Brackish Water

The experiment included 162 unvaccinated seawater-adapted Atlantic salmon with an initial average weight of 160 grams. During the challenge, the fish were kept in a tank supplied with particle filtered and UV treated brackish water (15 ‰ salinity, 12°C) up to 3 weeks post challenge (wpc). The fish were then switched to seawater (34 ‰ salinity, 12°C) for the remaining time of the study. The inoculum, consisting of pelleted blood cells, originated from a HSMI outbreak in 2012 and had undergone one previous passage in experimental fish. The inoculum has previously been described [31 (link)]; it contained high loads of PRV as analyzed by RT-qPCR, and was confirmed negative for infectious pancreatic necrosis virus (IPNV), infectious salmon anemia virus (ISAV), salmonid alphavirus (SAV) and piscine myocarditis virus (PMCV). Prior to inoculation, the blood pellet was diluted 1:1 in PBS.
A cohabitation challenge was run with 75 shedder fish injected intraperitoneally with 0.1 mL inoculum, marked by shortening of the adipose fin and placed in a tank containing 75 naïve fish (cohabitants). To estimate the peak of infection, heparinized blood was taken from the caudal vein of five cohabitant fish weekly, from 3 wpc to 6 wpc. At peak of infection, remaining fish were sampled on three subsequent occasions; 7 wpc (n = 16), 7 wpc + 2 days (n = 16) and 8 wpc (n = 24). At each sampling, two non-infected fish of the same population kept in a separate tank served as negative controls. In addition, blood samples were collected from five fish before initiation of the experiment (0 wpc).
The heparinized blood was immediately analyzed for PRV by flow cytometry, as described below. The blood was then centrifuged at 3000 × g for 5 min at 4°C, separating plasma and blood cells. Both were analyzed for PRV by RT-qPCR and the remaining samples were stored for PRV purification (details described below).

Four types of microsporidian sequences constituted our dataset: (i) sequences newly produced from our collection of infected G. roeselii individuals; (ii) published SSU sequences representing diversity and divergence of microsporidians already found to infect European freshwater or brackish water amphipods (we did not include sequences outside Europe, e.g. the recently published parasites from Lake Baikal [34 (link), 35 (link)]); (iii) published SSU sequences for microsporidians infecting other taxa, prioritising freshwater or brackish water invertebrates, when closely related amphipod sequences relative to newly produced sequences were not found; (iv) published SSU sequences representative of the five microsporidian clades (Clades I–V), as determined in the integrative phylogenies presented in literature [55 (link), 56 (link)]. All sequences were aligned using MAFFT7.388 software [57 (link)], with the E-IONS-I algorithm using legacy gap penalty option, incorporated in Geneious 10.2.2 [53 (link)]. Our dataset consisted of sequences of variable lengths depending on both the success in producing new sequences (from 180 to 826 bp) and on various length of the published ones (from 300 to 1448 bp for microsporidians, and 1786 bp for the fungus Basidiobolus ranarum used as an outgroup). All details, including sequence length, are given in Additional file 4: Table S3. Alignments are given in Additional file 5: Data S2. As some sequences were relatively short, reducing the full dataset to a standard size would, on the one hand, allow defining haplotypes but, on the other hand, would potential induce losing phylogenetic signal. Therefore, we attributed each sequence to haplogroups, defined in such a way that sequences belonged to distinct haplogroups if they differed by one or more variable sites, generating diagnostic features (Additional file 2: Table S2), whatever sequence length. Two sequences were clustered in one haplogroup, despite variable length, based on 100% pairwise identity, therefore sharing the same diagnostic sites. A limited set of newly produced sequences could be assigned to at least two haplogroups due to a combination of reduced length and lack of diagnostic features. Only the longest sequence representing each haplogroup was used for the phylogeny reconstruction (248 to 826 bp, noted in Additional file 4: Table S3; see also alignments in Additional file 5: Data S2).
Bayesian phylogeny reconstructions were performed with MrBayes [58 (link)] incorporated in Geneious 10.2.2. The best-fitting model of nucleotide substitution was determined with JModelTest-2.1.10. [59 (link)]. This was always the General Time Reversible (GTR) model with gamma-distributed rate heterogeneity (G) and a significant proportion of invariable sites (I). Four heated chains, each 1,100,000 iterations long, sampled every 200 iterations, were run. The runs reached satisfactory effective sampling sizes (ESS > 200), and the potential scale reduction factor values equalled 1 for all parameters. The 50% majority-rule consensus tree was constructed after the removal of 10% ‛burn-inʼ trees. Four phylogenetic trees were constructed. The first tree contained all haplogroups (i.e. sequences from this study and published sequences) using Basidiobolus ranarum (GenBank: AY635841) as the outgroup [55 (link)]. In this tree, we described novel parasites by conservatively using provisional names, e.g. Microsporidia sp. (hereafter abbreviated Msp) followed by the clade number (from I to V) sensu Vossbrinck et al. [55 (link)] and a superscript roman letter. The three other phylogenies represent detailed analyses for the already identified parasites of the microsporidian species of the genera infecting amphipods: Nosema [12 ], Cucumispora [9 (link)] and Dictyocoela [6 (link)]. Nosema antherae (GenBank: DQ073396), Vavraia culicis (GenBank: AJ252961), Dictyocoela cavimanum (GenBank: AJ438960) were used as outgroups for the Nosema, Cucumispora and Dictyocoela phylogenies, respectively. Following Grabner et al. [5 (link)], if a newly obtained sequence was > c.98% similar to a sequence for which a full taxonomic description was available, providing genus and species name, such name was ascribed to the new sequence. Alignments used for building these trees are provided in Additional file 5: Data S2).

A new taxon–character matrix (101 taxa × 245 characters) for Pleurodira was built using Mesquite v. 3.0 [36 ]. The character list is largely based on the extensive studies of Gaffney et al. [12 ,13 ], with additions from other sources (e.g. [14 (link),23 ,37 –42 ]), especially those focusing on chelids [17 ,29 ,43 ], and 18 new characters proposed here. Special effort was made to better sample post-cranial structures, resulting in 97 characters from that partition (39.5% of the total), more than in any previous study (e.g. [12 ] and [14 (link)] had 29.7% and 36.4% of post-cranial characters, respectively). Twelve characters were interpreted as forming morphoclines and ordered accordingly (see the electronic supplementary material for additional information about the phylogenetic analysis).
The taxon sample was conceived to incorporate all pleurodiran lineages, namely Chelidae, Pelomedusidae, Araripemydidae, Euraxemydidae, Bothremydidae and Podocnemididae, including 98 crown pleurodires as terminal taxa (see the electronic supplementary material for the complete taxon list). Previously, the largest matrices for pleurodires comprised 43 [8 (link)] and 91 [14 (link)] in-group taxa (although the latter study employed a reduced version, with 70 in-group taxa, in the main analyses). The non-Testudines Testudinata Proganochelys quenstedti and the stem-pleurodires Notoemys laticentralis and Platychelys oberndorferi composed the outgroup taxa.
The resulting matrix was analysed under the parsimony criterion in TNT v. 1.1 [44 ] via a heuristic search with the following settings: 2000 replicates of Wagner trees, random seed = 0, tree bisection reconnection (TBR) for branch-swapping, hold = 20 and collapse of zero-length branches according to rule ‘1' of TNT. The most parsimonious trees (MPTs) found in this first round of the analysis were the subject of a second round of TBR. A strict consensus tree, decay (Bremer support) and resampling (bootstrap and jackknife) values were obtained using implemented functions on TNT. The resampling values were calculated using 1000 replicates for absolute and difference of frequencies (group present/contradicted or GC in [45 ]). Consistency indexes (CI) and retention indexes (RI) were calculated using the script statsall (designed by Peterson L. Lopes, v. 1.3). The IterPCR script [46 ] was used to identify unstable taxa during preliminary analyses (i.e. taxa with multiple alternative positions) and to re-evaluate our scoring when the instability was caused by conflict of information rather than missing data. Additionally, considering that molecular-based phylogenetic analyses retrieve distinct arrangements for several extant taxa (e.g. [20 ,24 ,25 ,27 (link)]), we ran a second analysis (referred to hereafter as the ‘molecular constrained' analysis), following the same settings as the previous (referred to hereafter as the ‘original' analysis), except for setting constraints (see the electronic supplementary material) for the relations of the extant taxa based on the molecular phylogenetic hypothesis of Guillon et al. [25 ]. We also conducted three additional constrained analyses to evaluate how many steps were needed to achieve arrangements found in other alternative hypotheses.
In addition to the ‘original' and ‘molecular constrained' trees, we obtained two additional topologies to be employed in the subsequent analyses. The ‘non-marine taxa tree' was built for the biogeography analyses by pruning taxa previously considered marine or adapted to brackish water, i.e. Bothremydini & Stereogenyini [12 ,15 (link),16 (link),47 ] from the strict consensus tree of the ‘original' analysis. Further, an informal ‘supertree' was built for the diversity and diversification analyses by adding extinct and extant taxa not included in the ‘original' phylogenetic analysis to the strict consensus tree. The four topologies were time-scaled with the R [48 ] package strap [49 ], using information from the literature to define time ranges for each taxon (see the electronic supplementary material) and dividing branch lengths equally along the tree to avoid zero or close-to-zero values [50 (link)]. As the biogeographic analysis requires fully dichotomous topologies, the polytomies of the ‘original’ tree were manually resolved by deliberately choosing particular arrangements (see the electronic supplementary material for additional topologies).

Bryophytes, including mosses, are rarely tested for salt stress since they do not live in salt water. However, there are numerous species which have to cope with salt stress or even live in such environments as brackish water. With the aim of studying the response of mosses to salt and oxidative stress, we chose two phylogenetically unrelated species of bryo-halophytes, namely pottioid Hennediella heimii and funaroid Entosthodon hungaricus and a model moss Physcomitrium patens. For details on the already known biology of the species, please refer to Sabovljević et al. [5 ,31 ] and Ćosić et al. [50 (link)], while information on the axenic cultivation of bryophytes can be found in Sabovljević et al. [51 (link)].

Two levels of salinity were selected in the exposure experiment―0 ppt represents freshwater, and 10 ppt represents brackish water. Four days of exposure was determined due to no living larvae observed after 96 h of cultivation with 100 mg/L ZnO NPs under 0 ppt. Differences between treatments are calculated by one-way ANOVA and Tukey’s multiple comparison. Each result was displayed as mean ± standard error (SE), and p < 0.05 was considered as statistically significant. All data analysis and graphs were conducted with SigmaPlot 11.0.

A4-Membrane samples were prepared using a blend base polymer of polysulfone and polyacrylonitrile in the presence of graphene oxide nanoparticles as described briefly in [19 (link)]. This formulation was one of five different formulations regarding additives which were used in the polymeric dope solution and was found to be the optimum one and nominated “3rd family”. These membranes were prepared by dissolving 19% polysulfone, 1% polyacrylonitrile, and 0.25% nanographene oxide in N-N, dimethylacetamide (DMAc) for 24 h to form a homogeneous dope solution. The polymeric dope solution was cast on nonwoven polyester support, and the membranes were fabricated by non-solvent-induced phase separation (NIPS). The nominated 3 (blank membrane) was used to prepare a polyamide layer by interfacial polymerization of meta phenylene diamine and trimezoyl chloride and nominated 3T. Again, the blank membrane was coated with a hydrophilic, crosslinked layer of polyvinyl alcohol PVA-Glutaraldehyde GA and nominated 3P. A chemical grafted layer of poly-methacrylic acid/graphene oxide was formed for both 3T and 3P to form 3TG and 3PG, respectively.
These membranes were subjected to irradiation for surface modification by means of VUV and low-pressure plasma (Table 1) for membrane surface activation and grafting to present their effect on the performance of RO membranes in terms of brackish water desalination, chlorine resistance, and antifouling behavior. Membrane performance was measured by flat-sheet membrane testing cell with a constant feed pressure of 20 bar.