Most genes involved in excretion in nephrozoa are already present in non-nephrozoan and non-bilaterian animals

To get an overview of the presence of excretion-related genes in xenacoelomorphs and non-bilaterian animals, we first searched for the orthologous sequences of 20 nephrozoan candidate genes in the available transcriptomes and draft genomes of 13 xenacoelomorph species as well as in representatives of cnidarians, placozoans, and sponges (S1a and S2 Figs). We found that most of these genes were present in almost all groups with the exceptions of slc5 (a sodium glucose cotransporter), which was only present in Cnidaria and Bilateria, and the ultrafiltration component nephrin/kirre, which was only present in Bilateria (S1 Table, S1a Fig). This analysis also revealed that the last common ancestor of placozoans, cnidarians, and bilaterians already had at least two paralogs of amts (amt2/3 and am1/4) and one rhesus, with independent duplications of one or both of these genes in various animal lineages (S2 Fig). To identify potential excretory sites in xenacoelomorphs, we examined the expression of the entire set of these candidate excretion-related genes in the acoel Isodiametra pulchra and the nemertodermatid Meara stichopi, which differ in their digestive system composition (I. pulchra has a syncytial, lumenless gut, whereas M. stichopi has an epithelia-lined, cellular gut [49] (S1b Fig).

Genes encoding slit diaphragm–related components and SLCs are expressed broadly in I. pulchra and M. stichopi

Genes related to ultrafiltration sites (nephrin/kirre, cd2ap, zo1, stomatin/podocin) and SLCs (slc1, slc4, slc5, slc8, slc9, slc12, slc13, slc26) were broadly expressed within neural (brain and nerve cords), parenchymal/subepidermal, digestive, and gonadal-associated cells in both animals (S3 and S4 Figs). A summary of these expression patterns is summarized in Table 1. The broad expression of ultrafiltration-related components and SLCs in acoelomorphs shows that they are not part of defined excretory domains, as in nephrozoans, thus suggesting that the spatial arrangement of these genes resulting in the formation of specialized excretory compartments (e.g., nephridial compartments) took place in the nephrozoan lineage.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Summary of expression patterns of excretion-related genes in I. pulchra and M. stichopi.

https://doi.org/10.1371/journal.pbio.3000408.t001

The expression of a number of ammonia excretion–related genes and aquaporins suggests digestive-associated domains as putative excretion sites in I. pulchra and M. stichopi

Genes involved in ammonia excretion (rhesus/amts, nka, v-ATPase B, ca, hcn) and aquaporins were mainly demarcating neural, digestive-associated, and other parenchymal/subepidermal cells, as well as epidermal cells (Fig 2a, S4 and S5 Figs). A summary of these expression patterns is summarized in Table 1. The expression of ammonia excretion–related genes and aquaporins shows that these genes do not label demarcated excretory domains. However, because transcripts of the ammonium transporters rhesus, nka, and hcn (only in M. stichopi), the proton exchanger v-ATPase, as well as a number of cas and aquaporins, were found in association with the digestion system, the possibility that digestive-associated tissues could act as excretory sites was raised.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Excretion in acoelomorphs.

(a) WMISH of rhesus, v-ATPase, nka, and hcn in I. pulchra and M. stichopi. (b) Ammonia excretion rates of I. pulchra before (Ctrl) and after exposure for 2 hours to 50, 100, 200, and 500 μM and after exposure for 7 days in 1 mM NH₄Cl (boxplot). Excretion was measured over 2 hours following the HEA treatments in at least three independent biological replicates, each divided into two separate samples (six measurements in total). Bold horizontal bars in boxes indicate the median; lower and upper box borders indicate lower and upper quartile; and whiskers indicate minimum and maximum. Asterisks label significant changes (p < 0.02 in an unpaired, 2-tailed t test with unequal variance). (c) Quantitative relative expression of rhesus, nka, v-ATPase B, amts, aq, and ca after 7 days of exposure in HEA (1 mM NH₄Cl). Each circle indicates the average of three independent biological replicates, each with four technical replicates. Error bars indicate minimum and maximum of the biological replicates (averaged technical replicates). A 1-fold change represents no change; ≥2 indicates significantly increased expression level; ≤0.5 indicates significantly decreased expression level (red labels). (d) Effects of different inhibitors on ammonia excretion rates in I. pulchra (boxplot, with illustration and replicates similar to Fig 2b). The concentrations used were 5 μM Con-C as a v-ATPase A/B inhibitor, 1 mM azetazolamide as an inhibitor of the CA, 1 mM quabain as an NKA inhibitor, and 2 mM colchicine for inhibiting the microtubule network. Con-C was diluted in 0.5% DMSO for which we used an appropriate Ctrl with 0.5% DMSO. (e) Protein localization of Rhesus in I. pulchra and M. stichopi. Syncytium and gut are indicated in gray, and the magenta staining of the lumen in M. stichopi is false-positive staining of the gut content. Fluorescent pictures are projections of merged confocal stacks. The nervous system is stained green with tyr tubulin. (f) Double fluorescent WMISH of v-ATPase and nka, aq c and nka, v-ATPase and aq b, and v-ATPase and rhesus in I. pulchra. White areas in the first panel are the result of merged stacks and not of overlapping expression. Nuclei are stained blue with DAPI. Anterior is to the left. Scale bars are 50 μm for I. pulchra and 100 μm for M. stichopi. Values underlying panels b and d are provided in S6 Table, and values underlying panel c are provided in S4 Table. amt, ammonia transporter; aq, aquaporin; CA, carbonic anhydrase; Con-C, concanamycin C; Ctrl, control; DAPI, 4',6-diamidino-2-phenylindole; ds, digestive syncytium; gwc, gut-wrapping cell; HCN, K⁺[NH4⁺] channel; HEA, high environmental ammonia; NKA, Na⁺/K⁺[NH4⁺] ATPase; Rh, Rhesus glycoprotein; slc, solute carrier transporter; tyr, tyrosinated; v-ATPase, vacuolar H⁺-ATPase proton pump; WMISH, whole-mount in situ hybridization.

https://doi.org/10.1371/journal.pbio.3000408.g002

High environmental ammonia exposure indicates a diffusion mechanism in I. pulchra

To reveal the excretory mechanism in xenacoelomorphs, we conducted high environmental ammonia (HEA) incubation experiments, as previously performed in a large array of animals (summarized in [5,50]), using I. pulchra because of its availability in sufficient numbers. We first measured the pH of incubation mediums with different HEA concentrations (up to 1 mM) and found no difference in pH, which could otherwise have influenced any excretion rates. We then exposed animals to different HEA concentrations for a short period (2 hours) and measured the ammonia excretion during the following 2 hours, after bringing them back into normal conditions, to test excretion via diffusion. The ammonia excretion rates of exposed animals remained unchanged after exposure to 50 and 100 μM NH₄Cl, compared with the control conditions, but increased gradually after exposure to NH₄Cl concentrations of 200 and 500 μM NH₄Cl (Fig 2b). The increase in ammonia excretion rate could be explained by a concentration-dependent ammonia uptake during the HEA exposure and a subsequent release in normal conditions. These results suggest that ammonia excretion is concentration-dependent, which is indicative of a diffusion mechanism.

HEA exposure influences the expression of some excretion-related genes in I. pulchra

To test for a possible involvement of the excretion-related genes in the excretory mechanism of xenacoelomorphs, we tested for alteration of mRNA expression levels in chronically HEA-exposed animals by quantitative relative expression experiments (quantitative PCR [qPCR]) in I. pulchra. We first exposed animals to 1 mM HEA for 7 days, similar to conditions used in previous studies (summarized in [5,50]), and measured the ammonia excretion over 2 hours after bringing the animals into normal conditions. As expected, the ammonia excretion rates were strongly increased, in line with the short-term HEA-exposure experiments (Fig 2b). When we tested the expression level of excretion-related genes, we found that the expression of the passive ammonia transporters rhesus and three amts, as well as the active ammonia transporter nka, altered significantly (Fig 2c). Other differentially expressed genes were four aquaporins, the v-ATPase, and three cas (Fig 2c). These results indicate a putative role of these genes in ammonia excretion and suggest that acoels might not only excrete by diffusion and via passive transporters (rhesus, amts) but also by an alternative active transport mechanism (nka).

Inhibitor experiments support an active excretion mechanism via NKA transporter, as well as a passive vesicular transport mode, possibly mediated by Rhesus transporter

We further tested the involvement of NKA, V-ATPase A/B, and CA proteins in excretion, as well as a possible involvement of a vesicular transport mechanism, by conducting pharmacological inhibitor assays in I. pulchra, as previously demonstrated in other animals (summarized in [5,50]). Inhibition of the CA by azetazolamide did not show any significant change in ammonia excretion. Inhibition of the V-ATPase by concanamycin C seemed to lead to a decrease in ammonia excretion, although a 2-tailed t test did not support a significant change. In contrast, when perturbing the function of NKA with quabain, the ammonia excretion dropped significantly (Fig 2d), which further supports an active excretion mechanism via NKA, similar to what is described for many nephrozoans [10–12,14,17,19,20,25–27,31,51–53]. Interference with the vesicular transport using colchicine also led to a significant decrease in ammonia excretion, indicating a possible vesicular ammonia-trapping excretion mode (Fig 1c), as demonstrated in the midgut epithelium of the tobacco hornworm [26], the gills of the shore crab [27], and the integument of the nematode [14]. To test whether vesicular transport might occur through Rhesus transporter as shown in other studies (summarized in [5,50]), we revealed Rhesus protein localization by immunohistochemistry (Fig 2e). The protein localization mimicked the gene expression and revealed, apart from cells at the anterior tip and cells of the posterior epidermis, individual parenchymal cells affiliated with the digestive syncytium that extend ventrally. Higher magnification showed that the transporter was present in cytoplasmic vesicles and not on the cellular membrane (Fig 2e). This further indicated the presence of a vesicular transport mechanism, in which cellular ammonia moves via Rhesus transporters into vesicles and gets transferred to the membrane through the microtubule network [54] (Fig 1c). The antibody specificity was confirmed by an alignment of the epitope and the endogenous protein, as a well as a western blot analysis (S6 Fig). Similar vesicular protein localization was also observed in M. stichopi, suggesting a similar cytoplasmic–vesicular role of Rhesus transporter in gut-affiliated cells, also in nemertodermatids (Fig 2e). These data further supported the involvement of NKA and Rhesus transporters in ammonia excretion and also indicated the presence of a putative vesicular transport mode in I. pulchra.

Double fluorescent whole-mount in situ hybridization of differentially expressed genes shows similar spatial arrangement in gut-associated domains in I. pulchra and M. stichopi

To obtain a better resolution and understanding of the relative topology of the differentially expressed genes, double fluorescent whole-mount in situ hybridization (WMISH) was conducted for v-ATPase, nka, aquaporins b and c, and rhesus (Fig 2f, S4 Fig). Nka and v-ATPase were expressed in a mutually exclusive manner, with v-ATPase to be restricted in the ventral region digestive syncytium and nka in an adjacent parenchymal, distal sac-shaped gut-wrapping domain (Fig 2fA, S4C Fig). The expression of nka was not extending to the male gonads (testes) (S4A Fig). Aquaporin c was coexpressed with nka (Fig 2fB), and aquaporin b was partially overlapping with v-ATPase, with aquaporin b expression extending into the posterior region of the digestive syncytium (Fig 2fC). Finally, rhesus was partly coexpressed with v-ATPase in the ventral region of the digestive syncytium (Fig 2fD). Similar coexpression analysis of the orthologous genes was also conducted for M. stichopi and revealed striking similarities in their spatial arrangement to I. pulchra (S4 and S7 Figs). v-ATPase expression was not overlapping with nka, as v-ATPase was restricted to the gut epithelium and in two proximal lateral rows of subepidermal cells, whereas nka was limited to cells lining the distal part of the epithelial branches of the gut extending toward the subepidermis (S4E and S7 Figs). v-ATPase was partly coexpressed with rhesus in ventral gut-affiliated cells (S7 Fig). Overall, these data revealed a similar spatial arrangement in gut-associated domains in both animals, which seems to be unrelated to the presence of an epithelial gut or a syncytium. However, given the fact that I. pulchra has a lumenless digestive tissue, ammonia is probably accumulated intracellularly in the syncytium before it gets expelled via the mouth, whereas in the case of M. stichopi, ammonia gets released in the gut lumen.

Taken together, our findings suggest that I. pulchra uses different mechanisms for ammonia excretion that are also known from nephrozoans; an active ammonia excretion mechanism via NKA through the digestive system, as suggested by in situ hybridization, and a passive vesicular transport mechanism likely mediated by Rhesus, through digestive and likely also epidermal tissues. Given the commonalities in the expression of the involved genes in both animals, these excretory mechanisms could be plesiomorphic for acoelomorphs.

HEA experiments suggest a diffusion mechanism also in the cnidarian N. vectensis

Because our results showed the involvement of active and passive transport mechanisms across digestive tissues outside Nephrozoa, we also investigated a non-bilaterian species, the cnidarian N. vectensis (S1b Fig), to test whether this excretion mode might also be present outside Bilateria. The only available excretion studies in cnidarians are few morphological studies, which suggested that the septa filaments of the anthozoan mesenteries and the radial canals of medusozoans could serve as putative excretory sites [55], as well as some isotopic exchange experiments in Hydra oligactis that have shown that the gastrodermis seems to be involved in osmoregulation [56]. Moreover, there is evidence that Rh and AMT transporters are generally involved in ammonium excretion in corals [57], but localization studies that would suggest excretion sites are missing.

We first tested whether N. vectensis excretes via diffusion by exposing early-juvenile animals to HEA for 2 hours and measuring their ammonia excretion rates afterward, similar to the experiments performed with I. pulchra. We found that, also in N. vectensis, ammonia excretion increased significantly after HEA exposure starting at 200 and 500 μM NH₄Cl (Fig 3a). Measurements of the pH of each incubation medium showed that the pH dropped by 0.2 when the medium contained 500 μM NH₄Cl. However, when we measured the excretion of animals over 2 hours in a medium with an accordingly lowered pH, we found that a difference of 0.2 did not change the excretion rates (S2 Table). Therefore, the increase in ammonia excretion rates at 200 and 500 μM NH₄Cl indicates that ammonia excretion is concentration-dependent, supporting a diffusion mechanism also in N. vectensis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Excretion in N. vectensis.

(a) Ammonia excretion rates of N. vectensis before (Ctrl) and after exposure for 2 hours to 50, 100, 200, and 500 μM and after exposure for 7 days in 1mM NH₄Cl (boxplot). Excretion was measured over 2 hours following the HEA treatments in at least three independent biological replicates, each divided into two separate samples (six measurements in total). Bold horizontal bars in boxes indicate the median; lower and upper box borders indicate lower and upper quartile; and whiskers indicate minimum and maximum. Asterisks label significant changes. Significance, p < 0.02 (unpaired t test with unequal variance). (b) Quantitative relative expression of rhesus, nka, v-ATPase B, amts, and ca after exposure for 7 days in HEA (1 mM NH₄Cl). Each circle represents the average of five independent biological replicates, each with three technical replicates. A 1-fold change represents no change; ≥2 indicates increased expression level significantly; ≤0.5 indicates decreased expression level significantly (red labels). (c) Effects of different inhibitors on ammonia excretion rates in N. vectensis (boxplot, with illustration and replicates similar to Fig 2d). The concentrations used were 5–15 μM Con-C as a V-ATPase A/B inhibitor, 1–3 mM azetazolamide as an inhibitor of the CA, 1–5 mM quabain as an NKA inhibitor, and 2–10 mM colchicine for inhibiting the microtubule network. Quabain was diluted in 0.5% DMSO, for which we used an appropriate Ctrl with 0.5% DMSO. N = 3 for all treatments. (d) Whole-mount in situ hybridization of rh 1, rh 2, rh 3, v-ATPase, and amt1/4b in feeding primary polyps. Anterior is to the top. (e) Protein localization of Rh and v-ATPase in N. vectensis early-juvenile polyps. The muscle filaments are labeled green with phalloidin, and the nervous system is stained cyan with tyr tubulin. Every picture is a full projection of merged confocal stacks. Nuclei are stained blue with DAPI. The regions shown are indicated with dashed boxes in the illustrated animal. Values underlying panels a and c are provided in S6 Table, and values underlying panel b are provided in S4 Table. amt, ammonia transporter; CA, carbonic anhydrase; Con-C, concanamycin C; Ctrl, control; DAPI, 4',6-diamidino-2-phenylindole; ebw, endodermal body wall; HEA, high environmental ammonia; mes, mesenteries; nka, Na⁺/K⁺[NH4⁺] ATPase; ph, pharynx; rh, Rhesus glycoprotein; sf, septal filament; ten, tentacles; tyr, tyrosinated; v-ATPase, vacuolar H⁺-ATPase proton pump.

https://doi.org/10.1371/journal.pbio.3000408.g003

Quantitative gene expression and inhibitor experiments indicates an involvement of passive but not active transport mechanisms in N. vectensis

We then exposed animals to 1 mM HEA for 7 days and tested the expression of the orthologous genes altered in I. pulchra treatments (rh/amt, nka, v-ATPase B, and ca) by qPCR. As expected from the short-term HEA-exposure experiment, specimens exposed for 7 days in the HEA condition showed increased ammonia excretion rates (Fig 3a). In contrast to I. pulchra, none of the two nka transporters showed a significant change in gene expression in animals exposed to HEA for 7 days (Fig 3b). However, the expression of the passive transporters rhesus1, rhesus2, amt1/4b, and amt2/3e, as well as v-ATPase, altered significantly (Fig 3b), indicating the putative involvement of these transporters in excretion of N. vectensis. To test whether Rhesus acts via a vesicular transport mechanism, we conducted the same pharmacological experiment as in I. pulchra. Contrary to the results from acoels, we found that inhibition of vesicular transport did not alter the ammonia excretion (Fig 3c). We also inhibited the excretory function of V-ATPase and CA proteins and found that none of them showed any significant change in ammonia excretion rates (Fig 3c). Finally, when we perturbed the function of NKA, the ammonia excretion rates did not alter (Fig 3c), confirming the qPCR results (Fig 3b) and further supporting the non-involvement of the NKA transporter in excretion. These results suggest that the ammonia excretion of N. vectensis is likely mediated by the passive Rhesus and AMT transporters, but neither relies on active transport mechanism mediated by NKA or on vesicular ammonia-trapping excretion mode.

Gene expression of excretion-related genes reveals the gastrodermis as excretory site in N. vectensis

To understand whether these genes were expressed in gastrodermal or epidermal cells, we revealed the spatial expression of rhesus, amts, nka, and v-ATPase B by WMISH in feeding primary polyps. All genes were mainly demarcating gastrodermal domains, such as the endodermal body wall, the directive mesenteries, septal filaments, and the pharynx (Fig 3d, S8 Fig). Rhesus 1 was additionally expressed in the tentacular ectoderm (Fig 3dA). Protein localization of Rh, NKA, and V-ATPase B reflected the transcript expression patterns (Fig 3e, S9 Fig). High magnification of Rhesus antibody staining further revealed that the transporter was not expressed in cytoplasmic vesicles, supporting a non-vesicular transport mechanism, in agreement with the inhibitor experiments. Also, it showed that Rhesus was localized in individual cells of the tentacular ectoderm with clumped structures at the tentacle surface, which resembled gland cells [58] (Fig 3e). The NKA antibody was localized in endodermal neurons and individual cells of the mesenteries, likely neural precursors (S9 Fig), thus suggesting a non-excretory function of this transporter, as indicated already from the qPCR and inhibitor experiments. These data imply that gastrodermis-affiliated domains likely serve as excretory sites in N. vectensis.

Gene cloning and orthology assignment

Putative orthologous sequences of genes of interest were identified by tBLASTx search against the transcriptome (SRR2681926) of I. pulchra, the transcriptome (SRR2681155) and draft genome of M. stichopi, and the genome of N. vectensis (http://genome.jgi.doe.gov). Additional transcriptomes of Xenacoelomorpha species investigated were as follows: Childia submaculatum (Acoela) (SRX1534054), Convolutriloba macropyga (Acoela) (SRX1343815), Diopisthoporus gymnopharyngeus (Acoela) (SRX1534055), Diopisthoporus longitubus (Acoela) (SRX1534056), Eumecynostomum macrobursalium (Acoela) (SRX1534057), Hofstenia miamia (Acoela) (PRJNA241459), Ascoparia sp. (Nemertodermatida) (SRX1343822), Nemertoderma westbladi (Nemertodermatida) (SRX1343819), Sterreria sp. (Nemertodermatida) (SRX1343821), Xenoturbella bocki (Xenoturbella) (SRX1343818), and Xenoturbella profunda (Xenoturbella) (SRP064117). Data were deposited in the Dryad repository: https://doi.org/10.5061/dryad.bq068jr [72].

Sequences for the placozoan Trichoplax adhaerens, the sponge Amphimedon queenslandica, the ctenophore Mnemiopsis leidy, the protist Capsaspora owczarzaki, the amoeba Dictyostelium discoideum, and the nephrozoans Homo sapiens. Saccoglossus kowalevskii, Strongylocentrotus purpuratus, Xenopus laevis, Branchiostoma lanceolatum, Capitella teleta, Crassostrea gigas, Lottia gigantea, Schmidtea mediterranea, Tribolium castaneum, Caenorhabditis elegans, and Drosophila melanogaster were obtained from Uniprot and NCBI public databases. Gene orthology of genes of interest identified by tBLASTx was tested by reciprocal BLAST against NCBI Genbank and followed by phylogenetic analyses. Amino acid alignments were made with MUSCLE [73]. RAxML (version 8.2.9) [74] was used to conduct a maximum-likelihood phylogenetic analysis. Fragments of the genes of interest were amplified from cDNA of I. pulchra, M. stichopi, and N. vectensis by PCR using gene-specific primers. PCR products were purified and cloned into a pGEM-T Easy vector (Promega, Madison, WI, USA) according to the manufacturer’s instructions, and the identity of inserts was confirmed by sequencing. Gene accession numbers of the gene sequences are listed in the S3 Table.

Animal systems

Adult specimens of I. pulchra (Smith & Bush, 1991), M. stichopi Westblad, 1949, and N. vectensis Stephenson, 1935 were kept and handled as previously described [75–78].

WMISH

Animals were manually collected, fixed, and processed for in situ hybridization as described [79,80]. Labeled antisense RNA probes were transcribed from linearized DNA using digoxigenin-11-UTP (Roche, Basel, Switzerland) or labeled with DNP (Mirus Bio, Madison, WI, USA) according to the manufacturer’s instructions. For I. pulchra and M. stichopi, colorimetric WMISH was performed according to the protocol outlined in [79]. For N. vectensis, we followed the protocol described by [81]. Double fluorescent in situ hybridization (FISH) was performed as the colorimetric WMISH with the following modifications: after the posthybridization steps, animals were incubated overnight with peroxidase-conjugated antibodies at 4 °C (anti-DIG-POD [Roche, Basel, Switzerland], 1:500 dilution, and anti-DNP [Perkin Elmer, Waltham, MA, USA], 1:200 dilution) followed by the amplification of the signal with fluorophore-conjugated tyramides (1:100 TSA reagent diluents [Perkin Elmer, Waltham, MA, USA] TSA Plus Cy3 or Cy5 Kit). Residual enzyme activity was inhibited via 45-minute incubation in 0.1% hydrogen peroxide in PTW followed by four PTW washes prior to addition and development of the second peroxidase-conjugated antibody [82].

Whole-mount immunohistochemistry

Animals were collected manually, fixed in 4% paraformaldehyde in SW for 50 minutes, washed 3 times in PBT, and incubated in 4% sheep serum in PBT for 30 minutes. The animals were then incubated with commercially available primary antibodies (anti-RhAG [ab55911] rabbit polyclonal antibody, dilution 1:50 [Abcam, Cambridge, UK], anti-Na⁺/K⁺ ATPase a1 subunit rat monoclonal antibody, dilution 1:100 [Sigma-Aldrich, St. Louis, MO, USA], and anti-V-ATPase B1/2 [L-20] goat polyclonal antibody, dilution 1:50 [Santa Cruz Biotechnology, Dallas, TX, USA]) overnight at 4 °C, washed 5 times in PBT, and followed by incubation in 4% sheep serum in PBT for 30 minutes. Specimens were then incubated with a secondary antibody (anti-rabbit-AlexaFluor 555 [Invitrogen, Carlsbad, CA, USA] or anti-rat-AlexaFluor 555 and anti-goat-AlexaFluor 555) diluted 1:1,000 overnight at 4 °C followed by 10 washes in PBT. Nuclei were stained by incubation of animals in DAPI 1:1,000, and f-actin was stained by incubation in BODIPY-labeled phallacidin (5 U/ml) overnight.

Inhibitor and HEA experiments

For excretion experiments, approximately 300 I. pulchra (the number varied slightly between the biological replica but was similar in the corresponding controls and treatments) and 10 N. vectensis were placed into glass vials with 2 ml UV-sterilized natural seawater (1:4 diluted with distilled water for N. vectensis) containing the appropriate inhibitor or ammonia concentration. Animals were given 10 minutes to adjust to the medium before the solution was exchanged with 2 ml of fresh medium with the same appropriate condition. For the inhibitor experiment, the medium was removed after 2 hours and stored at −80 °C for later measurements. Animals from the short-term HEA experiments were incubated for 2 hours, rinsed five times over 20–30 minutes, and incubated for another 2 hours in fresh medium without additional ammonia, after which the medium was removed and frozen at −80 °C. We tested different inhibitor concentrations that were used in previous studies in other invertebrates [10,12,14,19,27]. The concentrations of 5–15 μM concanamycin C for inhibiting V-ATPase A/B, 1–3 mM azetazolamide as an inhibitor of the CA, 1–5 mM quabain to inhibit the NKA, and 2–10 mM colchicine for inhibiting the microtubule network were selected, as no other effects like shrinking or obvious changes in morphology or behavior were observed. After the inhibitor incubations, the animals were washed several times and monitored in normal conditions for several days to ensure that the inhibitors did not cause any unspecific permanent effects. Concanamycin C was diluted in DMSO with a final concentration of 0.5% DMSO per sample, for which we used an appropriate control with 0.5% DMSO. For the HEA experiments, we enriched seawater with NH₄Cl to the final ammonia concentrations of 50 μM, 100 μM, 200 μM, 500 μM, and 1 mM. We also measured the pH of both incubation mediums (HEA and control), and we found no difference. All experiments were independently repeated at least three times at different time points, and each repeat was divided into two samples. The values are provided in S6 Table.

Determination of ammonia excretion

Ammonia concentrations were measured with an ammonia-sensitive electrode (Orion, Thermo Scientific, Waltham, MA, USA) according to [52]. Samples were diluted 1:4 with distilled water to prevent salt precipitation (900 μl sample + 2.7 ml water), and total ammonia was transformed into gaseous NH₃ by adding 54 μl ionic strength adjusting solution (1.36 ml/l trisodiumcitrate dihydrate, 1 M NaOH). Because of the small ammonia concentrations, the electrode-filling solution was diluted to 10% with distilled water, as suggested in the electrode manual. In control conditions, we determined an average excretion of 44 pmol per adult animal per hour, although the excretion varied between different biological replicates from different generations (minimum 32 pmol/animal/hour, maximum 52 pmol/animal/hour), possibly because of slightly fluctuating conditions during long-term animal rearing. Solutions with defined concentrations of NH₄Cl for the standard curves were made together with the solutions used in the experiments and stored in a similar way at −80 °C. The differences in excretion rates were tested for significance with an unpaired, 2-tailed t test with unequal variance, and a p-value < 0.02 was seen as significant. Boxplots were created with “R.”

Quantitative gene expression

In total, 100 treated I. pulchra and 5 N. vectensis were collected after 7 days of incubation in HEA conditions (1 mM NH₄Cl) and tested for quantitative gene expression using the BIORAD CFX96 (Bio-Rad, Hercules, CA, USA) Real-time PCR detection system. ddCt values were calculated between treated and control animals and converted to fold differences. All experiments were repeated three to five times with different specimens (three biological replicates for I. pulchra and five biological replicates for N. vectensis), and two to four technical replicates were tested for each biological replicate (four biological replicates for I. pulchra and three biological replicates for N. vectensis). Fold changes were calculated using polyubiquitin, actin, and 18S as references for I. pulchra and ATPsynthase and EF1b as references for N. vectensis [83], and a threshold of 2-fold difference was chosen as a significant change. The Ct values are provided in S4 Table, and the primer sequences used are provided in S5 Table.

Western blot

Whole-animal extracts (50 I. pulchra adults and 5 N. vectensis juveniles) were fractionated by SDS-PAGE, loaded on Mini-PROTEAN TGX Stain-Free Precast Gels (Bio-Rad, Hercules, CA, USA), and transferred to a nitrocellulose membrane using a transfer apparatus according to the manufacturer’s protocols (Bio-Rad, Hercules, CA, USA). After incubation with 5% nonfat milk in TBST (10 mM Tris, [pH 8.0], 150 mM NaCl, 0.5% Tween 20) for 60 minutes, the membrane was washed once with TBST and incubated with antibodies against Rhesus (1:1,000) and NKA (1:500) at 4 °C for 12 hours. Membranes were washed three times for 10 minutes and incubated with a 1:5,000 dilution of horseradish peroxidase–conjugated anti-mouse or anti-rabbit antibodies for 2 hours. Blots were washed with TBST three times and developed with the ECL system (Amersham Biosciences, Little Chalfont, UK) according to the manufacturer’s protocols.

Documentation

Colorimetric WMISH specimens were imaged with a Zeiss AxioCam HRc mounted on a Zeiss Axioscope A1 equipped with Nomarski optics and processed through Photoshop CS6 (Adobe, San Jose, CA, USA). Fluorescent-labeled specimens were analyzed with a Leica SP5 confocal laser microscope (Leica Microsystems, Wetzlar, Germany) and processed by the Fiji software version 2.0.0-rc-42/1.50d [84]. Figure plates were arranged with Illustrator CS6 (Adobe, San Jose, CA, USA). Data were deposited in the Dryad repository: https://doi.org/10.5061/dryad.bq068jr [72].