Salt-Enrichment Impact on Biomass Production in a Natural Population of Peatland Dwelling Arcellinida and Euglyphida (Testate Amoebae)

Unicellular free-living microbial eukaryotes of the order Arcellinida (Tubulinea; Amoebozoa) and Euglyphida (Cercozoa; SAR), commonly termed testate amoebae, colonise almost every freshwater ecosystem on Earth. Patterns in the distribution and productivity of these organisms are strongly linked to abiotic conditions—particularly moisture availability and temperature—however, the ecological impacts of changes in salinity remain poorly documented. Here, we examine how variable salt concentrations affect a natural community of Arcellinida and Euglyphida on a freshwater sub-Antarctic peatland. We principally report that deposition of wind-blown oceanic salt-spray aerosols onto the peatland surface corresponds to a strong reduction in biomass and to an alteration in the taxonomic composition of communities in favour of generalist taxa. Our results suggest novel applications of this response as a sensitive tool to monitor salinisation of coastal soils and to detect salinity changes within peatland palaeoclimate archives. Specifically, we suggest that these relationships could be used to reconstruct millennial scale variability in salt-spray deposition—a proxy for changes in wind-conditions—from sub-fossil communities of Arcellinida and Euglyphida preserved in exposed coastal peatlands. Electronic supplementary material The online version of this article (10.1007/s00248-018-1296-8) contains supplementary material, which is available to authorized users.

Samples of the peatland surface (monoliths of 10 x 10 x 10 cm) were collected along the transect at 28 locations at 10 m intervals, with greater intervals for samples 24-28. Sample numbers were allocated sequentially (i.e. 1 = most coastal, 28 = furthest inland).
All samples were stored frozen and sealed within plastic bags prior to analysis. Conductivity and pH was recorded using a calibrated Hanna Instruments HI98129 meter from pore-water extracted by applying a compressive force to sub-samples of the surface 5 cm of each monolith. The relative level of salt-enrichment received by each sample was quantified by proxy of pore-water conductivity which is linearly related to salinity under these measurement conditions (Wagner et al, 2006). Physical properties of the substrate were measured in a separate sub-sample of 2 cm 3 (also from the surface 5 cm); bulk density was calculated by dividing the dry mass of each sample by its volume, and the change in sample mass after drying determined the moisture content (%).

Enumeration of Arcellinida and Euglyphida populations.
Tests (shells) of Arcellinida and Euglyphida were isolated from 1 cm 3 sub-samples collected from the surface (1 cm) of each monolith, and were concentrated for direct observation using a standard water-based protocol (Charman et al, 2000;Booth et al, 2010). One tablet of Lycopodium spores (Lund University; Batch number 1031) was added as an exotic marker in each sample (Stockmarr, 1971). Samples were boiled in 100 ml of de-ionised (DI) water for 10 minutes to disaggregate tests from the substrate, and allowed to cool to room temperature. The samples were then rinsed through sieves with DI water, and residues of the 15-300 µm fraction were retained for analysis. Supernatant water was removed by centrifuge (3000 rpm for 5 minutes).
Slides were prepared by diluting the residue with glycerol, and counts of taxon abundance were made at x200-400 magnification on a Zeiss AxioImager A1 light microscope. A minimum of two slides were analysed from each sample.
Counts included all individuals (i.e. living, encysted, and empty tests) to minimise possible assemblage bias caused by seasonal blooms (Barnett et al, 2013). A minimum of 200 individual tests were counted in each sample to detect subtle assemblage changes and to ensure that total diversity was accounted for (Payne & Mitchell, 2009). This threshold could not be reached for samples 2 and 3, where 12 and 50 individuals were observed respectively. A lower count total was considered sufficient because of the low diversity of taxa in the assemblage of sample 3 (Payne & Mitchell. 2009), and so these data were included in analysis. We also consider the low concentration of tests in sample 2 to be a valid result since the occurrence of Lycopodium spores rules out error in sample preparation, however these data were omitted from species-level analysis since the count total was deemed insufficient to accurately represent the taxonomic composition of the community.
Measurements of test dimensions were made using AxioVision software (version 4.8.2.0) coupled to a AxioCamHR3 camera. Tests are most commonly observed in broad lateral view during routine counting so measurements of test height (i.e. in the plane perpendicular to width and length) were made using a stereomicroscope. Literature values were used to supplement the data where it was not possible to obtain a sufficient number of measurements (Table S6).
Concentration of tests per cubic centimetre was calculated from the ratio of tests to Lycopodium spores, and converted to tests per dry gram using substrate bulk density (see Royles et al, 2016). Average biovolume of each taxon (Table S6) was calculated assuming standard geometric test shapes (Mitchell, 2004), and converted to biomass using the factor 1 µm 3 = 1.1 × 10 -7 µg C (Mitchell, 2004). For each sample, total biomass was calculated by multiplying these values by taxon abundance. Error in biomass estimates was quantified by comparing maximum and minimum estimates, each based on upper (75%) and lower quartile (25%) ideal individual test dimensions and maximum and minimum test concentration values, respectively.
Shannon-Weaver diversity index values (Sageman & Bina, 1997) were calculated for each sample using the equation: Where S is richness of taxa (alpha-diversity), Xi is the abundance of each taxon, and Ni is the total abundance of Arcellinida and Euglyphida within the sample. We assumed that values; 2.5-3.5, represent stable environmental conditions, 1.5-2.5, transitional, and 0.1-1.5 are indicative of stressed conditions dominated by a small number of taxa (Patterson & Kumar, 2002).
Morphological trait values associated with feeding ecology were measured for each taxon (Table S5). Community weighted means for trait prevalence in each sample were calculated by multiplying these values by taxon abundance using the add-on package for R FD (Laliberté et al, 2014). Relationships between the community weighted mean value for each trait and conductivity were then assessed.

Taxonomy.
Identification of taxa was based on Ogden and Hedley (1980), Charman et al, (2000), Mazei and Tsyganov (2006), and Meisterfeld (2002a,b) with additional use of photographs and descriptions of southern hemisphere taxa (Fernández et al, 2015;Zapata & Fernández, 2008;van Bellen et al, 2014) and saltmarsh taxa (Charman et al, 2002;Gehrels et al, 2006). For standardisation with existing studies, identification used morphological features of the test, including; composition, shape, ornamentation, size and colour. We adopted a conservative approach to taxonomy to ensure that taxa represent biological species as closely as possible, and to produce an accurate estimation of diversity. Following Charman et al (2002), complexes of taxa (e.g., the genus Centropyxis) that are difficult to identify by test morphology were divided to the lowest possible level whilst maintaining clear, consistent and convenient morphological criteria. Microphotographs of all observed taxa are shown in Figure S1, and a list of taxa is given in Table S1.

Analysis of community assemblages.
All observed taxa were included in statistical analysis. Prior to analysis, assemblage counts were converted to relative abundance (percent) and square-root transformed to reduce the influence of dominant taxa (see Vincke et al, 2004). Detrended correspondence analysis (DCA) was used to estimate the overall gradient length of the assemblage data by determining whether species responses were primarily unimodal (gradient length > 2 standard deviations) or monotonic (gradient length < 2) (Legendre & Birks, 2012). Gradient lengths exceeded 3.0 standard deviations (3.07 for square-root transformed assemblage data), and therefore unimodal ordination methods were used to analyse the relationship between assemblages and corresponding environmental variables (ter Braak & Prentice, 1988).
Principal component analysis (PCA) was used to determine the major gradients in the environmental data. Prior to analysis environmental variables were checked for skewness, centred and standardised. Skewness of conductivity was reduced by applying a log10 transformation. Both conductivity and pH were shown to correlate with distance inland from the coast (R 2 = -0.45, p = 0.05, and R 2 = -0.56, p = 0.01 respectively), which confirms that the salinity gradient is produced by spatially variable deposition of oceanic salt-spray (i.e. levels decrease with distance inland from the coast). pH and conductivity therefore did not represent independent ecological signals in this dataset. Instead, both reflect salt-enrichment of an otherwise low-pH freshwater environment. Since conductivity is linearly related to salinity, unlike pH, it was retained for further analysis as an indicator of saltenrichment level while pH was removed. Distance inland of coast was also removed from subsequent analysis since it is a synthetic (i.e. non-ecological) variable.
Canonical correspondence analysis (CCA) with Monte Carlo permutation tests was applied to assess the statistical significance of the measured microhabitat variables in explaining variations in assemblages between samples. Partial-CCA where each environmental variable was included as the only explanatory variable, against the assemblage dataset, was used to measure the significance of each environmental variable. Interaction between variables was estimated by sequentially including each as a co-variable in successive iterations (Table S4). To measure the strength of each environmental variable, in explaining changes in assemblages between samples, we calculated the ratio between the first constrained (CCA1) and first un-constrained (CA1) axis. A value > 1 indicates that the variable represents an important ecological gradient (Juggins, 2013).
All analyses were performed in R (version 3.4.1) (R core development team, 2017), and add-on package vegan (Oksanen et al, 2017).

Figure S1
Arcellinida and Euglyphida taxa (collectively testate amoebae) identified in this study. Abbreviations for taxon name correspond to those used in Fig.2 (in main text). Reference numbers correspond to the full list of taxa given in Table S1. Images were obtained using a Zeiss AxioImager A1 light microscope coupled with an AxioCamHR3 camera. Scale bars represent 20 µm in all images.

Figure S2
Percentage abundance of Arcellinida (blue shading) and Euglyphida (grey shading) taxa from the Kampkoppie peatland plotted against sample pore-water conductivity. Low conductivity values represent low levels of salt-enrichment. Taxa are ordered by conductivity optima calculated by weighted average. Taxon abbreviations refer to those given in Table S1.

Table S1
List of all Arcellinida and Euglyphida taxa observed from the Kampkoppie peatland on Marion Island and corresponding abbreviations used in Fig. 2 (in main text). Taxonomy follows Krashevska et al (2016). One taxon could not be identified accurately to genus level and has been termed Cr-Ps.spp to reflect this uncertainty. 'Spp.' was used where taxa could only be confidently identified to the genus level, and 'type' for taxa exhibiting several morphotypes that possibly represent distinct species. Microphotographs of each taxa are shown in supplementary figure S1. • Indicates taxa not previously identified on Marion Island in the existing study of testate amoebae by Grospietsch (1971 Table S4 a) Individual canonical correspondence analysis results. CCA1/CA1 is used to measure the strength of each explanatory variable in explaining changes in assemblages between samples, where a value >1 indicates that the corresponding variable represents an important ecological gradient (Juggins, 2013). b) Variance partitioning results. Cpore-water conductivity, Mmoisture content, and BDbulk density.  Table S5 Selected morphological traits associated with the feeding ecology of Arcellinida and Euglyphida. None of the relationships between the prevalence of traits and conductivity conditions were found to be statistically significant at the p ≤ 0.05 level. CWMcommunity weighted mean, Cpore-water conductivity.   Mitchell (2004). Full names corresponding to the taxon codes are given in Table S1.