-7 cells, we found an identical expression pattern to that described previously. The mutant spastin became associated with abnormal filamentous structures, which consist of thickened and elongated strands of microtubules (13). Significantly, we show that CHMP1B expression has dramatic effects on the expression pattern of ATPase-defective epitope-tagged spastin. The spastin-labelled filamentous structures disappeared from nearly all co-expressing cells, and instead the mutant spastin strongly co-localized with CHMP1B. These observations, taken with the fact that cells expressing low levels of ATPase-defective spastin do not show the filamentous phenotype, suggest that in these cell models mutant spastin might bind to microtubules only once sites of binding to other proteins, including CHMP1B, are saturated. Expression of exogenous CHMP1B might increase the available binding sites, preventing the filamentous phenotype from developing. Although unlikely, we cannot entirely exclude an alternative possibility that CHMP1B expression itself causes microtubule reorganization which in turn could have an indirect effect on spastin distribution.

We mapped the region of spastin which interacts with CHMP1B as lying between residues 80 and 196. Spastin's MIT domain lies between residues 116 and 194, constituting most of the CHMP1B interacting region, and it seems likely that the MIT domain is responsible for the interaction between spastin and CHMP1B. This is consistent with data on the binding of Vps4p to endosomes in yeast, which occurs via an MIT domain and depends on the presence of Vps2 and Vps24, homologues of human CHMP2 and CHMP3 (22,25). It therefore appears that in at least two cases, spastin and VPS4, binding to a CHMP family member is mediated by the MIT domain. It remains to be seen whether other MIT domain-containing proteins also interact with CHMP family members, and if so, whether differential or antagonistic interaction of different MIT domain-containing proteins with CHMP proteins could be an important mechanism of pathway regulation. The region of spastin's interaction with CHMP1B is separate from that required for interaction with microtubules, which has recently been identified as lying between residues 50 and 87 (11). The presence of distinct binding sites for CHMP1B and microtubule interaction is consistent with our finding that CHMP1B can still co-localize with ATPase-defective spastin on filaments, in those rare cells co-expressing both proteins that develop the filamentous phenotype.

The CHMP family of proteins are all involved in intracellular membrane traffic events, at least, but perhaps not exclusively, at the level of endosomes (24). The yeast Chm proteins and their human CHMP homologues constitute, or are closely associated with, the ESCRT–III complex, which may form the final component of a three part cascade that targets membrane-associated cargoes to the multivesicular body (22). It is not clear whether CHMP1B is part of the ESCRT-III complex, or whether it plays a regulatory role on complex function. At present, we can only speculate on the functional role that spastin's interaction with CHMP1B may take. The membrane association of the ESCRT–III CHMPs, including CHMP1B, is regulated by VPS4, an AAA protein that has similarities to spastin, because both contain an MIT domain and are members of the same AAA family subgroup (22–25). As spastin is structurally related to VPS4, it is possible that its interaction with CHMP1B might similarly take the form of regulation of the membrane association of this molecule. In this context, it is interesting that we found co-expression of spastin with myc–CHMP1B seemed to mobilize the CHMP1B protein from a perinuclear vesicular distribution. An alternative and perhaps more remote possibility is that CHMP1B may have functions associated with its interaction with spastin that are distinct from its role in membrane traffic.

We found some CHMP1B expression in the nucleus. This mixed nuclear/cytoplasmic expression pattern is similar to that described for the closely related CHMP1A molecule, for which a nuclear isoform may have a role in stable gene silencing (27). On the basis of the results of this study and several studies that have looked at the subcellular distribution of spastin (7,8,10,11), it seems likely that it also has a mixed nuclear/cytoplasmic distribution. Thus CHMP1B and spastin may fall into the group of molecules (reviewed in 27) that have separate functions in membrane traffic pathways and in the nucleus.

A role for spastin in membrane trafficking is compatible with the suggestion that spastin may have a role in microtubule regulation (11,13,15). Given the close interrelationship between traffic of vesicle-bound cargoes and transport by motor molecules on cytoskeletal filaments, it is conceivable that a molecule involved in intracellular membrane traffic events might also interact with microtubules or have secondary effects on microtubule dynamics, and indeed a few such molecules have been described (reviewed in 17). Against this background, it is interesting that we saw punctate foci of co-localized CHMP1B and spastin organized onto linear arrays, reminiscent of vesicles being transported on microtubules. It is also worthy of note that although Drosophila spastin had effects on axonal microtubules, the subcellular expression pattern of stable microtubules and D-spastin were distinct, with D-spastin instead co-localising with the synaptic vesicle pool marker synaptotagmin (15). This perhaps suggests that D-spastin may have other functions that are distinct to microtubule regulation. Further work will be required to dissect these issues.

If spastin has a role in the endocytic pathway, this raises the question of how abnormalities in this process might lead to the neurodegenerative HSP phenotype. The endocytic pathway has an important role in the regulation of membrane associated receptors, with receptor regulation being dependent on a balance between degradation in the lumenal compartment of the multivesicular body/lysosome versus recycling back to the plasma membrane (19). Abnormalities of sorting in the endocytic pathway may result in receptor up- or down-regulation. An abnormality in the endocytic pathway could therefore cause multiple abnormalities in plasma membrane receptor levels and consequent abnormalities in signalling, perhaps including activation of apoptotic pathways, or repression of survival pathways. The distal ends of very long axons might be especially vulnerable to this process, because downstream pathways might already be working at near threshold levels in such an extreme environment. Another possibility is that abnormalities in endosomal traffic might directly perturb downstream signalling, independent of plasma membrane receptor levels. Neuronal maintenance is dependent on neurotrophic survival factors released by innervation targets. These survival factors signal by activation of cognate receptors on the neuronal presynaptic membrane (35). In at least some neurons NGF signalling can be conveyed in specialized early endosomes that are transported to the neuronal cell body by retrograde axonal transport (35,36). Failure of trafficking of the signalling endosome could therefore lead to neurodegeneration, and again neurons with very long axons could be particularly vulnerable to abnormalities of this type of process.

There is growing evidence that defects in the related functions of intracellular transport and membrane traffic may be common mechanisms for the neurodegeneration seen in HSP. Several HSP proteins have been implicated in molecular transport, including the neuronal specific kinesin heavy chain KIF5A (37–39), or membrane traffic events, including alsin, atlastin, maspardin, spartin and NIPA1 (non-imprinted in Angelman's1) (16,40–46). More widely, analysis of the causes of genetic diseases of LMNs indicates that they are also vulnerable to defects in intracellular transport or membrane traffic, with the kinesin motor KIF1Bbeta, neurofilament light chain, myotubularin-related 2 protein phosphatase, gigaxonin and Rab7 proteins all being implicated in these processes (47–52).

In summary, we present data that strongly suggest an interaction in mammalian cells between spastin and CHMP1B. These results support the idea that spastin plays a role in membrane traffic, and contribute to an emerging recognition that defects in intracellular transport and membrane traffic mechanisms are a significant cause of motor neuron pathology.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
Construction of yeast two-hybrid and immunofluorescence vectors
The complete open reading frames of spastin and CHMP1B were amplified by proof-reading Pfx polymerase (Invitrogen) PCR from IMAGE clones 4829087 and 3928338, respectively, using Gateway^TM (Invitrogen) recombination cloning system-compatible primers designed from the reference cDNA sequence for each gene. To generate the spastinAAA, spastinN1 and spastinN2 mutants, Pfx PCR was carried out using appropriate Gateway (Invitrogen) recombination cloning system-compatible primers designed from within the reference cDNA. The T615I spastin mutant was generated by PCR using a reverse primer which incorporated the appropriate sequence change. Other mutant forms of spastin were generated using a splicing by overlap extension approach, as previously described, again using Gateway (Invitrogen) recombination cloning system-compatible flanking primers (53). Gateway recombination cloning system pENTR201 and/or pENTR207 entry vectors containing the amplified cDNAs were constructed according to the manufacturer's instructions. Sequence of entry vector constructs was verified by sequencing on an ABI377 or 3700 sequencer, using BigDyeDT chemistry (Applied Biosystems).

The spastin and CHMP1B wild-type transcripts were then subcloned into Gateway-compatible vectors pcDNA3-NG-myc, pE-GFP-N and pE-GFP-C, which contained 5'-myc, 3'-GFP or 5'-GFP epitopes, respectively. Spastin mutant constructs were subcloned into the Gateway-compatible vector pcDNA3-NG-myc. The orientation of the spastin and CHMP1B sequences within the constructs was verified by direct sequencing on an ABI377 or 3700 sequencer, or by PCR with appropriate combinations of vector and insert sequence specific primers.

For yeast two-hybrid experiments, a spastin-containing pENTR201 entry vector was subcloned into the Gateway-compatible bait vector pGBDU-G, which was constructed by cloning the Gateway reading frame cassette B into pGBDU-C1 (54) according to the manufacturer's instructions.

Yeast two-hybrid library screening and reporter assays
Yeast strain PJ69-4A (MATa trp1–901 leu2–3, 112 ura3–52, his3–200 gal4 gal80 LYS2 :: GAL1–HIS3 GAL2–ADE2 met2 :: GAL7–lacZ (54)) was transformed by electroporation (55) with the bait plasmid, which carry the URA3 marker. No self-activation of the bait strain reporter genes was obtained on synthetic drop-out (SD) –Ura/Ade and SD –Ura/His. The spastin bait was screened against a K562 erythroleukaemia cDNA Matchmaker library (Clontech), which contains the LEU2 marker, and was amplified and transformed into a mating type switched, MAT derivative of the PJ69-4A yeast strain. This library represents 3.5x10⁶ independent clones with an estimated cDNA insert size ranging from 0.4 to 3.8 kb, with an average insert size of 1.9 kb.

Bait yeast (10⁹ cfu) were individually mated to the prey yeast library (6x10⁸ cfu) using a plate mating procedure modified from a previously described method (56). This gave a mating efficiency of >2%, enabling a 4-fold coverage of the library. Mating mixtures were grown on 16x150 mm SD –Ura/Leu/Ade plates for up to 2 weeks for selection of clones expressing the ADE2 reporter. Colonies were picked in a 96-grid format onto SD –Ura/Leu/Ade plates such that all subsequent handling could be performed using a 96-pin replicator. Activation of the lacZ reporter was assayed by growing the yeast on filter paper on YPAD plates overnight, lysing the cells in liquid nitrogen and then incubating the filters at 37°C for 1 h on filters pre-soaked in 6 mL Z buffer (100 mM Na₂HPO₄, 40 mM NaH₂PO₄, 10 mM KCL, 1 mM MgSO₄, p. 7), 100 µl 4% X-gal and 11 µl ß-mercaptoethanol. Colonies were scored as positive for activation of lacZ if the coloration was greater than that of yeast containing the bait vector alone. We did not score for activation of the HIS3 gene because in our experience >99% of colonies that activate ADE2 also activate HIS3, as previously reported (54).

Identification of interacting proteins in yeast two-hybrid screens
Prey inserts were directly amplified from yeast using vector-specific primers. Yeast colonies were picked into 20 mM NaOH and lysed for 20 min at room temperature. A 1.5 µl aliquot of lysis mix was added to a standard 25 µl PCR reaction containing vector-specific primers. The PCR reaction was heated for 5 min at 95°C followed by 35 cycles of 1 min denaturation at 95°C, 1 min annealing at 58°C and 3.5 min extension at 72°C. Four microlitres of each PCR reaction was run on a 0.8% agarose gel to estimate the insert size. To identify the prey inserts, 3–5 µl of each PCR product was sequenced as described above.

Reconfirmation and specificity testing for yeast two-hybrid assays
To retest each interaction in fresh yeast, each prey PCR product was cloned using Gateway recombination cloning system into the Gateway compatible pGAD-G vector, as described above. Competent MAT PJ69-4A yeast cells were transformed by electroporation (55) with each of these prey-vector containing plasmids and plated directly onto SD –Leu plates. After growth at 30°C for 3 days, they were mated by replica plating onto a lawn of MATa yeast containing either the relevant bait, two different irrelevant baits or an empty bait vector, and then incubated for >5 h on YPAD plates. Diploids were selected on SD –Ura/Leu prior to testing for reporter activation by replicating onto SD –Ura/Leu/Ade.

Computational analysis
Prey sequences were searched against locally held versions of the Homo sapiens Unigene and the EMBLminus databases using an automated BLAST (34) algorithm. Custom-built Perl modules and scripts were used to prefilter and format the raw BLAST output, and determine whether the 5' sequence read overlapped with the protein coding region of a gene. Only matches to known protein coding regions and to ESTs that have no defined open reading frame were included as positives. Matches to 3' UTRs and genomic DNA with no associated gene prediction were excluded. Singleton, unspliced ESTs that did not correspond to any gene predictions were also excluded, as they probably represent genomic contamination of cDNA libraries.

Antibodies
Monoclonal anti-myc primary antibody (mouse clone 4A6) was obtained from Clontech. Monoclonal anti-microtubule antibody (rat clone YL1/2) and anti-GFP rabbit polyclonal antibodies were obtained from abcam (Cambridge). EEA1, M6PR and Lamp1 antibodies were a kind gift from Paul Luzio (Cambridge). Peroxidase conjugated secondary antibodies for western blotting were obtained from Sigma. Secondary antibodies for immunofluorescence were obtained from Jackson ImmunoResearch, Inc.

Cell culture, transfection and immunofluorescence
Cos-7 cells were seeded onto sterile glass coverslips in six well plates (2.5x10⁵ cells/well). After 24 h the cells were transfected with 400 ng of vector DNA, using the Effectene^® transfection reagent (Qiagen), according to the manufacturer's instructions. Following transfection, cells were incubated for 24 or 48 h in Dulbecco's modified Eagle's medium (DMEM) (Life Technologies) containing 10% fetal bovine serum (FBS) (Labtech International), 100 U/ml penicillin and 100 µg/ml streptomycin (Life Technologies). Undifferentiated PC12 cells were cultured in DMEM containing 10% horse serum (HS), 5% FBS and 1% penicillin and streptomycin (Life Technologies). Cells were seeded at 1x10⁵ into six well plates containing glass coverslips pre-coated with Poly-D-Lysine (Sigma). Undifferentiated cells were then either transiently transfected with Effectene^® transfection reagent (Qiagen), according to the manufacturer's instructions for 48 h or transiently transfected followed by differentiation with 50 ng/ml NGF (Sigma) in low serum DMEM containing 1% HS and 0.5% FBS. Differentiated cells were cultured in NGF containing media for at least 48 h after transfection.

Following transfection, cells were washed in PBS (Life Technologies) and fixed with 3.8% paraformaldehyde and 4% sucrose in PBS at room temperature for 15 min. Fixed cells were washed in PBS before being permeabilized in PBS containing 0.1% Triton X-100 (Sigma) or 0.05% saponin (Sigma) at room temperature for 5 min. Permeabilized cells were then washed three times in PBS, incubated for 15 min in a blocking solution (PBS, 10% FBS, ±0.05% saponin) and then transferred to blocking solution containing the appropriate epitope-specific antibody at an appropriate dilution. After 60 min incubation, coverslips were washed three times in blocking solution and were then incubated for 60 min in blocking buffer containing the appropriate secondary antibodies, at a concentration of 1/100. Cover slips were then washed three times in PBS and once in distilled H₂O, after which they were mounted in Vectorshield^TM (Vector Laboratories Inc.) medium on a glass slide. Stained samples were analysed on a Nikon Eclipse E800 microscope and Bio-rad microradience confocal analysis equipment. Images were recorded using Lazer-sharp-OS2 software and data were subsequently processed using Adobe Photoshop and Microsoft PowerPoint programmes.

ß-Lactamase protein fragment complementation assay
ß-Lactamase cDNA constructs.
The ampicillin resistance gene (bla) was used as a target to PCR two fragments of ß-lactamase, BLF1 (aminoacids 24–197) and BLF2 (aminoacids 198–288), using forward and reverse primers containing restriction sites PstI and XhoI, respectively. Complementary oligonucleotides containing several unique restriction sites (HindIII, HpaI, PstI and XhoI) and coding in frame for a 15 amino-acid flexible polypeptide linker (GGGGS)₃ were hybridized together and ligated into pcDNA3.1/Zeo(+) linearized with HindIII and XhoI. BLF1 and BLF2 fragments of ß-lactamase were ligated between PstI and XhoI sites downstream and in frame with the (GGGGS)₃ linker to give two separate vectors expressing different fragments of the ß-lactamase gene (pcDNA3.1/Zeo(+)(GGGGS)₃ BLF1 and pcDNA3.1/Zeo(+)(GGGGS)₃ BLF2). An upstream HpaI site was used to ligate a Gateway reading frame cassette (Rf B). This Gateway reading frame cassette was used to ligate either full-length human CHMP1B or spastin upstream of BLF1 in pcDNA3.1/Zeo(+)(GGGGS)₃ or BLF2 in pcDNA3.1/Zeo(+) (GGGGS)₃. All constructs were verified in frame and confirmed by DNA sequencing.

ß-Lactamase protein-assisted complementation colorimetric assay.
The assay was based on the previously described method (30,31), with minor modifications. HEK293T cells were split 24 h prior to transfection and seeded at 2.5x10⁵ in 12 well plates (Nunc^TM) in DMEM containing 10% FBS and 1% penicillin and streptomycin. Cells were transiently transfected with 300 ng of plasmid DNA using Effectene® transfection reagent (Qiagen) according to the manufacturer's instructions. Forty-eight hours after transfection cells were washed twice with PBS and re-suspended in 100 µl of 100 mM phosphate buffer pH 7.4 then lysed by three freeze–thaw cycles by freezing in dry ice/ethanol for 10 min and thawing in a 37°C water bath for 10 min. Total cell lysate was then used to test for ß-lactamase activity in 96 well plates (Falcon^®). One hundred microlitres of 100 mM phosphate buffer pH 7.4 was added to each well of a 96 well plate to a final concentration of 60 mM containing 20 µl of total cell lysate and 2 µl 10 mM Nitrocefin (Becton Dickinson and Company, Sparks, MD; final concentration 100 µM) and diluted to 200 µl with deionized water. Samples were assayed in triplicate using a Fusion^TM Alpha Universal Plate Reader (Perkin Elmer^TM Instruments) fitted with a 485 nm (20 nm bandwidth) filter in the absorption mode.

Nitrocefin hydrolysis rates were calculated from plots of the linear range of increasing absorbance monitored over 60 min. Data was normalized against total cell lysate protein content determined by the Bio-Rad Protein assay (Bio-Rad) based on the Bradford method (57).

In vivo ß-lactamase fluorescence protein fragment complementation assay.
HEK293T cells were split 24 h prior to transfection and seeded at 2.5x10⁵ per well in a 2-well glass chamber slide (Nunc^® Lab-tech II Chamber Slide System) in DMEM containing 10% FBS and 1% penicillin and streptomycin. Cells were transiently transfected with 300 ng plasmid DNA using Effectene^® transfection reagent (Qiagen) according to the manufacturer's instructions. Forty-eight hours later cells were washed once with PBS and resuspended in 750 µl Hank's balanced salt solution (HBSS) (Life Technologies). ß-Lactamase activity was then detected using the GeneBLAzer^TMdetection Kit (Invitrogen) according to the manufacturer's instructions. Following the addition of the fluorogenic cell permeable substrate CCF2/AM, chamber slides were incubated for 1 h protected from light and then visualized for fluorescence. Fluorescence microscopy of live HEK293T cells on glass chamber slides was performed using an Olympus IX81 inverted epi-fluorescence microscope with a 40x objective (Olympus). Simultaneous acquisition of two-colour fluorescence in real-time was performed using a 400DF15 excitation filter and the Dual View^TM imaging system (Optical Insights), containing a filter cube that had 450DF65 and 535DF35 emission filters (OMEGA Optical). CellR imaging software (Olympus) was used for analysis of multicolour images. Quantification of ß-lactamase activity in vivo was performed by counting the number of blue and green cells in 10 fields of view from each putative protein–protein interaction pairing.

Co-immunoprecipitation analysis of Spastin CHMP1B interactions
HEK293T cells transfected with spastin–Myc and CHMP1B–GFP alone or co-transfected with spastin–Myc and CHMP1B or spastin–Myc and EGFP-C1 were washed twice with PBS and lysed in 600 µl NP40 detergent buffer containing 1% (v/v) NP-40, 150 mM NaCl, 10 mM Tris–HCl (pH7.5), 0.02% sodium azide, 1 mg/ml BSA and 0.5% protease inhibitor cocktail. An aliquote of 100 µl was removed from a whole cell lysate fraction and the remaining 500 µl centrifuged at 13 000 g for 15 min at 4°C. The soluble fraction was then removed and extracts were incubated overnight at 4°C with rabbit polyclonal anti-GFP primary antibody (abcam) at a concentration of 1 : 250. Fifty microlitres of 50% Protein A–Sepharose beads was then added and incubated with protein–antibody complexes for 2 h at 4°C. Beads were then isolated by brief centrifugation and washed three times in NP-40 buffer containing BSA with a final wash in NP-40 buffer without BSA. Beads were then resuspended in 5x SDS–PAGE sample buffer and heated for 5 min at 95°C. Bound proteins were resolved by SDS–PAGE and immunoblotted with mouse monoclonal anti-Myc clone 4A6 primary antibody at a concentration of 1 : 250 (Upstate).

Subcellular fractionation
Cos-7 cells were split 24 h prior to transfection and seeded at 0.75x10⁶ per plate in 10 cm culture dishes. Twenty-four hours later they were transfected using Effectene® transfection reagent (Qiagen). Forty-eight hours after transfection, subcellular fractionation was carried out using a subcellular proteome extraction kit (Calbiochem), according to the manufacturer's protocol for adherent cells. Where some cells became non-adherent during the protocol, the cytosolic, membrane and nuclear fractions were spun at 750 g, 5500 g and 6800 g, respectively, for 10 min at 4°C, to remove any contamination from later fractions. Proteins were resolved by SDS–PAGE and immunoblotted with appropriate antibodies.

PCR and sequencing of genomic DNA
Ethical approval was granted by the Addenbrooke's NHS Trust Ethical Committee. The coding sequence, with flanking genomic sequence, of CHMP1B's single exon was amplified by PCR and sequenced in two overlapping fragments. PCR and sequencing was carried out as previously described (37). Primer sequences used to amplify the two CHMP1B fragments from genomic DNA were: F1: GCTCTGACGTCACCACCTG, R1: CGAATTTGTCCATCAAAGCA, F2: AACATGGAAGTTGCGAGGAT, R2: GGTAGAAGGGCATTTCAGCA.

	SUPPLEMENTARY MATERIAL

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 
Supplementary Material is available at HMG Online.

	ACKNOWLEDGEMENTS

 
We thank the families who kindly agreed to participate in our research on HSP. We thank Paul Luzio for helpful discussions. The pGBDU-C series of vectors and the PJ69-4A MATa and MAT yeast strains were kindly provided by Dr. Philip James (Department of Biomolecular Chemistry, University of Wisconsin, Madison, WI 53706-1532, USA). This work has been supported by grants from the Wellcome Trust and Action Research to E.R., who is a Wellcome Trust Advanced Clinical Fellow.

	FOOTNOTES

 
The authors wish it to be known that, in their opinion, the first two authors should be regarded as joint First Authors.

	REFERENCES

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS SUPPLEMENTARY MATERIAL REFERENCES

 

1. Barr, M.L. and Kiernan, J.A. (1983) The Human Nervous System. Harper and Row, Philadelphia, PA.

2. Reid, E. (2003) Science in motion: common molecular pathological themes emerge in the hereditary spastic paraplegias. J. Med. Genet., 40, 35–39.

3. Fink, J.K. (2002) Hereditary spastic paraplegia. In Rimoin, D.L., Pyeritz, R.E., Connor, J.M. and Korf, B.R. (eds) Emory and Rimoin's Principles and Practice of Medical Genetics, 4th edn. Harcourt, London, UK, 3124–3145.

4. Harding, A.E. (1984) The Hereditary Ataxias and Related Disorders. Churchill Livingstone, Edinburgh.

5. Hazan, J., Fonknechten, N., Mavel, D., Paternotte, C., Samson, D., Artiguenave, F., Davoine, C.S., Cruaud, C., Dürr, A., Wincker, P. et al. (1999) Spastin, a new AAA protein, is altered in the most frequent form of autosomal dominant spastic paraplegia. Nat. Genet., 23, 296–303.[CrossRef][Web of Science][Medline]

6. Sauter, S., Miterski, B., Klimpe, S., Bonsch, D., Schols, L., Visbeck, A., Papke, T., Hopf, H.C., Engel, W., Deufel, T. et al. (2002) Mutation analysis of the spastin gene (SPG4) in patients in Germany with autosomal dominant hereditary spastic paraplegia. Hum. Mutat., 20, 127–132.[CrossRef][Web of Science][Medline]

7. Charvin, D., Cifuentes-Diaz, C., Fonknechten, N., Joshi, V., Hazan, J., Melki, J. and Betuing, S. (2003) Mutations of SPG4 are responsible for a loss of function of spastin, an abundant neuronal protein localized in the nucleus. Hum. Mol. Genet., 12, 71–78.[Abstract/Free Full Text]

8. Wharton, S.B., McDermott, C.J., Grierson, A.J., Wood, J.D., Gelsthorpe, C., Ince, P.G. and Shaw, P.J. (2003) The cellular and molecular pathology of the motor system in hereditary spastic paraparesis due to mutation of the spastin gene. J. Neuropathol. Exp. Neurol., 62, 1166–1177.[Web of Science][Medline]

9. Svenson, I.K., Ashley-Koch. A.E., Gaskell, P.C., Riney, T.J., Cumming, W.J., Kingston, H.M., Hogan, E.L., Boustany, R.M., Vance, J.M., Nance, M.A. et al. (2001) Identification and expression analysis of spastin gene mutations in hereditary spastic paraplegia. Am. J. Hum. Genet., 68, 1077–1085.[CrossRef][Web of Science][Medline]

10. Beetz, C., Brodhun, M., Moutzouris, K., Kiehntopf, M., Berndt, A., Lehnert, D., Deufel, T., Bastmeyer, M. and Schickel, J. (2004) Identification of nuclear localization sequences in spastin (SPG4) using a novel Tetra-GFP reporter system. Biochem. Biophys. Res. Commun., 318, 1079–1084.[CrossRef][Web of Science][Medline]

11. Errico, A., Claudiani, P., D'Addio, M. and Rugarli, E.I. (2004) Spastin interacts with the centrosomal protein NA14, and is enriched in the spindle pole, the midbody, and the distal axon. Hum. Mol. Genet., 13, 2121–2132.[Abstract/Free Full Text]

12. Patel, S. and Latterich, M. (1998) AAA team: related ATPases with diverse functions. Trends Cell Biol., 8, 65–71.[Web of Science][Medline]

13. Errico, A., Ballabio, A. and Rugarli, E.I. (2002) Spastin, the protein mutated in autosomal dominant hereditary spastic paraplegia, is involved in microtubule dynamics. Hum. Mol. Genet., 11, 153–163.[Abstract/Free Full Text]

14. McDermott, C.J., Grierson, A.J., Wood, J.D., Bingley, M., Wharton, S.B., Bushby, K.M. and Shaw, P.J. (2003) Hereditary spastic paraparesis: disrupted intracellular transport associated with spastin mutation. Ann. Neurol., 54, 748–759.[CrossRef][Web of Science][Medline]

15. Trotta, N., Orso, G., Rossetto, M.G., Daga, A. and Broadie, K. (2004) The hereditary spastic paraplegia gene, spastin, regulates microtubule stability to modulate synaptic structure and function. Curr. Biol. 14, 1135–1147.[CrossRef][Web of Science][Medline]

16. Patel, H., Cross, H., Proukakis, C., Hershberger, R., Bork, P., Ciccarelli, F.D., Patton, M.A., McKusick, V.A. and Crosby, A.H. (2002) SPG20 is mutated in Troyer syndrome, an hereditary spastic paraplegia. Nat. Genet., 31, 347–348.[Web of Science][Medline]

17. Ciccarelli, F.D., Proukakis, C., Patel, H., Cross, H., Azam, S., Patton, M.A., Bork, P. and Crosby, A.H. (2003) The identification of a conserved domain in both spartin and spastin, mutated in hereditary spastic paraplegia. Genomics, 81, 437–441.[CrossRef][Web of Science][Medline]

18. Mellman, I. and Warren, G. (2000) The road taken: past and future foundations of membrane traffic. Cell, 100, 99–112.[CrossRef][Web of Science][Medline]

19. Katzmann, D.J., Odorizzi, G. and Emr, S.D. (2002) Receptor downregulation and multivesicular-body sorting. Nat. Rev. Mol. Cell. Biol., 3, 893–905.[CrossRef][Web of Science][Medline]

20. Katzmann, D.J., Babst, M. and Emr, S.D. (2001) Ubiquitin-dependent sorting into the multivesicular body pathway requires the function of a conserved endosomal protein sorting complex, ESCRT-I. Cell, 106, 145–155.[CrossRef][Web of Science][Medline]

21. Babst, M., Katzmann, D.J., Snyder, W.B., Wendland, B. and Emr, S.D. (2002) Endosome-associated complex, ESCRT-II, recruits transport machinery for protein sorting at the multivesicular body. Dev. Cell, 3, 283–289.[CrossRef][Web of Science][Medline]

22. Babst, M., Katzmann, D.J., Estepa-Sabal, E.J., Meerloo, T. and Emr, S.D. (2002) Escrt-III: an endosome-associated heterooligomeric protein complex required for mvb sorting. Dev. Cell, 3, 271–282.[CrossRef][Web of Science][Medline]

23. von Schwedler, U.K., Stuchell, M., Muller, B., Ward, D.M., Chung, H.Y., Morita, E., Wang, H.E., Davis, T., He, G.P., Cimbora, D.M. et al. (2003) The protein network of HIV budding. Cell, 114, 701–713.[CrossRef][Web of Science][Medline]

24. Howard, T.L., Stauffer, D.R., Degnin, C.R. and Hollenberg, S.M. (2001) CHMP1 functions as a member of a newly defined family of vesicle trafficking proteins. J. Cell Sci., 114, 2395–2404.[Abstract/Free Full Text]

25. Babst, M., Wendland, B., Estepa, E.J. and Emr, S.D. (1998) The Vps4p AAA ATPase regulates membrane association of a Vps protein complex required for normal endosome function. EMBO J., 17, 2982–2993.[CrossRef][Web of Science][Medline]

26. Scheuring, S., Rohricht, R.A., Schoning-Burkhardt, B., Beyer, A., Muller, S., Abts, H.F. and Kohrer, K. (2001) Mammalian cells express two VPS4 proteins both of which are involved in intracellular protein trafficking. J. Mol. Biol., 312, 469–480.[CrossRef][Web of Science][Medline]

27. Stauffer, D.R., Howard, T.L., Nyun, T. and Hollenberg, S.M. (2001) CHMP1 is a novel nuclear matrix protein affecting chromatin structure and cell-cycle progression. J. Cell Sci., 114, 2383–2393.[Abstract/Free Full Text]

28. Lehner, B., Semple, J.I., Brown, S.E., Counsell, D., Campbell, R.D. and Sanderson, C.M. (2004) Analysis of a high-throughput yeast two-hybrid system and its use to predict the function of intracellular proteins encoded within the human MHC class III region. Genomics, 83, 153–167.[CrossRef][Web of Science][Medline]

29. Peck, J.W., Bowden, E.T. and Burbelo, P.D. (2004) Structure and function of human Vps20 and Snf 7 proteins. Biochem. J., 377, 693–700.[Web of Science][Medline]

30. Galarneau, A., Primeau, M., Trudeau, L.-E. and Michnick, S.W. (2002) ß-Lactamase protein fragment complementation assays as in vivo and in vitro sensors of protein-protein interactions. Nat. Biotechnol., 20, 619–622.[CrossRef][Web of Science][Medline]

31. Wehrman, T., Kleaveland, B., Her, J.H., Balint, R.F. and Blau, H.M. (2002) Protein-protein interactions monitored in mammalian cells via complementation of beta-lactamase enzyme fragments. Proc. Natl Acad. Sci. USA 99, 3469–3474.[Abstract/Free Full Text]

32. Lee, H.K., Dunzendorfer, S. and Tobias, P.S. (2003) Cytoplasmic domain-mediated dimerizations of toll-like receptor 4 observed by beta-lactamase enzyme fragment complementation. J. Biol. Chem., 279, 10564–10574.[Medline]

33. Xu, T., Vasilyeva, E. and Forgac, M. (1999) Subunit interactions in the clathrin-coated vesicle vacuolar (H(+))-ATPase complex. J. Biol. Chem., 274, 28909–28915.[Abstract/Free Full Text]

34. Altschul, S.F., Gish, W., Miller, W., Myers, E.W. and Lipman, D.J. (1990) Basic local alignment search tool. J. Mol. Biol., 215, 403–410.[CrossRef][Web of Science][Medline]

35. Sofroniew, M.V., Howe, C.L. and Mobley, W.C. (2001) Nerve growth factor signaling, neuroprotection, and neural repair. Annu. Rev. Neurosci., 24, 1217–1281.[CrossRef][Web of Science][Medline]

36. Delcroix, J.D., Valletta, J.S., Wu, C., Hunt, S.J., Kowal, A.S. and Mobley, W.C. (2003) NGF signaling in sensory neurons: evidence that early endosomes carry NGF retrograde signals. Neuron 39, 69–84.[CrossRef][Web of Science][Medline]

37. Reid, E., Kloos, M., Ashley-Koch, A., Hughes, L., Bevan, S., Svenson, I.K., Lennon Graham, F., Gaskell, P.C., Dearlove, A., Pericak-Vance, M.A. et al. (2002) A kinesin heavy chain (KIF5A) mutation in hereditary spastic paraplegia (SPG10). Am. J. Hum. Genet., 71, 1189–1194.[CrossRef][Web of Science][Medline]

38. Kanai, Y., Okada, Y., Tanaka, Y., Harada, A., Terada, S. and Hirokawa, N. (2000) KIF5C, a novel neuronal kinesin enriched in motor neurons. J. Neurosci., 20, 6374–6384.[Abstract/Free Full Text]

39. Xia, C.H., Roberts, E.A., Her, L.S., Liu, X., Williams, D.S., Cleveland, D.W. and Goldstein, L.S. (2003) Abnormal neurofilament transport caused by targeted disruption of neuronal kinesin heavy chain KIF5A. J. Cell Biol., 161, 55–66.[Abstract/Free Full Text]

40. Hadano, S., Hand, C.K., Osuga, H., Yanagisawa, Y., Otomo, A., Devon, R.S., Miyamoto, N., Showguchi-Miyata, J., Okada, Y., Singaraja, R. et al. (2001) A gene encoding a putative GTPase regulator is mutated in familial amyotrophic lateral sclerosis 2. Nat. Genet., 29, 166–173.[CrossRef][Web of Science][Medline]

41. Yang, Y., Hentati, A., Deng, H.X., Dabbagh, O., Sasaki, T., Hirano, M., Hung, W.Y., Ouahchi, K., Yan, J., Azim, A.C. et al. (2001) The gene encoding alsin, a protein with three guanine-nucleotide exchange factor domains, is mutated in a form of recessive amyotrophic lateral sclerosis. Nat. Genet. 29, 160–165.[CrossRef][Web of Science][Medline]

42. Otomo, A., Hadano, S., Okada, T., Mizumura, H., Kunita, R., Nishijima, H., Showguchi-Miyata, J., Yanagisawa, Y., Kohiki, E., Suga, E. et al. (2003) ALS2, a novel guanine nucleotide exchange factor for the small GTPase Rab5, is implicated in endosomal dynamics. Hum. Mol. Genet., 12, 1671–1687.[Abstract/Free Full Text]

43. Zhao, X., Alvarado, D., Rainier, S., Lemons, R., Hedera, P., Weber, C.H., Tukel, T., Apak, M., Heiman-Patterson, T., Ming, L. et al. (2001) Mutations in a newly identified GTPase cause autosomal dominant hereditary spastic paraparesis. Nat. Genet., 29, 326–331.[CrossRef][Web of Science][Medline]

44. Zhu, P.P., Patterson, A., Lavoie, B., Stadler, J., Shoeb, M., Patel, R. and Blackstone, C. (2003) Cellular localization, oligomerization, and membrane association of the hereditary spastic paraplegia 3A (SPG3A) protein atlastin. J. Biol. Chem., 278, 49063–49071.[Abstract/Free Full Text]

45. Simpson, M.A., Cross, H., Proukakis, C., Pryde, A., Hershberger, R., Chatonnet, A., Patton, M.A. and Crosby, A.H. (2003) Maspardin is mutated in mast syndrome, a complicated form of hereditary spastic paraplegia associated with dementia. Am. J. Hum. Genet., 73, 1147–1156.[CrossRef][Web of Science][Medline]

46. Rainier, S., Chai, J.H., Tokarz, D., Nicholls, R.D. and Fink, J.K. (2003) NIPA1 gene mutations cause autosomal dominant hereditary spastic paraplegia (SPG6). Am. J. Hum. Genet., 73, 967–971.[CrossRef][Web of Science][Medline]

47. Zhao, C., Takita, J., Tanaka, Y., Setou, M., Nakagawa, T., Takeda, S., Yang, H.W., Terada, S., Nakata, T., Takei, Y. et al. (2001) Charcot-Marie-Tooth disease type 2A caused by mutation in a microtubule motor KIF1Bbeta. Cell, 105, 587–597.[CrossRef][Web of Science][Medline]

48. Mersiyanova, I.V., Perepelov, A.V., Polyakov, A.V., Sitnikov, V.F., Dadali, E.L., Oparin, R.B., Petrin, A.N. and Evgrafov, O.V. (2000) A new variant of Charcot-Marie-Tooth disease type 2 is probably the result of a mutation in the neurofilament-light gene. Am. J. Hum. Genet., 67, 37–46.[Medline]

49. Previtali, S.C., Zerega, B., Sherman, D.L., Brophy, P.J., Dina, G., King, R.H., Salih, M.M., Feltri, L., Quattrini, A., Ravazzolo, R. et al. (2003) Myotubularin-related 2 protein phosphatase and neurofilament light chain protein, both mutated in CMT neuropathies, interact in peripheral nerve. Hum. Mol. Genet., 12, 1713–1723.[Abstract/Free Full Text]

50. Bomont, P., Cavalier, L., Blondeau, F., Ben Hamida, C., Belal, S., Tazir, M., Demir, E., Topaloglu, H., Korinthenberg, R., Tuysuz, B. et al. (2000) The gene encoding gigaxonin, a new member of the cytoskeletal BTB/kelch repeat family, is mutated in giant axonal neuropathy. Nat. Genet., 26, 370–374.[CrossRef][Web of Science][Medline]

51. Ding, J., Liu, J.J., Kowal, A.S., Nardine, T., Bhattacharya, P., Lee, A. and Yang, Y. (2002) Microtubule-associated protein 1B: a neuronal binding partner for gigaxonin. J. Cell Biol. 158, 427–433.[Abstract/Free Full Text]

52. Verhoeven, K., De Jonghe, P., Coen, K., Verpoorten, N., Auer-Grumbach, M., Kwon, J.M., FitzPatrick, D., Schmedding, E., De Vriendt, E., Jacobs, A. et al. (2003) Mutations in the small GTP-ase late endosomal protein RAB7 cause Charcot–Marie–Tooth type 2B neuropathy. Am. J. Hum. Genet., 72, 722–727.[CrossRef][Web of Science][Medline]

53. Horton, R.M., Hunt, H.D., Ho, S.N., Pullen, J.K. and Pease, L.R. (1989) Engineering hybrid genes without the use of restriction enzymes: gene splicing by overlap extension. Gene, 77, 61–68.[CrossRef][Web of Science][Medline]

54. James, P., Halladay, J. and Craig, E.A. (1996) Genomic libraries and a host strain designed for highly efficient two-hybrid selection in yeast. Genetics, 144, 1425–1436.[Abstract]

55. Grey, M. and Brendel, M. (1992) A ten-minute protocol for transforming Saccharomyces cerevisiae by electroporation. Curr. Genet., 22, 335–336.[CrossRef][Web of Science][Medline]

56. Fromont-Racine, M., Rain, J.C. and Legrain, P. (1997) Toward a functional analysis of the yeast genome through exhaustive two-hybrid screens. Nat. Genet., 16, 277–282.[CrossRef][Web of Science][Medline]

57. Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem., 72, 248–254.[CrossRef][Web of Science][Medline]

CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				B. McDonald and J. Martin-Serrano No strings attached: the ESCRT machinery in viral budding and cytokinesis J. Cell Sci., July 1, 2009; 122(13): 2167 - 2177. [Abstract] [Full Text] [PDF]

				S. W. Eastman, M. Yassaee, and P. D. Bieniasz A role for ubiquitin ligases and Spartin/SPG20 in lipid droplet turnover J. Cell Biol., March 23, 2009; 184(6): 881 - 894. [Abstract] [Full Text] [PDF]

				S. Vergarajauregui, P. S. Connelly, M. P. Daniels, and R. Puertollano Autophagic dysfunction in mucolipidosis type IV patients Hum. Mol. Genet., September 1, 2008; 17(17): 2723 - 2737. [Abstract] [Full Text] [PDF]

				J. L. Burman, L. Bourbonniere, J. Philie, T. Stroh, S. Y. Dejgaard, J. F. Presley, and P. S. McPherson Scyl1, Mutated in a Recessive Form of Spinocerebellar Neurodegeneration, Regulates COPI-mediated Retrograde Traffic J. Biol. Chem., August 15, 2008; 283(33): 22774 - 22786. [Abstract] [Full Text] [PDF]

				C. Yorikawa, E. Takaya, Y. Osako, R. Tanaka, Y. Terasawa, T. Hamakubo, Y. Mochizuki, H. Iwanari, T. Kodama, T. Maeda, et al. Human Calpain 7/PalBH Associates with a Subset of ESCRT-III-related Proteins in its N-terminal Region and Partly Localizes to Endocytic Membrane Compartments J. Biochem., June 1, 2008; 143(6): 731 - 745. [Abstract] [Full Text] [PDF]

				P. E. Row, H. Liu, S. Hayes, R. Welchman, P. Charalabous, K. Hofmann, M. J. Clague, C. M. Sanderson, and S. Urbe The MIT Domain of UBPY Constitutes a CHMP Binding and Endosomal Localization Signal Required for Efficient Epidermal Growth Factor Receptor Degradation J. Biol. Chem., October 19, 2007; 282(42): 30929 - 30937. [Abstract] [Full Text] [PDF]

				P. Gissen and E. R Maher Cargos and genes: insights into vesicular transport from inherited human disease J. Med. Genet., September 1, 2007; 44(9): 545 - 555. [Abstract] [Full Text] [PDF]

				J. C. Bakowska, H. Jupille, P. Fatheddin, R. Puertollano, and C. Blackstone Troyer Syndrome Protein Spartin Is Mono-Ubiquitinated and Functions in EGF Receptor Trafficking Mol. Biol. Cell, May 1, 2007; 18(5): 1683 - 1692. [Abstract] [Full Text] [PDF]

				V. M. Olkkonen and E. Ikonen When intracellular logistics fails - genetic defects in membrane trafficking J. Cell Sci., December 15, 2006; 119(24): 5031 - 5045. [Abstract] [Full Text] [PDF]

				A. Tarrade, C. Fassier, S. Courageot, D. Charvin, J. Vitte, L. Peris, A. Thorel, E. Mouisel, N. Fonknechten, N. Roblot, et al. A mutation of spastin is responsible for swellings and impairment of transport in a region of axon characterized by changes in microtubule composition Hum. Mol. Genet., December 15, 2006; 15(24): 3544 - 3558. [Abstract] [Full Text] [PDF]

				K. Evans, C. Keller, K. Pavur, K. Glasgow, B. Conn, and B. Lauring Interaction of two hereditary spastic paraplegia gene products, spastin and atlastin, suggests a common pathway for axonal maintenance PNAS, July 11, 2006; 103(28): 10666 - 10671. [Abstract] [Full Text] [PDF]

				F. Crippa, C. Panzeri, A. Martinuzzi, A. Arnoldi, F. Redaelli, A. Tonelli, C. Baschirotto, G. Vazza, M. L. Mostacciuolo, A. Daga, et al. Eight Novel Mutations in SPG4 in a Large Sample of Patients With Hereditary Spastic Paraplegia. Arch Neurol, May 1, 2006; 63(5): 750 - 755. [Abstract] [Full Text] [PDF]

				J. M. Lottridge, A. R. Flannery, J. L. Vincelli, and T. H. Stevens Vta1p and Vps46p regulate the membrane association and ATPase activity of Vps4p at the yeast multivesicular body PNAS, April 18, 2006; 103(16): 6202 - 6207. [Abstract] [Full Text] [PDF]

				C Depienne, C Tallaksen, J Y Lephay, B Bricka, S Poea-Guyon, B Fontaine, P Labauge, A Brice, and A Durr Spastin mutations are frequent in sporadic spastic paraparesis and their spectrum is different from that observed in familial cases J. Med. Genet., March 1, 2006; 43(3): 259 - 265. [Abstract] [Full Text] [PDF]

				J. Schickel, C. Beetz, C. Frommel, G. Heide, A. Sasse, P. Hemmerich, and T. Deufel Unexpected pathogenic mechanism of a novel mutation in the coding sequence of SPG4 (spastin) Neurology, February 14, 2006; 66(3): 421 - 423. [Abstract] [Full Text] [PDF]

				C. M. Sanderson, J. W. Connell, T. L. Edwards, N. A. Bright, S. Duley, A. Thompson, J. P. Luzio, and E. Reid Spastin and atlastin, two proteins mutated in autosomal-dominant hereditary spastic paraplegia, are binding partners Hum. Mol. Genet., January 15, 2006; 15(2): 307 - 318. [Abstract] [Full Text] [PDF]

				A.-L. Mahul-Mellier, F. J. Hemming, B. Blot, S. Fraboulet, and R. Sadoul Alix, Making a Link between Apoptosis-Linked Gene-2, the Endosomal Sorting Complexes Required for Transport, and Neuronal Death In Vivo J. Neurosci., January 11, 2006; 26(2): 542 - 549. [Abstract] [Full Text] [PDF]

				A. Scott, J. Gaspar, M. D. Stuchell-Brereton, S. L. Alam, J. J. Skalicky, and W. I. Sundquist Structure and ESCRT-III protein interactions of the MIT domain of human VPS4A PNAS, September 27, 2005; 102(39): 13813 - 13818. [Abstract] [Full Text] [PDF]

				K. J. Evans, E. R. Gomes, S. M. Reisenweber, G. G. Gundersen, and B. P. Lauring Linking axonal degeneration to microtubule remodeling by Spastin-mediated microtubule severing J. Cell Biol., February 14, 2005; 168(4): 599 - 606. [Abstract] [Full Text] [PDF]

This Article Abstract FREE Full Text (PDF) Supplementary Material All Versions of this Article: 14/1/19    most recent ddi003v1 Alert me when this article is cited Alert me if a correction is posted Services Email this article to a friend Similar articles in this journal Similar articles in ISI Web of Science Similar articles in PubMed Alert me to new issues of the journal Add to My Personal Archive Download to citation manager Search for citing articles in: ISI Web of Science (58) Request Permissions Google Scholar Articles by Reid, E. Articles by Sanderson, C. M. Search for Related Content PubMed PubMed Citation Articles by Reid, E. Articles by Sanderson, C. M. Social Bookmarking          What's this?

Online ISSN 1460-2083 - Print ISSN 0964-6906

Copyright © 2009 Oxford University Press

Oxford Journals Oxford University Press

• Site Map
• Privacy Policy
• Frequently Asked Questions

Other Oxford University Press sites: Oxford University Press Oxford Journals China Oxford Journals Japan American National Biography Booksellers' Information Service Children's Fiction and Poetry Children's Reference Corporate & Special Sales Dictionaries Dictionary of National Biography Digital Reference English Language Teaching Higher Education Textbooks Humanities International Education Unit Law Medicine Music Online Products Oxford English Dictionary Reference Rights and Permissions Science School Books Social Sciences Very Short Introductions World's Classics