Involvement of a membrane-bound form of glutamate dehydrogenase in the association of lysosomes to microtubules.

A 50-kDa membrane protein corresponding to a membrane-bound isoform of glutamate dehydrogenase was proposed as a molecular species that could mediate lysosome-microtubule interactions. This protein, isolated from purified lysosome membranes, is a peripheral membrane protein with an ATP-dependent microtubule binding activity. We have produced antibodies against the purified 50-kDa protein to investigate its role in the association of lysosomes to microtubules using a cell-free reconstitution assay and cell microinjection. Pretreatment of purified lysosomes with the antibodies inhibited the association of these vesicles to microtubules. The blocking effect of antibodies was demonstrated by a differential sedimentation method and negative staining electron microscopy, allowing us to quantify the amount of microtubules interacting with lysosomes and the proportion of lysosomes bound to microtubules, respectively. Affinity-purified antibodies microinjected into intact cells altered the distribution of lysosomes that appeared less clustered in the vicinity of nuclei. The antibody-induced lysosome dispersion was assessed by quantitative videomicroscope analyses. These data show that the 50-kDa membrane protein could act, through its microtubule binding activity, as a molecule of attachment of lysosomes to microtubules. This membrane-bound isoform of glutamate dehydrogenase could be involved in the microtubule-dependent perinuclear localization of lysosomes.

A 50-kDa membrane protein corresponding to a membrane-bound isoform of glutamate dehydrogenase was proposed as a molecular species that could mediate lysosome-microtubule interactions. This protein, isolated from purified lysosome membranes, is a peripheral membrane protein with an ATP-dependent microtubule binding activity. We have produced antibodies against the purified 50-kDa protein to investigate its role in the association of lysosomes to microtubules using a cellfree reconstitution assay and cell microinjection. Pretreatment of purified lysosomes with the antibodies inhibited the association of these vesicles to microtubules. The blocking effect of antibodies was demonstrated by a differential sedimentation method and negative staining electron microscopy, allowing us to quantify the amount of microtubules interacting with lysosomes and the proportion of lysosomes bound to microtubules, respectively. Affinity-purified antibodies microinjected into intact cells altered the distribution of lysosomes that appeared less clustered in the vicinity of nuclei. The antibody-induced lysosome dispersion was assessed by quantitative videomicroscope analyses. These data show that the 50-kDa membrane protein could act, through its microtubule binding activity, as a molecule of attachment of lysosomes to microtubules. This membrane-bound isoform of glutamate dehydrogenase could be involved in the microtubule-dependent perinuclear localization of lysosomes.
It is generally accepted that the distribution and the intracellular motion of organelles, among which are endosomes and lysosomes, require the integrity of the microtubule network. In polarized cells, the movement of endosomal vesicles from the cell periphery to the perinuclear region of the cell is abolished after a treatment with microtubule inhibitors (1)(2)(3). Thus, the endocytic vesicles move centripetally in a microtubule-dependent manner in order to deliver their content to lysosomes (4,5).
The key microtubule-dependent step could be the transfer of material via carrier vesicles into late endosomes (6 -8). Microtubules play also a key role in the clustering of late endosomes or prelysosomes and lysosomes in the perinuclear region of the cell (1, 9 -12). In most cells, prelysosomes and lysosomes accumulate around the nucleus, often in a region near the microtubule-organizing center (MTOC). 1 The positioning of lyso-somes depends on their association to microtubules. When microtubules are depolymerized by a treatment of cells with nocodazole, lysosomes disperse into the cytoplasm. Upon removal of the microtubule disrupting agent, microtubules progressively reassemble from the MTOC, and lysosomes return to their normal perinuclear localization. When cells are treated with taxol, after removal of nocodazole, microtubules reassemble as short filaments over the entire cytoplasm, without relation to the MTOC, and lysosomes remain scattered throughout the cell (10). Microtubules are also required for the formation of tubular extensions from endosomes and lysosomes (9,13,14). According to a previous tentative model (15), interaction of lysosomes and prelysosomes with microtubules could bring into play two types of microtubule-binding proteins, organelle linker and cytoplasmic motor proteins. The microtubule-based motor proteins, kinesin (anterograde motor) and cytoplasmic dynein (retrograde motor), are involved in the endosomal vesicles or lysosomes movement along microtubules (16 -18). In polarized Madin-Darby canine kidney cells, the microtubuledependent fusion of either apical or basolateral endosomes requires kinesin and dynein (7). Kinesin is also involved in the formation of tubular extensions of lysosomes in macrophages (19,20). Lysosome/endosome-microtubule interactions mediated by kinesin (19,21) or by cytoplasmic dynein (22) have been well documented. In addition to cytoplasmic motors, molecules acting as a receptor, kinectin (23) and accessory factors such as dynactin (24,25), could be part of the organelle-motor complex.
Proteins mediating "stable" interactions between endosomes or lysosomes and microtubules have been identified in cell-free reconstitution assays (26,27). Proteins of the vesicle membrane could play the role of anchoring factors of the vesicles onto microtubules. The binding of endocytic carrier vesicles to microtubules depends on a microtubule-binding protein, CLIP-170 (28), which is not a motor protein (27). CLIP-170, colocalized in vivo with endocytic vesicles, could be involved in the capture of early endosomes by microtubules (29). The ATPsensitive binding of this protein to microtubules is regulated by phosphorylation (30). Another protein, MP50 (membrane protein of 50 kDa), was proposed as a factor of anchoring of lysosomes to microtubules (31). MP50, isolated from highly purified preparations of prelysosomes and lysosomes (32,33), exhibits an ATP-dependent microtubule binding activity. This protein has been identified to a membrane-bound isoform of a ubiquitous enzyme, glutamate dehydrogenase (GDH) (34). MP50 is a peripheral membrane protein with the same isoelectric point as the isoprotein 2 of GDH. The present study aims at * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  demonstrating that MP50 present on the lysosome/prelysosome membrane is involved in the association of lysosomes and prelysosomes to microtubules and that MP50 could contribute to the maintenance of these organelles in the perinuclear region of the cells. We have raised polyclonal antibodies against MP50 and used them to block the activity of potential microtubule binding sites on lysosome membranes. Studies have been conducted in a cell-free reconstitution assay (26) and in intact cells using cell microinjection of affinity-purified antibodies. We also raised anti-GDH antibodies to differentiate the properties and/or functions of the membrane-bound form of GDH, i.e. MP50, from those of soluble GDH.

EXPERIMENTAL PROCEDURES
Purification and Labeling of MP50 -MP50 was purified from pig liver vesicle membranes according to the method described by Mithieux and Rousset (31), slightly modified by Rajas and Rousset (34). Purified liver MP50 and purified GDH from bovine liver (Sigma) were labeled with Na 125 I using lactoperoxidase and glucose-glucose oxidase as the H 2 O 2 generating system, as described previously (34). The specific radioactivity of the labeled proteins were 2-5 Ci/g of protein.
Preparation and Purification of Antibodies-Two polyclonal anti-MP50 immune sera (Ab1 and Ab2) and one polyclonal anti-GDH immune serum (Ab3) were produced from New Zealand White male rabbits by multipoint subcutaneous injections of 500 g of liver MP50 or purified bovine liver GDH (Sigma) in 1 ml of complete Freund's adjuvant. After 3 weeks, rabbits were given a booster injection. Blood was collected every 48 h, starting 5 days after the second injection, and sera were stored at Ϫ20°C. The IgG fractions were purified by precipitation with ammonium sulfate and anion-exchange chromatography on DEAE-cellulose. Sera and IgG fractions were tested for their antibody titer and specificity by soluble phase radioimmunoassay. 125 I-Labeled protein (MP50 or GDH) (40,000 cpm/tube, about 5 ng) was incubated with the immune serum in the absence or in the presence of unlabeled protein (MP50 or GDH) (1-10,000 ng) in a total volume of 500 l of 10 mM phosphate, 0.15 M NaCl, PBS buffer containing 10 mg/ml BSA, pH 7.4. Incubation was carried out for 1 h at 37°C. Immune complexes were collected by centrifugation (1,500 ϫ g for 15 min) after a 45-min incubation with protein A adsorbent (Staphylococcus aureus, Pansorbin, Calbiochem) at room temperature. Immunoprecipitated 125 Iprotein was measured in a Packard Instruments ␥ counter.
Anti-MP50 IgG was purified by affinity chromatography on immobilized MP50. Purified MP50 (200 -250 g of protein/ml of packed gel) was coupled to Affi-Gel 10 (N-hydroxysuccinimide ester of a succinylated aminoalkyl-agarose gel from Bio-Rad) as indicated by the manufacturer. The coupling efficiency was 80 -90%. The IgG fraction purified from the immune sera was incubated with MP50/Affi-Gel 10 matrix (1 mg of IgG/ml of packed gel) for 4 h at 4°C under gentle agitation. The gel was then packed in a column, and the incubation solution was recycled 5 times. After extensive washing with PBS, bound IgG was eluted from the column with 0.1 M glycine, pH 2.5. Fractions were collected in 1 M Tris base (containing 20 mg/ml BSA) to obtain a neutral pH. Fractions containing anti-MP50 antibodies were pooled and stored at Ϫ20°C.
Purification and Labeling of Microtubule Proteins-Microtubule proteins were purified from pig brain by two cycles of temperature-dependent polymerization-depolymerization using the procedure of Shelanski et al. (37). Purification was carried out in buffer A (100 mM MES, 0.5 mM MgCl 2 , 1 mM EGTA, 0.1 mM GTP, pH 6.4) supplemented with 4 M glycerol and 1 mM GTP in polymerizing steps. Twice-cycled microtubule proteins were radioiodinated by conjugation with the 125 I-Bolton-Hunter reagent (Amersham Corp., Les Ulis, France) according to the method of Carlier et al. (38). Polymerization-competent labeled microtubule proteins were selected by an additional assembly-dissassembly cycle, as described previously (26). The specific radioactivity of the labeled protein was 0.1-0.2 Ci/g of microtubule protein. More than 95% of 125 I-labeled iodine was recovered on tubulin and equally distributed between ␣and ␤-subunits.
Purification of Lysosomes/Prelysosomes and Pretreatment with Antibodies-Lysosomes/prelysosomes were purified from pig thyroid glands by differential and isopycnic centrifugations on Percoll gradient (32,33) prepared in hypertonic sucrose medium (26). Purified lysosomes/prelysosomes (1.5 mg of protein) were incubated with immune sera or the corresponding purified IgG fractions or affinity-purified anti-MP50 antibodies in 4 ml of 10 mM Tris/HCl, 0.5 M sucrose, pH 7.4, for 1 h at 20°C. Normal rabbit serum or purified IgG (from preimmune sera) were used as controls. The mixtures were then centrifuged at 26,000 ϫ g for 15 min at 4°C. Lysosomes were washed by resuspending the pellets in the same buffer and were sedimented by centrifugation. The final lysosome/prelysosome pellets were resuspended in the reconstitution assay medium, i.e. buffer A, containing 0.5 M sucrose (buffer B) to obtain a final concentration of 1.2-1.4 mg of protein/ml.
Measurement of the Binding of Microtubules to Lysosomes/Prelysosomes-Radiolabeled microtubules were prepared by mixing unlabeled microtubule protein with 125 I-labeled tubulin in buffer B. The specific radioactivity of 125 I-labeled microtubule protein ranged from 100 to 150 cpm/g. After 30 min of incubation at 37°C in the presence of 1 mM GTP, microtubules were sheared by five passages through a 25 G needle of a syringe and kept for 5 min at 25°C. Aliquots of the lysosome/ prelysosome suspension (0.12 mg of protein) preequilibrated at 25°C for 5 min were mixed with preassembled labeled microtubules (0.2-1.2 mg of protein) and incubated for 30 min at 25°C. Experiments were carried out in a total volume of 200 l of buffer B. At the end of the incubation period, the lysosome-microtubule mixture was layered on a 100-l sucrose cushion (buffer A ϩ 0.75 M sucrose), placed in an Eppendorf tube, and centrifuged at 10,000 ϫ g for 5 min to separate free from lysosome-bound microtubules. The supernatant containing free microtubules and the cushion were removed, and the pellet (containing lysosomes and microtubule-lysosome complexes) was counted for radioactivity. The amount of microtubules recovered in the pellet was calculated from radioactivity measurements and the specific radioactivity of microtubule protein.
Negative Staining Electron Microscope Analyses of the Association of Microtubules to Lysosomes/Prelysosomes-Lysosomes/prelysosomes (0.05 mg of protein) were incubated with preassembled microtubules (0.2 mg) in a total volume of 0.5 ml of buffer B for 30 min at 25°C. Aliquots of 20 l of the reconstitution mixtures were laid as drops on Parafilm. Grids (400 mesh), previously coated with Formvar and carbon, were placed upside down on the drops for 1 min at room temperature. Subsequent treatments and washings were carried out by transferring grids from drop to drop of different solutions. Material adsorbed on the grids was first washed with buffer B and then fixed with 1% glutaraldehyde in buffer B for 1 min. After two washings in water (1 min each), grids were negatively stained for 1 min with 2% uranyl acetate in water and finally air-dried. Grids were examined in a Jeol 1200 EX electron microscope (Center Commun de Microscopie Electronique, Faculté de Médecine Lyon-RTH Laënnec, Lyon, France).
Cell Culture and Immunofluorescence Labeling-Thyroid cells, isolated from pig thyroid glands by a discontinuous trypsin treatment, were cultured in F-12 medium supplemented with 5% calf serum in Petri dishes under 95% air, 5% CO 2 atmosphere at 37°C (39). Cells cultured at a density of 0.5 ϫ 10 6 cells/cm 2 formed monolayers. After 3 days of culture, cells were fixed with 4% paraformaldehyde for 30 min and permeabilized with 1% Triton X-100 in PBS for 30 min at room temperature. Fixed cells were incubated overnight at 4°C with the first antibody. The antibodies used were: Ab1, Ab2, Ab3, polyclonal rabbit antibodies raised against soluble lysosomal proteins (SLP) (10) or against arylsulfatase A (Ars-A) (40,41), or a monoclonal anti-␤-tubulin (Amersham). Immune complexes were detected using either goat antimouse Ig antibodies conjugated to rhodamine or donkey anti-rabbit Ig antibodies conjugated to Texas Red (Amersham) for 30 min at room temperature. Observations were made using a Zeiss Axiophot fluorescence microscope. Fluorescence images, taken with a silicon intensified target camera (LHESA, Cergy Pontoise, France), were transferred to a crystal sapphire image processor (Quantel, Montigny-le-Bretonneux, France). Numerized images were stored in a Bernoulli box for subsequent analysis. Photomicrographs were obtained using a UP-5000 P video printer from Sony (Tokyo, Japan). To visualize mitochondria, living cells were incubated with 10 g/ml rhodamine 123 (Sigma) in F-12 medium. After 30 min at room temperature, cells were washed three times and observed with a Zeiss Axiophot fluorescence microscope equipped with the filter combination used for fluorescein.

Microinjection of Antibodies in Living Cells and Videomicroscope Analyses of the Intracellular Distribution of Lysosomes/Prelysosomes-
Affinity-purified anti-MP50 IgG or control IgG were dialyzed against a 5-fold diluted microinjection buffer (2 mM phosphate, 29 mM KCl, pH 7.0) overnight at 4°C and then concentrated 5 times using a SpeedVac concentrator. The microinjected solutions contained antibodies at a concentration of 5 mg/ml and BSA-Lucifer Yellow conjugate (Molecular Probes Inc., Eugene, OR) at a concentration of 5 mg/ml used to identify the injected cells. Before microinjection, the solutions were clarified by centrifugation at 10,000 ϫ g for 10 min. Microinjection was performed on an inverted microscope (Axiovert 35M from Zeiss) using a micromanipulator 5170 and a microinjector 5242 from Eppendorf (Hamburg, Germany). The procedure for microinjection in the cytoplasm was similar to that described by Munari-Silem et al. (39). Thyroid cells cultured as monolayer in a Petri dish (6 cm in diameter) were penetrated by the capillary in the vicinity of the nucleus. The microinjected volume (10 -50 fl) remained below 10% of the cell volume. The concentration of IgG in the cytoplasm should range from 1 to 6 M. Microinjections were made at room temperature within a time interval of 15-20 min in order to prevent changes in medium pH. After injection, the culture medium was changed, and cells were incubated at 20°C under 95% air, 5% CO 2 atmosphere for 4 h. At the end of the incubation period, cells were fixed and immunostained for lysosomes with anti-SLP antibodies (10) as mentioned above. Immune complexes were detected using donkey antirabbit Ig antibodies conjugated to Texas Red (Amersham). Fluorescence images, taken with the silicon intensified target camera installed on the Zeiss Axiophot microscope, were analyzed using the crystal sapphire image processor mentioned above and a built-in program for surface and luminance analyses. The parameters were obtained as follows. The outlines of the microinjected cell defined by Lucifer Yellow fluorescence were delimited on a screen by an optical pen device. The surface (number of pixels) of the delimited field, as well as the average luminance (fluorescence intensity) of the field, were automatically obtained. The surface in which Texas Red-labeled lysosomes were found and the mean luminance of this surface were determined in the same way. The ratio between the surface containing the lysosomes and the total surface of the cell was taken as a parameter of lysosome dispersion. This videomicroscope double fluorescence analysis was performed, in doubleblind, on two separate experiments. In each experiment, 40 -50 cells in 8 -10 different Petri dishes were microinjected with affinity-purified anti-MP50 or control IgGs. Fluorescence measurements were made on Lucifer Yellow-labeled cells taken at random. The values of surface and luminance were ordered according to the internal code of the experiment, and the corresponding means Ϯ S.E. were calculated for each group of cells. Statistical analysis was performed by the Student's t test.
SDS-PAGE and Western Blot-Proteins were separated by SDS-PAGE according to Laemmli (35) on 9% acrylamide slab minigels and then transferred onto polyvinylidene difluoride/Immobilon membrane. The membrane was preincubated in PBS containing 2% (w/v) fat-free dry milk, 0.2% (v/v) Tween 20 for 1 h at 20°C and then incubated with anti-MP50 or anti-GDH antibodies diluted in the PBS/milk/Tween buffer. Antigen-antibody complexes were detected using goat anti-rabbit Igs conjugated to alkaline phosphatase (Sigma) in the PBS/milk/ Tween buffer and bromochloroindolyl phosphate/nitro blue tetrazolium as substrate (36).

RESULTS
Characterization of Antibodies-Two rabbit immune sera raised against purified MP50 (Ab1 and Ab2) and one immune serum raised against purified GDH (Ab3) were tested for their antibody titer in soluble phase radioimmunoassay using 125 Ilabeled MP50 or 125 I-labeled GDH as antigen. The antibody titer (defined as the highest dilution of the immune serum giving a signal significantly higher than that of the preimmune serum) was 1:3,000, 1:10,000, and 1:30,000 for Ab1, Ab2, and Ab3, respectively (Fig. 1, A and B). The relative affinity of these antibodies for MP50 and GDH was studied by competitive binding experiments. The binding of 125 I-labeled MP50 to Ab1 and Ab2 was decreased by increasing amounts of unlabeled MP50 (Fig. 1C). The apparent affinity of Ab2 antibodies was about 5-fold higher than that of the Ab1 antibodies. The binding of 125 I-labeled MP50 to either Ab1 or Ab2 antibodies was also inhibited by unlabeled GDH. The reactivities of Ab1 and competition curves between 125 I-labeled GDH and unlabeled MP50 or GDH for binding to Ab3 antibodies. Labeled MP50 or GDH were incubated with Ab1 (E, q, respectively) or Ab2 (छ, ࡗ, respectivly) immune serum at a 1:300 dilution or with Ab3 immune serum (Ç, å, respectively) at a 1:3,000 dilution, in the presence of increasing concentrations of unlabeled purified MP50 (closed symbols) or purified liver GDH (open symbols). Immune complexes were collected using protein A adsorbent. The amount of labeled antigen, B, immunoprecipitated in the presence of different amounts of competitor is expressed as the percentage of the maximum amount of the labeled antigen, Bo, immunoprecipitated in the absence of competitor. Each symbol represents the mean of triplicate values.
Ab2 antibodies toward GDH and MP50 were almost the same. Binding of 125 I-labeled GDH to Ab3 antibodies was affected differently by GDH and MP50 (Fig. 1D). The competition curves were not parallel; this indicates that GDH and MP50 present distinct epitopes. Fifty-fold more MP50 than GDH was required to decrease the binding of 125 I-labeled GDH to Ab3 antibodies by 50%. Fig. 2B shows that Ab1 antibodies reacted with both MP50 and GDH on a Western blot and only labeled a 50-kDa band when tested on thyroid lysosome/prelysosome membrane preparations. This band corresponds to the MP50 initially identified, purified, and characterized from these purified organelles ( Fig. 2A). Ab2 antibodies were unable to detect any molecular species on the Western blot (data not shown), and thus, the specificity of these antibodies could not be documented. Ab3 antibodies reacted with both GDH and MP50, and as observed with Ab1 antibodies, Ab3 antibodies recognized a single band of 50 kDa in thyroid lysosome/prelysosome membrane preparations (Fig. 2C).
Validation of the Cell-free Reconstitution Assay-Lysosomes/ prelysosomes used in this study were purified 50 -75-fold (as compared with tissue homogenate) on the basis of acid phosphatase activity measurements. Purified organelles with a buoyant density from 1.08 to about 1.12 g/ml and higher than 1.12 g/ml correspond to secondary lysosomes or prelysosomes and lysosomes, respectively; their biochemical and morphological characteristics have been previously described (32,33,41). Prelysosomes have been defined as organelles positive for cation-independent mannose 6-phosphate receptor and arylsulfatase-A (Ars-A), and lysosomes have been identified as mannose 6-phosphate receptor-negative but Ars-A-positive organelles (41).
The cell-free reconstitution assay of the microtubule-lysosome association was conducted in the presence of 0.5 M sucrose as microtubule stabilizer. Microtubule protein concentrations varied from 1.0 to 6.0 mg/ml; in these conditions, the assembly state of preformed microtubules remained constant for the duration of the assay. Depending on the microtubule protein concentration, the proportion of assembled tubulin in the assay conditions varied from 55% to about 70%. Preassembled microtubules were sheared (by passages through the needle of a syringe) to decrease the possible formation of bundles or networks that could nonspecifically entrap organelles.
The formation of microtubule-lysosome complexes was monitored by measuring the amount of 125 I-labeled microtubules that sedimented at 10,000 ϫ g in the presence of lysosomes. A centrifugation at 10,000 ϫ g allowed 90 -100% of the lysosomes to pellet. The use of a sucrose cushion allowed obtaining a clear separation of free microtubules from lysosome-bound microtubules. The groundings of the assay are reported in Fig. 3A. In the presence of lysosomes, up to 70 -80% of 125 I-labeled microtubules introduced into the incubation mixture sedimented at 10,000 ϫ g. The lack of labeled microtubules in the 10,000 ϫ g pellet, when microtubules were incubated in the absence of lysosomes, documents the appropriateness of the sedimentation procedure as evidence of the microtubule-lysosome association. The absence of labeled tubulin in the 10,000 ϫ g pellet, when lysosomes were incubated with non-assembled tubulin, shows the specificity of the interaction of lysosomes with microtubules. More detailed analyses of the assay conditions and the formation of microtubule-lysosome complexes have been previously reported (15,26).
Antibodies That Specifically React with MP50 Prevent the Association of Lysosomes/Prelysosomes to Preassembled Microtubules-Lysosomes/prelysosomes preincubated with Ab1 antibodies had a reduced capacity to interact with microtubules (Fig.3B).TheblockingeffectofAb1antibodieswasconcentrationdependent; at a 1:10 dilution, Ab1 immune serum decreased the formation of lysosome-microtubule complexes by more than 75%. Purified IgG from Ab1 exhibited the same blocking effect (Fig. 3C). The preincubation of lysosomes with NRS or the corresponding IgG fraction did not influence the capacity of lysosomes to interact with microtubules. The blocking effect of Ab1 antibodies was only related to the interaction of antibodies with lysosomes during the preincubation period since lysosomes/prelysosomes were extensively washed before their introduction in the reconstitution mixture. It was verified that anti-MP50 Ab1 neither caused the lysis nor altered the structural integrity of lysosomes. The amount of lysosomes/prelysosomes recovered at the end of the preincubation period was the same whatever the treatment, by the buffer alone, the immune serum, or the IgG fractions. Furthermore, the release of soluble lysosomal enzymes (assessed by acid phosphatase activity measurements) during the incubation period was very low and not dependent upon the pretreatment of lysosomes. Ab2 antibodies, whatever the serum dilution up to 1:10, were devoid of blocking activity. In spite of its high antibody titer (1:30,000 in the soluble phase radioimmunoassay) and its capacity to react with MP50 on Western blot, anti-GDH antibodies of Ab3, at dilutions up to 1:10, did not alter the capacity of lysosomes/ prelysosomes to interact with microtubules (Fig. 3B). To ascertain the specificity of the blocking effect of Ab1 immune serum and IgG, antibodies from Ab1 were purified by affinity chromatography on immobilized MP50. Affinity-purified antibodies inhibited the binding of microtubules to lysosomes/prelysosomes with the same potency as that of unfractionated antibodies (Fig. 3C).
The blocking activity of anti-MP50 antibodies from Ab1 was further documented by varying the reconstitution assay conditions. Using a fixed amount of lysosomes, the formation of lysosome-microtubule complexes increased with the concentration of microtubules and then reached a plateau. As illustrated in Fig. 4, a saturation of the formation of microtubule-lysosome  (2), and lysosome/prelysosome membrane proteins (M) were fractionated by SDS-PAGE and transferred onto Immobilon-P membrane, which was then incubated with the Ab1 immune serum (B) at a 1:300 dilution or with the Ab3 immune serum (C) at a 1:1,000 dilution. Immune complexes were detected using a second antibody conjugated to alkaline phosphatase. The amounts of protein loaded were 1-2 g for G and 1 and 40 g for M. Arrows indicate the mobility of MP50. The position of proteins of known molecular mass (expressed in kDa) is indicated on the left side of the figure. complexes was clearly observed using a microtubule protein concentration higher than 4 mg/ml. Anti-MP50 antibodies from Ab1 neutralized 80 -90% of microtubule binding sites on lysosomes. The residual fraction of non-blocked microtubule binding sites kept the capacity to bind microtubules as a function of the microtubule concentration. It is worth noting that the amount of lysosomes recovered in the pellet under the different experimental conditions was constant (Fig. 4, dotted line).
The ability of anti-MP50 antibodies to block the interaction of lysosomes/prelysosomes with microtubules was analyzed by negative staining electron microscopy. The conditions of the reconstitution assay were the same as those used above except that the concentration of the reactants was reduced to get a good spreading of material on the grids and to avoid artifactual superimposition of free microtubules and free lysosomes. Association of microtubules to lysosomes and prelysosomes could be clearly visualized on negatively stained samples (Fig. 5A). Microtubules of reduced length (due to the shearing by passages through a needle of a syringe) appeared attached to rather round and small vesicles or to more irregular and larger structures that probably corresponded to lysosomes and prelysosomes, respectively. Besides vesicle-bound microtubules, one could observe free microtubules. Quantitative measurements of the proportion of free and microtubule-bound organelles revealed that about 80% of lysosomes/prelysosomes interacted with at least one microtubule in control conditions (Fig. 5B). After a preincubation with anti-MP 50 antibodies from Ab1, the proportion of lysosomes/prelysosomes that did not interact with any microtubule increased from 20 to about 60%. As previously observed (15), an inhibition of the microtubule-vesicle interaction of the same amplitude was obtained when 1 mM ATP was added to the reconstitution mixture.
Immunolocalization of MP50 in Intact Cells-Indirect immunofluorescence experiments have been performed to localize MP50 and GDH in thyroid cells in culture. Purified IgG or affinity-purified anti-MP50 IgG from Ab1 labeled perinuclear structures and elongated cytoplasmic elements (Fig. 6D). This labeling was totally suppressed when the antibody solution incubation mixtures were then layered on 100 l of 0.75 M sucrose and centrifuged at 10,000 ϫ g for 5 min. The radioactivity recovered in the pellets was converted in absolute amounts of tubulin (from the known specific radioactivity of tubulin in the microtubule protein solution). Aliquots of 125 I-labeled microtubule protein (either non-polymerized or in the form of microtubules) incubated in the absence of lysosomes for 30 min at 25°C were centrifuged at 150,000 ϫ g for 60 min to determine the polymerization state of microtubule protein. B and C, analyses of the blocking effects of anti-MP50 and anti-GDH antibodies. Purified lysosomes/prelysosomes were preincubated with either normal rabbit serum (NRS) or Ab1, Ab2, or Ab3 immune serum; purified IgG from NRS (NRIgG) or purified IgG from Ab1 immune serum (IgG(Ab1)); or affinity chromatography fractions F0, F1, and F2 defined below. The dilution of sera and the concentration of purified IgG are given at the bottom of the histograms. Anti-MP50 antibodies from Ab1 were purified by affinity chromatography on immobilized MP50. Fractions F0, F1, and F2 correspond to the complete serum, the serum fraction that did not bind to the column, and the purified antibody fraction eluted from the column at pH 2.5, respectively. The dilution of fractions F1 and F2 was adjusted to correspond to that of F0 (unfractionated Ab1 at 1:30 dilution). At the end of the preincubation period, lysosomes/prelysosomes were collected by centrifugation, washed, and resuspended in the cell-free reconstitution assay buffer (buffer B). Aliquots of the lysosome/ prelysosome suspensions (0.12 mg of protein) were mixed with preassembled 125 I-labeled microtubules (0.4 mg) in a total volume of 200 l of buffer B and incubated for 30 min at 25°C. The incubation mixtures were centrifuged at 10,000 ϫ g for 5 min on a sucrose cushion, and radioactivity of the pellets was counted. The absolute amount of microtubules present in the pellets was calculated from radioactivity values and the specific radioactivity of tubulin. Results are expressed in percent of control values, i.e. values obtained with lysosomes/prelysosomes preincubated in control conditions (buffer alone). Columns and vertical bars represent the mean Ϯ S.E. of triplicate values. was preincubated with purified MP50. MP50 appeared widely distributed on intracellular membranes and located not only on lysosomes/prelysosomes (Fig. 6, compare panels A and C with D). Anti-GDH antibodies from Ab3-labeled elongated and/or filamentous structures scattered throughout the cytoplasm (Fig. 6B); these organelles probably corresponded to mitochondria. Indeed, a very similar staining pattern was obtained with the fluorescent dye rhodamine 123 (42), which specifically labels mitochondria (results not shown).
Immunoassay of MP50 in Lysosome/Prelysosome Membrane Fractions-As immunofluorescence stainings did not allow us to visualize MP50 on lysosomes/prelysosomes, we tried to detect MP50 present on a lysosome membrane using another immunological technique, soluble phase radioimmunoassay (Fig. 7). Increasing amounts of CHAPS-solubilized lysosomal membrane protein displaced the binding of 125 I-labeled MP50 to antibodies from Ab1. The parallelism between the standard curve generated with purified MP50 and the curve obtained with the lysosomal membrane protein extracts allowed the quantification of MP50. MP50 represented 0.10 Ϯ 0.02% (mean Ϯ S.E. of three determinations) of protein solubilized by 0.5% CHAPS from lysosome membrane preparations (obtained by osmotic pressure-dependent lysis of purified lysosomes/ prelysosomes).
Anti-MP50 Antibodies Alter the Intracellular Distribution of Lysosomes/Prelysosomes in Intact Cells-Affinity-purified anti-MP50 antibodies were microinjected into thyroid cells cultured as monolayers, and their ability to alter lysosome-micro-tubule interactions was assessed by analyzing the intracellular distribution of lysosomes. The immunofluorescence labeling of lysosomes/prelysosomes was carried out using previously characterized antibodies (10) named anti-SLP antibodies; these antibodies as well as anti-Ars-A antibodies labeled organelles mainly located on one side of the nuclei, probably in the vinicity of the MTOC, or all around the nuclei (Fig. 6C). Microinjection of non-immune IgG modified neither the cell morphology nor the perinuclear localization of lysosomes (Fig. 8C). In contrast, microinjection of affinity-purified anti-MP50 IgG altered the organization of lysosomes/prelysosomes that appeared more scattered in the cytoplasm (Fig. 8D). The microtubule network . The lysosome/ prelysosome-microtubule mixtures were layered on a sucrose cushion and centrifuged at 10,000 ϫ g for 5 min. The amount of microtubules recovered in the pellets (E, q) was determined from the radioactivity values. Pellets were then resuspended in 10 mM Tris, pH 7.4, to induce the lysis of lysosomes/prelysosomes. The lysosome lysates were centrifuged at 10,000 ϫ g for 5 min, and absorbance measurements at 440 nm (Ⅺ, f) were made on the supernatants. The assay of lysosomal chromogenic compounds (with a maximal absorption around 440 nm), released by osmotic pressure-dependent lysis, was used to estimate the amount of lysosomes/prelysosomes sedimented in each experimental condition. Each symbol and vertical bar represent the mean Ϯ S.E. of triplicate values.
FIG . 5. A, association of microtubules to lysosomes/prelysosomes visualized by negative staining electron microscopy. Lysosomes/prelysosomes (0.05 mg of protein) were incubated with preassembled microtubules (0.2 mg) in a total volume of 0.5 ml of buffer B for 30 min at 25°C. At the end of the incubation period, samples of 20 l of the reconstitution mixtures were layered on the grids and processed as indicated under "Experimental Procedures." Bar, 500 nm. B, analyses of the blocking effect of anti-MP50 antibodies on the association of microtubules to lysosomes/prelysosomes. Lysosomes/prelysosomes were preincubated with NRS [1] and [3] or with Ab1 immune serum [2] at a 1:10 dilution. Lysosomes/prelysosomes were collected by centrifugation, washed, resuspended in buffer B and incubated with preassembled microtubules in the conditions described above. 1 mM ATP was added in the incubation mixture of condition [3]. Three grids were prepared with each incubation mixture and processed for negative staining. The proportion of lysosomes/prelysosomes interacting with one or several microtubules (hatched columns) and the proportion of free lysosomes/ prelysosomes (open columns) were determined by examining at least 100 vesicles taken at random on each grid. Columns and vertical bars represent the mean Ϯ S.E. of the values obtained on three grids.
(visualized using a monoclonal anti-␤-tubulin antibody) of cells microinjected with either control or affinity-purified IgG was not different from that of non-injected cells (data not shown). The effect of anti-MP50 antibodies is also visualized in Fig. 8B; the lysosome distribution of the antibody-injected cell appears clearly distinct from that of the neighboring non-injected ones. The distribution of lysosomes/prelysosomes in anti-MP50 IgGand control IgG-injected cells was analyzed by quantitative videomicroscope analyses. Four parameters were determined: (a) the surface labeled by Lucifer Yellow (S1) and the mean luminance (mean fluorescence intensity) over this surface, and (b) the cell surface (S2) over which Texas Red-immunolabeled lysosomes were found and the mean luminance of this delimited area. Results of these experiments are summarized in Table I. The average Lucifer Yellow-labeled surface (S1), i.e. the cell surface, was similar in control IgG-and anti-MP50 IgG-injected cells. The cell spreading varied somewhat from culture to culture, the average cell surface was 1.5-fold higher in experiment 2 than in experiment 1. The surface occupied by labeled lysosomes (S2) as well as the ratio S2:S1 were significantly higher in anti-MP50 IgG-injected cells than in control IgG-injected cells. Very similar results were obtained in the two series of microinjections. In each experiment, S2 values of antibody-injected cells were about 20% higher than those of control IgG-injected cells. It is worth noticing that the increase of S2 in antibody-injected cells was accompanied by a statistically significant decrease of the mean luminance of S2 (as compared with that of control cells). Thus, total fluorescence of labeled lysosomes (S2 ϫ mean luminance of S2) was approximately the same in the two groups of cells. These data show that anti-MP50 antibodies microinjected into living cells cause a shift in the positioning of lysosomes from a rather compact organization near the nucleus to a more random distribution in the cytoplasm. DISCUSSION We report here that the peripheral membrane protein, MP50, has a function of microtubule-lysosome linker protein and probably plays a role in the microtubule-dependent positioning of lysosomes in living cells. The association of lysosomes/prelysosomes to microtubules can be blocked by antibodies that specifically recognize MP50 on the lysosome/ prelysosome membrane. The antibodies probably act by neutralizing the microtubule binding activity of MP50 (31,34). In the cell-free reconstitution assay, only anti-MP50 antibodies from Ab1 exerted a potent concentration-dependent inhibitory effect. Despite their reactivity or cross-reactivity with native MP50 (in solution), Ab2 and Ab3 antibodies were repeatedly devoid of activity. These antibodies either are directed against a part of MP50 molecule not involved in the binding of microtubules or recognize conformational epitopes inaccessible or absent on MP50 attached to the lysosome membrane. The lack of effect of anti-GDH antibodies from Ab3 was not unexpected since these antibodies poorly reacted with MP50 in solution. Differences in reactivity between antibodies produced against MP50 and GDH were also found in immunofluorescence experiments. This indicates that MP50 and GDH differ in some structural and/or conformational aspects.
MP50 is probably linked to the organelle membrane in the form of oligomers; we previously reported (34) that purified MP50 (as GDH) was mostly in the form of hexamers. Due to its high microtubule binding affinity (K a , ϳ0.5 ϫ 10 8 M Ϫ1 ), MP50 by itself should be capable of mediating the interaction and, therefore, could play the role of a direct anchoring factor. As anti-MP50 antibodies are able to neutralize the large majority of microtubule binding sites on lysosomes, one can assume that lysosomes possess a limited number of MP50 molecules on their membrane. The "density" of MP50 molecules on lysosome/ prelysosome membranes was estimated from data of the MP50 immunoassay. We know from previous studies that the average diameter of isolated thyroid lysosomes is about 200 nm (32). The apparent diameter of prelysosomes or late endosomes was found to be higher and more variable, from 0.2 to 1 m (41). The calculations that follow have been made for a lysosomal vacuole with an average diameter of 0.5 m. The membrane surface of such a vacuole considered as a sphere would be 0.8 m 2 . Knowing that there are approximately 5 ϫ 10 6 lipid molecules (mostly glycerophospholipids) in a 1 m ϫ 1 m area of lipid bilayer (43) and taking an average molecular mass of 850 Da for glycerophospholipids, one can estimate the lipid content of the lysosomal membrane at 6 ϫ 10 Ϫ15 g. Assuming that membrane proteins represent about 50% of the membrane mass and considering that MP50 represents 0.10% of lysosomal membrane protein (Fig. 7) and is mostly in the form of hexamers, we arrive to the conclusion that the membrane of the lysosomal vacuole would bear 12 MP50 hexamers. Given their size variability, lysosomes and prelysosomes would possess from Ͻ5 to up to 50 MP50 hexamers. The possibility that antibodies complexed to a low number of MP50 molecules could limit the access of microtubules to other potential linker proteins on lysosomes by a steric hindrance effect appears thus most unlikely. A low "density" of MP50 molecules on the lysosome/prelysosome membrane would be in keeping with two experimental observations: (a) the possibility to reach the saturation of microtubule binding sites on lysosomes/prelysosomes at high polymerized microtubule protein concentration (Fig. 4) and (b) the lack of selective labeling of lysosomes with anti-MP50 antibodies in intact cells.
The biological relevance of the data obtained in the cell-free reconstitution system is reinforced by the observations made on living cells. Affinity-purified anti-MP50 IgG microinjected into the cytoplasm of thyroid cells caused a definite change in the distribution of lysosomes. By inducing the scattering of lysosomes in the cytoplasm, anti-MP50 antibodies produced an effect comparable with but not as marked as that obtained with a nocodazole treatment or an exposure to cold (1, 10, 44). The disorganization of lysosomes resulting from the microtubule disassembly was more extensive than the dispersion of lyso- FIG. 8. Effects of affinity-purified anti-MP50 antibodies, microinjected into thyroid cells, on the intracellular localization of lysosomes/prelysosomes. Thyroid cells in primary culture for 3 days were microinjected with either affinity-purified anti-MP50 IgG from Ab1 (A, B, and D) or with control IgG purified from normal rabbit serum (C). The Ig concentration of the microinjected solution was 5 mg/ml. In addition, the microinjected solution contained BSA conjugated to Lucifer Yellow at a concentration of 5 mg/ml. After microinjection, cells were washed and incubated for 4 h at 20°C. Cells were then fixed and immunostained using anti-SLP antibodies (immune serum at a 1:100 dilution) that label thyroid lysosomes and donkey anti-rabbit Ig antibodies conjugated to Texas Red. Panel A shows the Lucifer Yellow fluorescence image of the cell identified by an arrow in panel B. This cell was microinjected with anti-MP50 IgG; its lysosome distribution pattern was different from that of the neighboring non-injected cells. The anti-MP50 antibody-induced changes in lysosome distribution are evidenced by comparing panels C and D. Cells microinjected with control IgG (C) present the typical lysosome organization at one side of the nucleus, whereas antibody-treated cells (D) have their lysosomes more dispersed in the cytoplasm.

TABLE I Quantitative videomicroscope analyses of the distribution of lysosomes in cells microinjected with affinity-purified anti-MP50 IgG
Thyroid cells cultured in petri dishes for 3 days were microinjected with affinity-purified anti-MP50 IgG or purified IgG from normal rabbit serum. The microinjected solutions contained IgG (5 mg/ml) and BSA-Lucifer Yellow conjugate (5 mg/ml) in 10 mM phosphate, 145 mM KCl, pH 7.0. After microinjection, cells were washed and incubated for 4 h at 20°C and then fixed and immunostained with anti-SLP antibodies and a Texas Red-labeled second antibody. Lucifer Yellow and Texas Red fluorescence images were analyzed as described under "Experimental Procedures." Results are expressed as the mean Ϯ S.E. of values obtained from 12 to 30 microinjected cells. Probability values were obtained from the Student's t test. somes caused by anti-MP50 antibodies. The incomplete effect of the anti-MP50 antibodies could be due to a limitation of the amount of antibodies available to block microtubule binding sites on lysosomes/prelysosomes. Indeed, microinjected antibodies probably interact with MP50 molecules present on other intracellular membranes. Alternately, MP50 could only represent a component of a more complex molecular machinery accounting for the association of lysosomes to microtubules in the perinuclear region of the cell. This machinery could include other linker proteins and/or motility factors still operative after inactivation of MP50. Cytoplasmic dynein, involved in the centripetal movement of organelles toward the minus-end of microtubules (22), probably contributes to the clustering of lysosomes in the perinuclear region of the cell. Opposite forces due to kinesin would tend to pull lysosomes in the outward direction. The location of lysosomes near the MTOC or around the nucleus could be explained by a local inactivation of the kinesindependent movement and/or to the implication of additional molecular partners such as adhesion or anchoring factors mediating lysosome-microtubule interactions. The former hypothesis has not been documented yet; we bring experimental groundings in favor of the latter. The attachment of lysosomes and prelysosomes to microtubules via anchor proteins such as MP50 would provide a mechanism for immobilization or retention of these organelles in the perinuclear region of the cell. The efficiency of the anchoring of lysosomes onto microtubules would be favored in the centrosomal region of the cell where the density of microtubules is high. Lysosomes exposing several MP50 molecules on their surface could interact with several microtubules. MP50-mediated lysosome-microtubule interactions would be subject to regulation by nucleotides. We previously found that the association of lysosomes to microtubules was inhibited by ATP (16) and that the binding of MP50 to microtubules was also ATP-dependent (31,34). Interestingly, the two ATP effects were observed in the same concentration range (between 0.1 and 2 mM). Understanding how microtubules and linker and motor proteins control the "geography" of organelles is a challenge that has begun to be addressed. Our data support the idea that the intracellular positioning of lysosomes and prelysosomes not only brings into play movements along microtubules but also depends on processes of anchoring to microtubules. With MP50 being present on membranes other than lysosome membranes, one may consider the possibility that MP50 could mediate other microtubule-organelle interactions. This question deserves further investigations using other purified cell organelles such as mitochondria or endoplasmic reticulum-derived vesicular fractions. Finally, the present work assigns a new function to GDH. In the last ten years, there have been a number of reports describing new functions for known proteins. Protein disulfide isomerase and prolyl hydroxylase (45), the cation-independent mannose 6-phosphate receptor and the insulin-like growth factor II receptor (46), ␣-enolase, and plasminogen receptor (47) have all been shown to reside in the same polypeptide. The best example is given by the crystallin protein family, the major structural protein of the lens (48). Some of the crystallins were found to be essentially identical to enzymes (lactate dehydrogenase B4, ␣-enolase, hydroxylacyl-CoA dehydrogenase, alcohol dehydrogenases, etc.). Similarly, GDH could have cellular functions other than that of an enzyme. An activity of RNA-binding protein has recently been assigned to GDH (49).