Spatial pattern of floral morphology: a possible insight into the effects of pollinators on plant distribution

Pollination syndromes involve convergent evolution towards phenotypes composed of specific scents, colours or floral morphologies that attract or restrict pollinator access to reward. How these traits might influence the distributions of plant species in interaction with pollinators has rarely been investigated. We sampled 870 vegetation plots in the western Swiss Alps and classified the plant species into seven blossom types according to their floral morphology (wind, disk, funnel, tube, bilabiate, head or brush). We investigated the environmental features of plots with functional diversity (FD) lower than expected by chance alone to detect potential pollination filtering and related the proportions of the seven blossom types to a combination of environmental descriptors. From these results, we inferred the potential effect of the pollinator on the spatial distribution of plant species. The vegetation plots with significantly lower FD of blossom types than expected by chance were found at higher altitudes, and the proportions of blossom types were strongly patterned along the same gradient. These results support a biotic filtering effect on plant species assemblages through pollination: disk blossoms became dominant at higher altitudes, resulting in a lower FD. In harsh conditions at high altitudes, pollinators usually decrease in activity, and the openness of the disk blossom grants access to any available pollinator. Inversely, bilabiate blossoms, which are mostly pollinated by bees, were more abundant at lower elevations, which are characterised by greater abundance and diversity of bees. Generalisation through openness of the blossom could be advantageous at high elevations, while specialisation could be a successful alternative strategy at lower elevations. The approach used in this study is purely correlative, and further investigations should be conducted to infer the nature of the causal relationship between plant and pollinator distributions.

The TetR-type MfsR protein of the integrative and conjugative element (ICE) ICEclc controls both a putative efflux system and initiation of ICE transfer.

Integrative and conjugating elements (ICE) are self-transferable DNAs widely present in bacterial genomes, which often carry a variety of auxiliary genes of potential adaptive benefit. One of the model ICE is ICEclc, an element originally found in Pseudomonas knackmussii B13 and known for its propensity to provide its host with the capacity to metabolize chlorocatechols and 2-aminophenol. In this work, we studied the mechanism and target of regulation of MfsR, a TetR-type repressor previously found to exert global control on ICEclc horizontal transfer. By using a combination of ICEclc mutant and transcriptome analysis, gene reporter fusions, and DNA binding assays, we found that MfsR is a repressor of both its own expression and that of a gene cluster putatively coding for a major facilitator superfamily efflux system on ICEclc (named mfsABC). Phylogenetic analysis suggests that mfsR was originally located immediately adjacent to the efflux pump genes but became displaced from its original cis target DNA by a gene insertion. This resulted in divergence of the original bidirectional promoters into two separated individual regulatory units. Deletion of mfsABC did not result in a strong phenotype, and despite screening a large number of compounds and conditions, we were unable to define the precise current function or target of the putative efflux pump. Our data reconstruct how the separation of an ancestor mfsR-mfsABC system led to global control of ICEclc transfer by MfsR.

Chronic malnutrition favours smaller critical size for metamorphosis initiation in Drosophila melanogaster.

Critical size at which metamorphosis is initiated represents an important checkpoint in insect development. Here, we use experimental evolution in Drosophila melanogaster to test the long-standing hypothesis that larval malnutrition should favour a smaller critical size. We report that six fly populations subject to 112 generations of laboratory natural selection on an extremely poor larval food evolved an 18% smaller critical size (compared to six unselected control populations). Thus, even though critical size is not plastic with respect to nutrition, smaller critical size can evolve as an adaptation to nutritional stress. We also demonstrate that this reduction in critical size (rather than differences in growth rate) mediates a trade-off in body weight that the selected populations experience on standard food, on which they show a 15-17% smaller adult body weight. This illustrates how developmental mechanisms that control life history may shape constraints and trade-offs in life history evolution.

Individual caspase-10 isoforms play distinct and opposing roles in the initiation of death receptor-mediated tumour cell apoptosis.

The cysteine protease caspase-8 is an essential executioner of the death receptor (DR) apoptotic pathway. The physiological function of its homologue caspase-10 remains poorly understood, and the ability of caspase-10 to substitute for caspase-8 in the DR apoptotic pathway is still controversial. Here, we analysed the particular contribution of caspase-10 isoforms to DR-mediated apoptosis in neuroblastoma (NB) cells characterised by their resistance to DR signalling. Silencing of caspase-8 in tumour necrosis factor-related apoptosis-inducing ligand (TRAIL)-sensitive NB cells resulted in complete resistance to TRAIL, which could be reverted by overexpression of caspase-10A or -10D. Overexpression experiments in various caspase-8-expressing tumour cells also demonstrated that caspase-10A and -10D isoforms strongly increased TRAIL and FasL sensitivity, whereas caspase-10B or -10G had no effect or were weakly anti-apoptotic. Further investigations revealed that the unique C-terminal end of caspase-10B was responsible for its degradation by the ubiquitin-proteasome pathway and for its lack of pro-apoptotic activity compared with caspase-10A and -10D. These data highlight in several tumour cell types, a differential pro- or anti-apoptotic role for the distinct caspase-10 isoforms in DR signalling, which may be relevant for fine tuning of apoptosis initiation.

Feedback control of DnaA-mediated replication initiation by replisome-associated HdaA protein in Caulobacter.

Chromosome replication in Caulobacter crescentus is tightly regulated to ensure that initiation occurs at the right time and only once during the cell cycle. The timing of replication initiation is controlled by both CtrA and DnaA. CtrA binds to and silences the origin. Upon the clearance of CtrA from the cell, the DnaA protein accumulates and allows loading of the replisome at the origin. Here, we identify an additional layer of replication initiation control that is mediated by the HdaA protein. In Escherichia coli, the Hda protein inactivates DnaA after replication initiation. We show that the Caulobacter HdaA homologue is necessary to restrict the initiation of DNA replication to only once per cell cycle and that it dynamically colocalizes with the replisome throughout the cell cycle. Moreover, the transcription of hdaA is directly activated by DnaA, providing a robust feedback regulatory mechanism that adjusts the levels of HdaA to inactivate DnaA.

A spatial accommodation by neighboring cells is required for organ initiation in Arabidopsis.

Lateral root formation in plants can be studied as the process of interaction between chemical signals and physical forces during development. Lateral root primordia grow through overlying cell layers that must accommodate this incursion. Here, we analyze responses of the endodermis, the immediate neighbor to an initiating lateral root. Endodermal cells overlying lateral root primordia lose volume, change shape, and relinquish their tight junction-like diffusion barrier to make way for the emerging lateral root primordium. Endodermal feedback is absolutely required for initiation and growth of lateral roots, and we provide evidence that this is mediated by controlled volume loss in the endodermis. We propose that turgidity and rigid cell walls, typical of plants, impose constraints that are specifically modified for a given developmental process.

The major transcription initiation site of the SV40 late promoter is a potent thyroid hormone response element.

Thyroid hormone receptors (TRs) are members of the nuclear hormone receptor superfamily, which act as transcription factors upon binding to specific DNA sequences called thyroid hormone (T3) response elements (TREs). Such elements are found in the upstream regulatory region of promoters as well as in intragenic sequences of T3-responsive genes. In this report, we demonstrate that SV40 late (SVL) promoter activity is strongly down-regulated by TR in the absence of ligand. Addition of T3 releases this repression, but does not further induce SVL promoter activity. Electrophoretic mobility shift analyses reveal a TR binding element that overlaps with the SV40 major late transcription initiation site. This element closely fits the consensus TRE, formed of two hexanucleotides organized in a tandem repeat separated by 4 nt, and is able to confer T3 responsiveness on a heterologous promoter. We further show that, although the presence of TR leads to quantitatively modified expression of an SVL-driven reporter gene, neither displacement of the site of transcription initiation nor modification of the splicing pattern of the primary transcripts occur.

Induced down-regulation of neuropeptide Y-Y1 receptors delays initiation of kindling.

Neuropeptide Y appears to modulate epileptic seizures differentially according to the receptor subtypes involved. In the hippocampus, neuropeptide Y expression and release are enhanced in different models of epileptogenesis. On the contrary, the expression of Y1 receptors is decreased and it has been shown that activation of these receptors has pro-convulsant effects. The aim of our study was to investigate the role of Y1 receptors during hippocampal kindling epileptogenesis using (i) knock-out mice lacking Y1 receptors and (ii) intrahippocampal infusion of Y1 antisense oligodeoxynucleotide in rats. Y1 knock-out mice showed similar susceptibility to seizure induction and presented no difference in kindling development as compared with their control littermates. Conversely, local hippocampal down-regulation of Y1 receptors during the first week of hippocampal kindling, induced by a local infusion of a Y1 antisense oligodeoxynucleotide, significantly increased seizure threshold intensity and decreased afterdischarge duration. A reverse effect was observed during the week following the infusion period, which was confirmed by a significant decrease in the number of hippocampal stimulations necessary to evoke generalized seizures. At the end of this second week, an up-regulation of Y1 receptors was observed in kindled rats infused with the antisense as compared with the mismatch-treated controls. Our results in the rat suggest that the down-regulation of Y1 receptors in the hippocampus participates in the control of the initiation of epileptogenesis. The lack of an effect of the deficiency of Y1 receptors in the control of kindling development in Y1 knock-out mice could be due to compensatory mechanisms.

Initiation de l'adhésion leucocytaire : rôle des fucosyltransférases et des ligands des sélectines

RESUME La dissémination extramédullaire des cellules blastiques est une complication majeure des leucémies myéloïdes (LMA) ou lymphoïdes aiguës (LLA). La migration des cellules blastiques dépend de mécanismes semblables à ceux qui régulent la migration des leucocytes dans un site d'inflammation. Parmi ceux-ci, les oligosaccharides fucosylés décorant les ligands des sélectines jouent un rôle clé en interagissant avec les sélectines. PSGL-1 (P-Selectin Glycoprotein Ligand-1) est une protéine de 240 kD, exprimée à la surface des leucocytes, permettant de soutenir le roulement leucocytaire sur les sélectines, le long de la paroi vasculaire. L'interaction de PSGL-1 avec les sélectines nécessite des modifications post-traductionnelles de type sialylation, sulfatation , N et 0-glycosylation. Parmi les enzymes impliqués, les α1,3-fucosyltransférases jouent un rôle important dans la biosynthèse d'oligosaccharides fucosylés, ligands des sélectines (sLex, Lex, VIM-2, CLA). Comme l'expression des α1,3-fucosyltransférases par les cellules blastiques leucémiques n'a pas été étudiée précédemment, nous l'avons recherchée dans 120 cas de leucémies aiguës. Les ARNm des FucT-IV et -VII ont été détectés, par RT-PCR, dans tous les cas testés. L'ARNm de la FucT-IX n'a été observé que dans 40% des leucémies aiguës (48/120). L'ARNm de la FucT-IX est détecté dans 65% des LMA (47/72) et, moins fréquemment, dans 26% des LLA (11/42). A noter que les cas de LLA exprimant la FucT-IX correspondent essentiellement à des LLA secondaires à la transformation d'une leucémie myéloïde chronique ou des LLA de la lignée B de type leucémie/lymphome de Burkitt. L'expression de PSGL-1 et des oligosaccharides fucosylés par les blastes varie significativement parmi les LMA et les LLA : Lex, VIM-2 et sLex étant exprimés plus fréquemment par les myéloblastes que par les lymphoblastes. Le rôle des FucT-IV, -VII et -IX dans la synthèse des Lex, VIM-2, CLA et sLex a été examiné en exprimant l'ADNc de chaque FucT dans des cellules CHO. L'immunophénotypisation des transfectants indique que la FucT-VII synthétise sLex et CLA, mais pas Lex et VIM-2. Lex et VIM-2 sont générés par la FucT-IV. La FucT-IX ne participe qu'à la synthèse de Lex, sa capacité de synthèse de VIM-2 dans les cellules CHO est très faible. Le rôle de la FucT-IX dans la régulation du roulement cellulaire dépendant des sélectines a été testé dans des conditions de flux. Les vitesses de roulement des cellules CHO co-exprimant la FucT-LX, la core-2 01,6-N-acetylglucosaminyltransferase et PSGL-1 sont très élevées sur la P-sélectine (médiane : 497.95 µm/s, n=96) alors qu'elles sont beaucoup plus lentes sur la E-sélectine (médiane 7 µm/s, n=64). Les recrutements sur la E-sélectine des cellules CHO-C2F9PSGL¬1 et des CHO-C2F7PSGL-1 sont similaires (moyenne ± SEM : 127.44 ± 4.38 vs. 151.16 ± 3.16 cellules/min/mm2, n=5). Celui des cellules CHO-C2F4PSGL-1 est par contre plus faible (54.20 ± 2.13 cellules/min/mm2, n=5). Ces résultats indiquent que la FucT-IX est impliquée dans la biosynthèse de Lex, VIM-2 et CLA et qu'elle régule l'interaction des cellules CHO avec la E-sélectine. Contrairement aux FucT-IV et -VII, la FucT-IX ne joue qu'un rôle mineur dans la régulation du roulement cellulaire sur la L- et la P-sélectine. L'expression fréquente de la FucT-IX par les myéloblastes suggère qu'elle pourrait participer avec les FucT-IV et -VII à la régulation de la migration cellulaire dépendant de la E-sélectine. Finalement, ce travail de thèse a été étendu à l'identification des protéines cytoplasmiques qui interagissent avec le domaine cytoplasmique de PSGL-1 et qui pourraient être impliquées dans la transmission de signaux intracellulaires. Les ligands intracellulaires de PSGL-1 seront identifiés par la technique du double hybride qui nous a déjà permis de confirmer que syk et la N-moésine se lient au domaine cytoplasmique de PSGL-1. Des ligands supplémentaires seront identifiés employant une librairie provenant des cellules souches hématopoïétiques comme proie. ABSTRACT Blast cell dissemination is a major complication of acute myeloblastic (AML) and lymphoblastic leukemia (ALL). Blast cell migration is dependent on mechanisms that are similar to those which regulate leukocyte migration into inflammatory lesions. Among them, fticosylated oligosaccharides that decorate selectin ligands play a key role by interacting with selectins. PSGL-1 (P-Selectin Glycoprotein Ligand-1) is a 240 kD glycoprotein constitutively expressed on leucocytes and which supports leukocyte rolling on selectins. PSGL-1 interaction with selectins is dependent on post-translational modifications such as sialylation, sulfation, N- and 0-glycosylation. Among the involved enzymes, the α1,3-fucosyltransferases (FucT) play a major role in generating cell surface glycoconjugates carrying fucosylated oligosaccharides which interact with selectins (sLex, Lex, VIM-2, CLA). Since no information is available on the expression of α1,3-fucosyltransferases by leukemic blast cells, we examined it in 120 cases of acute leukemia. FucT-IV and -VII mRNAs were detected, by RT-PCR, in all tested cases. In contrast, the presence of FucT-IX mRNA was shown in only 40% of patients with acute leukemia (48/120). FucT-IX mRNA was detected in 65% of AML (47/72) and, less frequently, in 26% of ALL (11/42). Importantly, all ALL cases expressing FucT-IX were either secondary leukemia resulting from the transformation of chronic myelocytic leukemia in acute lymphoblastic leukemia or mature B-ALL (FAB L3 subtype or Burkitt lymphoma/leukemia according to WHO classification). FucT-IX was not detected in precursor B or T-ALL. The expression of PSGL-1 and fucosylated epitopes was significantly different among AML and ALL, Lex, VIM-2 and sLex being more frequently expressed by myeloblasts than by lymphoblasts. The role of FucT-IV, -VII and -IX in the biosynthesis of Lex, VIM-2, CLA and sLex was examined by expressing the cDNA of each α1,3-FucT in CHO cells. Immunophenotypic analysis of CHO transfectants indicated that FucT-VII synthesizes sLex and CLA but not Lex or VIM-2. Lex and CLA were generated by both FucT-IV and -IX. FucT-IV and FucT-IX differed in their ability to synthesize VIM-2, FucT-IX being less efficient than FucT-IV. The role of FucT-IX in regulating selectin-dependent rolling was assessed under hydrodynamic flow conditions. P-selectin-dependent interactions were transient and occurred at high velocities (median: 497.95 1,µm/s, n=96). In contrast, much slower rolling velocities were observed on E-selectin (median: 7 µm/s, n=64). The recruitment of CHO-C2F9PSGL-1 and CHO-C2F7PSGL-1 cells was similar on E-selectin (mean ± SEM: 127.44 ± 4.38, n=5 vs 151.16 ± 3.16 cells/min/mm2, n=5). In the other hand, CHO-C2F4PSGL-1 cells were less efficiently recruited on E-selectin (54.20 ± 2.13 cells/min/mm2, n=5). This results indicate that FucT-IX is involved in the biosynthesis of Lex, VIM-2 and CLA and that it confers E-selectin binding activity to CHO cells. By contrast to FucT-IV and -VII, FucT-IX had a minor role in regulating P- and L-selectin-dependent rolling on CHO transfectants. The frequent expression of FucT-IX in myeloblasts suggests that it may participate with FucT-IV and -VII in regulating E-selectin-dependent cell migration into tissues. Finally, this thesis work was extended to the identification of the cytoplasmic proteins interacting with cytoplasmic domain of PSGL-1 that may be involved in transducing intracellular signals. We planned to identify these intracellular ligands of PSGL-1 by using the double hybrid technique and already confirmed that syk and N-moesin bind to the cytoplasmic domain of PSGL-1. Additional PSGL-1 ligands will be sought by the same technique using a CD34+ stem cell library as pray. RESUME DESTINE A UN LARGE PUBLIC : L'adhésion et la migration leucocytaire sont nécessaires à de nombreux processus cellulaires comme la régulation de l'hématopoïèse, mais aussi dans la pathogenèse de l'artériosclérose, des maladies inflammatoires et de la métastatisation des cellules cancéreuses. Les molécules impliquées constituent depuis peu des cibles pour la thérapie du cancer. La migration leucocytaire vers un site d'inflammation dépend de mécanismes complexes, se déroulant en plusieurs étapes, nécessitant l'interaction séquentielle de molécules d'adhésion leucocytaires et endothéliales. Ainsi, chronologiquement, suite à un stimulus inflammatoire, les leucocytes « roulent » sur les cellules endothéliales, sont activées, s'arrêtent et traversent la paroi endothéliale (diapédèse) pour migrer dans les tissus environnants inflammés selon un gradient chimiotactique. La première étape de roulement met en jeu deux molécules principales : PSGL-1 (P-Sélectine Glycoprotéine Ligand-1) du coté des leucocytes et les sélectines du coté de l'endothélium de la paroi vasculaire. L'interaction entre ces deux molécules nécessite des décorations de ces protéines par des sucres, des résidus sulfates et des acides sialiques. Le sucre essentiel à la liaison demeure le fucose qui est attaché aux protéines grâce à des enzymes de la famille des fucosyltransferases. Actuellement, neuf fucosyltransférases humaines ont été identifiées et désignées sous FucT-I à IX. La FucT-IX, dernière fucosyltransférase clonée, a un faible degré d'homologie avec les autres fucosyltransférases mais sa séquence est extrêmement conservée entre les espèces. Ceci traduit son importance par une forte résistance à la pression évolutive. L'examen de son expression au sein de 120 cas de leucémies aiguës a mis en évidence son comportement atypique. En effet, alors que les autres FucTs sont toujours présentes, la FucT¬IX ne s'exprime que dans un cas sur deux en moyenne avec une préférence plus importante pour les leucémies myéloïdes. Ainsi, une étude plus approfondie de cet enzyme à mis en évidence sa capacité à induire une interaction cellulaire plus spécifique de la E-sélectine. Elle décore non seulement des protéines de surface, mais aussi certainement les glycolipides constituant la membrane cellulaire.

Sequence homologies in the region preceding the transcription initiation site of the liver estrogen-responsive vitellogenin and apo-VLDLII genes.

In the liver of oviparous vertebrates vitellogenin gene expression is controlled by estrogen. The nucleotide sequence of the 5' flanking region of the Xenopus laevis vitellogenin genes A1, A2, B1 and B2 has been determined. These sequences have been compared to each other and to the equivalent region of the chicken vitellogenin II and apo-VLDLII genes which are also expressed in the liver in response to estrogen. The homology between the 5' flanking region of the Xenopus genes B1 and B2 is higher than between the corresponding regions of the other closely related genes A1 and A2. Four short blocks of sequence homology which are present at equivalent positions in the vitellogenin genes of both Xenopus laevis and chicken are characterized. A short sequence with two-fold rotational symmetry (GGTCANNNTGACC) was found at similar positions upstream of the five vitellogenin genes and is also present in two copies close to the 5' end of the chicken apo-VLDLII gene. The possible functional significance of this sequence, common to liver estrogen-responsive genes, is discussed.

Late Cretaceous to Eocene geology of the South Central American forearc area (southern Costa Rica and western Panama) : initiation and evolution of an intra-oceanic convergent margin

Résumé : L'arc volcanique du sud de l'Amérique Centrale se situe sur la marge SW de la Plaque Caraïbe, au-dessus des plaques subduites de Cocos et Nazca. Il s'agit de l'un des arcs intra-océaniques les plus étudiés au monde, qui est généralement considéré comme s'étant développé à la fin du Crétacé le long d'un plateau océanique (le Plateau Caraïbe ou CLIP) et se trouvant actuellement dans un régime de subduction érosive. Au cours des dernières décennies, des efforts particuliers ont été faits pour comprendre les processus liés à la subduction sur la base d'études géophysiques et géochimiques. Au sud du Costa Rica et à l'ouest du Panama, des complexes d'accrétions et structures à la base de l'arc volcanique ont été exposés grâce à la subduction de rides asismiques et de failles transformantes. Des affleurements, situés jusqu'à seulement 15 km de la fosse, offrent une possibilité unique de mieux comprendre quelques uns des processus ayant lieu le long de la zone de subduction. Nous présentons de nouvelles contraintes sur l'origine de ces affleurements en alliant une étude de terrain poussée, de nouvelles données géochimiques, sédimentaires et paléontologiques, ainsi que des observations structurales effectuées en télédétection. Une nouvelle stratigraphie tectonique entre le Campanien et l'Éocène est définie pour la région d'avant-arc située entre la Péninsule d'Osa (Costa Rica) et la Péninsule d'Azuero (Panama). Nos résultats montrent que la partie externe de la marge est composée d'un arrangement complexe de roches ignées et de séquences sédimentaires de recouvrement qui comprennent principalement le socle de l'arc, des roches d'arc primitif, des fragments de monts sous-marins accrétés et des mélanges d'accrétion. Des preuves sont données pour le développement de l'arc volcanique du sud de l'Amérique Centrale sur un plateau océanique. Le début de la subduction le long de la marge SW de la Plaque Caraïbe a eu lieu au Campanien et a généré des roches d'arc primitif caractérisées par des affinités géochimiques particulières, globalement intermédiaires entre des affinités de plateau et d'arc insulaire. L'arc était mature au Maastrichtien et formait un isthme essentiellement continu entre l'Amérique du Nord et l'Amérique du Sud. Ceci a permis la migration de faunes terrestres entre les Amériques et pourrait avoir contribué à la crise fin Crétacé -Tertiaire en réduisant les courants océaniques subéquatoriaux entre le Pacifique et l'Atlantique. Plusieurs unités composées de fragments de monts sous-marins accrétés sont définies. La nature et l'arrangement structural de ces unités définissent de nouvelles contraintes sur les modes d'accrétion des monts sous-marins/îles océaniques et sur l'évolution de la marge depuis la formation de la zone de subduction. Entre la fin du Crétacé et l'Éocène moyen, la marge a enregistré plusieurs épisodes ponctuels d'accrétion de monts sous-marins alternant avec de la subduction érosive. A l'Éocène moyen, un événement tectonique régional pourrait avoir causé un fort couplage entre les plaques supérieure et inférieure, menant à des taux plus important d'accrétion de monts sous-marins. Durant cette période, la situation le long de la marge était très semblable à la situation actuelle et caractérisée par la présence de monts sous-marins subductants et l'absence d'accrétion de sédiments. L'enregistrement géologique montre qu'il n'est pas possible d'attribuer une nature érosive ou accrétionnaire à la marge dans le passé ou -par analogie- aujourd'hui, parce que (1) les processus d'accrétion et érosifs varient fortement spatialement et temporellement et (2) il est impossible d'évaluer la quantité exacte de matériel tectoniquement enlevé à la marge depuis le début de la subduction. Au sud du Costa Rica, certains fragments de monts sous-marins accrétés sont représentatifs d'une interaction entre une ride et un point chaud dans le Pacifique au Crétacé terminal/Paléocène. L'existence de ces fragments de monts sous-marins et la morphologie du fond de l'Océan Pacifique indiquent que la formation de la ride de Cocos-Nazca s'est formée au moins ~40 Ma avant l'âge proposé par les modèles tectoniques actuels. Au Panama, nous avons identifié une île océanique d'âge début Éocène qui a été accrétée à l'Éocène moyen. L'accrétion a eu lieu à très faible profondeur par détachement de l'île dans la fosse, et a mené à une exceptionnelle préservation des structures volcaniques. Des affleurement comprenant aussi bien des parties basses et hautes de l'édifice volcanique on été étudiées, depuis la phase sous-marine bouclier jusqu'à la phase subaérienne post-bouclier. La stratigraphie nous a permis de différencier les laves de la phase sous-marine de celles de la phase subaérienne. La composition des laves indique une diminution progressive de l'intensité de la fusion partielle de la source et une diminution de la température des laves produites durant les derniers stades de l'activité volcanique. Nous interprétons ces changements comme étant liés à l'éloignement progressif de l'île océanique de la zone de fusion ou point chaud. Abstract The southern Central American volcanic front lies on the SW edge of the Caribbean Plate, inboard of the subducting Cocos and Nazca Plates. It is one of the most studied intra-oceanic convergent margins around the world, which is generally interpreted to have developed in the late Cretaceous along an oceanic plateau (the Caribbean Large Igneous Province or CLIP) and to be currently undergoing a regime of subduction erosion. In the last decades a particular effort has been made to understand subduction-related processes on the basis of geophysical and geochemical studies. In southern Costa Rica and western Panama accretionary complexes and structures at the base of the volcanic front have been exposed in response to subduction of aseismic ridges and transforms. Onland exposures are located as close as to 15 km from the trench and provide a unique opportunity to better understand some of the processes occurring along the subduction zone. We provide new constraints on the origins of these exposures by integrating a comprehensive field work, new geochemical, sedimentary and paleontological data, as well as structural observations based on remote imaging. A new Campanian to Eocene tectonostratigraphy is defined for the forearc area located between the Osa Peninsula (Costa Rica) and the Azuero Peninsula (Panama). Our results show that the outer margin is composed of a complicated arrangement of igneous complexes and overlapping sedimentary sequences that essentially comprise an arc basement, primitive island-arc rocks, accreted seamount fragments and accretionary mélanges. Evidences are provided for the development of the southern Central American arc on the top an oceanic plateau. The subduction initiation along the SW edge of the Caribbean Plate occurred in the Campanian and led to formation of primitive island-arc rocks characterized by unusual geochemical affinities broadly intermediate between plateau and arc affinities. The arc was mature in the Maastrichtian and was forming a predominantly continuous landbridge between the North and South Americas. This allowed migration of terrestrial fauna between the Americas and may have contributed to the Cretaceous-Tertiary crisis by limiting trans-equatorial oceanic currents between the Pacific and the Atlantic. Several units composed of accreted seamount fragments are defined. The nature of the units and their structural arrangement provide new constraints on the modes of accretion of seamounts/oceanic islands and on the evolution of the margin since subduction initiation. Between the late Cretaceous and the middle Eocene, the margin recorded several local episodes of seamount accretion alternating with tectonic erosion. In the middle Eocene a regional tectonic event may have triggered strong coupling between the overriding and subducting plates, leading to higher rates of seamount accretion. During this period the situation along the margin was very similar to the present and characterized by subducting seamounts and absence of sediment accretion. The geological record shows that it is not possible to ascribe an overall erosive or accretionary nature to the margin in the past and, by analogy, today, because (1) accretionary and erosive processes exhibit significant lateral and temporal variations and (2) it is impossible to estimate the exact amount of material tectonically eroded from the margin since subduction initiation. In southern Costa Rica, accreted seamount fragments point toward a plume-ridge interaction in the Pacific in the late Cretaceous/Paleocene. This occurrence of accreted seamount fragments and morphology of the Pacific Ocean floor is indicative of the formation of the Cocos-Nazca spreading system at least ~40 Ma prior to the age proposed in current tectonic models. In Panama, we identified a remarkably-well preserved early Eocene oceanic island that accreted in the middle Eocene. The accretion probably occurred at very shallow depth by detachment of the island in the trench and led to an exceptional preservation of the volcanic structures. Exposures of both deep and superficial parts of the volcanic edifice have been studied, from the submarine-shield to subaerial-postshield stages. The stratigraphy allowed us to distinguish lavas produced during the submarine and subaerial stages. The lava compositions likely define a progressive diminution of source melting and a decrease in the temperature of erupted melts in the latest stages of volcanic activity. We interpret these changes to primarily reflect the progressive migration of the oceanic island out of the melting region or hotspot.