An adaptable mesocosm platform for performing integrated assessments of
nanomaterial risk in complex environmental systems
Mélanie Auffan1,2,3*, Marie Tella1,2, Catherine Santaella2,4, Lenka Brousset2,5, Christine
Paillès1,2, Mohamed Barakat2,4, Benjamin Espinasse2,3, Ester Artells2,5, Julien Issartel2,5,
Armand Masion1,2, Jérôme Rose1,2,3, Mark R. Wiesner2,3, Wafa Achouak2,4, Alain Thiéry2,5,
Jean-Yves Bottero1,2,3
1
CNRS, Aix-Marseille Université, CEREGE UM34, UMR 7330, 13545 Aix en Provence,
France.
2
International Consortium for the Environmental Implications of Nanotechnology iCEINT,
Aix en Provence, France.
3
Center for the Environmental Implications of NanoTechnology CEINT, Duke University,
Durham, North Carolina 27707, USA.
4
CNRS, Aix-Marseille Université, CEA Cadarache DSV / IBEB / SBVME, Lab Ecol Microb
Rhizosphere & Environ Extrem (LEMiRE), UMR 7265, Saint Paul lez Durance, France.
5
CNRS, Aix-Marseille Université, CNRS, Institut Méditerranéen de Biodiversité et
d’Ecologie marine et continentale (IMBE), UMR 7263, Marseille, France
Corresponding author: Melanie Auffan [email protected]
Water composition
Volvic® water composition: pH 7, 11.5 mg/L Ca2+, 13.5 mg/L Cl-, 71 mg/L HCO3-, 8 mg/L
Mg2+, 6.3 mg/L NO3-, 6.2 mg/L K+, 11.6 mg/L Na+
Pond water characteristic: pH 7, 413 S/cm, 9.5 mg/L dissolved O2,
Determination of the concentration of picoplankton and algae
Picoplankton and algae concentrations are determined at the surface of the sediment (0.5 ± 0.1
mm depth) and in the water column (10 cm below the surface close to the area where water
was recycled) on a weekly basis. Five mL of water and 15 mL of sediment are sampled,
treated with formaldehyde (3.7%), and stored at 4°C before counting.
Before picoplankton counting, 1 mL of each water column sample was centrifuged (5.9  g at
4°C for 15 min), and 200 µL of each sediment sample was treated with 800 µL of 0.1 mM
sterile tetrasodium pyrophosphate and vortexed with a steel ball for 30 seconds. For the
counting, 10 µL of each sample was mixed with 5µL of 3µM SYTO® 9 Green Fluorescent
Nucleic Acid Stain and dropped on a glass slide. Concentration of picoplankton was the mean
± standard deviation of five counts.
Of the 15 mL sampled sediment, 50 µL were subsampled, diluted with distilled water (2450
µL) and sonicated for few seconds for algae counting. An aliquot of the suspension was laid
on a Nageotte counting cell and diatom and other micro-algae were counted at X400
magnification using an Olympus-CHS where only alive individuals with chloroplasts were
counted along 3 transects. Five mL water samples were centrifuged for 10 min at 2 863  g.
Four mL of supernatant were removed. An aliquot of the remaining concentrated water
sample laid on a Nageotte counting cell. Diatoms and other algae in the water column were
counted following the same procedure. Concentration of algae was the mean ± standard
deviation of three counts.
Glass coverslip suspended at mid height of the water column or laid on the surface of the
sediment were weekly removed to measure the growth of biofilms over time. A FluoView
Olympus confocal laser scanning microscope BX50 equipped with krypton-argon laser (488
nm and 545 nm lines), a CCD Camera and objectives X40 and X100 was used for
microscopic observations. Emissions on green channel were observed between 510–560 nm.
Three to five samples were observed.
Determination of the bacterial diversity
The water column and the sediments were sampled in the 9 mesocosms at the end of Stage I.
The surficial sediments were collected at 3 different locations per mesocosm and pooled. The
limnetic column (100 mL) was sampled at about 10 cm below water surface. The similarity of
microbial community compositions was assessed by pyrosequencing of the 16S rRNA gene.
Microbial diversity in the mesocosms was considered in terms of richness and phylogenetic
distance. The number of OTUs observed and the Chao1 estimator best characterize the
microbial richness of the sample, while the Bray-Curtis dissimilarity describes the
phylogenetic distance between two samples. Two communities that share no OTU have a
maximum Bray-Curtis dissimilarity of 1.
Sediments were allowed to settle and excess water was removed. Bacterial cells were
centrifuged from the water columns at 5.9 x g at 4°C for 30 min. DNA was extracted from 0.5
mL of sediment and from the water column pellets using the Fast DNA spin kit for soil and a
MP Bio FastPrep® equipment, according to the protocol of the manufacturer. DNAs from 3
replicates were pooled and analyzed by high throughput 16S rRNA gene pyrosequencing 1.
Operational taxonomic units (OTUs) were defined after removal of singleton sequences,
clustering at 3% divergence (97% similarity). Bioinformatics treatment of data was done
using QIIME (www.qiime.org/) and R tools for diversity (www.r-project.org/).
Figure S1. (a) Number of particles in water column at 10 cm below water surface. (b)
Concentration of algae in surficial sediments (depth: 0.5 to 1 mm). (c) Concentration of
picoplankton in water column (grey bars) and in surficial sediments (black bars). The data
represent the mean  standard deviation of the bacteria and algae counting in the 9
mesocosms. Different letters represent a significant difference between samples (p <0.05).
Figure S2. Confocal laser scanning micrographs (section thickness 0.3 µm) and optical
transmission images of microorganisms in the mesocosms. (part I) micro-organisms
developing in the water column and (part II) in the biofilm at the sediment surface. Organisms
were labeled with SYTO9® that stains DNA. Excitation wavelengths 488 and 645 nm;
emission wavelengths 510-560 nm on green channel. Part I: (a, b) unidentified bacterial cells,
(c) Spirochetes, (d,e,h) unidentified organism, (f, g) Incerta sedis, (k, l) diatoma, (l) zoospore
of a Chlorophyce. Part II: (a, c) confocal laser scanning micrographs. (b, d) optical
transmission images of the same fields. Bacteria developed around mineral particles and
phototrophic organisms. Scale bar: 20 µm
Figure S3. X-ray diffractogram of the natural inoculum after sieving. The X-ray
diffractometer (Panalytical X’Pert Pro MPD) used was equipped with Co Kα radiation (1.79
Å) and operated at 40 kV and 40 mA current.
References
1
Dowd, S. E., Sun, Y., Wolcott, R. D., Domingo, A. & Carroll, J. A. Bacterial tag-
encoded FLX amplicon pyrosequencing (TEFAP) for microbiome studies: bacterial
diversity in the ileum of newly weaned Salmonella-infected pigs. Foodborne Pathog Dis 5,
459-472 (2008).