Votre navigateur n’affiche pas l’aperçu PDF. Ouvrir le PDF →

Résumé

The study of food webs and their dynamics is fundamental to understand how the feeding habits of the different species can affect the community, thus improving our understanding of the functioning of the ecosystem as a whole. Furthermore, the study of feeding ecology becomes crucial when it concerns endangered species since a precise knowledge of their diet is to be gathered when designing reliable conservation strategies. A wide range of methodologies have been proposed for diet analysis, including simple ones, as visual observation of foraging behavior, and more complex ones such as Near Infrared Reflectance Spectroscopy and DNA based methods. DNA barcoding, i.e. species identification using a standardized DNA region or markers, has recently received much attention and is being further developed through an international initiative called "Consortium for the Barcode of Life". When using DNA barcoding for diet analysis, the choice of the markers is crucial. The ideal DNA barcoding marker should meet several criteria. It should be variable among species, standardized, with enough phylogenetic information, extremely robust, and short enough to allow amplification of degraded DNA. In this study we propose the trnL (UAA) intron as marker for plant DNA barcoding. The power and the limitations of this system were evaluated as well as the possibility of species identification with highly degraded DNA. The main limitation of this system is its relatively low resolution in discriminating closely related species. Despite the relatively low resolution, it has many advantages: the primers are highly conserved, the amplification system is very robust and it is able to work with much degraded DNA samples. This system has been coupled with massively parallel pyrosequencing technique. We demonstrate the efficiency of this new approach by analyzing the diet of various herbivorous species. The whole chloroplast trnL (UAA) intron (254–767 bp) and a shorter fragment of this intron (the P6 loop, 10–143 bp) were used in this study. For the whole trnL intron 67.3% of the species retrieved from GenBank were unambiguously identified and 19.5% for the P6 loop. The resolution is much higher after calibration of specific contexts using species originating from the same ecosystem. Furthermore, the trnL approach was coupled with individual and sex identification using microsatellites polymorphism in the Himalayan brown bear (Ursus arctos isabellinus). Among world brown bears populations; those in Asia are the most endangered and least studied. Here, populations have declined by more than half in the past century owing to habitat loss and fragmentation and human activity. Presently in Pakistan brown bear occur sparsely in seven small populations, with the largest isolate in the Deosai National Park. We examined this population using a combination of fécal DNA analysis and field data for which geographical location and date of sampling were available, with the aim to study individual and sexual differentiation in the diet, and also temporal and geographical variations. Twenty-eight individuals (16 male, 10 females and 2 unknown sex) were identified in this study with microsatellites markers. Only eight plant species were found represented in more than 50% of individual feces. Temporal differences were found with more energetic food detected before the hibernation periods. L'étude des réseaux trophiques et leur dynamique est fondamentale pour comprendre comment les habitudes alimentaires des différentes espèces peuvent influencer la communauté, afin d'améliorer notre compréhension du fonctionnement de l'écosystème dans son ensemble. En outre, l'étude de l'écologie alimentaire devient cruciale lorsqu'il s'agit des espèces en voie de disparition, une connaissance précise de leur alimentation doit être acquise lors de la conception des stratégies de conservation. Un large éventail de méthodes a été proposé pour l’analyse du régime alimentaire, y compris