Unveiling biogeographical patterns of the ichthyofauna in the Tuichi basin, a biodiversity hotspot in the Bolivian Amazon, using environmental DNA.

To date, more than 2400 valid fish species have been recorded in the Amazon basin. However, some regions remain poorly documented. This is the case in the Beni basin and in particular in one of its main sub-basins, the Tuichi, an Andean foothills rivers flowing through the Madidi National Park in the Bolivian Amazonia. The knowledge of its ichthyological diversity is, however, essential for the management and protection of aquatic ecosystems, which are threatened by the development of infrastructures (dams, factories and cities), mining and deforestation. Environmental DNA (eDNA) has been relatively little used so far in the Amazon basin. We sampled eDNA from water in 34 sites in lakes and rivers in the Beni basin including 22 sites in the Tuichi sub-basin, during the dry season. To assess the biogeographical patterns of the amazonian ichthyofauna, we implemented a metabarcoding approach using two pairs of specific primers designed and developed in our laboratory to amplify two partially overlapping CO1 fragments, one of 185bp and another of 285bp. We detected 252 fish taxa (207 at species level) among which 57 are newly identified for the Beni watershed. Species compositions are significantly different between lakes and rivers but also between rivers according to their hydrographic rank and altitude. Furthermore, the diversity patterns are related to the different hydro-ecoregions through which the Tuichi flows. The eDNA approach makes it possible to identify and complete the inventory of the ichthyofauna in this still poorly documented Amazon basin. However, taxonomic identification remains constrained by the lack of reference barcodes in public databases and does not allow the assignment of all OTUs. Our results can be taken into account in conservation and management strategies and could serve as a baseline for future studies, including on other Andean tributaries.

Species-level ichthyoplankton dynamics for 97 fishes in two major river basins of the Amazon using quantitative metabarcoding.

The Amazon basin holds the world's largest freshwater fish diversity. Information on the intensity and timing of reproductive ecology of Amazonian fish is scant. We use a metabarcoding method by capture using a single probe to quantify species-level ichthyoplankton dynamics. We sampled the Marañón and the Ucayali rivers in Peru monthly for 2 years. We identified 97 species that spawned mainly during the flood start, the flood end or the receding periods, although some species had spawning activity in more than one period. This information was new for 40 of the species in the Amazon basin and 80 species in Peru. Most species ceased spawning for a month during a strong hydrological anomaly in January 2016, demonstrating the rapidity with which they react to environmental modifications during the breeding season. We also document another unreported event in the Amazon basin, the inverse phenology of species belonging to one genus (Triportheus). Overall larval flow in the Marañón was more than twice that of the Ucayali, including for most commercial species (between two and 20 times higher), whereas the Ucayali accounts for ~80% of the fisheries landings in the region. Our results are discussed in the light of the main anthropogenic threats to fishes, hydropower dam construction and the Hidrovía Amazónica, and should serve as a pre-impact baseline.

Delimiting species by reproductive isolation: the genetic structure of epigean and hypogean Trichomycterus spp. (Teleostei, Siluriformes) in the restricted area of Torotoro (Upper Amazon, Bolivia).

Genetic variability of Trichomycterus from the region of Torotoro (Bolivia, Upper Amazon), distributed in the same watershed where the habitat is structured by waterfalls, canyons and a cave, was studied by allozyme (twelve putative loci) and RFLP-mtDNA (DLoop and cytochrome b) analyses. Alloenzymatic variation studied by Correspondence Analysis and Maximum Likelihood Analysis revealed a four-group structure, which was largely congruent with the distribution of the 14 mtDNA haplotypes. Two of these four clusters (I and II) were differentiated by two diagnostic loci (IDH and G3PDH), two semi-diagnostic loci (PGM and 6PGDH) and consequently a very high F(st )value (estimator theta = 0.77). Therefore, clusters I and II are reproductively isolated. The distribution limit of these two (sibling) species does not correspond to those of the morphological species of Trichomycterus identified in this region: the epigean T. cf. barbouri and the hypogean T. chaberti. However, hypogean fish exhibited two mtDNA haplotypes, a private one and another shared with the epigean Trichomycterus from upstream reaches.