Adhesion of freshwater sponge cells mediated by carbohydrate-carbohydrate interactions requires low environmental calcium

Marine ancestors of freshwater sponges had to undergo a series of physiological adaptations to colonize harsh and heterogeneous limnic environments. Besides diminished salinity, river-lake systems also have calcium contents far lower than seawater. Cell adhesion in sponges is mediated by calcium-dependent multivalent self-interactions of sulfated polysaccharides components of membrane-bound proteoglycans named aggregation factors. Cells of marine sponges have already been shown to require seawater average calcium concentration (10 mM) to sustain adhesion promoted by aggregation factors. We demonstrate here that the freshwater sponge Spongilla alba can thrive in a calcium-poor aquatic environment and that their cells are able to aggregate and to form primmorphs at calcium concentrations 40-fold lower than that required by cells of marine sponges. We also find that their specialized resistant reproductive bodies named gemmules need calcium and other micronutrients to hatch and generate new sponges. The sulfated polysaccharide composed of glucose and mannose units found in S. alba has molecular size notably lower than those present in aggregation factors of marine sponges. Assessments with atomic force microscopy/single-molecule force spectroscopy revealed that the low-molecular-size sulfated polysaccharide from S. alba self-interacts more efficiently at low calcium concentrations (1 mM) than that from the marine sponge Desmapsamma anchorata . Such an ability to retain multi-cellular morphology with low exogenous calcium contents must have been a crucial evolutionary step for freshwater sponges to successfully colonize inland waters.

digestion with papain and then partially purified with cethylpyridinium and ethanol precipitations (Stelling et al. 2019). Crude polysaccharide extracts (~ 50 mg) were applied into a Q-Sepharose XL column (GE healthcare), linked to a HPLC system (Shimadzu), equilibrated with 20 mM Tris-HCl and 1 mM EDTA and then eluted through a linear gradient of 0→3 M NaCl. Fractions of 0.5 mL were collected, checked for metachromasy and then pooled as two distinct fractions named G8 and G20, dialyzed against distillated water, lyophilized and stored at -20 °C for further utilization. SP from D. anchorata (DaSP) was extracted and purified as described elsewhere (Vilanova et al. 2016).
Electrophoresis was performed at 100 V for approximately 40 min at room temperature and then the gel was stained with 0.1% Toluidine Blue in 1% acetic acid.

Chemical analysis
Chemical analyses were performed as described elsewhere (Vilanova et al. 2009).
Total hexose contents of S. alba SPs were estimated by phenol-H2SO4 reaction.
Presence of hexuronic acid or pyruvated sugars was evaluated with carbazole reaction.
After acid hydrolysis of G8 and G20 with 6.0 M trifluoroacetic acid for 5 h at 100 °C, their sulfate contents were measured by BaCl2-gelatin method. Monosaccharide compositions were determined via gas-chromatography/mass-spectroscopy (GC/MS; Shimadzu) by analyzing alditol acetate derivatives produced by each fraction.

Single-molecule force spectroscopy assays
Self-binding forces of G20 from S. alba and the DaSP from D. anchorata at different concentrations of calcium were measured with a MFP-3D atomic force microscope (Asylum Research). SPs were immobilized on mica substrates and silicon nitride tips (NP-S) cantilevers (Bruker) by using a N-hydroxysuccinimide-poly-(ethylene glycol)maleimide (NHS-PEG-MAL) linker of 4,750 Da (Iris Biotech GmbH), as described elsewhere (Vilanova et al. 2016). SMFS assays were performed at room temperature by approaching and separating the cantilevers and substrates functionalized with G20 or DaSP in CMFFW-T (2 mM NaCl, 0.2 mM Na2SO4, 0.1 mM KCl and 20 mM Tris, pH 8.0) or CMFSW-T (490 mM NaCl, 7 mM Na2SO4, 11 mM KCl, 2 mM NaHCO3 and 20 mM Tris, pH 8.0), respectively, supplemented with 5 mM EDTA or different concentrations of CaCl2 (0.1→10 mM). Spring constants of the cantilevers were determined by the thermal noise method (Lévy and Maaloum 2002). Approximately 2000 force curves were recorded with constant approach and retract velocities (2000 nm/s) and afterwards analyzed with custom-made MATLAB-based software (Math Works) in order to calculate molecular adhesion forces and elasticity. Force histograms were fitted as normal curves to obtain the average dissociation forces (Fmax) by using the software Origin 8.0.

Comentado [J1]
: This is the first time this abbreviation appears in the manuscript. It should be explained what it stands for (calcium and magnesium free sea water).

Results and Discussion
Freshwater sponges are able to thrive in calcium-poor aquatic environments Considering that the freshwater sponge S. alba (Fig. 1a), which was employed as a model in our study, was collected in a costal lake that is separated from the sea by a narrow sand bank (Fig. 1b), we analyzed the lake´s water to evaluate whether that proximity to seawater could influence its chemical composition. Seawater has average salinity of 35 ppt (mg/ton) whereas the salinity of Lake Carapebus was trace when expressed as ppt (Table 1) and, therefore, it is a genuine freshwater lake. Such a reduced salinity was due to low concentrations of major constituents of seawater, including chloride, sulfate, sodium, magnesium, calcium and potassium (Table 1). Its calcium content (~7 mg/L) was approximately 50-fold lower than that of seawater (~400 mg/L) and thus the Lake Carapebus can be considered as a calcium-poor freshwater environment. Chemical parameters presented in Table 1 were used as a basis to prepare the media (CMFFW, M-medium and CMFSW) employed in our in vitro and AFM/SMFS assays.
Although data on the calcium content of rivers and lakes harboring freshwater sponges are scarce and uncertain, partial information available in different reports allowed us to trace some correlations. Nine species of freshwater sponges, including Metania reticulate, Trochospongilla paulula and Oncosclera navicella, were found inhabiting waters with very low calcium content (~0.25 mg/L) at the River Negro (Brazil) (Küchlera et al 2000;Volkmer-Ribeiro 2012). Similarly, the freshwater sponge Lubomirskia baikalensis, along with other 14 species that comprise the family Lubomirskidae, which are endemic of the Lake Baikal (Russia), can also thrive in calcium-deficient (15-18 mg/L) waters (Rahmi et al. 2008;Khanaev et al. 2018).
Besides the examples outlined above, the reduced calcium content (1-2 mg/L), which is commonly seen in inland waters (Potasznik and Szymczyk 2015), allows us to speculate that most species of freshwater sponges should be also adapted to colonize calcium-poor aquatic environments.

S. alba cells aggregate at low calcium concentrations
Once the water chemistry of S. alba´s habitat was determined, we were able to prepare media (CMFFWs) suitable to evaluate the calcium content required to promote in vitro approximately 40-fold less calcium than cells of marine sponges to aggregate in an effective manner. We also found that the minimum calcium concentration (0.25 mM CaCl2) that was necessary to yield consistent aggregates in our in vitro assays is in strict accordance with the calcium content (~0.2 mM) measured in the water of Lake Carapebus.

S. alba cells form primmorphs at low calcium concentrations
After ascertaining the calcium requirements to promote cell aggregation, we assessed Then, we followed the initial development of the sponges generated by S. alba gemmules. On the first 2 days after hatching (Fig. 4a), archeocytes and histocytes migrated outwards and proliferated actively on the substrate around the gemmules (Fig. 4b). Over the next four days (Fig. 4c-d), histocytes became basopinacocytes and archeocytes differentiated into a variety of cell types (e.g. choanocytes and sclerocytes), which began the formation of the basal pinacoderm (Fig. 4e), aquiferous system ( Fig. 4f-g) and siliceous skeleton (spiculogenesis) (Fig. 4h-i). Afterwards, the choanosome and skeleton underwent further organization and an ectosome lining the outer surface (Fig. 4j) and an osculum (Fig. 4k) formed, giving origin to fully functional miniature sponges 6 days after the hatching. We also observed that the sponges in formation released a great number of motile archeocytes toward the adjacent substrate ( Fig. 4l), which in turn must act as phagocytes in the defense against pathogens and foreign body invasion during their development (Johnston and Hildemann 1982).
Our results revealed that the development of new sponges from gemmules of S.
alba is similar to that described for other species of freshwater sponges such as E.

Sulfated polysaccharides from S. alba have low molecular size
As stated above, cell-cell adhesion in sponges is mediated by calcium-dependent selfbinding between SPs (e.g. Bucior et al. 2004;Misevic et al. 2004;Vilanova et al. 2009); therefore, we investigated in further detail the role of calcium in the adhesion of S. alba cells by evaluating the self-interactions of its purified SP. Crude polysaccharide extracts from S. alba were applied into a Q-Sepharose column and then eluted through a linear NaCl gradient, yielding chromatograms with two peaks, which were collected as distinct fractions identified as G8 and G20 (Fig. 5a). Agarose gel electrophoresis of G8 revealed a polydisperse polysaccharide with low metachromasy whereas G20 is homogeneous and strongly stained by Toluidin Blue (Fig. 5b). Molecular size estimations using polyacrylamide gel electrophoresis confirmed that G20 has a mass of approximately 20 kDa and a homogenous molecular weight distribution; on the other hand, G8 presented a highly polydisperse distribution and molecular weight around 8 kDa (Fig. 5c). Chemical analysis showed that both fractions are composed of glucose and mannose units and that G20 is more sulfated than G8 (Table 2). Its high polydispersity, small size and low sulfate content suggested that G8 is a degraded product of G20 rather than an intact and functional SP expressed in the cells of the sponge; for this reason, we designated G20 as the putative SP component of S. alba

AF.
Besides containing only sulfated glucose and/or mannose units, G20 does not contain hexuronic acid or pyruvated sugars and thus it has a less complex composition than SPs from marine sponges such as the highly branched DaSP component of D.
anchorata AF (Vilanova et al. 2008;Vilanova et al. 2016). G20 has a sulfate content slightly higher than DaSP (molar ratio ~0.80 and ~0.66, respectively); however, SPs from other marine sponges, such as that found in Chondrilla nucula, were already shown to be more sulfated (molar ratio ~1.5) (Vilanova et al. 2008). On the other hand, G20 has a molecular size significantly smaller than the SPs of most marine sponges.
Both DaSP and the SP component of C. prolifera AF, which has already been used in several cell adhesion studies, have molecular masses ca. 10-fold higher than G20 (Garcia-Manyes et al. 2009;Vilanova et al. 2016). Therefore, our results demonstrate that the sulfated glucomannan deprived of carboxylated sugars from S. alba has comparable sulfate content but lower molecular size than SPs components of AFs of marine sponges. Although the presence of AFs in E. fluviatilis has been described almost 50 years ago, the chemical structures of their constituents have never been analyzed in detail, and thus it is the first report of the composition of a freshwater sponge SP (Fernàndez-Busquets and Burger 1999). our assays conducted in lower calcium (1 and 5 mM CaCl2) and, therefore, requires seawater average calcium concentration for self-interacting in an efficient manner. On the other hand, G20 from S. alba already reached its maximum dissociation force at 1 mM CaCl2 and thus has shown to be able to self-interact efficiently with reduced contents of exogenous calcium.
Despite "calcium bridges" between the SP component of AFs are the major forces driving cell-cell adhesion in sponges, self-interactions in the absence of calcium alike to those observed for DaSP had been reported (Fernàndez-Busquets et al. 2009;Vilanova et al. 2016). These weaker interactions have been attributed to hydrogen bonds between hydroxyl groups on the SPs, while the repulsive forces promoted by their anionic sulfate and/or carboxyl epitopes are neutralized by the high content of Na + Na ++ ions present in seawater (Spillmann and Burger 2000). Considering that the CMFFW-T employed to assess self-interactions of G20 from S. alba has a reduced Na + Na ++ content, the small binding forces revealed in our AFM/SMFS assays in the presence of EDTA was likely due to the repulsion between the highly anionic sulfate groups on their glucose and/or mannose units. Nevertheless, such calcium-free interactions mediated by hydrogen bonds are unable to promote physiological adhesion of sponge cells, as has been demonstrated for more than 100 years by dissociating tissues of marine sponges with artificial seawater deprived of calcium (Wilson 1907). The increased proportion of weak self-interactions observed for DaSP at calcium concentrations lower than 10 mM suggests that they are mediated by hydrogen bonds and not by "calcium bridges" and thus might not be able to sustain cell adhesion. On the other hand, self-interactions between the low-molecular-size SPs from the freshwater sponge S. alba at reduced calcium concentrations are strong enough to promote cell-cell adhesion.

Comentado [J2]:
The discussion on the origin of the binding forces between SPs of DaSP at CaCl2 concentrations below 10 mM is valid. The formation of hydrogen bonds when the electrostatic repulsion between sulfate groups is screened by the high NaCl concentration could explain the binding forces observed. However, it seems difficult to explain why the hydrogen-bond forces are not able to sustain cell adhesion like calcium-bridge forces, because they both have similar magnitudes (NS in Fig. 6e means not significant difference, right?). In our previous manuscript (Vilanova et al. 2016) we observed that, although the magnitude of the binding forces could be similar, the mean lifetime of the bonds could be very different, and that seemed to make the difference for the adhesion of the sponge cells in 10 mM CaCl2 and not in calcium-free medium. Maybe we can include a comment about that here, referring to the paper Vilanova et al. 2016.