Changes in D-serine levels and localization during postnatal development of the rat vestibular nuclei.

The patterns of development of the vestibular nuclei (VN) and their main connections involving glutamate neurotransmission offer a good model for studying the function of the glial-derived neuromodulator D-serine in synaptic plasticity. In this study we show that D-serine is present in the VN and we analyzed its distribution and the levels of expression of serine racemase and D-amino acid oxidase (D-AAO) at different stages of postnatal (P) development. From birth to P21, high levels of D-serine were detected in glial cells and processes in all parts of the VN. This period corresponded to high expression of serine racemase and low expression of D-AAO. On the other hand, in the mature VN D-serine displayed very low levels and was mainly localized in neuronal cell bodies and dendrites. This drop of D-serine in adult stages corresponded to an increasing expression of D-AAO at mature stages. High levels of glial D-serine during the first 3 weeks of postnatal development correspond to an intense period of plasticity and synaptogenesis and maturation of VN afferents, suggesting that D-serine could be involved in these phenomena. These results demonstrate for the first time that changes in D-serine levels and distribution occur during postnatal development in the central nervous system. The strong decrease of D-serine levels and the glial-to-neuronal switch suggests that D-serine may have distinct functional roles depending on the developmental stage of the vestibular network.

Functional analysis of Plasmodium vivax VIR proteins reveals different subcellular localizations and cytoadherence to the ICAM-1 endothelial receptor.

The subcellular localization and function of variant subtelomeric multigene families in Plasmodium vivax remain vastly unknown. Among them, the vir superfamily is putatively involved in antigenic variation and in mediating adherence to endothelial receptors. In the absence of a continuous in vitro culture system for P. vivax, we have generated P. falciparum transgenic lines expressing VIR proteins to infer location and function. We chose three proteins pertaining to subfamilies A (VIR17), C (VIR14) and D (VIR10), with domains and secondary structures that predictably traffic these proteins to different subcellular compartments. Here, we showed that VIR17 remained inside the parasite and around merozoites, whereas VIR14 and VIR10 were exported to the membrane of infected red blood cells (iRBCs) in an apparent independent pathway of Maurer's clefts. Remarkably, VIR14 was exposed at the surface of iRBCs and mediated adherence to different endothelial receptors expressed in CHO cells under static conditions. Under physiological flow conditions, however, cytoadherence was only observed to ICAM-1, which was the only receptor whose adherence was specifically and significantly inhibited by antibodies against conserved motifs of VIR proteins. Immunofluorescence studies using these antibodies also showed different subcellular localizations of VIR proteins in P. vivax-infected reticulocytes from natural infections. These data suggest that VIR proteins are trafficked to different cellular compartments and functionally demonstrates that VIR proteins can specifically mediate cytoadherence to the ICAM-1 endothelial receptor.

Linking supply to demand: the neuronal monocarboxylate transporter MCT2 and the alpha-amino-3-hydroxyl-5-methyl-4-isoxazole-propionic acid receptor GluR2/3 subunit are associated in a common trafficking process.

MCT2 is the major neuronal monocarboxylate transporter (MCT) that allows the supply of alternative energy substrates such as lactate to neurons. Recent evidence obtained by electron microscopy has demonstrated that MCT2, like alpha-amino-3-hydroxyl-5-methyl-4-isoxazole-propionic acid (AMPA) receptors, is localized in dendritic spines of glutamatergic synapses. Using immunofluorescence, we show in this study that MCT2 colocalizes extensively with GluR2/3 subunits of AMPA receptors in neurons from various mouse brain regions as well as in cultured neurons. It also colocalizes with GluR2/3-interacting proteins, such as C-kinase-interacting protein 1, glutamate receptor-interacting protein 1 and clathrin adaptor protein. Coimmunoprecipitation of MCT2 with GluR2/3 and C-kinase-interacting protein 1 suggests their close interaction within spines. Parallel changes in the localization of both MCT2 and GluR2/3 subunits at and beneath the plasma membrane upon various stimulation paradigms were unraveled using an original immunocytochemical and transfection approach combined with three-dimensional image reconstruction. Cell culture incubation with AMPA or insulin triggered a marked intracellular accumulation of both MCT2 and GluR2/3, whereas both tumor necrosis factor alpha and glycine (with glutamate) increased their cell surface immunolabeling. Similar results were obtained using Western blots performed on membrane or cytoplasm-enriched cell fractions. Finally, an enhanced lactate flux into neurons was demonstrated after MCT2 translocation on the cell surface. These observations provide unequivocal evidence that MCT2 is linked to AMPA receptor GluR2/3 subunits and undergoes a similar translocation process in neurons upon activation. MCT2 emerges as a novel component of the synaptic machinery putatively linking neuroenergetics to synaptic transmission.

Cloning, functional expression, and chromosomal localization of the human pancreatic islet glucose-dependent insulinotropic polypeptide receptor.

Glucose-dependent insulinotropic polypeptide (GIP) is a hormone secreted by the endocrine K-cells from the duodenum that stimulates glucose-induced insulin secretion. Here, we present the molecular characterization of the human pancreatic islet GIP receptor. cDNA clones for the GIP receptor were isolated from a human pancreatic islet cDNA library. They encoded two different forms of the receptor, which differed by a 27-amino acid insertion in the COOH-terminal cytoplasmic tail. The receptor protein sequence was 81% identical to that of the rat GIP receptor. When expressed in Chinese hamster lung fibroblasts, both forms of the receptor displayed high-affinity binding for GIP (180 and 600 pmol/l). GIP binding was displaced by < 20% by 1 mumol/l glucagon, glucagon-like peptide (GLP-I)(7-36) amide, vasoactive intestinal peptide, and secretin. However exendin-4 and exendin-(9-39) at 1 mumol/l displaced binding by approximately 70 and approximately 100% at 10 mumol/l. GIP binding to both forms of the receptor induced a dose-dependent increase in intracellular cAMP levels (EC50 values of 0.6-0.8 nmol/l) but no elevation of cytoplasmic calcium concentrations. Interestingly, both exendin-4 and exendin-(9-39) were antagonists of the receptor, inhibiting GIP-induced cAMP formation by up to 60% when present at a concentration of 10 mumol/l. Finally, the physical and genetic chromosomal localization of the receptor gene was determined to be on 19q13.3, close to the ApoC2 gene. These data will help study the physiology and pathophysiology of the human GIP receptor.

Visualizing olfactory receptor expression and localization in Drosophila.

Odor detection and discrimination by olfactory systems in vertebrates and invertebrates depend both on the selective expression of individual olfactory receptor genes in subpopulations of olfactory sensory neurons, and on the targeting of the encoded proteins to the exposed, ciliated endings of sensory dendrites. Techniques to visualize the expression and localization of olfactory receptor gene products in vivo have been essential to reveal the molecular logic of peripheral odor coding and to permit investigation of the developmental and cellular neurobiology of this sensory system. Here, we describe methods for detection of olfactory receptor transcripts and proteins in the antennal olfactory organ of the fruit fly, Drosophila melanogaster, an important genetic model organism. We include protocols both for antennal cryosections and whole-mount antennae. These methods can be adapted for detection of receptor expression in other olfactory and gustatory tissues in Drosophila, as well as in the chemosensory systems of other insects.

Evolution of a polymineralic mantle shear zone and the role of second phases in the localization of deformation

The influence of second phases (e.g., pyroxenes) on olivine grain size was studied by quantitative microfabric analyses of samples of the Hilti massif mantle shear zone (Semail ophiolite, Oman). The microstructures range from porphyroclastic tectonites to ultramylonites, from outside to the center of the shear zone. Starting at conditions of ridge-related flow, they formed under continuous cooling leading to progressive strain localization. The dependence of the average olivine grain size on the second-phase content can be split into a second-phase controlled and a dynamic recrystallization-controlled field. In the former, the olivine grain size is related to the ratio between the second-phase grain size and volume fraction (Zener parameter). In the latter, dynamic recrystallization manifested by a balance between grain growth and grain size reduction processes yields a stable olivine grain size. In both fields the average olivine and second-phase grain size decreases with decreasing temperature. Combining the microstructural information with deformation mechanism maps suggests that the porphyroclastic tectonites (similar to 1100 degrees C) and mylonites (similar to 800 degrees C) formed under the predominance of dislocation creep. Since olivine-rich layers are intercalated with layer parallel, polymineralic bands in the mylonites, nearly equiviscous conditions can be assumed. In the ultramylonites, diffusion creep represents the major deformation mechanism in the polymineralic layers. It is this switch in deformation mechanism from dislocation creep to diffusion creep that forces strain to localize in the fine-grained polymineralic domains at low temperatures (<similar to 700 degrees C), underlining the role of the second phases on strain localization in cooling mantle rocks.

Cross validation of experts versus registration methods for target localization in deep brain stimulation.

In the last five years, Deep Brain Stimulation (DBS) has become the most popular and effective surgical technique for the treatent of Parkinson's disease (PD). The Subthalamic Nucleus (STN) is the usual target involved when applying DBS. Unfortunately, the STN is in general not visible in common medical imaging modalities. Therefore, atlas-based segmentation is commonly considered to locate it in the images. In this paper, we propose a scheme that allows both, to perform a comparison between different registration algorithms and to evaluate their ability to locate the STN automatically. Using this scheme we can evaluate the expert variability against the error of the algorithms and we demonstrate that automatic STN location is possible and as accurate as the methods currently used.

Knot localization in proteins.

The backbones of proteins form linear chains. In the case of some proteins, these chains can be characterized as forming linear open knots. The knot type of the full chain reveals only limited information about the entanglement of the chain since, for example, subchains of an unknotted protein can form knots and subchains of a knotted protein can form different types of knots than the entire protein. To understand fully the entanglement within the backbone of a given protein, a complete analysis of the knotting within all of the subchains of that protein is necessary. In the present article, we review efforts to characterize the full knotting complexity within individual proteins and present a matrix that conveys information about various aspects of protein knotting. For a given protein, this matrix identifies the precise localization of knotted regions and shows the knot types formed by all subchains. The pattern in the matrix can be considered as a knotting fingerprint of that protein. We observe that knotting fingerprints of distantly related knotted proteins are strongly conserved during evolution and discuss how some characteristic motifs in the knotting fingerprints are related to the structure of the knotted regions and their possible biological role.

Nuclear localization of mouse fibroblast growth factor 2 requires N-terminal and C-terminal sequences.

In vertebrates, different isoforms of fibroblast growth factor 2 (FGF2) exist, which differ by their N-terminal extension. They show different localization and expression levels and exert distinct biological effects. Nevertheless, genetic inactivation of all FGF2 isoforms in the mouse results in only mild phenotypes. Here, we analyzed mouse FGF2, and show that, as in the human, mouse FGF2 contains CTG-initiated high molecular-weight (HMW) isoforms, which contain a nuclear localization signal, and which mediate localization of this isoform to the nucleus. Using green fluorescent protein-FGF2 fusions, we furthermore observed, that C-terminal deletions disable nuclear localization of the short low-molecular-weight (LMW) 18-kDa isoform. This loss of specific localization is accompanied by a loss in heparin binding. We therefore suggest that, first, localization of mouse FGF2 is comparable to that in other vertebrates and, second, FGF2 contains at least two sequences important for nuclear localization, a nuclear localization sequence at the N terminus which is only contained in the HMW isoform, and another sequence at the C terminus, which is only required for localization of the LMW 18-kDa isoform.

Anthracene-tethered ruthenium(II) arene complexes as tools to visualize the cellular localization of putative organometallic anticancer compounds.

Anthracene derivatives of ruthenium(II) arene compounds with 1,3,5-triaza-7-phosphatricyclo[3.3.1.1]decane (pta) or a sugar phosphite ligand, viz., 3,5,6-bicyclophosphite-1,2-O-isopropylidene-α-d-glucofuranoside, were prepared in order to evaluate their anticancer properties compared to the parent compounds and to use them as models for intracellular visualization by fluorescence microscopy. Similar IC(50) values were obtained in cell proliferation assays, and similar levels of uptake and accumulation were also established. The X-ray structure of [{Ru(η(6)-C(6)H(5)CH(2)NHCO-anthracene)Cl(2)(pta)] is also reported.

Whole-mount confocal immunofluorescence of mammalian CNS.

Bright-field wholemount labeling techniques applied to the mammalian central nervous system (CNS) offer advantages over conventional methods based on sections since an immediate and three-dimensional view of the stained components is provided. It thereby becomes possible to survey and count large number of cells and fibers in their natural relationships. The ability of confocal laser scanning microscopy to visualize in one focal plane the fluorescence associated with multiple markers could be most valuable by the availability of reliable wholemount fluorescent techniques. Accordingly, based in our previously published bright-field wholemount protocols [Brain Res. Prot. 2 (1998) 165-173], we have devised an effective immmunofluorescence wholemount procedure. We show that reliable wholemount fluorescent staining can be obtained using isolated complete CNS aged up to rat embryonic day 17, with antibodies penetration in the millimeter range. Examples are shown of preparations in which colocalization can be observed in nerve cells of cytoskeletal and calcium-binding proteins.

Histone H1 expression varies during the Leishmania major life cycle.

The deduced amino acid sequence of Leishmania major sw3 cDNA reveals the presence of characteristic histone H1 amino acid motifs. However, the open reading frame is of an unusually small size for histone H1 (105 amino acids) because it lacks the coding potential for the central hydrophobic globular domain of linker histones present in other eukaryotes. Here, we provide biochemical evidence that the SW3 protein is indeed a L. major nuclear histone H1, and that it is differentially expressed during the life cycle of the parasite. Due to its high lysine content, the SW3 protein can be purified to a high degree from L. major nuclear lysates with 5% perchloric acid, a histone H1 preparative method. Using an anti-SW3 antibody, this protein is detected as a 17 kDa or as a 17/19 kDa doublet in the nuclear subfraction in different L. major strains. The nuclear localization of the SW3 protein is further supported by immunofluorescence studies. During in vitro promastigote growth, both the sw3 cytoplasmic mRNA and its protein progressively accumulate within parasites from early log phase to stationary phase. Within amastigotes, the high level of H1 expression is maintained but decreases when amastigotes differentiate into promastigotes. Together, these observations suggest that the different levels of this histone H1 protein could influence the varying degrees of chromatin condensation during the life-cycle of the parasite, and provide us with tools to study this mechanism.

Les aquaporines 4 et 9 dans le système nerveux central : expression lors de la différenciation cellulaire et implication dans le métabolisme énergétique astrocytaire "in vitro"

RESUME : Les aquaporines (AQPs) sont des protéines membranaires perméables à l'eau (aquaporines strictes) et, pour certaines d'entre elles, également au glycérol (aquaglycéroporines). Ces protéines sont présentes dans les bactéries, les plantes et les différents organes des mammifères. Dans le cerveau, la moindre augmentation de volume hydrique peut avoir de graves conséquences sur son fonctionnement, d'où l'importance de la régulation de l'homéostasie de l'eau grâce aux AQPs. L'AQP4, une aquaporine stricte, est présente dans les astrocytes et est impliquée dans la formation et la résorption des oedèmes cérébraux. En revanche, l'AQP9 est une aquaglycéroporine, qui est localisée non seulement dans les astrocytes mais également dans les neurones catécholaminergiques. Bien que la distribution de l'AQP4 dans le cerveau soit clairement établie, la présence de l'AQP9 est toujours une donnée controversée et son rôle fonctionnel dans le système nerveux central n'est pas connu. Par ailleurs, aucune donnée n'existe sur l'expression des AQP4 et 9 lors de la différenciation de cellules souches neurales foetales (CSNf) en astrocytes ou en neurones catécholaminergiques. Dans la première partie de ce travail, un protocole a été mis au point permettant de différencier des CSNf de souris en astrocytes et neurones, dont des neurones catécholaminergiques. La caractérisation des cultures de CSNf et des cultures mixtes par immunofluorescence a permis de montrer que l'immunomarquage AQP9 est présent dans les CSNf et est conservé lors de leur différenciation en astrocytes ou en neurones catécholaminergiques. Les résultats obtenus ont mis en évidence une très bonne corrélation entre l'expression de la TH (tyrosine hydroxylase: enzyme limitante de la synthèse des catécholamines) et celle de l'AQP9 lors de la différenciation des CSNf en neurones catécholaminergiques. Par contre, l'immunomarquage AQP4 n'est pas présent dans les CSNf alors qu'il est observé dans les astrocytes. De plus, aucun immunomarquage AQP4 ou AQP9 n'a été observé dans les neurones NIAP2-positifs. Dans la deuxième partie de ce travail, l'expression des AQP4 et 9 a été quantifiée dans les CSNf ainsi que dans trois populations d'astrocytes présentant des propriétés métaboliques différentes. Ces trois populations astrocytaires sont issues de la différenciation des CSNf par le CNTF, le LIF ou le sérum de veau foetal. Les analyses par RTPCR quantitative et western blot ont montré une augmentation de l'expression de l'AQP9 et de l'AQP4 corrélée à l'acquisition de propriétés métaboliques spécifiques des astrocytes matures. Dans la dernière partie, la technique d'ARN interférents a permis d'étudier le rôle fonctionnel de l'AQP9 dans le modèle de culture pure d'astrocytes différenciés par le sérum. L'inhibition de l'expression d'AQP9 entraîne une diminution de la perméabilité au glycérol et une augmentation de l'utilisation de glucose, corrélée à une stimulation du métabolisme oxydatif astrocytaire. En revanche, 1a baisse d'expression d'AQP9 n'a aucun effet sur la glycolyse anaérobie ni sur la libération du lactate. En conclusion, dans ce modèle in vitro, seule l'AQP9 est exprimée dans les CSNf et les neurones catécholaminergiques alors que dans Ies astrocytes, à la fois l'AQP9 et l'AQP4 sont exprimées. Cette distribution est identique à celle observée in vivo et confirme la localisation spécifique de l'AQP9 dans les neurones catécholaminergiques. De plus, ces résultats montrent, pour la première fois, l'implication de l'AQP9 dans la perméabilité des astrocytes au glycérol et son implication dans le métabolisme énergétique astrocytaire. ABSTACT : Aquaporins (AQPs) are membrane proteins permeable to water (orthodoxes aquaporins) and some of them are also permeable to glycerol (aquaglyceroporins). These proteins are widely expressed in bacteria, plants and mammals. AQP water homeostasis regulation in brain is of primary importance as the brain volume cannot increase. AQP4, an orthodoxe aquaporin, is present in astrocytes and seems to be involved in edema formation and resorption. On the other hand, AQP9 is an aquaglyceroporin which is localised not only in astrocytes but also in catecholaminergic neurons. Although AQP4 distribution in brain is clearly established, the presence of AQP9 is still a discussed data and its functional role in the central nervous system is unknown. In addition, no data exists on AQP4 or AQP9 expression during fetal neural stem cells (fNSC) differentiation into astrocytes or catecholaminergic neurons. In the first part of this work, a protocol was developed to differentiate mouse fNSC into astrocytes and neurons, with the aim to obtain catecholaminergic neurons. By immunefluorescence, we have shown that AQP9 is expressed in fNSC cultures and also in astrocytes and catecholaminergic neurons in mixt cultures. The results obtained highlighted a very good correlation between TH expression (tyrosin hydroxylase being a limiting enzyme of catecholamines synthesis) and AQP9 in fNSC and all along their differentiation into catecholaminergic neurons. On the other hand, AQP4 immunolabelling is not observed in fNSC whereas it is in astrocytes. Moreover, neitheir AQP4, nor AQP9 immunoreactivity was observed in MAP2-positive neurons. In the second part of this work, AQP4 and AQP9 expression was quantified in fNSC and in three populations of astrocytes presenting different metabolic properties. These three astrocyte populations result from fNSC differentiation by addition of CNTF, LIF or fetal calf serum. Quantitative RT-PCR and western blot analyses have shown an increase in both AQP4 and AQP9 expression, correlated with the acquisition of specific metabolic properties of mature astrocytes. In the last part, siRNA were used to study the functional role of AQP9 in the pure astrocyte culture model differentiated by addition of fetal calf serum. Inhibition of AQP9 expression leads to a decrease of glycerol uptake and to an increase of glucose uptake, correlated with a stimulation of the astrocyte oxydative metabolism. On the other hand, inhibition of AQP9 expression does not have any effect on anaerobic glycolysis nor on lactate release. In conclusion, in this in vitro model, only AQP9 is expressed in fNSC and in catecholaminergic neurons whereas in astrocytes, both AQP9 and AQP4 are expressed. This distribution is identical to that observed in vivo and confirms the specific AQP9 localization in catecholaminergic neurons. IVloreover, these results show, for the first time, that AQP9 is implicated in glycerol uptake and in astrocyte energetic metabolism. Résumé large public : Les aquaporines, des protéines localisées dans les membranes cellulaires sont, comme leur nom l'indique, des canaux à eau. Pendant longtemps, il a été considéré que l'eau diffusait librement dans et à travers les cellules; la caractérisation des AQPs a révolutionné la vision des scientifiques concernant les mouvements d'eau entre les différents compartiments infra et extracellulaires, et a d'ailleurs valu le Prix Nobel à Peter Agre en 1992. Certaines AQPs, dites "strictes", laissent passer uniquement l'eau et participent au contrôle du volume hydrique. Ce contrôle est particulièrement important pour le bon fonctionnement du cerveau en raison de la présence de la boîte crânienne qui limite les variations de volume. D'autres AQPs, les aquaglycéroporines, sont perméables non seulement à l'eau mais également à d'autres molécules comme le glycérol. Elles facilitent, par exemple, la sortie du glycérol des cellules graisseuses et sa capture par les cellules du foie afin de produire du glucose en période de jeûne. Le cerveau est principalement composé de deux types de cellules: les neurones et les cellules gliales, majoritairement des astrocytes. L'AQP4, une AQP stricte, est présente dans les astrocytes et joue un rôle dans la formation et la résorption des oedèmes cérébraux. L'AQP9, une aquaglycéroporine, est également présente dans les astrocytes et dans une population spécifique de neurones, les neurones catécholaminergiques, touchés dans la maladie de Parkinson. A ce jour, la présence de l'AQP9 dans le cerveau est une donnée controversée et son rôle fonctionnel est inconnu. Ce travail de thèse a permis de montrer que l'AQP9 est bien présente d'une part dans les cellules souches neurales foetales et d'autre ,part dans les astrocytes et neurones catécholaminergiques issus de leur différenciation. De plus, ces expériences ont mis en évidence un rôle de l'AQP9 dans l'entrée du glycérol dans les astrocytes, ce qui pourrait être bénéfique dans des conditions d'ischémie. Enfin, les .résultats de cette étude suggèrent également un rôle de l'AQP9 dans le métabolisme énergétique des astrocytes. L'ensemble de ces travaux démontre le rôle important de l'AQP9 dans le cerveau et ouvre de nouvelles perspectives quant aux rôles des AQPs dans des situations pathologiques telles que l'ischémie cérébrale ou encore la maladie de Parkinson.

Tumor localization of radiolabeled antibodies against carcinoembryonic antigen in patients with carcinoma: a critical evaluation.

Purified, [131I]-labeled goat antibodies against carcinoembryonic antigen, which have been shown to localize in human carcinoma in nude mice, were injected into 27 patients with carcinoma. Patients were scanned with a scintillation camera at various intervals. In 11 patients, radioactivity was detectable in the tumor 48 hours after injection. Computerized subtraction of blood-pool radioactivity provided clearer pictures in positive cases, but in 16 patients the scans remained doubtful or negative. To study the specificity of [131I]-antibody localization, we gave some patients simultaneous injections of [125I]-labeled normal IgG. Both isotopes were measured by means of scintillation counting in tumors and normal tissues recovered after surgery. The results demonstrated that only the anti-CEA antibodies localized in tumors. However, the total antibody-derived radioactivity in the tumor was only about 0.001 of the injected dose. We conclude that, despite the present demonstration of specificity, this method of tumor detection is not yet clinically useful.

Intracellular targeting of GLUT4 in transfected insulinoma cells: evidence for association with constitutively recycling vesicles distinct from synaptophysin and insulin vesicles.

In adipocytes and muscle cells, the GLUT4 glucose transporter isoform is present in intracellular vesicles which continuously recycle between an intracytoplasmic location and the plasma membrane. It is not clear whether the GLUT4-vesicles represent a specific kind of vesicle or resemble typical secretory granules or synaptic-like microvesicles. To approach this question, we expressed GLUT4 in the beta cell line RINm5F and determined its intracellular localization by subcellular fractionation and by immunofluorescence and immunoelectron microscopy. GLUT4 was not found in insulin granules but was associated with a subpopulation of smooth-surface vesicles present in the trans-Golgi region and in vesicular structures adjacent to the plasma membrane. In the trans-Golgi region, GLUT4 did not colocalize with synaptophysin or TGN38. Incubation of the cells with horseradish peroxidase (HRP) led to colocalization of HRP and GLUT4 in some endosomal structures adjacent to the plasma membrane and in occasional trans-Golgi region vesicles. When cells were incubated in the presence of Bafilomycin A, analysis by confocal microscopy revealed GLUT4 in numerous large spots present throughout the cytoplasm, many of which costained for TGN38 and synaptophysin. By immunoelectron microscopy, numerous endosomes were observed which stained strongly for GLUT4. Together our data demonstrate that ectopic expression of GLUT4 in insulinoma cells reveals the presence of a subset of vesicular structures distinct from synaptic-like vesicles and insulin secretory granules. Furthermore, they indicate that GLUT4 constitutively recycles between the plasma membrane and its intracellular location by an endocytic route also taken by TGN38 and synaptophysin.