Biology of the Cell 91 (1999) 227-280 277 ANALYSIS OF CELL MIGRATION BY VIDEOMICROSCOPY ZAHM Jean-Marie, POLETTE Myriam, CORAUX Christelle and BONNET Noel lNS~fV& U514, /Fi?53, 45 rue Cognacq-Jay, 5 1092, Reims, France. Cell migration is an important mechanism involved in many physiopathological processes such as embryogenesis, wound healing or tumoral invasion. The development of in vitro models associated to videomicroscopic techniques allows to analyze the migratory behavior of cells during these different processes. The aim of the present work was to study, by using videomicroscopic techniques, the migration of respiratory epitheliai cells cultured in different experimental conditions. A phase contrast inverted microscope was equiped with a chamber conuoling the air temperature and humidity and with a video- camera driven by a personal computer. This device allowed to record time-lapse video sequences of airway epithelial cells at times varying from seconds to hours. When the cells were plated at high density (~10 6 cells/ml), they adhered to the culture dish and rapidly grouped together to form clusters from which they proliferated to constitute a confluent monolayer. Creating a chemical wound within the confluent cell monolayer induced the cell spreading at the wound edge. The cell spreading was visualized by the highly dynamic formation of IameIIipodia and was the onset of the cell migration. In this in vitro wound repair model, the cell migration was characterized by the uniformity and unidirectionality of the migration speed. At the opposite when the cells were plated at low density, after their adherence onto the culttu~ dish, we observed that the cells extended lamelIipodia in alI directions but did not exhibit any preferential migration. Interestingly when placing a confluent cell layer in a collagen gel sandwich, the cells moved to form 3 dimensional branched structures within several hours. These data demonstrate that building a video sequence from digitized images allows to visualize the migratory behavior of cells in culture and to correlate this behavior with the environmental conditions in which the cells are placed. COORDINATION OF Ca*+ OSCILLATIONS BETWEEN CONNECTED HEPATOCYTES: WHAT IS RESPONSIBLE ? CLAIR Caroline, Tordjmann Thier Christophe, POGGIOLI Josiane’, ~C$&rn;aukil,e*, ERNEUX+ COMBETTES Laurent INSERM U44.2, UPS, b,f. 443, 91405 Orsay, ‘INSERM U356, lnstitut des Cordeliers, 75270 Paris, France and +IRIBHN, ULt3, Bruxelles, Belgium. In the liver, intercellular communication is a key process for ensuring a coherent organisation of cellular and whole organism functions. The stimulation of hepatocytes by glycogenolytic agonists induces intracellular Ca” ([Ca”]i) oscillations which are tightly coordinated when hepatocytes are connected by gap junctions (GJ). However, mechanisms ensuring this coordination are not precisely known and may involve au intercellufar diffusion of Cal’ and/or inositol 1.4.5 trisphosphate (Imp,) through GJ. The putative Ca-” intercellular diffusion was tested by microinjecting heparin in one cell of fun&loaded hepatocytes doublets. Heparin, which inhibits lnsPt-induced Cal* release, cannot diffuse through GJ. Addition of maximal concentration of vasopressin (Vp, 1OuM) or noradrenaline (Nor, ION) elicited a [Ca’*]i increase only in the non- injected cell, whereas [Ca’+]i remained at a low basal level in the heparin- injected cell. As revealed by fluorescence recovery after photobleaching WRAP) experiments, GJ permeability was not significantly affected in these conditions. InsP, intercellular diiusion was tested by microinjecting type I InsP, Sphosphatase in the intermediate cell of a triplet. This enzyme plays a major role in regulating InsP? levels and its over-expression in CHO cells inhibits CaZ+ movements (De Smcdt et al., (1997) J.Biol.Chem. 272, 17367-17375). Addition of Vp (O.lnM) or Nor (O.lpM) elicited [CY+]i oscillations only in the non-injected cells, and were not coordinated. These results confirm that Cti does not diffuse from cell to cell in detectable amounts during hormone stimulation and that intercellular transfer of lns4,, may be involved in the synchronization of agonist- induced [Ca’+]i oscillations. Finally, addition of low concentration of Vp (0.021&Q or Nor (0.05pM) to coMeCted cells, induced non coordinated [Cti]i oscillations which become coordinated at higher agonist concentrations. .4s the amplitude of [Ca’+]i oscillations does not depend on the level of InsP,, the mount of Ca2* diffusing through GJ remains the same for different levels of hormonal stimulation. Thus, [Ca’+]i oscillations are either coordinated, or do not occur at all because diffusion of InsP, through GJ is too low. These resuk lead us to the conclusion that InsP, diffusion through GJ play a predominant role in the coordination of Cb’ oscillations among connected hepauxytes. TIME-RESOLVED STORAGE AND RETRIEVAL OF CA” SIGNALS BY MITOCHONDRIA DEPENDS ON MATRIX CA** HOTSPOTS ICHAS FranGoi+ DE GIORGI Francesaab, DUVERT Michel”. MANNELLA Carmen A? MARK0 Michael COLONNA Raffaele’ and BERNARDI Paold * CNR Center for the Study of Biomembranes, De artment of Biomedical Sciences, University of Padova, Via/e G. Co/o mto3, I-35127 Padova, Hal b INSERM- l .9929, Mitochondrial Physiology, Victor Segalen-Bordeaux 2 University, F-33076 Bordeaux cedex, France c New York State Department of Health, Wadsworth Center, Division of Molecular Medicine, Albany NY12201-0509, USA Mitochondria actively contribute to intracellular Ca” signaling processes by taking uo. and subsenuentlv releasinn into the cvtosol lame amounts of Ca’*. We report here t&t the chronblogy’of mitociondrial C& uptake-is recorded by mitochot&ia and retrieved during Ca’+ release. Indeed, ca:’ is released from the organelle in a temporally-ordered manner tbat mirrors the chronology of Ca’+ uptake, cations accumulated first being released last. Such memorv oersists for minutes. indicating that mitochondrial C&is sorted and stored on a tern&&al basis, within an organize; submitochondrial compartment. We find that the latter is specificaily accessible to the hvdroohobic Ca” or&e chlortetracvcline. and that most of the Ca*’ taken &released b$ m&chondria. 1s directly chanieled &from this submitochondrial c&parunent capable of chronological Ca’* storage. Using ultrastructural imaging and spectroscopy, we find that the submitochondrial distribution of Cd’ is indeed uneven and that Ca” is stored at high concentration in specialized matrix regions appearing as dense granules of variable morphology, often associated with the inner mitochondrTa1 membrane. These resulls not only indicate that mitochondria can memorize the chronology of Ca” signals, but also show that Ca” is compartmentalized at high local concentration within the organelle. HIGH RESOLUTION STUDY OF MICROTUBULES MEURER-GROB P., THOUVENIN E., HEWAT E. and WADE R.H. lnstitut de Bioiogie Structurale JP EbeL 41, rue Jules Horowi&; 38027 Grenoble cedex 1, France Microtubules, as well as other filamentary structures, are elements of the cytoskebton. In association with the motor proteins, they play a major role in the organisation of the cytoplasm of most of the eucatyotic cells, and in cell division. Microtubules are formed from the tubulin heterodimer composed of two domains, u and p,of about 55kDa and JO A in diameter each. Tubulin assembles! via a GTP- induced process in a head to tail fashion to form protofilaments wluch associate laterally into hollow tubes of about 250 A in diameter and several micrometer lona: the microtubules. Tubulin assembly can be carried out in vitro in the presence, or not. of stabilisine agents such as the antimitotic drues tax01 and taxotere. or the non hydrolizablean~ogue of GTP, GMPCPP. This g&s rise to a variety of imtofilament numbers and different organisations of the tubulin heterodimer within the microtubular wall. In particular, microtubules can be strictly helical or the surface lattice can be interrupted by a packing discontinuity (Wade R.H., Chr&ien D., Job D. (1990) I Mol Biol 212, 775-786: Amal I. , Wade R.H. (1995) Current Biol 5,900. 908 ; Hyman A.A., Chr&ien D., Arnal I., Wade R.H. (1995) J Cell Biol 128, 117. 125). An image of a microtubule observed in vitreous ice by cryoelectron microscopy is a projection of this object. Hence, provided that the original object has a truly helical symmetry, it can be reconstructed in 3D fmm a singIe micrograph, using the helical reconstruction method. The same approach can be used to study the interaction between microtubules and motor proteins. Kinesins are excellent candidates to study this motor activity, since one single molecule is able to move along a microtubule and the motor domain of these aroteins has about 340 amino acids, a well adapted scale for this kind of study (also-compared to members of other famiiies like Si myosin and dynein, respectively three times and ten times bigger). These proteins can decorate the micmtubule surface lattice. and different states of the microtitbule/motor interaction can be ‘frozen’ and studied’in three dimensions as for microtubules alone (Amal I., Metoz F., DeBonis S.. Wade R.H. (1996). Current Biol 6, 1265-1170). There are many questions concerning the details of the interaction between the different constituents of these complexes. For example, if a-n and BP interactions are more stable than a-P, interactions, it is likely that the seam between two protofilaments will be a weak link in the micmtubule wall. Another question concerns microtubules with skewed protofilaments, which is the case with any protofilament numbers different from the ‘canonical’ 13. How does the energy increase due to wotofilament curvature comoare to the enerev saved bv .renairinn’ the mismatchedLmicrotubular lattice ? what ls the impact of6 diifere& st~bilisi~g molecules on these energies ? How do motor proteins interact with the tubulin, according to the different stages of ATP hydrolysis, used to fuel the movement along the microtubule ? X-ray and electron crystallography have provided atomic scale structures of the different constituents of these macromolecular complexes (Nogales E., Wolf S.G., Downing K.H. (1998). Nature 391, 199-203 ; Kozielski et al (1997) Cell 91, 985. 994). By docking these structures in the envelopes obtained by cryomicroscopy, an atomic model of the micmtubule, decorated or not, can be build, and these questions can be quantitatively addressed by a molecular dynamics approach. An important effort has ken made to improve the resolution of the reconstructions obtained by cryomicroscopy and data processing, so that the atomic scale structures fit tightly in the% envelopes, to give a precise idea of these interactions. A first structure of microtubule as!emble in the presence of GMPCPP baa been obtained at a resolution better than 17 A. The atomic structure of the tubulin dimer has been fitted in this envelope without ambiguity by a cross-correlation program developed in the laboratory. The interacting zones of the tubulin molecules can be clearly identified. The structural study of the microtubule/motor protein can be engaged now by applying the same techniques. 3e ColloqueSFp, Paris-Sud ‘99 LicolePofytechnique,Palaiseau, 28 juin2 juillet 1999