Proteomic characterization of a foraminiferal test’s organic matrix Authors: Fiorella Prada, Liti Haramaty, Oded Livnah, Racheli Shaul, Sigal Abramovich, Tali Mass, Yair Rosenthal, Paul G. Falkowski Published: 2024 ABSTRACT SignificanceOur extensive proteomic analysis of foraminiferal tests suggests that biomineralization in these organisms involves both vesicular pathways and transmembrane transporters, opening different avenues for understanding the molecular basis of the “vital effects” seen in foraminifera. Furthermore, the detection of several proteins that are shared across diverse calcifying taxa has wide-reaching implications for understanding common biomineralization pathways across various taxa. Foraminifera are single-celled protists, widely distributed across marine environments, with a fossil record dating back to the Early Cambrian (1). Their calcium carbonate tests provide valuable geochemical archives for reconstructing past climate and ocean conditions (2). Indeed, as ions needed to build their carbonate structures are sourced from seawater (3), the trace element chemistry and isotopic composition of foraminiferal tests may reflect physicochemical conditions, such as the temperature, salinity, carbonate, and major ion chemistry of the seawater in which they develop (4–6). Nonetheless, the elemental and isotopic composition of the tests may deviate from concentrations expected from equilibrium processes and observed in inorganic precipitates under known environmental conditions (7). This suggests that their incorporation in the shells is mediated by biological processes, collectively defined as “vital effects”, which may modify the environmental signals as known for carbon and oxygen isotopes (8), elemental proxies (e.g., (5), and metal isotopes (e.g., Ca, (9)), complicating their interpretation as geochemical archives. These effects can occur during the acquisition of ions from seawater into the foraminiferal cell and/or during the intracellular test biomineralization process. The majority of foraminifera precipitate calcite tests, which are formed by the addition of consecutive interconnected chambers (10). In general, each new chamber is formed by first producing an organic matrix, representing ca. 0.2% of the shell weight (11), often called the primary organic sheet (POS) composed of polysaccharides and acidic proteins (12, 13), and further mineralizing it with calcite crystals from both the inner and outer sides (14). Our understanding of the POS structure and formation is limited, though analogous structures in other organisms (15, 16) suggest that polysaccharides provide stability to the organic framework, and acidic proteins play a role in initiating calcite formation (17). Proteins rich in acidic amino acids are likely involved at multiple stages of mineralization, by temporarily stabilizing amorphous calcium carbonate or triggering the nucleation of aragonite and calcite under optimal conditions (18). Fluids surrounding the POS are believed to be derived from seawater and undergo unknown modifications prior to calcification (3, 19). Studies of various benthic foraminifera have shown that a major mode of ion uptake is via the endocytosis of seawater into vacuoles that enter into the cytoplasm (10). These vacuoles provide the ions to the fluid that is in intimate contact with the site of mineralization and are therefore a source of calcium and carbonate ions (19). The diffusion of ions alone cannot explain, however, the discrepancy between seawater metal composition and Me/Ca ratios in foraminiferal calcite (20). Therefore, it has been suggested that Ca2+ and HCO3- concentrations are actively enriched by transmembrane pumps, channels, and enzymes such as carbonic anhydrase during foraminiferal biomineralization (21, 22). Additionally, new research proposes that CO2 uptake by benthic foraminifera occurs through proton pumping (23). Many of these seminal studies have been carried out on Amphistegina lobifera Larsen, 1976, a diatom-bearing large benthic calcitic radial-perforate foraminifer frequently found in shallow areas (0 to 12 m depth). This species, common throughout the central and western Pacific, Indian Ocean, and Red Sea (24), has successfully spread across the Mediterranean since the Suez Canal’s opening in 1869 (25). Although foraminifera are responsible for a large part of the global calcium carbonate production (26), our understanding of the macromolecular components (such as proteins, polysaccharides, and lipids) in the organic matrix of foraminiferal tests and how they guide the biomineralization process is currently very limited. While genomic approaches reveal potential gene expression, proteomics identifies the actual proteins being produced (27). Here, we apply a proteomics approach to describe the test organic matrix proteins in A. lobifera (Fig. 1). A. lobifera specimens collected from Tel Shikmona on the Israeli Mediterranean coast. (A and B) Scanning Electron Microscope (SEM) images of A. lobifera dorsal and ventral view, respectively, taken with a Phenom ProX desktop SEM (C and D). Dorsal and ventral view, respectively, of live specimens taken using Leica M205 C stereo microscope. Using liquid chromatography–tandem mass spectrometry (LM–MS/MS) protein sequencing we identify proteins in the calcite test that may participate in the biomineralization process in A. lobiphera. Understanding how the biomineralization process works in foraminifera is particularly important because i) foraminifera account for ~25% of calcium carbonate production in the present oceans, making them crucial players in marine carbon cycling and ocean-atmosphere CO2 exchange (26) and ii) it will help us understand the connections between test chemistry and environmental parameters used for paleoclimate reconstruction (19). Results and Discussion Our comprehensive proteomic analysis of the test organic matrix in A. lobifera identified three hundred and seventy-three proteins by LC–MS/MS analysis (Dataset S1), providing significant insights into the biochemical complexity of foraminiferal calcification (Fig. 2). Ninety-three proteins showed hits (e-value ≤ 10−5) to proteins with documented functions and processes in other calcifying organisms (SI Appendix, Table S1), while ninety-three proteins have unknown roles in biomineralization (Dataset S1). Six proteins exhibit blast hits to proteins involved in photosynthesis, remnant proteins from the symbionts, as previously observed in tests of the large tropical benthic foraminifer Schlumbergerella floresiana (28). Finally, one hundred and eighty-one candidate proteins exhibit no blast hits in National Center for Biotechnology Information (NCBI), suggesting that these may be proteins involved in foraminiferal calcitic biomineralization containing sequences that are unique to these unicellular micro-organisms. Schematic model describing the calcification mechanism in bilamellar perforate foraminifera. The process of CaCO3 precipitation involves a dual ion transport mechanism in which initial seawater vacuolization is followed by active release of Ca2+ ions from the vacuoles and uptake into smaller, newly formed Ca-rich vesicles that subsequently migrate to the calcification fluid. This could result in Mg-rich vesicles that may passively leak Mg2+ (squiggly dashed arrow), possibly accounting for the very low Mg/Ca ratio observed in foraminiferal tests compared to inorganic calcite. Supply of inorganic carbon for calcification likely depends on both enzymatic catalysis (APaseD and CA) and pH-driven speciation from CO2 to HCO3− and CO32−. Mitochondria are shown in close association with the cell membrane surface immediately adjacent to the test wall, as observed in several studies [e.g., (29)]. Proteins detected in the current study and involved in the process are shown inside the white boxes. Question marks indicate uncertain ion pathways and potential localization of detected proteins. CDK5: cyclin-dependent kinase 5, SLC24A3: sodium/potassium/calcium exchanger 3-like isoform ×1, APaseD: phosphodiesterase/alkaline phosphatase D precursor, SLC26: solute carrier family 26 protein, CA: carbonic anhydrase, V-ATPase: V-type ATPase, A subunit, H+-PPase: vacuolar-type H pump pyrophosphatase-like protein. The majority of these proteins are highly acidic, as indicated by the higher than standard frequencies in aspartic and glutamic acids. Proteins from A. lobifera contain up to 7.5% and 8.4%, while standard proteins have 5.3% and 6.2%, respectively (30, 31). Acidic proteins are a feature shared by most – if not all – mineralizing phyla (31). They play crucial roles in the biomineralization process by either promoting crystal nucleation when absorbed onto an organic insoluble template, or inhibiting crystal growth when they are free in solution (32). It has also been suggested that, in some cases, they might stabilize amorphous calcium carbonate (33). Proteins Controlling Foraminiferal Chamber Formation. We identify five putative carbonic anhydrases (CA) (i.e., carbonic anhydrase 2, one carbonic anhydrase 3, two alpha carbonic anhydrase 6 [Corallium rubrum], and a carbonic anhydrase, partial [synthetic construct]; SI Appendix, Table S1 and Dataset S1). This suggests intracellular HCO3− supply for calcification in A. lobifera via the enzymatic interconversion of H2O and CO2 (34), as previously proposed for A. lessonii based on incubation experiments (21) and transcriptomic analyses (22). This does not exclude that also a pH-driven speciation of CO2 into HCO3- and subsequently CO32− is taking place in the calcification fluid, as observed in Ammonia sp. (23). Previous studies have documented the involvement of α-CAs in mollusk, sea urchin, coral, and foraminifera calcification (22, 35). Four out of the five sequences have the suitable length to be regarded as CAs (~300 amino acids; Dataset S1). Furthermore, multiple sequence alignment (SI Appendix, Fig. S1) shows that only three of the four sequences contain the 3 essential Histidine residues and are therefore capable of binding Zn2+, making them catalytically active. These sequences were subject to three-dimensional structural prediction through AlphaFold 3 (36) using both the protein sequences and a single Zn2+ cation per each run. Finally, the resulting structures were analyzed and compared for their overall fold by PyMol where the predicted position of the Zn2+ ion and its coordination indicate the functionality of the structures (SI Appendix, Fig. S2). All three CAs are identified in Bigelowiella natans with 26 to 30% sequence similarity (SI Appendix, Table S1). An additional source of intracellular HCO3- for calcification could be through the activity of a phosphodiesterase/alkaline phosphatase D precursor (APaseD) (Fig. 2). The latter has been hypothesized to catalyze HCO3− formation and facilitate calcium carbonate crystal formation in the nacreous layer of the pearl oyster and has been previously detected in the mantle of Patinopecten yessoensis (37) and Chlamys farreri (38). The intracellular HCO3− is then likely released to the calcification fluid via the solute carrier family 26 proteins (SLC26; Fig. 2 and SI Appendix, Table S1), anion transporters that exchange intracellular HCO3−, and extracellular chloride ions (Cl−) (22). SLC26 proteins have been associated with calcification processes as they are up-regulated during adult mussel shell repair (39) and during mussel larval development and shell deposition (40). Members of the same protein family are involved in accumulation of HCO3− from seawater across various calcifying organisms, including corals (41), sea urchins (42), coccolithophores (43), crustaceans (44), and mammals (45). Another ion transporter identified in the current study is the sodium/potassium-transporting ATPase subunit alpha isoform X1 which was previously found to be upregulated during Mytilus edulis larval development (40). Low pH vesicles have been observed in newly formed Ammonia beccarii chambers, suggesting that H+ are released into the cytosol upon catalysis by CA (22). Hydrogen ions are then sequentially recruited into H+ vesicles through two proteins identified in the current study a V-type ATPase, A subunit (V-ATPase), and a vacuolar-type H pump pyrophosphatase-like protein (H+-PPase) (Fig. 2, SI Appendix, Table S1). The former plays a key role in pH up regulation in foraminiferal calcification (23) and in aragonite production in corals (46) and was shown to be highly expressed during calcification in A. beccarii (22). The latter is an electrogenic proton pump that acidifies vacuoles in plant cells (47). V-type ATPases are ubiquitous in eukaryotic cells, playing crucial roles in organellar acidification and various physiological processes (43). As chamber formation involves active Ca2+ uptake, tight control of intracellular Ca2+ concentrations (~100 nM during resting conditions) is necessary to prevent cytotoxic phenomena (48). Various signaling pathways and metabolic processes are stimulated by cytosolic Ca2+ (48), as previously shown by the significant expression of calcium/CALM-dependent protein kinase in A. beccarii (22), which is also identified in the current study (i.e., Calcium/calmodulin-dependent protein kinase cmkA). By regulating essential processes like Ca2+ transport and secretion, cell proliferation and differentiation, aragonite nucleation, and calcite growth, calmodulin and calmodulin-like proteins play a central role in numerous signaling pathways (37). Vesicles function as reservoirs for calcium compounds in eukaryotic cells (49). Proteins related to vesicle generation, specifically polyubiquitin and clathrin heavy chain 1-like proteins (SI Appendix, Table S1), which indicate clathrin-dependent endocytosis at the onset of vesicle formation (50), are detected in the current study. We also document heat shock 70-kDa proteins (SI Appendix, Table S1) which are responsible for uncoating the clathrin-coated vesicles. The latter are then transported to endosomes for lysosome biogenesis via activation of Rab GTPase (51). Gene expression profiles conducted during A. beccarii calcification report upregulation of Ca2+ pumps located on the vesicle membranes, that may transport and export free Ca2+ between the cytosol and vesicles, and calcium/proton antiporters, which could indicate the uptake of Ca2+ into vesicles (22). These gene expressions suggest the formation of Ca2+ vesicles. The same study also showed Ca2+-storing vesicles and their transportation toward the calcification site by Rhod-3 AM image (22). P-type Ca2+ transporter type 2B, which is involved in Ca2+ excretion outside the cytoplasm (48), is also highly expressed during A. beccarii calcification (22). This transporter is stimulated by protein kinase A, which is in turn regulated by cyclin-dependent kinase 5 (52), which is detected in our proteomic analysis (CDK5; Fig. 2, SI Appendix, Table S1). We identify a solute carrier family 24 protein (i.e., SLC24A3, sodium/potassium/calcium exchanger 3-like isoform ×1; Fig. 2), which is a distinct type of Na+/Ca2+ exchanger that cotransports K+ and a sodium-dependent transporter (SI Appendix, Table S1). These membrane transporters likely have a similar function as the solute carrier family 8 observed in A. lessonii (22), which is to pump Ca2+ toward the calcification site. The solute carrier family 24 has been previously found to be involved in bone morphogenic processes (53) and is up-regulated during enamel maturation (54). Our findings combined with previously published data strongly suggest that the biomineralization process in these organisms is regulated by a well-coordinated Ca2+ and HCO3− trafficking system (22) together with vesicular pathways (19, 55). Elevated Ca2+ in the mitochondrial matrix stimulates ATP synthesis to provide sufficient energy for chamber formation (22). We identify several mitochondria-related proteins, including a predicted protein (TRINITY_DN27079_c0_g1_i1.p1) associated with proton transmembrane transport across the inner mitochondrial membrane (56). This protein shows a ∼ 57% sequence similarity with a cytochrome b-c1 complex subunit Rieske-4, mitochondrial-like protein in B. natans and 36% with ubiquinol-cytochrome c reductase iron–sulfur subunit protein in R. filosa (SI Appendix, Table S1), which are both key components of the electron transport chain in mitochondria (57). We also identify a mitochondrial ATP transmembrane transport protein (i.e., ADP, ATP carrier protein 3, mitochondrial-like) which transports ADP into the mitochondrial matrix for ATP synthesis and ATP out to fuel the cell (58). Proteins associated with respiratory H+ pumps across the inner mitochondrial membrane and ATP transport from the mitochondrial matrix to the cytoplasm were seen to be highly expressed during foraminiferal calcification in A. beccarii (22). These previously observed upregulations in addition to the proteins detected in the current study suggest active mitochondrial ATP synthesis during chamber formation, as foraminiferal cells may require substantial energy for reticulopodia assembly and cell volume expansion. We also identify a monovalent cation:proton antiporter-2 family, which plays a crucial role in maintaining ion homeostasis and regulating the protonmotive force in mitochondria (59). Our results support previous suggestions that a dense network of reticulopodia precisely regulates chamber deposition (23, 60). Indeed, we detect several proteins related to cytoskeleton and microtubule-related processes (SI Appendix, Table S1 and Dataset S1). Such proteins have previously been found to be highly expressed during chamber formation in the Rotaliida species A. beccarii (22). Specifically, we detect actin type 2, actin-binding proteins (i.e., putative coronin, coronin-2B-like isoform X1, fimbrin-like proteins), actin gamma 1, actin-related proteins 2/3 complex, which generally regulate the dynamic structure of the cytoskeleton, influencing cell shape, motility, division, and various other cellular processes in eukaryotes (61). We also identify actin polymerization/depolymerization proteins (i.e., gelsolin-like protein 2, component of SCAR regulatory complex), tubulin-α, and tubulin-β, which are related to microtubule polymerization and organization (62). Actins and tubulins allow for flexibility of the size and shape of the calicoblastic space during aragonite crystal growth in corals (16) and could perform a similar function in foraminifera. We also find several Ras-related proteins (Ras family small GTPase, Ras protein let-60, ras-related protein Rab-5C), responsible for signal transduction, which may play roles in exocytosis, phagocytosis, or programmed cell death (63). Proteins Potentially Serving Dual Roles in Foraminifera Immunity and Biomineralization. One function of the foraminiferal test may be to serve a solid protective barrier against invaders and pathogens. To ensure the protist's protection from parasites and pathogens, the test likely includes antimicrobial molecules and toxins, as we see in the skeleton of corals (64). We detect several test organic matrix proteins involved in innate immune response (SI Appendix, Table S1). Specifically, we identify an endoplasmin homolog isoform X1 that might be involved in stress response in A. lobifera (63). We also detect proteases, which have been hypothesized to contribute to the digestion and formation of the skeletal organic matrix in stony corals as well as to the processing and activation of other bioactive molecules and proteins necessary for biomineralization (64). Specifically, we found serine proteases (i.e., serine proteases, subtilase family protein, subtilisin-like protein, hypothetical protein SARC_03076, S8 family serine peptidase, serine carboxypeptidase-like), which are crucial mediators of the innate immune response, functioning as processing enzymes for proinflammatory cytokines and other inflammation-related enzymes (65). We also find proteinase inhibitors (i.e., three zinc metalloproteases, ImpA family metalloprotease, M23 family metallopeptidase, inter alpha-trypsin inhibitor, heavy chain 4-like, inter-alpha-trypsin inhibitor heavy chain H3, serpin family protein, and WD repeat- and FYVE domain-containing protein 4) (66). Many proteases, including matrix metallopeptidases and protease inhibitors, are expressed in calicoblasts, specialized cells responsible for coral skeletal CaCO3 precipitation, in the stony coral Stylophora pistillata (67). Protease inhibitors are related to organic matrix immunity and protective mechanisms in mollusk shells (68). Metalloproteases regulate the innate immune response through the NF-kB signaling in mice (69) and are highly expressed in the adult calicoblasts of S. pistillata (67). Because metallopeptidases can hydrolyze and process a wide variety of substrates, they may be involved in remodeling the test organic matrix scaffold (64). Moreover, some metalloproteases could play important roles in the biomineralization process by activating and processing highly acidic proteins responsible for mineral nucleation and be involved in factors affecting the innate immune response, including proinflammatory cytokines, chemokines, growth factors, and other receptor ligands (64). We also find a staphylococcal nuclease domain-containing protein 1, which has been linked with bivalve immune functions, being up-regulated by the Manila clam (Venerupis philippinarum) following Vibrio infection (70). Finally, we identify several Arylsulfatase-related proteins (i.e., Arylsulfatase B, Arylsulfatase B-like, arylsulfatase I, and arylsulfatase I-like isoform ×1). Arylsulfatase B is responsible for breaking down glycosaminoglycans by removing sulfate groups (71). Glycosaminoglycans are a major component of the organic matrix deposits in the tests of benthic and planktic foraminifera (11). They are found in organic linings, organic cements, and secretion products and are assumed to be of structural importance in the foraminifer test (72). Arylsulfatase B has been potentially related to biomineralization also in the mantle transcriptome of the Yesso Scallop P. yessoensis (37). Implications for Biomineralization and Geochemical Proxies in Foraminifera. Calcite precipitated by most hyaline species like A. lobifera contains low Mg/Ca ratios (1 to 20 mmol/mol) compared to miliolid species that build tests with high Mg content (100 to 150 mmol/mol), comparable to calcites precipitated inorganically from seawater (73). Furthermore, the nonlinear relationship with seawater Mg/Ca in both groups (74) highlights the need to better understand the main mechanisms responsible for ion transport to the site of calcification. Prior studies indicate that the production of an intracellular Ca-pool with a very low Mg/Ca ratio, that is then used for precipitating new calcite, can only be accomplished by effectively discriminating between Mg2+ from Ca2+ following seawater vacuolization (75). It has also been suggested that this difference may be attributed to the active excretion of Mg2+ (3, 10). Based on the Rayleigh fractionation model, it has been suggested that the main ion source for calcification is via seawater vacuolization (74). The low vesicle activity observed during normal calcification in Ammonia tepida, however, renders endocytosis insufficient as the major transport mechanism for Ca2+ from seawater to the site of calcification (76). Therefore, it has been suggested that initial seawater vacuolization may be followed by Ca2+-transport in very small vesicles that bud off from the larger ones, but are invisible even when using confocal microscopy (76) The absence of Mg2+-related proteins in the current study concurs with previous transcriptomic analyses revealing no significant Mg2+-related gene expression during A. beccarii calcification (22). Thus, the identification of proteins involved in vesicle generation and active Ca2+ transport and the apparent absence of similar functionality for Mg2+ suggested by the current study align with a dual ion transport mechanism involving both vacuolization (10) and transmembrane transport (77). Specifically, we propose that initial seawater vacuolization is followed by active release of Ca2+ ions from the vacuoles and then uptaken into smaller, newly formed vesicles that subsequently migrate to the calcification site (Fig. 2). The cytoplasmic Ca2+ pool in adult A. lobifera is, indeed, the largest recorded so far, storing up to 90% of the calcium used for building a new chamber (78). Moreover, observations using fluorescent indicators in A. tepida and A. beccarii (76, 79) and high expression of genes involved in Ca2+ vesicle formation during calcification in the latter species (22) support our hypothesis. This dual mechanism is further supported by the presence of MgCa-particles, characterized by varying Mg/Ca ratios in the A. lobifera cytoplasm, as observed by cryo-scanning electron microscopy/energy dispersive spectroscopy (55). Given these lines of evidence, we speculate that this process could result in Mg-rich vesicles which may passively leak Mg2+, though with much lower concentration than in seawater, thus explaining the very low Mg/Ca ratio in the foraminiferal test relative to inorganic calcite. Mg2+ ions can leak from vesicles in the presence of Na+ or Ca2+ (80). We note that this hypothesis may also explain the higher temperature sensitivity of the Mg/Ca ratio in low-Mg foraminifera, in contrast to inorganic CaCO3 (4). Indeed, this could result from increased Mg2+ leakage from the vesicles, potentially influenced by the general increase in diffusion constant and permeability driven by membrane conformational changes with temperature (81). This hypothesis aligns with the suggestions of ref. 82. The Mg-rich vesicles could also be involved in the formation of the Mg-rich amorphous calcium carbonate (Mg-ACC) nanoparticles observed by confocal autofluorescence, elemental, and Electron Backscatter Diffraction analysis during chamber formation in Amphistegina by Dubicka et al. (83). The same study also shows a significant release of Mg2+ ions from the ACC phase during the final stage of shell crystallization, where the ACC is assembled, aggregated, and transformed into a crystalline structure. In summary, while random incorporation of proteins within the mineral fraction cannot be excluded, likewise there are many macromolecules that are involved in the calcification process that are not components of the organic matrix. Thus, the occluded matrix proteins represent only the “tip of the iceberg” (84) in terms of the molecular machinery involved in the biomineralization process. Nonetheless, this comprehensive proteomic study of foraminifera tests highlights a matrix of protein functions and interactions, serving as a valuable resource for deeper exploration of the molecular mechanisms behind biomineralization. Specifically, we propose a foraminiferal calcification model (Fig. 2) involving 1) seawater vacuoles in the cytoplasm that serve as ion reservoirs, 2) active Ca2+ excretion from the vacuoles into newly formed Ca-rich vesicles that transport Ca2+ to the calcification fluid, 3) formation of Mg-rich vesicles, 4) enzymatic catalysis and supply of intracellular bicarbonate to the calcification fluid, and 5) formation of proton enriched vesicles upon carbonic anhydrase catalysis. The Mg-rich vesicles could both neutralize the inhibition effects of Mg2+on the kinetics of calcite formation and stabilize ACC precursors during the initial phases of chamber formation (83). Gene editing techniques, which are potentially possible in this symbiotic foraminiferan, and protein labeling experiments to monitor their expression during calcification would provide a clearer understanding of specific protein functions and their role(s) in vital effects. Materials and Methods Sample Collection and Handling. On September 5th 2022, 4140 A. lobifera specimens were collected from Tel Shikmona on the Israeli Mediterranean coast which is known to harbor abundant foraminifera. Macroalgal samples were washed over a 500 to 2,000 µm sieve and A. lobifera specimens were picked under the microscope. Live specimens were identified by cytoplasm color and cleaned by brushing. Tests were oxidized overnight in 1:1 of 30% H2O2: 3% NaClO to remove organic material on the exterior of the test, copiously rinsed in ultrapure water, and dried overnight at 6 °C. Bleached tests were sent to the Department of Marine and Coastal Sciences at Rutgers University for proteomic analyses. Test Organic Matrix Extraction. Dried tests were ground to <150 μm diameter and then reoxidized, rinsed, and dried again to further remove any external organics. Cleaning was sufficient to remove contaminant proteins, as determined by sonicating the cleaned powder at room temperature in filter-sterilized phosphate-buffered saline (PBS, pH 7.4) for 30 min, pelleting the powder at 8,000×g for 30 min at room temperature, concentrating the supernatant on 3 kDa Centricon YM-3 ultracentrifugal filter units (Millipore) and running bichronoic acid assays (BCA) and 8 to 16% SDS- PAGE TGX Stain-Free gels (Bio-Rad) (Fig. 3A). Approximately 1.5 g of cleaned test powder was decalcified using a 16 cm-long osmotic tube for dialysis (MWCO = 3.5 kDa Thermo Scientific) with 5 mL of 0.2 µm filtered milli-Q water. The sealed membrane was placed into 2 L of 0.1 M CH3COOH (glacial acetic acid) solution under stirring for at least 72 h. The membrane containing the dissolved organic matrix (OM) was dialyzed against 0.2 µm filtered milli-Q water until the final pH was about 6. The obtained aqueous solution containing the OM was centrifuged at 15,000×g for 30 min at 4 °C to separate the soluble (SOM) and the insoluble (IOM) OM fractions. SOM and IOM proteins were separated by SDS/PAGE, and bands were visualized by silver staining (Pierce silver stain for mass spectrometry) (Fig. 3B). (A) Concentrated PBS that had been soaked on foraminifer A. lobifera powder oxidized three times and run on a 4 to 20% gel by SDS-PAGE. (B) A. lobifera test organic matrix proteins separated by SDS/PAGE. Bands visualized by silver staining. MW (kDa): molecular weight, SOM: soluble organic matrix, IOM: insoluble organic matrix Liquid Chromatography–Tandem Mass Spectrometry (LC–MS/MS). Amphistegina lobifera samples were analyzed by LC–MS using Nano LC–MS/MS (Dionex Ultimate 3,000 RLSCnano System, Thermofisher) interfaced with Eclipse (ThermoFisher) at the Rutgers Center for Advanced Biotechnology and Medicine (CABM), following established procedures used for similar samples from coccolithophores and corals (16, 85). Specifically, three microliters of 12.5 µL in-gel was digested using trypsin. Sample P was loaded on to a fused silica trap column (Acclaim PepMap 100, 75UM×2CM, ThermoFisher). After washing for 5 min at 5 µL/min with 0.1% TFA, the trap column was brought in-line with an analytical column (Nanoease MZ peptide BEH C18, 130A, 1.7 µm, 75um×250mm, Waters) for LC–MS/MS. Peptides were fractionated at 300 nL/min using a segmented linear gradient 4 to 15% B in 30 min (where A: 0.2% formic acid, and B: 0.16% formic acid, 80% acetonitrile), 15 to 25% B in 40min, 25 to 50% B in 44min, and 50 to 90% B in 11min. Solution B then returns at 4% for 5 min for the next run. The scan sequence began with an MS1 spectrum (Orbitrap analysis, resolution 120,000, scan range from M/Z 375 to 1,500, automatic gain control (AGC) target 1E6, maximum injection time 100 ms). The top S (3 s) duty cycle scheme was used to determine the number of MS/MS performed for each cycle. Parent ions of charge 2 to 7 were selected for MS/MS and dynamic exclusion of 60 s was used to avoid repeat sampling. Parent masses were isolated in the quadrupole with an isolation window of 1.2 m/z, AGC target 1E5, and fragmented with higher-energy collisional dissociation with a normalized collision energy of 30%. The fragments were scanned in Orbitrap with resolution of 15,000. The MS/MS scan ranges were determined by the charge state of the parent ion, but lower limit was set at 110 amu. MS Database Search. The peak list of the LC–MS/MS was generated by Thermo Proteome Discoverer (v. 2.1) into MASCOT Generic Format (MGF) and searched against the A. lobifera transcriptome (86) translated protein database and common lab contaminants using an in-house version of X!Tandem (GPM Fury) (87). Search parameters were chosen based on previous work on similar systems (16, 85): fragment mass error, 20 ppm; parent mass error, ± 7 ppm; fixed modification, no fixed modification; variable modifications: methionine mono-oxidation for the primary search, asparagine deamination, tryptophan oxidation and dioxidation, methionine dioxidation, and glutamine to pyro-glutamine were considered at the refinement stage. Protease specificity: trypsin (C-terminal of R/K unless followed by P with 1 missed cleavage during the preliminary search and five missed cleavages during refinement. Minimum acceptable peptide and protein expectation scores were set at 10 to 2 and 10 to 4, respectively. The overall peptide false positive rate (88) was 0.07%. Bioinformatics. Despite the inclusion of a contaminants database, several proteins likely of human origin were sequenced and attributed to A. lobifera. To filter out these potential contaminants from our final list of test-specific proteins, we BLASTed all sequences against the “Primates” database in NCBI (taxa: 9443, Primates) using Blast2GO (number of blast hits = 20). We then examined NCBI-generated sequence alignments of coral versus Homo sapiens proteins with e-values lower than e−50. We removed from our final list of test proteins any sequences with i) e-values lower than e−100, ii) percent similarity greater than 80%, and iii) three or more peptides each of seven or more amino acids in length that were identical between A. lobifera and humans. All remaining test-specific proteins identified by the LC–MS\MS analysis were filtered to those with at least two peptides or at least one peptide with at least 10 spectra. A list of all proteins detected with e-values ≤ 10−10 was used for blast analysis against NCBI and to query sequences translated using TransDecoder v.5.5.0 (89, 90) in the OmicsBox 3.2.4 from the naked basal foraminifer Reticulomixa filosa (PRJNA29155), the endomyxans Spongospora subterranea (GCA_900404475) and Plasmodiophora brassicae (GCA_003833335) and the cercozoan Bigelowiella natans (GCA_000320545), which are all members of Rhizaria, a major eukaryotic supergroup that includes Foraminifera, Radiolaria, and Cercozoa (91, 92). The latter analysis was performed to narrow down and maximize our search for protein homologs within organisms closely related to A. lobifera. Predicted proteins with similarities greater than 40% and e-values ≤ 10−10 were retained. We then sorted all skeletal proteins by their gene ontology (GO) terms and grouped them by terms of interest: cytoskeleton and microtubule-based processes, signal transduction and regulation, metabolic and biosynthetic processes, proteolysis and autophagy, translation and related processes, protein folding, phosphorylation and transport, mitochondrial transmembrane transport, Ca2+ transportation, vesicle generation, CaCO3 precipitation, innate immune system, photosynthesis-related, unknown function, no hit Supplementary Material Appendix 01 (PDF) Dataset S01 (XLSX)