Myotubularin Lipid Phosphatase Binds the Hvps15/ Hvps34 Lipid Kinase Complex on Endosomes

# 2007 The Authors Journal compilation # 2007 Blackwell Publishing Ltd Traffic 2007; 8: 1052–1067 Blackwell Munksgaard doi: 10.1111/j.1600-0854.2007.00586.x Myotubularin Lipid Phosphatase Binds the hVPS15/ hVPS34 Lipid Kinase Complex on Endosomes

Canhong Cao1, Jocelyn Laporte2,3,4,5,6, Received 14 July 2006, revised and accepted for publica- Jonathan M. Backer7, Angela Wandinger- tion 20 April 2007, uncorrected manuscript published online 24 April 2007 Ness1,* and Mary-Pat Stein8

1Molecular Trafficking Laboratory, Department of Myotubularins define a diverse family of lipid phospha- Pathology, University of New Mexico School of tases that when mutated result in severe human diseases Medicine, Albuquerque, NM 87131, USA 2 affecting different tissues such as muscle in X-linked Department of Molecular Pathology, IGBMC (Institut de myotubular myopathy (MTM1)(1,2) and the peripheral Ge´ ne´ tique et de Biologie Mole´ culaire et Cellulaire), Illkirch F-67400, France nervous system in Charcot-Marie-Tooth diseases type 3Inserm, U596, Illkirch F-67400, France 4B1(3) and type 4B2 (2,4). The MTM family consists of 4CNRS, UMR7104, Illkirch F-67400, France 14 highly conserved dual-specificity protein tyrosine phos- 5Universite´ Louis Pasteur, Strasbourg F-67000, France phatase (PTP)-like enzymes: eight are catalytically active 6 Colle` ge de France, Chaire de Ge´ ne´ tique Humaine, and six contain mutations in their HCX5R PTP catalytic Illkirch F-67400, France motif, rendering them phosphatase dead (5,6). Although 7 Department of Molecular Pharmacology, Albert Einstein homologues do not compensate for individual family College of Medicine, Bronx, NY 10461, USA members, cooperation between catalytically active and 8Department of Biology, California State University Northridge, Northridge, CA 91330, USA inactive myotubularins has been suggested to occur *Corresponding author: Angela Wandinger-Ness, (2,7–9). In addition to a PTP domain, each family member [email protected] has a Suvar 3-9, Enhancer of zeste, Trithorax-interacting domain (10). Most members also have a coiled-coil domain that mediates homodimerization and/or heterodimerization (11,12), and one or more phosphoinositide-binding do- Myotubularins constitute a ubiquitous family of phos- mains that facilitate myotubularin membrane localization phatidylinositol (PI) 3-phosphatases implicated in several neuromuscular disorders. Myotubularin [myotubular (11,13). The human myotubularin family appears to be myopathy 1 (MTM1)] PI 3-phosphatase is shown associ- expressed in all tissues with the exception of hMTMR7, ated with early and late endosomes. Loss of endosomal which is only expressed in brain (14). Thus, differential phosphatidylinositol 3-phosphate [PI(3)P] upon overex- expression levels, intracellular localization or activation of pression of wild-type MTM1, but not a phosphatase-dead distinct myotubularin enzymes may account for different MTM1C375S mutant, resulted in altered early and late disease states observed when closely related myotubularin endosomal PI(3)P levels and rapid depletion of early family members are mutated. Understanding how mutant endosome antigen-1. Membrane-bound MTM1 was myotubularins cause disease necessitates a more detailed directly complexed to the hVPS15/hVPS34 [vacuolar protein sorting (VPS)] PI 3-kinase complex with binding understanding of their cellular localizations, functions and mediated by the WD40 domain of the hVPS15 (p150) interacting partners. adapter protein and independent of a GRAM-domain point mutation that blocks PI(3,5)P2 binding. The WD40 Although first described as dual-specificity PTP enzymes, domain of hVPS15 also constitutes the binding site for the primary substrates for myotubularins in yeast and Rab7 and, as shown previously, contributes to Rab5 mammalian cells are the phosphoinositides PI(3)P (15,16) binding. In vivo, the hVPS15/hVPS34 PI 3-kinase complex and PI(3,5)P (17–19). These phosphoinositide species are forms mutually exclusive complexes with the Rab 2 localized to early and late endosomes and play important GTPases (Rab5 or Rab7) or with MTM1, suggesting a competitive binding mechanism. Thus, the Rab roles in membrane trafficking (20,21). In particular, PI(3)P is GTPases together with MTM1 likely serve as molecular required for the recruitment of factors containing Fab1, switches for controlling the sequential synthesis and YOTB/ZK632.12, Vac1 and early endosome antigen-1 degradation of endosomal PI(3)P. Normal levels of endo- (EEA1) FYVE or phox (PX) domains (22–25). These factors somal PI(3)P and PI(3,5)P2 are crucial for both endosomal include tethering factors (EEA1), phosphoinositide kinases morphology and function, suggesting that disruption of (PIKfyve/Fab1p) and intralumenal vesicle forming machin- endosomal sorting and trafficking in skeletal muscle ery (hepatocyte growth factor-regulated tyrosine kinase when MTM1 is mutated may be a key factor in pre- substrate, Hrs/Vps27p) recruited to early and late endo- cipitating X-linked MTM. somes (26,27). The binding of FYVE/PX domain proteins Key words: Charcot-Marie-Tooth neuropathy, endocyto- to PI(3)P-containing endosomes provides a platform for sis, phosphatidylinositol 3-phosphatase or 3-kinase, Rab the recruitment and assembly of multiprotein complexes GTPase, X-linked myotubular myopathy that promote endosomal tethering, docking, fusion and

1052 www.traffic.dk MTM1 Binds hVPS15/hVPS34 PI 3-Kinase membrane remodeling, thereby facilitating endocytic hVPS34 complex. In addition, we demonstrate that both transport and receptor sorting. This has been demon- Rab7 and MTM1 bind to hVPS15 (formerly called p150), strated most clearly by numerous studies on the sorting the hVPS34 adapter molecule (42) that regulates hVPS34 and degradation of epidermal growth factor receptor activity through its interactions with Rab5 (38) and Rab7 (EGFR), which is highly dependent on PI(3)P and (32,41). The WD40 domain of hVPS15 mediates binding to

PI(3,5)P2 (13,28,29). Increased degradation of PI(3)P and Rab5, Rab7 and MTM1. Consequently, binding of the PI(3,5)P2 by overexpression of MTM1 has been shown to hVPS15/hVPS34 complex to the Rab GTPases and cause aberrant endosomal sorting and result in a profound MTM1 is mutually exclusive. The data suggest that tight inhibition of epidermal growth factor (EGF)-stimulated regulation and rapid turnover of the phosphoinositide- receptor degradation (13). Thus, myotubularins play an mediated signals may be accomplished through the cou- important role in integrating cellular phosphoinositide pling of proteins involved in the synthesis and turnover of regulation, membrane trafficking and growth control. PI(3)P on endosomes.

In mammalian cells, early and late endosomal PI(3)P are generated by the PI 3-kinase hVPS34 [vacuolar protein Results sorting (VPS)] (30–33), and the PI(3)P 5-kinase PIKfyve generates PI(3,5)P2 from PI(3)P on late endosomes (34). MTM1 is recruited to Rab5- and Rab7-positive early These phosphoinositide species are thought to be short and late endosomes lived because of rapid modification by kinases, dephos- Overexpressed MTM1 has been observed at the plasma phorylation by phosphatases such as the myotubularins or membrane and diffusely dispersed in the cytosol (15,16) degradation by lipases. In yeast, genetic studies demon- and has been shown to diminish early endosomal PI(3)P strate that PI(3)P is transported to vacuoles (the mamma- (43). Recently, however, EGF-dependent recruitment of lian equivalent of lysosomes), where it is either converted overexpressed MTM1 to kinetically defined late endo- into PI(3,5)P2 by Fab1p (35) or degraded by vacuolar somes has been reported (13). As a first step toward hydrolases (36). In mammalian cells, conversion of PI(3)P assessing the endosomal localization and recruitment of to PI(3,5)P2 by PIKfyve on late endosomes is thought to MTM1 more precisely, endogenous and overexpressed trigger the inward invagination of intraluminal vesicles into MTM1 and a phosphatase-dead mutant of MTM1 multivesicular bodies (MVB) and to facilitate sequestration (MTM1C375S) were localized relative to specific early of cargo such as EGFR into MVB (13,37). Hence, balanced and late endosomal markers. synthesis and consumption of PI(3)P and PI(3,5)P2 species on endosomal compartments ensures proper and efficient The endosomal distribution of endogenous MTM1 was endocytic transport. evaluated relative to markers of early and late endosomes (Figure 1A). To reveal the membrane-bound pool of The recruitment and activity of the hVPS34 PI 3-kinase on MTM1, cells were saponin extracted, a method well early and late endosomes depends on the upstream ac- known to preserve membrane ultrastructure while extract- tivation of the small GTPases Rab5 and Rab7, respectively ing cytosolic content (44,45). Endogenous MTM1 was (32,38). Through their organelle-specific localization and consistently observed outlining early endosomal mem- tightly regulated nucleotide binding and hydrolysis cycle, branes that were positive for Rab5. The Pearson’s corre- Rab GTPases are ideally poised as pivotal regulators of lation coefficient (range 1toþ1) for MTM1 and Rab5 was vesicular trafficking (reviewed in 64). Thus, a delicate 0.8 0.03, indicating a high degree of colocalization. The balance between Rab GTPases, lipid kinases and lipid incomplete overlap between EEA1 and MTM1 may reflect phosphatases must exist to regulate endosomal transport. localization of EEA1 to discrete early endosomal mem- Indirect evidence exists for the co-ordinate regulation of brane domains and/or competition between the two PI(3)P levels in Saccharomyces cerevisiae and S. pombe proteins for binding to the same PI(3)P-enriched regions. through the actions of hVPS34 and myotubularin phospha- Endogenous MTM1 was also observed on a subset of tases (15,39). In mammalian cells, Rab5 has been reported Rab7-positive late endosomes with a Pearson’s correlation to regulate PI(3)P levels at the plasma membrane and coefficient of 0.4 0.1. The MTM1 staining often gave the possibly continuing en route to early endosomes through impression of vesicle chains associated with the cytoskel- interactions with two PI 3-kinases, as well as PI 4- and eton, consistent with the known association between the 5-phosphatases (40). Our own work showed the direct endosomes and the microtubule network. Tyramide ampli- binding of Rab7 to the hVPS15/hVPS34 PI 3-kinase com- fication was necessary to definitively identify the endo- plex and demonstrated the importance of PI(3)P in late somal localization of the endogenous, membrane-bound endocytic events (32,41). This prompted us to characterize MTM1. The peroxidase-mediated deposition of tyramide the interfaces of the hVPS15/hVPS34 PI 3-kinase with during the amplification process resulted in a strong MTM1 PI 3-phosphatase on early and late endosomes. nuclear staining that was not observed in the absence of amplification (data not shown) or when overexpressed Here, we clearly establish that MTM1 localizes to both MTM1 was immunolocalized (Figures 1B,S1). Therefore, early and late endosomes and associates with the hVPS15/ the nuclear staining was not pursued further.

Traffic 2007; 8: 1052–1067 1053 Cao et al.

Figure 1: MTM1 localizes to Rab5-positive early endosomes and Rab7-positive late endosomes. A) Endogenous MTM1 localization. A431 cells were fixed and stained for endogenous MTM1 and EEA1, Rab5 or Rab7. MTM1 was localized with a rabbit polyclonal antibody directed against the human protein and Tyramide Signal Amplification (green channel). EEA1 was localized with a human antibody directed against EEA1 and a Cy5-conjugated donkey anti-human IgG secondary antibody (pseudocolored red). Rab5 was localized using a mAb directed against canine Rab5 and a rhodamine- conjugated donkey anti-mouse IgG secondary antibody. Rab7 was localized using a chicken IgY directed against canine Rab7 and a Cy5-conjugated donkey anti-chicken IgY secondary antibody (pseudocolored red). Arrows and inset panels highlight the colocalization. Colocalization of MTM1 with Rab5 and Rab7 was measured using SLIDEBOOK 4.1 IMAGE ANALYSIS software and determining the Pear- son’s correlation coefficient. B) Overexpressed MTM1wt and MTM1CS localization. The BHK cells expressing wtMTM1 or MTM1C375S-FLAG in combination with Rab5 (wt or constitutively active Q79L) or Rab7wt were permeabilized with saponin, fixed and stained. MTM1-FLAG and MTM1CS- FLAG were localized with a mouse mAb directed against the FLAG epitope. Rab5 and Rab7 were detected as in panel A. Fluorescently conjugated, species-specific secondary antibodies raised in donkey were used to visualize the primary antibody staining. Shown are images (pseudocolored as indicated) taken from one of three independent experiments, performed independently by two of the coauthors. Single optical sections, 0.5-mm thick, were collected on Zeiss LSM 510 using a plan-Neofluar 40/1.30 oil objective and scan zoom of 3.0. Insets are enlarged 2.3 times. Bars, 10 mm.

1054 Traffic 2007; 8: 1052–1067 MTM1 Binds hVPS15/hVPS34 PI 3-Kinase

Similar to the endogenous MTM1, overexpressed wild- the remaining diffuse EEA1 staining most likely repre- type (wt) MTM1 was found membrane associated and sents a cytosolic EEA1 pool that is incompletely ex- partially colocalized with early and late endosomes (Fig- tracted by saponin. The MTM1 was also transfected Hrs ure 1B). Because PI(3)P and PI(3,5)P2 facilitate membrane with 2xFYVE -myc as a second reporter for endo- binding of MTM1 and also serve as substrates, an MTM1- somal PI(3)P, and samples were stained for endogenous phosphatase-dead mutant (MTM1C375S) was used to Rab7. The 2xFYVEHrs-myc staining (red) was detected on further evaluate localization in the absence of catalytic Rab7-positive late endosomes (blue) in the absence of function (Figure 1B). The MTM1C375S was consistently MTM1 overexpression or in samples transfected with and extensively colocalized with both Rab5 (Rab5wt) and MTMC375S (seen as magenta in the merged image), but Rab7 (Rab7wt). The combined data demonstrate that decreased 10-fold when wtMTM1 was overexpressed MTM1 is present on early and late endosomes. Further- (Figure 2B, absence of any magenta and Figure 2D, more, the ectopically expressed MTM1 is shown to reflect Slidebook quantification of overlap). The combined data the distribution of the native protein, making it a useful tool demonstrate that wtMTM1 not only localizes to endo- to further dissect the function and regulation of the somes but also functions to decrease early and late phosphatase. endosomal PI(3)P levels.

Overexpression of MTM1 wild type (MTM1wt) or MTM1 colocalizes with and binds the hVPS34/ MTM1C375S was observed to cause changes in early hVPS15-kinase complex in vivo and late endosome morphology. In particular, myotubu- MTM1 displays enzymatic activity for PI(3)P and larin wt and mutant were often colocalized with Rab5 PI(3,5)P2, phosphoinositide species that are localized to and Rab7 on enlarged endosomes (Figure 1, inset early and late endosomal compartments, and is specifi- panels). Coexpression of MTM1C375S with the activated cally recruited to membranes in response to the presence Rab5Q79L caused extensive dilation of early endosomes of these phosphoinositides (13). In previous work, we that was more marked than when Rab5Q79L was ex- demonstrated that the synthesis of late endosomal PI(3)P pressed alone and enabled the detection of Rab5 and is governed by the hVPS15/hVPS34 PI 3-kinase complex MTM1 in discrete puncta while on the same dilated early bound to Rab7 and that hVPS34-kinase activity was endosome (Figure S1). Our observations suggest that the dependent on active Rab7 (32). The hVPS15/hVPS34 PI Rab proteins and the MTM1 phosphatase are likely pres- 3-kinase complex is also present on early endosomes ent in distinct domains and MTM1 impacts endosome where it is activated by Rab5 (31,38). Because specific form and function by affecting endosomal PI(3)P levels. To membrane recruitment of lipid-binding proteins is often further address these issues, we examined the impact of dependent on dual lipid–protein and protein–protein inter- MTM1 expression on PI(3)P pools and FYVE-domain pro- actions, it was of interest to determine if any direct tein binding. We also undertook to dissect the protein– association between the MTM1 and the proteins respon- protein interactions involved in the membrane association sible for coordinating PI(3)P synthesis on late endosomes of MTM1. exists.

Immunofluorescence analysis of wtMTM1 or the phos- Overexpression of wtMTM1 results in alteration of phatase-dead MTM1C375S mutant with hVPS34 revealed endosomal PI(3)P pools and loss of FYVE-domain their colocalization on endosomes (Figure 3A). To test if protein binding MTM1 was part of the hVPS15/hVPS34 complex, co- Activation of Rab5 is known to cause activation of the immunoprecipitation experiments were performed. The hVPS15/hVPS34 PI 3-kinase complex. The resulting en- BHK cells overexpressing FLAG-epitope (DYKDDDDK octa- richment in PI(3)P results in the recruitment of cytosolic peptide from gene-10 product of bacteriophage T7) tagged proteins with PI(3)P-binding motifs such as EEA1 (46). wtMTM1 or MTM1C375S with hVPS34 were lysed and Therefore, EEA1 membrane association is a useful MTM1 proteins were immunoprecipitated. Coimmuno- measure of early endosomal PI(3)P levels. Overexpres- precipitated hVPS34 was detected by immunoblotting sion of wtMTM1 in BHK cells resulted in the obvious (Figure 3B). The hVPS34-kinase was bound to wtMTM1 and apparently complete depletion of EEA1 staining as well as MTM1CS, but the association was more pro- from early endosomes relative to adjacent untrans- nounced with wtMTM1. Individually transfected control fected cells in the same field (Figure 2A). Cells express- samples demonstrate the specificity of the coimmunopre- ing the inactive MTM1C375S had no effect on EEA1 cipitation and whole cell lysate controls confirmed uniform levels compared with adjacent untransfected cells in the protein expression across samples. Analogous results same field (Figure 2A). The results are in agreement were obtained in reciprocal immunoprecipitations using with those reported using COS7 cells infected with hVPS34 or MTM1 antibodies and radiolabeled lysates (data adenoviruses encoding mutant MTM1 (43). Quantifica- not shown). Furthermore, the interaction was also dem- tion of the decreased endosomal EEA1 staining sug- onstrated between endogenous MTM1 and hVPS34 gests a two-fold decrease in early endosomal PI(3)P (Figure 3C). The data indicate that MTM1 associates with pools (Figure 2C). This is a minimum estimate because the hVPS34 PI 3-kinase on endosomes.

Traffic 2007; 8: 1052–1067 1055 Cao et al.

Figure 2: Legend on next page.

1056 Traffic 2007; 8: 1052–1067 MTM1 Binds hVPS15/hVPS34 PI 3-Kinase

The association of hVPS34 with Rab5 and Rab7 is mediated exhibited nearly one to one stoichiometry, confirming by the hVPS15 adapter protein (31,32,38,42). Therefore, that the lack of binding of hVPS15DWD40 was not cells expressing MTM1 and hVPS34 with or without because of misfolding of the mutant protein. There were hVPS15 were used for coimmunoprecipitation experi- also no statistically significant differences between ments. Lysates were prepared from radiolabeled samples wthVPS15 and either of the two other deletion mutants and the relative amounts of hVPS34 coimmunoprecipitated (hVPS15DPKD, 0.67 and hVPS15DHEAT, 0.80). The com- with MTM1 were determined. As shown in Figure 4A and bined data suggest that hVPS15 is pivotal in binding 4B, endogenous hVPS15 is sufficient for the coimmuno- MTM1 and likely important in MTM1 recruitment to sites precipitation of hVPS34 with MTM1, however, the ratio of of hVPS34 localization. Furthermore, specifically the hVPS34 coimmunoprecipitated with MTM1 increased upon WD40 domain appeared important for MTM1 binding to coexpression of exogenous hVPS15. This result suggests the hVPS15. The findings raised the question as to the importance of stoichiometry in the interaction between whether or not MTM1 may compete with Rab5 and the hVPS15/hVPS34-kinase complex and the MTM1 and Rab7 for binding to the hVPS15/hVPS34-kinase complex. hinted that the hVPS15 adapter protein might facilitate MTM1 binding to the hVPS15/hVPS34 complex. Rab7 and MTM1 bind to the hVPS15-WD40 domain in vitro Deletion constructs of hVPS15 were previously used to An in vitro binding assay was used to confirm the interac- map the binding site for Rab5 to the hVPS15-WD40 domain tion between hVPS15 and MTM1 that was observed (38). Rab7 was also shown to bind to wthVPS15 both in vivo in vivo (41). Wild-type hVPS15, hVPS15 deletion mutants and in vitro (32,41). Coimmunoprecipitation experiments and the hVPS15-WD40 domain were transcribed and performed using radiolabeled cell lysates overexpressing translated in vitro. Binding of the 35S-labeled hVPS15 individual hVPS15 deletion mutants and wtMTM1 revealed proteins to equimolar amounts of purified glutathione S- that the WD40 domain also plays an important role in transferase (GST), GST–Rab7 or GST–MTM1 was analyzed MTM1 binding to hVPS15 (Figure 4C,D). The WD40 as detailed in Materials and Methods (Figure 5A–D). Data hVPS15 deletion mutant was consistently coprecipitated presented are expressed as the fold increase in hVPS15 with MTM1 at 2.5- to 3-fold lower ratios than wthVPS15 or binding to specific GST fusion proteins relative to binding other deletion mutants (Figure 4D) despite similar levels to GST alone (a value of one indicates no binding above of expression of all hVPS15 constructs (Figure 4C, lower that detected for GST alone). Deletion of the WD40 panel). As a control for proper folding of the deletion domain of hVPS15 reduced binding to both Rab7 (Figure 5 constructs, various hVPS15 deletion mutants were tested A) and MTM1 (Figure 5B) to background levels. Deletion of for retention of their ability to bind to hVPS34 relative to either the HEAT [Huntington, Elongation Factor 3, PR65/A, the wthVPS15 (Figure 4E,F). The hVPS34-kinase was TOR (HEAT)] or the PKD (protein kinase domain) domain immunoprecipitated with a polyclonal antibody directed of hVPS15 actually increased binding of mutant hVPS15 against hVPS34 and the coprecipitation of the hVPS15 proteins above the levels observed for wthVPS15 to both adapter was scored by immunoblotting with an antibody Rab7 and MTM1. Both Rab7 and MTM1 bound to the directed against the V5-epitope tag. The wthVPS15 and WD40 domain expressed alone. The suboptimal binding of hVPS15DWD40 mutant were indistinguishable in their the WD40 domain to Rab7 may suggest that additional binding to hVPS34 (0.85 and 0.90, respectively) and interactions between hVPS15 and Rab7 contribute to

Figure 2: MTM1 regulates early and late endosomal PI(3)P levels. A) MTM1 regulates EEA1 membrane association. The BHK cells overexpressing MTM1wt-FLAG or MTM1C375S-FLAG were costained for MTM1-FLAG and endogenous EEA1. Samples were permeabilized with saponin, fixed and stained. The MTM1 was localized with a mouse mAb directed against the FLAG epitope and a rhodamine-conjugated donkey anti-mouse IgG secondary antibody (pseudocolored green), while EEA1 was stained with a human antibody directed against EEA1 and a Cy5-conjugated donkey anti-human antibody (pseudocolored red). Shown are images taken from one of three independent experiments, performed independently by two of the coauthors. Single optical sections, 0.5-mm thick, were collected on Zeiss LSM 510 using a plan-Neofluar 40/1.30 oil objective and scan zoom of 3.5. Arrows and inset panels highlight the colocalization. Transfected cell in each panel is outlined to distinguish it from the adjacent untransfected cell. Bar, 10 mm. B) MTM1 regulates late endosomal 2XFYVE membrane association. The BHK cells overexpressing 2xFYVEHrs-myc and MTM1wt-FLAG or MTM1C375S-FLAG were costained for MTM1-FLAG, myc and endogenous Rab7. The MTM1 was localized with a rabbit polyclonal antibody directed against the FLAG epitope and a FITC-conjugated donkey anti-rabbit IgG secondary antibody (green); 2XFYVE was stained with a mouse mAb directed against myc and a rhodamine-conjugated donkey anti-mouse antibody (red). Rab7 was detected as above with Cy5-conjugated secondary antibodies (blue). Shown are images taken from one of three independent experiments, performed independently by two of the coauthors. Single optical sections, 0.5-mm thick, were collected on Zeiss LSM 510 using a plan-Neofluar 40/1.30 oil objective and scan zoom of 3.5. Bar, 10 mm. C) MTM1 regulates early endosomal PI(3)P levels. Early endosomal PI(3)P levels (EEA1 levels) were quantified using SLIDEBOOK 4.1 IMAGE ANALYSIS software. Quantification was performed on 20 cells from each transfected sample derived from one of three independent experiments. Average fluorescence intensities are plotted with standard error of the mean calculated using GRAPHPAD PRISM software. ***Decrease in early endosomal PI(3)P (t-test, p < 0.0001). D) MTM1 regulates late endosomal PI(3)P levels. Late Hrs endosomal PI3P levels (2xFYVE -myc/Rab7 colabel) were quantified using SLIDEBOOK 4.1 IMAGE ANALYSIS software as above. ***Decrease in late endosomal PI(3)P (t-test, p < 0.0001).

Traffic 2007; 8: 1052–1067 1057 Cao et al.

Figure 3: MTM1 colocalizes with and is in a complex with the hVPS34 PI 3-kinase. A) MTM1 colocalizes with hVPS34 PI 3-kinase. The BHK cells expressing MTM1wt-FLAG or MTM1C375S-FLAG and hVPS34 were permeabilized with saponin, fixed and stained. The MTM1 was localized using a mouse mAb directed against the FLAG epitope and a rhodamine-conjugated donkey anti-mouse IgG secondary antibody (red). Human VPS34 was visualized using a rabbit polyclonal antibody (pAb) directed against hVPS34 and a FITC- conjugated donkey anti-rabbit IgG secondary antibody (green). Shown are images taken from one of three independent experiments, performed independently by two of the coauthors. Single optical sections, 0.5-mm thick, were collected on Zeiss LSM 510 using a plan- Neofluar 40/1.30 oil objective and scan zoom of 3.0. Bar, 10 mm. B) Overexpressed MTM1 is in a complex with the hVPS34 PI 3-kinase. The BHK cells expressing MTM1wt-FLAG or MTM1C375S-FLAG together with hVPS34 PI 3-kinase were saponin extracted. Lysates were solubilized in RIPA buffer and MTM1-FLAG immunoprecipitates were collected using a mouse mAb directed against the FLAG epitope. Immunoprecipitates were separated by SDS–PAGE and overexpressed MTM1-FLAG and hVPS34 detected by Western blotting using a mouse mAb directed against FLAG and a rabbit pAb directed against hVPS34. Representative data from one of two independent experiments, performed independently by two of the coauthors. C) Endogenous MTM1 binds to the hVPS34 PI 3-kinase. The A431 cells were lysed in RIPA buffer and immunoprecipitates were collected using a mouse mAb directed against human MTM1. Immunoprecipitates were separated by SDS–PAGE and endogenous MTM1 and hVPS34 detected by Western blotting with rabbit pAb directed against human MTM1 or hVPS34. Representative data from one of two independent experiments. IP, immunoprecipitates; WCL, whole cell lysates. wthVPS15 binding. Alternatively, the isolated WD40 (40). Similarly, the results presented here suggest that domain may adopt an altered conformation that is less hVPS15/hVPS34 PI 3-kinase activity may be directly effective in binding to Rab7. The results of the in vitro coupled to PI(3)P degradation. To test if a three-way binding studies confirm the binding of hVPS15 to MTM1 complex between the Rab GTPases, the PI 3-kinase and Rab7 observed in vivo and suggest that Rab7, MTM1 complex and MTM1 might exist in vivo (as opposed to and Rab5 (38) share the hVPS15-WD40 domain as a com- mutually exclusive MTM1/hVPS15/hVPS34 and Rab/ mon binding site. hVPS15/hVPS34 complexes), coimmunoprecipitation experiments were performed with cells simultaneously Binding of Rab GTPases or MTM1 to the hVPS15/ expressing Rab5 or Rab7, together with MTM1 and hVPS34 complex is mutually exclusive hVPS15. Immunoprecipitation using anti-Rab5 or anti- The established interaction of Rab5 with the hVPS15/ Rab7 antibodies demonstrated that the Rab GTPases hVPS34 PI 3-kinase complex, as well as with PI 5- and PI and hVPS15 are in a complex that does not include 4-phosphatases demonstrates that phosphoinositide gen- MTM1 (Figure 6). Conversely, immunoprecipitation of eration can be directly linked to phosphoinositide turnover MTM1 with an anti-FLAG antibody demonstrated that

1058 Traffic 2007; 8: 1052–1067 MTM1 Binds hVPS15/hVPS34 PI 3-Kinase

MTM1 bound to hVPS15 but not to Rab5 or Rab7. Because mutant, V49F, blocked PI(3,5)P2 binding (13). GRAM- hVPS15 is virtually completely bound to hVPS34 in vivo domain-mediated binding to PI(5)P and PI(3,4,5)P3 have (see Figure 4E–F), the results suggest that the PI 3-kinase also been reported (11,47). The importance of GRAM- complex binds to the Rab GTPases and MTM1, but not to domain-mediated phosphoinositide binding in endosomal both simultaneously. membrane localization and hVPS15 binding was therefore tested. Coimmunoprecipitation assays show hVPS15 bind- MTM1 GRAM domain-mutant lacking PI(3,5)P2 ing of the V49F mutant was identical to that of the binding is membrane localized and retains wthVPS15 (Figure 7A). Early and late endosomal localiza- hVPS15-binding activity tion of the mutant V49F protein was also unperturbed as The GRAM domain of MTM1 has been shown to facilitate evidenced by the depletion of EEA1 membrane localization

PI(3)P- and PI(3,5)P2-dependent membrane association in in cells expressing MTM1V49F and the colocalization with liposome-binding assays and a single GRAM-domain point Rab7 (Figure 7B).

Figure 4: Legend on next page.

Traffic 2007; 8: 1052–1067 1059 Cao et al.

Discussion in turn serves to recruit MTM1 to turn off the PI(3)P signal and downstream effector actions. Degradation of the Endosomal PI(3)P is a critical signaling molecule required PI(3)P signal may also contribute to the release of the for effective cargo sorting and transport on the endocytic MTM1 from membranes because of decreased GRAM- pathway. Thus, PI(3)P levels must be tightly regulated. dependent phosphoinositide binding. Although the mechanistic details are still incomplete, the demonstration that Rab5 interacts with PI 4- and The present data support the idea that the MTM1 phos- 5-phosphatases at the cell cortex suggests coordinated phatase is localized to both early and late endosomes control of phosphoinositide synthesis and degradation through dual binding specificities for the hVPS15/hVPS34 (40). In the present study, we provide new evidence that protein complex and the phosphoinositide lipids, PI(3)P endosomal PI(3)P levels are regulated through coordi- and PI(3,5)P2. Specific lipid binding was previously shown nated synthesis and degradation brought about by the to be mediated by the GRAM domain of MTM1 (11,13,47). formation of a specific complex between the myotublarin Deletion of the GRAM domain or substitution of valine at MTM1 lipid phosphatase and the hVPS34/hVPS15 lipid position 49 for phenylalanine (V49F) resulted in a failure of kinase. Biosynthesis of endosomal PI(3)P requires activa- stimulated membrane recruitment in response to activated tion of the Rab5 and Rab7 GTPases on early and late EGFR (13). In the present work, we reveal for the first time endosomes and their direct association with the hVPS15/ that both endogenous and overexpressed MTM1 are hVPS34 PI 3-kinase complex (38,41). Interaction of both endosome associated even under basal, unstimulated Rab GTPases with the lipid kinase is mediated principally conditions. The importance of phosphoinositide binding through the WD40 domain of the hVPS15 adapter protein for stable MTM1 membrane recruitment is underscored by [present study and (38)]. Although Rab5 also interacts our findings that active MTM1 depleted endosomal PI(3)P with the hVPS15 HEAT domain (38), this was not ob- pools and exhibited reduced membrane binding relative to served for Rab7. As shown here, the same WD40 domain the inactive MTMC375S mutant. Under conditions where of hVPS15 is also responsible for binding to the myotubu- endosomal PI(3)P synthesis was simultaneously enhanced larin (MTM1) 3-phosphatase. Consequently, the hVPS15/ through coexpression of hVPS34 or activated Rab7 (not hVPS34 complex associates with the Rab GTPases, Rab5 shown), membrane retention of active MTM1 was and Rab7, or the MTM1 phosphatase in a mutually exclu- enhanced. In contrast to what was observed in EGF- sive manner. Thus, on early and late endosomes, PI(3)P stimulated cells, the V49F mutant was indistinguishable synthesis and degradation are most likely coordinately and from wtMTM1 in its ability to associate with endosomes possibly sequentially regulated. We envision that the under unstimulated conditions. Most likely, this is because activated Rab5 and Rab7 GTPases bind to and activate as shown in the present work, the V49F mutant retains its the hVPS15/hVPS34 complex leading to the production of hVPS15-binding activity and therefore can still be recruited PI(3)P. The presence of active hVPS15/hVPS34 and PI(3)P to endosomal membranes. The membrane-associated

Figure 4: MTM1 binds to the PI 3-kinase adapter protein hVPS15 through the hVPS15-WD40 domain in vivo. A and B) Immunoprecipitation of MTM1/hVPS15/hVPS34 protein complexes formed in vivo. The BHK cells expressing MTM1wt-FLAG or MTM1C375S-FLAG in combination with hVPS34 alone or with epitope tagged, wthVPS15 (hVPS15wt-V5) were radiolabeled for 30 minutes with 35S-Trans label prior to collection. Saponin extracted lysates were solubilized in RIPA buffer and immunoprecipitates were collected using a mouse mAb directed against the FLAG epitope. Immunoprecipitates were separated by SDS–PAGE and A) coimmunoprecipitated radiolabeled proteins (top panel) and total cell lysates (bottom panel) were visualized by phosphorimage analysis on a Molecular Dynamics instrument. Because protein expression is driven by vaccinia virus infection/transfection, MTM1-FLAG, hVPS15wt and hVPS34 are the major newly synthesized proteins seen in the total cell lysate samples. Data are representative of one of two independent experiments. B) The relative amount of hVPS34 coimmunoprecipitated with MTM1-FLAG in the presence or absence of coexpressed hVPS15 adapter protein was expressed as a ratio of hVPS34 to MTM1 immunoprecipitated. Data represent average of duplicate experiments. C and D) Immunoprecipitation of MTM1/hVPS15 protein complexes formed in vivo with wthVPS15 or various deletion mutants. Cells expressing MTM1-FLAG with V5-epitope tagged, wthVPS15 (hVPS15wt-V5) or domain deletion mutant proteins (hVPS15DHEAT-V5, hVPS15DWD40- V5, hVPS15DPKD-V5) were radiolabeled as above. Saponin extracted lysates were solubilized in RIPA buffer and immunoprecipitates were collected as described using the anti-FLAG mAb. C) Coimmunoprecipitated radiolabeled proteins (top panel) and total cell lysates (bottom panel) were separated by SDS–PAGE and visualized by phosphorimage analysis. Data are representative of one of three independent experiments. D) The relative amounts of full-length wthVPS15 and each domain deletion mutant of hVPS15 (DHEAT, DWD40 and DPKD) coimmunoprecipitated with wtMTM1-FLAG was expressed as a ratio of hVPS15 to MTM1 immunoprecipitated. Results from three experiments were averaged and standard errors of the mean were calculated using GRAPHPAD PRISM software. E and F) Immunoprecipi- tation of hVPS34/hVPS15 protein complexes formed in vivo with wthVPS15 or various deletion mutants. Wild type, V5-tagged hVPS15 (hVPS15) and specific domain deletion mutant hVPS15 proteins (DPKD, DHEAT and DWD40) were coexpressed with hVPS34. Lysates were immunoprecipitated with a rabbit polyclonal antibody directed against hVPS34 and immunoblotted with a mouse mAb directed against the V5 epitope to detect hVPS15. F) The relative amount of full-length hVPS15 and each domain deletion mutant of hVPS15 (DHEAT, DWD40 and DPKD) coimmunoprecipitated with hVPS34 was expressed as ratio of hVPS15 to hVPS34 immunoprecipitated. Quantification was performed by densitometry. Neg. Cont, negative control; IP, immunoprecipitates; WCL, whole cell lysates.

1060 Traffic 2007; 8: 1052–1067 MTM1 Binds hVPS15/hVPS34 PI 3-Kinase

Figure 5: Rab7 and MTM1 bind to the WD40 domain of hVPS15 in vitro. A–D) In vitro binding of wt and domain deletion mutants of hVPS15 to Rab7 and MTM1. A) GST–Rab7 and B) GST–MTM1 were expressed in E. coli and bound to glutathione–Sepharose beads. The hVPS15wt and hVPS15 deletion mutant proteins (DHEAT, DWD40 and DPKD) and a hVPS15-WD40 domain protein were 35S-labeled by in vitro transcription and translation. Equimolar amounts of individual hVPS15 proteins and GST–Rab7 or GST–MTM1 were incubated at 48C for 2 h. Glutathione–Sepharose beads were pelleted and washed three times. Proteins were resolved by SDS–PAGE and quantified on a Molecular Dynamics phosphorimager. Binding of hVPS15 proteins to GST–Rab7 or GST–MTM1 is expressed as the fold increase over the binding of the same hVPS15 proteins to GST alone (denoted by dashed line). Results from three independent experiments were averaged and the standard errors of the mean were calculated using GRAPHPAD PRISM software (t-test, *p < 0.05, **p < 0.01). C) Input radiolabeled hVPS15 proteins resolved on SDS–PAGE and visualized by phosphorimage analysis. Results from one of three independent experiments are shown. D) Input purified GST, GST–Rab7 and GST–MTM1 resolved on SDS–PAGE and visualized by Western blotting using a mouse mAb directed against GST. Results from one of three independent experiments are shown.

Figure 6: Binding of hVPS15 to MTM1 and Rab5 or Rab7 is mutually exclusive. Saponin extracted lysates from BHK cells expressing MTM1wt-FLAG, hVPS15wt-V5 and Rab5 or Rab7 were solubilized in RIPA buffer. Immunoprecipitates were collected using a mAb anti- Rab5, a rabbit anti-Rab7 antibody or a mAb anti-FLAG antibody. Immunoprecipitates were separated by SDS–PAGE and coimmunoprecipitated proteins were visualized by Western blotting using anti-V5 (hVPS15), anti-FLAG (MTM1) and anti-Rab5 or anti-Rab7 antibodies. Total cell lysates confirm similar total expression levels of hVPS15-V5, MTM1-FLAG and Rab5 or Rab7 in each sample.

Traffic 2007; 8: 1052–1067 1061 Cao et al.

Figure 7: MTM1 GRAM-domain mutant lacking PI(3)P binding is membrane localized and retains hVPS15-binding activity. A) The MTM1V49F GRAM-domain mutant binds to hVPS15. The BHK cells overexpressing EGFP-MTM1wt or EGFP- MTM1V49F with hVPS15wt-V5 were saponin extracted. Lysates were solubilized in RIPA buffer and MTM1 immunoprecipitates were collected using an anti-GFP antibody. Immunoprecipi- tates were separated by SDS–PAGE and visualized by Western blotting using anti-GFP and anti-V5 antibodies. B) MTM1 GRAM-domain mutant retains membrane association. The BHK cells overexpressing EGFP-MTM1wt or EGFP-MTM1V49F in combination with Rab7wt were permeabilized with saponin and fixed. Endogenous EEA1 and overexpressed Rab7wt were detected as above with Cy5-conjugated secondary antibodies (pseudocolored red). Single optical sections, 0.5-mm thick, were collected on Zeiss LSM 510 using a plan-Neofluar 40/1.30 oil objective and scan zoom of 3.0. Arrows highlight the colocalization. Bar, 10 mm.

pool in unstimulated cells was not detected by Tsujita and brane binding of MTM1 reported by Tsujita et al. (13). coworkers because of the large cytosolic pool of MTM1 Taken together, the data suggest that both phosphoinosi- present in the absence of saponin pre-extraction. Yet, upon tide and protein–protein interactions are important in the ligand stimulation, the greatly expanded PI(3,5)P2 levels endosomal membrane association of the myotubularin appear to contribute significantly to the enhanced mem- MTM1 phosphatase.

1062 Traffic 2007; 8: 1052–1067 MTM1 Binds hVPS15/hVPS34 PI 3-Kinase

Accumulating evidence indicates that endosomal sorting dilated endosome morphology upon overexpression of and cargo selection entails segregation into discrete mem- MTM1 was attributed to depletion of the substrate or brane domains that are marked by different Rab GTPases. product of the PIKfyve PI 5-kinase by MTM, leading to For example, Rab5, Rab4 and Rab11 cooperatively medi- impaired endosome invagination (13). Late endosomal ate early endosomal sorting of recycling cargo (48). Recent PI(3)P is known to facilitate the recruitment of the PIKfyve real-time studies along with our previous work suggest PI 5-kinase and is a requisite substrate in the synthesis of that Rab7 is recruited to Rab5-positive endosomes and PI(3,5)P2. In turn, PI(3,5)P2 is necessary for late endosomal participates in the selection of cargo destined for late intraluminal vesicle formation and growth factor receptor endosomes and lysosomes (40,49–51). When present on sorting (13,34). The recruitment of MTM1 provides a nega- Rab5-positive endosomes, Rab7 is always observed in tive feedback mechanism at sites where PIKfyve is active, discrete patches that do not contain Rab5 (51). Because reducing the concentration of PI(3,5)P2 and generating the Rab7 association with early endosomes appears to be MTM1 allosteric activator PI(5)P on endosomal mem- extremely transient, detection is enhanced upon disruption branes (47). Thus, endosome form and function are highly of microtubules or following the expression of mutant dependent on PI(3)P and PI(3,5)P2 levels both of which are Rab5Q79L (51). Expression of Rab5Q79L also has been substrates of MTM1. shown to cause the accumulation of hVPS15/hVPS34 on early endosomes, usually on budding structures where In conclusion, the present studies demonstrate for the first lower levels of Rab5 are observed [(38) and our own time an interaction between a lipid kinase complex and unpublished observations]. Therefore, it is notable that a lipid phosphatase, mediated by the kinase adapter the association of wtMTM1 with early endosomes ap- hVPS15. Furthermore, it appears that a dynamic equilib- peared to be transient and was enhanced by expression of rium must exist between PI 3-kinase activation, mediated by the phosphatase-dead MTM1C375S mutant or Rab5Q79L. the endosomal Rab5 and Rab7 GTPases, and PI 3-kinase Rab7 itself is shown to be recruited to early endosomes inactivation, mediated by association with the MTM1 lipid and to bud and move away from Rab5-positive structures phosphatase, which in turn promotes endosomal phos- (52). Based on the cumulative data, Rab7 may facilitate the phoinositide homeostasis. segregation of the hVPS15/hVPS34-kinase complex and consequently the MTM1, into discrete early endosome domains destined for transport to late endosomes. Such a Materials and Methods mechanism may ensure that these components are present at the observed high levels on late endosomes where Cells and reagents continued tight regulation of PI(3)P synthesis and degra- Cell lines were purchased from the American Tissue Culture Collection (ATCC, dation is essential. Because Rab5 interacts directly with PI Rockville, MD, USA). The BHK cell line BHK21 was cultured in Glasgow’s minimal essential medium containing 10% FBS, L-glutamine, penicillin, 4- and 5-phosphatases (40), the data also imply that there streptomycin and tryptose-phosphate broth. The A431 cells were cultured may be a hierarchy of coordination of phosphoinositide in DMEM supplemented with 10% FBS, 2 mM L-glutamine, penicillin/ kinases and phosphatases. streptomycin. For vaccinia virus infection and transfection assays, cells were plated in six-well plates or 100-mm dishes at 80–90% confluence 1 day prior Our present studies and those of others highlight the to use. For transfection alone, cells were plated at 60% confluence 1 day prior importance of MTM1 in endosome function through its to transfection. All reagents were from Sigma unless otherwise noted. impact on PI(3)P levels. The EEA1 is recruited to membranes by active Rab5 and is sensitive to endosomal PI(3)P Vectors levels because of the specific recognition of PI(3)P by its Vectors containing FLAG-tagged wtMTM1 and FLAG-tagged MTM1C375S FYVE domain. Expression of wtMTM1 was previously were as described (54). The FLAG-tagged constructs were subcloned into shown to result in dissociation of EEA1 from early endo- pcDNA3.1 for use in infection/transfection experiments. N-terminally somes and PI(3)P consumption (43,53). We have repeated tagged MTM1wt was prepared as described (16). The pEGFP [enhanced green fluorescent protein (eGFP)] human MTM1wt and V49F mutant have this result and show that MTM1 also depletes late endo- been described elsewhere (13,43). The FLAG-tagged MTM1 and the eGFP- somal PI(3)P levels. Changes in endosomal PI(3)P levels tagged MTM1 proteins localized identically in all experiments performed. most likely affect membrane recruitment not only of EEA1 The pGEM-Rab5 constructs were the generous gift of Dr Marino Zerial but also of Hrs and other factors required for early endo- (Max Planck Institute for Molecular Cell Biology and Genetics, Dresden, Hrs some fusion (24). The resulting changes in the trafficking Germany). The pGEM-2xFYVE -myc vector was the generous gift of and fusion machinery in turn precipitate alterations in Dr Harald Stenmark (Norwegian Radium Hospital, Oslo, Norway). The hVPS15 constructs in pcDNA3.1-V5-His (Invitrogen), (hVPS15wt, hVPS15DHEAT, endosome morphology and function. Coexpression of hVPS15DWD40, hVPS15DPKD and specific domain constructs) were MTM1C375S and Rab5Q79L resulted in massively dilated constructed as described (38). The Rab7 and hVPS34 mammalian expres- early endosomes most likely because of the enhanced sion vectors were constructed as described (32,55). The N-terminal GST– fusion precipitated by the increased synthesis and failure Rab7 (canine) fusion construct was prepared in pGEX-5X-2 (Amersham to degrade PI(3)P precipitated under these conditions (data Biosciences, Piscataway, NJ, USA) as described (41). Human MTM1 was amplified by polymerase chain reaction using 50TCCCCCGGGACAATGGC- not shown). Tsujita et al. showed that overexpression of TTCTGCATCAACTTC and 50ATAAGAATGCGGCCGCTCAGAAGTGAGTT- wtMTM1 caused the dilation of an unidentified endosomal TGCACATG as primers and inserted into the pGEX5x.2 vector with SmaI compartment and impaired EGFR degradation (13). The and NotI such that GST was fused to the N-terminus of MTM1.

Traffic 2007; 8: 1052–1067 1063 Cao et al.

Antibodies Pearson’s correlation coefficient for two markers in 10 cells and determin- Endogenous or overexpressed Rab7 was detected using a custom chicken ing the average and standard deviation. anti-Rab7 antibody generated against a C-terminal Rab7 peptide and Coimmunoprecipitation from radiolabeled lysates prepared by Aves Laboratories (Tigard, OR, USA) or a polyclonal rabbit Lysates from cells radiolabeled and overexpressing MTM1wt-FLAG or antibody (50). Rab5 was detected using a polyclonal rabbit antibody that MTM1C375S-FLAG with hVPS34, and/or hVPS15wt or hVPS15 mutants was the kind gift of Dr Marino Zerial (Max Planck Institute for Molecular Cell were permeabilized in 0.05% saponin in 80 mM PIPES buffer for 5 minutes, Biology and Genetics, Dresden, Germany) or a monoclonal antibody (mAb) washed again with PBS and lysed in ice-cold RIPA (radioimmunoprecipita- 4F11 prepared by our laboratory (56,57). A custom polyclonal rabbit anti- tion assay) buffer (150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, hVPS34 antibody was generated against recombinant protein purified from 0.1% SDS, 50 mM Tris, pH 8.0) containing protease inhibitors. Lysates Escherichia coli [all animal work performed by HRP, Inc. (Denver, PA, USA)] were clarified by centrifugation at 20 817 g in an Eppendorf microfuge for and used to monitor the endogenous and overexpressed protein as 10 minutes at 48C. Monoclonal anti-FLAG (clone M2) antibody was added to described previously (32,41). Human autoimmune serum against EEA1 each lysate and samples were incubated rotating for 1.5 h at 48C. A rabbit was from Dr Ban-Hok Toh (Monash Medical School, Prahran, Victoria, anti-mouse immunoglobulin G (IgG) linker antibody was added for 30 Australia) (58). Mouse mAb 9E10 to the myc-epitope tag was purchased minutes prior to the addition of protein A–Sepharose to those lysate from Santa Cruz Biotechnologies Inc. (Santa Cruz, CA, USA). A mouse mAb samples immunoprecipitated with mAb anti-FLAG. Samples were washed directed against GST was from Santa Cruz Biotechnologies Inc. A rabbit and solubilized in SDS–PAGE sample buffer. Samples were resolved on 7.5, anti-GFP antibody was from Molecular Probes Inc. (Eugene, OR, USA). 10, 12.5 and 4–16% gradient gels, dried and bands were visualized using Monoclonal (IG6) and polyclonal (R929) antibodies to detect human MTM1 a Molecular Dynamics Storm Phosphorimager (Sunnyvale, CA, USA). were as described (54). Immunofluorescence and immunoprecipitation of Quantification of coimmunoprecipitated proteins was performed using MTM1-FLAG constructs were accomplished using the M2 clone of the MOLECULAR DYNAMICS IMAGEQUANT software. All values represent the ratio monoclonal anti-FLAG antibody or a polyclonal rabbit anti-FLAG antibody of coimmunoprecipitated protein as a fraction of the specifically immuno- purchased from Sigma (St Louis, MO, USA). A mouse mAb to the V5- precipitated protein and average ratios and error bars denoting the standard epitope tag was purchased from Invitrogen Life Technologies (Carlsbad, error of the mean were generated using GRAPHPAD PRISM software (San CA, USA). Primary antibodies were detected using fluorescein isothiocya- Diego, CA, USA). nate (FITC)-, Rhodamine- or Cy-5 (bis-OSU, N,N’-biscarboxypentyl-5,5’- disulfonatoindodicarbocyanine)-conjugated donkey secondary antibodies Immunoprecipitation and Western blot purchased from Jackson Immunoresearch Laboratories (West Grove, Cells were plated in 60- or 100-mm dishes and infected/transfected as PA, USA). described above. After 6 h transfection, cells were washed once with PBS, placed on ice and lysed in ice-cold RIPA buffer containing protease Vaccinia virus infection and transient transfection of inhibitors. The lysate was clarified by centrifugation for 10 minutes at BHK cells 20 817 g in an Eppendorf microfuge at 48C. The mAb directed against the BHK21 cells were infected with the vaccinia virus (vTF7.3) and then FLAG epitope (Clone M2), a rabbit anti-GFP antibody, a monoclonal anti- transfected with plasmids containing complementary DNA under the Rab5, a polyclonal anti-hVPS34 antibody or a rabbit anti-Rab7 antibody control of the T7 promoter as previously described (55). After 5–6 h of and protein A–Sepharose were individually added to supernatants to transfection, cells were washed with PBS and collected for immunofluo- immunoprecipitate protein complexes. Immunoprecipitated samples rescence or biochemical analysis. Cells were incubated in DMEM medium were washed repeatedly with RIPA buffer and solubilized in SDS–PAGE without cysteine or methionine for 30 minutes, then radiolabeled in the sample buffer for analysis. Proteins were resolved on 10% SDS–PAGE same medium containing 100 mCi/mL Trans 35S-label (ICN Biomedicals Inc., and transferred to nitrocellulose. Filters were blocked with 5% milk in Costa Mesa, CA, USA) for 30 minutes and analyzed. Tris-buffered saline (TBS) prior to incubation with the primary antibody. After extensive washing in TBS containing 1% Tween-20 (TBST), blots were incubated with horseradish peroxidase-conjugated anti-rabbit or Immunofluorescence microscopy anti-mouse antibody in milk-TBST, and bands were visualized using the As observed in previous reports (15,16), visualization of membrane-bound Pierce SuperSignal chemiluminescent detection kit according to the manu- MTM1 in cells overexpressing MTM1 that were fixed prior to permeabiliza- facturer’s instructions (Pierce Biotechnology Inc. Rockford, IL, USA). tion was not possible because of a significant cytosolic distribution (data not shown). Therefore, we used permeabilization with low levels of saponin GST–Rab7 and GST–MTM1 fusion protein expression prior to fixation of cells, which is known to preserve ultrstructural and purification morphology and allow functional studies yet sufficiently extracts cytosolic Recombinant Rab7 and MTM1 were expressed as a GST fusion proteins in proteins to allow for optimal visualization of membrane-bound pools of Rab E. coli BL21 as described (41). Briefly, cells were grown to A600 of 0.5, GTPases, lipid kinases and phosphatases, among others (32,59–62). For chilled on ice for 20 minutes and then induced with 50 mM IPTG [isopropyl immunofluorescence microscopy, cells were permeabilized for 5 minutes b-D-1 thiogalactopyranoside overnight at room temperature (63). Bacteria in 0.05% saponin in 80 mM PIPES buffer then fixed in 3% paraformalde- were harvested and lysed in GST Lysis Buffer (25 mM Tris pH 7.4, 2 mM hyde. Care must be taken with saponin extraction because the membrane ethylenediaminetetraacetic acid, 137 mM NaCl, 2.6 mM KCl, 1 mM dithio- pool of MTM1 may also be extracted by increased time or concentration of threitol, 1 mM phenylmethylsulfonyl fluoride, 10 mg/mL each chymostatin, saponin treatment. Fixed samples were quenched in 50 mM NH4Cl in PBS leupeptin, antipain, pepstatin) by sonication. Samples were Triton-X-100 for 10 minutes, then treated with 0.1% Triton-X-100 in PBS containing solubilized by addition of Triton-X-100 to a final concentration of 1% (v/v) 0.2% gelatin, 0.9 mM CaCl2 and 1 mM MgCl2 for 5 minutes. Cells were and incubation for 30 minutes at 48C while rotating. The GST fusion incubated with primary antibodies diluted in PBS containing 0.2% gelatin, proteins were affinity purified using Glutathione–Sepharose 4B (Amersham 0.9 mM CaCl2 and 1 mM MgCl2 for 30 minutes and primary antibodies were Biosciences). Protein concentration was measured using BioRad DC Pro- detected with fluorophore-conjugated secondary antibodies for 30 minutes. tein Assay (BioRad Laboratories Inc., Hercules, CA, USA) according to Endogenous MTM1 was detected using the Tyramide signal amplification manufacturer’s instruction. fluorescence system according to the manufacturer’s protocols (PerkinElmer, Boston, MA, USA). Coverslips were mounted with Mowiol (41), and hVPS15-binding assay samples were viewed on a Zeiss LSM 510 confocal microscope unless Full-length (hVPS15wt) deletion mutants or isolated hVPS15 domains (38) otherwise noted. All images were exported as tiff files and compiled in were synthesized in vitro by TNT Quick Coupled Transcription/Translation ADOBE PHOTOSHOP. Quantification of immunofluorescence images was Systems (Promega, Madison, WI, USA) as previously described (41). In vitro performed using SLIDEBOOK software (Intelligent Imaging Innovations Inc., translated hVPS15 samples were added to equimolar GST (as a control), GTP- Denver, CO, USA). Colocalization was determined by measuring the bound GST–Rab7 or GST–MTM1 immobilized on Glutathione–Sepharose

1064 Traffic 2007; 8: 1052–1067 MTM1 Binds hVPS15/hVPS34 PI 3-Kinase beads and incubated at 48C for 2 h. Beads were washed with GST Lysis Charcot-Marie-Tooth type 4B is caused by mutations in the gene Buffer, boiled in SDS/PAGE sample buffer and proteins were resolved by encoding myotubularin-related protein-2. Nat Genet 2000;25:17–19. SDS/PAGE. Bound proteins were quantified using a Molecular Dynamics 4. Azzedine H, Bolino A, Taieb T, Birouk N, Di Duca M, Bouhouche A, Phosphorimager. Benamou S, Mrabet A, Hammadouche T, Chkili T, Gouider R, Ravazzolo R, Brice A, Laporte J, LeGuern E. Mutations in MTMR13, a new pseudophosphatase homologue of MTMR2 and Sbf1, in two Acknowledgments families with an autosomal recessive demyelinating form of Charcot- Marie-Tooth disease associated with early-onset glaucoma. Am J Hum Genet 2003;72:1141–1153. We thank Dr Rebecca Lee and Ms Genevieve Phillips for expert technical 5. Laporte J, Bedez F, Bolino A, Mandel JL. Myotubularins, a large support and equipment maintenance that allowed us to acquire the disease-associated family of cooperating catalytically active and inactive images presented in this study. We also gratefully acknowledge Melanie phosphoinositides phosphatases. Hum Mol Genet 2003;12:R285–R292. Lenhart and Elsa G. Romero for their expert technical assistance and all 6. Wishart MJ, Dixon JE. PTEN and myotubularin phosphatases: from those who kindly provided plasmids and antibodies for this project. This 3-phosphoinositide dephosphorylation to disease. Trends Cell Biol work was generously supported by the National Science Foundation 2002;12:579–585. under grant numbers MCB9982161 and MCB MCB0446179 to A. W. N. 7. Kim SA, Vacratsis PO, Firestein R, Cleary ML, Dixon JE. Regulation of This work was also supported by grants from the American Diabetes myotubularin-related (MTMR)2 phosphatidylinositol phosphatase by Association and National Institute of Diabetes and Digestive Diseases MTMR5, a catalytically inactive phosphatase. Proc Natl Acad Sci U S A (DK070679) to J. M. B. and by INSERM, CNRS, Colle` ge de France and 2003;100:4492–4497. grants from the Association Francxaise contre les Myopathies to J. L. 8. Mochizuki Y, Majerus PW. Characterization of myotubularin-related Partial salary support was provided to M. P. S. through a grant from the protein 7 and its binding partner, myotubularin-related protein 9. Proc University of New Mexico (UNM) Cancer Center while a postdoctoral Natl Acad Sci U S A 2003;100:9768–9773. fellow. Images in this article were generated in the UNM Fluorescence 9. Nandurkar HH, Layton M, Laporte J, Selan C, Corcoran L, Caldwell KK, Microscopy Facility, which received extramural support from National Mochizuki Y, Majerus PW, Mitchell CA. Identification of myotubularin as Center for Research Resources (P20RR11830, S10RR14668 and the lipid phosphatase catalytic subunit associated with the 3-phosphatase S10RR016918), NEM-sensitive factor (MCB9982161), National Cancer adapter protein, 3-PAP. Proc Natl Acad Sci U S A 2003;100:8660–8665. Institute (R24 CA88339) and intramural funding from the University of 10. Cui X, De Vivo I, Slany R, Miyamoto A, Firestein R, Cleary ML. New Mexico Health Sciences Center and the University of New Mexico Association of SET domain and myotubularin-related proteins modu- Cancer Center. lates growth control. Nat Genet 1998;18:331–337. 11. Berger P, Schaffitzel C, Berger I, Ban N, Suter U. Membrane associ- Supplementary Material ation of myotubularin-related protein 2 is mediated by a pleckstrin homology-GRAM domain and a coiled-coil dimerization module. Proc Natl Acad Sci U S A 2003;100:12177–12182. Figure S1: MTM1 and Rab5Q79L are in discrete domains on exten- 12. Robinson FL, Dixon JE. The phosphoinositide-3-phosphatase MTMR2 sively dilated endosomes. The BHK cells were transfected with associates with MTMR13, a membrane-associated pseudophospha- MTM1C375S-FLAG and Rab5Q79L. Cells were permeabilized with sapo- tase also mutated in type 4B Charcot-Marie-Tooth disease. J Biol Chem nin, fixed and stained as described in Materials and Methods. The 2005;280:31699–31707. MTM1C375S-FLAG was detected with a mAb directed against the FLAG 13. Tsujita K, Itoh T, Ijuin T, Yamamoto A, Shisheva A, Laporte J, Takenawa epitope and a FITC-conjugated donkey anti-mouse IgG. The EEA1 was T. Myotubularin regulates the function of the late endosome through stained with human anti-EEA1 and a Texas Red-conjugated donkey anti- the GRAM domain-phosphatidylinositol 3,5-bisphosphate interaction. human IgG. Rab5 was localized using a rabbit polyclonal antibody directed J Biol Chem 2004;279:13817–13824. against Rab5 and Cy5-conjugated donkey anti-rabbit IgG. Insets are 14. Laporte J, Blondeau F, Buj-Bello A, Tentler D, Kretz C, Dahl N, enlarged 2.2 times. Bar, 5 mm. Mandel JL. Characterization of the myotubularin dual specificity phosphatase gene family from yeast to human. Hum Mol Genet 1998;7: Supplemental materials are available as part of the online article at http:// 1703–1712. www.blackwell-synergy.com 15. Blondeau F, Laporte J, Bodin S, Superti-Furga G, Payrastre B, Mandel JL. Myotubularin, a phosphatase deficient in myotubular myopathy, acts on phosphatidylinositol 3-kinase and phosphatidylinositol 3-phosphate pathway. Hum Mol Genet 2000;9:2223–2229. References 16. Taylor GS, Maehama T, Dixon JE. Myotubularin, a protein tyrosine phosphatase mutated in myotubular myopahty, dephosphorylates the 1. Laporte J, Guiraud-Chaumeil C, Vincent MC, Mandel JL, Tanner SM, lipid second messenger, phosphatidylinositol 3-phosphate. Proc Natl Liechti-Gallati S, Wallgren-Pettersson C, Dahl N, Kress W, Bolhuis PA, Acad Sci U S A 2000;97:8910–8915. Fardeau M, Samson F, Bertini E. Mutations in the MTM1 gene 17. Walker DM, Urbe S, Dove SK, Tenza D, Raposo G, Clague MJ. implicated in X-linked myotubular myopathy. ENMC International Characterization of MTMR3. An inositol lipid 3-phosphatase with novel Consortium on Myotubular Myopathy. European Neuro-Muscular substrate specificity. Curr Biol 2001;11:1600–1605. Center. Hum Mol Genet 1997;6:1505–1511. 18. Berger P, Bonneick S, Willi S, Wymann M, Suter U. Loss of phosphatase 2. Dang H, Li Z, Skolnik EY, Fares H. Disease-related myotubularins activity in myotubularin-related protein 2 is associated with Charcot- function in endocytic traffic in Caenorhabditis elegans. Mol Biol Cell Marie-Tooth disease type 4B1. Hum Mol Genet 2002;11:1569–1579. 2004;15:189–196. 19. Tronchere H, Laporte J, Pendaries C, Chaussade C, Liaubet L, Pirola L, 3. Bolino A, Muglia M, Conforti FL, LeGuern E, Salih MA, Georgiou DM, Mandel JL, Payrastre B. Production of phosphatidylinositol 5-phos- Christodoulou K, Hausmanowa-Petrusewicz I, Mandich P, Schenone phate by the phosphoinositide 3-phosphatase myotubularin in mam- A, Gambardella A, Bono F, Quattrone A, Devoto M, Monaco AP. malian cells. J Biol Chem 2004;279:7304–7312.

Traffic 2007; 8: 1052–1067 1065 Cao et al.

20. Gillooly DJ, Morrow IC, Linsay M, Gould R, Bryant NJ, Gaullier J-M, of Rab5 effectors regulates phosphoinositide turnover in the endocytic Parton RG, Stenmark H. Localization of phosphatidylinositol 3-phos- pathway. J Cell Biol 2005;170:607–618. phate in yeast and mammalian cells. EMBO J 2000;19:4577–4588. 41. Stein M-P, Cao C, Tessema M, Feng Y, Romero E, Welford A, 21. Cooke FT. Phosphatidylinositol 3,5-bisphosphate: metabolism and Wandinger-Ness A. Interaction and functional analyses of human function. Arch Biochem Biophys 2002;407:143–151. VPS34/p150 phosphatidylinositol 3-kinase complex with Rab7. In: 22. Gaullier JM, Simonsen A, D’Arrigo A, Bremnes B, Stenmark H, Aasland Balch W, Hall A, editors. Methods in Enzymology. New York: Elsevier R. FYVE fingers bind PtdIns(3)P. Nature 1998;394:432–433. Inc.; 2005, pp. 628–649. 23. Stenmark H, Aasland R, Toh BH, D’Arrigo A. Endosomal localization of 42. Panaretou C, Domin J, Cockcroft S, Waterfield MD. Characterization of the autoantigen EEA1 is mediated by a zinc-binding FYVE finger. J Biol p150, an adapter protein for the human phosphatidylinositol (PtdIns) Chem 1996;271:24048–24054. 3-kinase. Substrate presentation by phosphatidylinositol transfer pro- 24. Stenmark H, Aasland R. FYVE-finger proteins-effectors of an inositol tein to the p150.Ptdins 3-kinase complex. J Biol Chem 1997;272: lipid. J Cell Sci 1999;112:4175–4183. 2477–2485. 25. Misra S, Miller GJ, Hurley JH. Recognizing phosphatidylinositol 3-phos- 43. Chaussade C, Pirola L, Bonnafous S, Blondeau F, Brenz-Verca S, phate. Cell 2001;107:559–562. Tronchere H, Portis F, Rusconi S, Payrastre B, Laporte J, Van Obber- 26. Bache KG, Brech A, Mehlum A, Stenmark H. Hrs regulates multi- ghen E. Expression of myotubularin by an adenoviral vector demon- vesicular body formation via ESCRT recruitment to endosomes. J Cell strates its function as a phosphatidylinositol 3-phosphate [PtdIns(3)P] Biol 2003;162:435–442. phosphatase in muscle cell lines: involvement of PtdIns(3)P in insulin- 27. Clague MJ, Urbe S. Hrs function: viruses provide the clue. Trends Cell stimulated glucose transport. Mol Endocrinol 2003;17:2448–2460. Biol 2003;13:603–606. 44. Willingham MC. An alternative fixation-processing method for preem- 28. Futter CE, Collinson LM, Backer JM., Hopkins CR. Human VPS34 is bedding ultrastructural immunocytochemistry of cytoplasmic antigens: required for internal vesicle formation within multivesicular endo- the GBS (glutaraldehyde-borohydride-saponin) procedure. J Histochem somes. J Cell Biol 2001;155:1251–1264. Cytochem 1983;31:791–798. 29. Petiot A, Faure J, Stenmark H, Gruenberg J. PI3P signaling regulates 45. Goldenthal KL, Hedman K, Chen JW, August JT, Willingham MC. receptor sorting but not transport in the endosomal pathway. J Cell Biol Postfixation detergent treatment for immunofluorescence suppresses 2003;162:971–979. localization of some integral membrane proteins. J Histochem Cyto- 30. Volinia S, Dhand R, Vanhaesebroeck B, MacDougall LK, Stein R, chem 1985;33:813–820. Zvelebil MJ, Domin J, Panaretou C, Waterfield MD. A human phos- 46. Patki V, Virbasius J, Lane WS, Toh BH, Shpetner HS, Corvera S. phatidylinositol 3-kinase complex related to the yeast Vps34p-Vps15p Identification of an early endosomal protein regulated by phosphatidy- protein sorting system. EMBO J 1995;14:3339–3348. linositol 3-kinase. Proc Natl Acad Sci U S A 1997;94:7326–7330. 31. Christoforidis S, Miaczynska M, Ashma K, Wilm M, Zhao L, Yip S-C, 47. Schaletzky J, Dove SK, Short B, Lorenzo O, Clague MJ, Barr FA. Waterfield MD, Backer JM, Zerial M. Phosphatidylinositol-3-OH Phosphatidylinositol-5-phosphate activation and conserved substrate kinases are Rab5 effectors. Nat Cell Biol 1999;1:249–252. specificity of the myotubularin phosphatidylinositol 3-phosphatases. 32. Stein MP, Feng Y, Cooper KL, Welford AM, Wandinger-Ness A. Curr Biol 2003;13:504–509. Human VPS34 and p150 are Rab7 interacting partners. Traffic 2003; 48. Sonnichsen B, De Renzis S, Nielsen E, Rietdork J, Zerial M. Distinct 4:754–771. membrane domains on endosomes in the recycling pathway visualized 33. Siddhanta U, McIlroy J, Shah A, Zhang Y, Backer JM. Distinct roles for by multicolor imaging of rab4, rab5, and rab11. J Cell Biol 2000;149: the p110alpha and hVPS34 phosphatidylinositol 3’-kinases in vesicular 901–913. trafficking, regulation of the actin cytoskeleton, and mitogenesis. J Cell 49. Feng Y, Press B, Chen W, Zimmerman J, Wandinger-Ness A. Biol 1998;143:1647–1659. Expression and properties of Rab7 in endosome function. In: Hall A, 34. Ikonomov OC, Sbrissa D, Shisheva A. Mammalian cell morphology and editor. Methods in Enzymology. San Diego: Academic Press; 2001, pp. endocytic membrane homeostasis require enzymatically active phos- 175–187. phoinositide 5-kinase PIKfyve. J Biol Chem 2001;276:26141–26147. 50. Press B, Feng Y, Hoflack B, Wandinger-Ness A. Mutant Rab7 causes 35. Gary JD, Wurmser AE, Bonangelino CJ, Weisman LS, Emr SD. Fab1p is the accumulation of cathepsin D and cation-independent mannose essential for PtdIns(3)P 5-kinase activity and the maintenance of 6-phosphate receptor in an early endocytic compartment. J Cell Biol vacuolar size and membrane homeostasis. J Cell Biol 1998;143:65–79. 1998;140:1075–1089. 36. Wurmser AE, Emr SD. Phosphoinositide signaling and turnover: 51. Vonderheit A, Helenius A. Rab7 associates with early endosomes to PtdIns(3)P, a regulator of membrane traffic, is transported to the mediate sorting and transport of Semliki forest virus to late endo- vacuole and degraded by a process that requires lumenal vacuolar somes. PLoS Biol 2005;3:e233. hydrolase activities. EMBO J 1998;17:4930–4942. 52. Rink J, Ghigo E, Kalaidzidis Y, Zerial M. Rab conversion as a mecha- 37. Ikonomov OC, Sbrissa D, Foti M, Carpentier JL, Shisheva A. PIKfyve nism of progression from early to late endosomes. Cell 2005;122: controls fluid phase endocytosis but not recycling/degradation of 735–749. endocytosed receptors or sorting of procathepsin D by regulating 53. Kim SA, Taylor GS, Torgersen KM, Dixon JE. Myotubularin and multivesicular body morphogenesis. Mol Biol Cell 2003;14:4581–4591. MTMR2, phosphatidylinositol 3-phosphatases mutated in myotubular 38. Murray JT, Panaretou C, Stenmark H, Miaczynska M, Backer JM. Role myopathy and type 4B Charcot-Marie-Tooth disease. J Biol Chem of Rab5 in the recruitment of hVps34/p150 to the early endosome. 2002;277:4526–4531. Traffic 2002;3:416–427. 54. Laporte J, Blondeau F, Gansmuller A, Lutz Y, Vonesch JL, Mandel JL. 39. Parrish WR, Stefan CJ, Emr SD. Essential role for the myotubularin- The PtdIns3P phosphatase myotubularin is a cytoplasmic protein that related phosphatase Ymr1p and the synaptojanin-like phosphatases also localizes to Rac1-inducible plasma membrane ruffles. J Cell Sci Sjl2p and Sjl3p in regulation of phosphatidylinositol 3-phosphate in 2002;115:3105–3117. yeast. Mol Biol Cell 2004;15:3567–3579. 55. Feng Y, Press B, Wandinger-Ness A. Rab 7: an important regulator of 40. Shin HW, Hayashi M, Christoforidis S, Lacas-Gervais S, Hoepfner S, late endocytic membrane traffic. J Cell Biol 1995;131:1435–1452. Wenk MR, Modregger J, Uttenweiler-Joseph S, Wilm M, Nystuen A, 56. Bucci C, Wandinger-Ness A, Lutcke A, Chiariello M, Bruni CB, Zerial M. Frankel WN, Solimena M, De Camilli P, Zerial M. An enzymatic cascade Rab5a is a common component of the apical and basolateral endocytic

1066 Traffic 2007; 8: 1052–1067 MTM1 Binds hVPS15/hVPS34 PI 3-Kinase

machinery in polarized epithelial cells. Proc Natl Acad Sci U S A 61. Estrada-Garcia T, Craxton A, Kirk CJ, Michell RH. A salt-activated 1994;91:5061–5065. inositol 1,3,4,5-tetrakisphosphate 3-phosphatase at the inner sur- 57. Qiu Y, Xu X, Wandinger-Ness A, Dalke DP, Pierce SK. Separation of face of the human erythrocyte membrane. Proc Biol Sci 1991;244: subcellular compartments containing distinct functional forms of MHC 63–68. class II. J Cell Biol 1994;125:595–605. 62. Ohsaki Y, Maeda T, Fujimoto T. Fixation and permeabilization protocol 58. Waite RL, Sentry JW, Stenmark H, Toh BH. Autoantibodies to a novel is critical for the immunolabeling of lipid droplet proteins. Histochem early endosome antigen 1. Clin Immunol Immunopathol 1998;86: Cell Biol 2005;124:445–452. 81–87. 63. Taylor GS, Liu Y, Baskerville C, Charbonneau H. The activity of Cdc14p, 59. Willingham MC. Electron microscopic immunocytochemical localiza- an oligomeric dual specificity protein phosphatase from Saccharomy- tion of intracellular antigens in cultured cells: the EGS and ferritin bridge ces cerevisiae, is required for cell cycle progression. J Biol Chem procedures. Histochem J 1980;12:419–434. 1997;272:24054–24063. 60. Chavrier P, Parton RG, Hauri HP, Simons K, Zerial M. Localization of low 64. Stein M-P, Dong J, Wandinger-Ness A. Rab proteins and endocytic molecular weight GTP binding proteins to exocytic and endocytic trafficking: potential targets for therapeutic intervention. Advanced compartments. Cell 1990;62:317–329. Drug Delivery Reviews 2003;55:1421–1437.

Traffic 2007; 8: 1052–1067 1067