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

river sites where the width of the river was greater than 20 m, sampling was carried out by boat at three points, one in the middle and the other two a few metres from each bank. Otherwise a single point was sampled in the high-flow part of the river. The average sampling depth is 1.15 m SD = 0.6 (see S1 Table for details). The sampling points on the tributaries of the Beni and Tuichi rivers were carried out at least 200 m from their confluence. In lakes, sampling was car- ried out by boat and designed to integrate as much as possible spatial heterogeneity of eDNA distribution, as previously suggested in lentic ecosystems [32, 33]: collecting points were dis- tributed over the entire area of the lake (see S1 Table). The total number of sampling points per lake ranged from seven for the smallest lake, Ruta, to 90 for lake Tumichuca (S1 Table). The sampling kit is composed of a drill driver (AEG GmbH -BSB12cli202c), a peristaltic pump (Watson-Marlow 313-DW) in which the marpene suction hose was connected upstream to 5–6 m of a polyethylene suction hose and downstream to a polypropylene hose connection ensuring connection to the filter cartridge. Lead sinkers and a suction strainer (2 mm mesh) were added to the end of the suction pipe to ensure ballasting and prevent the entry of large particles into the cartridge. All materials were handled with sterile gloves and between each site, tubings, hose fittings and suction strainer were decontaminated in a 9% sodium hypochlorite solution for 10 min

[34, 35], then rinsed three times with commercial water. The driver unit was decontaminated using DNA away solution (Thermo Scientific™). In order to control possible contamination of equipment and cross contamination between sites, 7 negative control cartridges interspersed throughout the sampling campaign were made by filtering, for each cartridge, 20 l of commer- cial water in the field (see details in S1 Table). Each eDNA sample was taken by filtration of water pumped directly into the medium, applying an initial flow of approximately 1 l/min. Filtration was carried out until the filter was clogged and during the whole sampling phase the cartridges were protected from light by alu- minium foil. The final volume filtered for each cartridge varied between 6.5 and 75 l, depend- ing on the clogging speed as a function of the turbidity level at each site. Total volume per site filtered ranged from 11 to 180 l. At sampling completion, cartridges were filled with 150 ml of conservation buffer (Tris- HCl 0.1 M, EDTA 0.1 M, NaCl 0.01 M and N-lauroyl sarcosine 1%, pH 7.5–8) and stored at 4˚C until DNA extraction between 1–2 months after sampling. A total of 58 cartridges were processed, including seven negative field controls performed on average every 7 cartridges (SD = 2.8; commercial water filtered in the field) and one alien control consisting of four spe- cies (Alburnus alburnus, Carassius gibelio, Cyprinus carpio, Gambusia holbrooki) never recorded in our study area. This control was collected in an ornemental concrete pond at IRD

Montpellier (France), which contains only species not recorded in the Beni basin (Cyprini- formes and Cyprinodontiformes), then none should be detected in our samples if cross-con- tamination and hopping index are controlled. Laboratory analysis. DNA extraction. The DNA extractions and the PCR amplification steps were carried out in dedicated and physically separated rooms. To minimize the risk of cross contamination, all PCR steps were prepared under a workstation (GuardOne1 StarLab) allowing UV decontamination. Three independent extractions were performed for each cartridge as follows: 5.4 ml of a 2.5 M NaCl solution were added to each cartridge to obtain a concentration of 0.1 M and 3 g of SDS to a final concentration of 2%. After a 2 h 30 lysis at 56˚C with stirring, the cartridge was emptied into 10 tubes of 50 ml. An additional volume of 7.5 ml of 3M K acetate is added to precipitate the proteins. After centrifugation for 30 min at 4˚C at 6,200 g the supernatant was collected by pipetting to which 10 ml of isoamyl chloroform (24:1) was added, mixed by reverse (50 times), then centrifuged for 15 min at 15˚C and 6,000 g. The aqueous phase is PLOS ONE Fish diversity and distribution in the Tuichi basin PLOS ONE | https://doi.org/10.1371/journal.pone.0262357 January 4, 2022 5 / 25