Na⁺/H⁺ exchangers differentially contribute to midgut fluid sodium and proton concentration in the sea urchin larva

In contrast to the acidic gastric conditions found in most vertebrates, larval stages of many hemichordate and echinoderm species have highly alkaline conditions in their digestive system, with pH values ranging from pH 8.5 to 10.5 (Stumpp et al., 2015; Hu et al., 2017). Highly alkaline conditions are well described in the gut of lepidopteran and dipteran larvae as well as in some termite species (Brune and Kühl, 1996; Boudko et al., 2001a). In insects, these alkaline conditions are believed to be favorable for the digestion of a plant- and lignocellulose-rich diet (Boudko et al., 2001a,b; Brune and Kühl, 1996) and may serve as the first line of immunity against ingested pathogens (Clark et al., 1999). Also, in the sea urchin larva, highly alkaline conditions of the midgut are involved in the breakdown of microalgae and serve as an important mechanism to protect against ingested pathogens (Schwenzfeier et al., 2011; Stumpp et al., 2013, 2020). In contrast to vertebrate and insect models, the ectoderm of the sea urchin larva is permeable for molecules up to 4 kDa (Stumpp et al., 2012). Consequently, the ion composition of the extracellular matrix (ECM) resembles the ion composition of the surrounding seawater and the cells of the midgut epithelium are directly exposed to pH and ionic changes of the environment. Acidification experiments demonstrated that the midgut epithelium has high capacities to maintain midgut fluid pH under a wide range of seawater pH conditions (Lee et al., 2019). However, chronic CO₂-induced acidification down to pH 7.4 directly translated into a small decrease of 0.3 to 0.5 pH units in midgut pH, leading to lower digestion efficiencies due to a shift in the pH optimum for major digestive proteases (Stumpp et al., 2013). Additionally, a previous study demonstrated that strict regulation of highly alkaline midgut conditions (pH 10.5) leads to a high sensitivity of hemichordate larvae to CO₂-induced seawater acidification (Hu et al., 2017).

These studies suggested that energy allocations to the maintenance of a proton gradient up to three orders of magnitude may represent a key trait determining the sensitivity of marine larvae to reductions in environmental pH. Accordingly, the existence and ability to defend midgut pH against large proton gradients across the midgut epithelium indicates the presence of well-developed ion-transporting properties in the larval midgut (Stumpp et al., 2013). Ultrastructural and molecular analyses suggested the presence of at least two different cell types in the midgut epithelium of the sea urchin larva. One cell type is clustered around the area of the cardiac sphincter and seems to be responsible for mediating digestive processes, seen in high expression levels of homologues to mammalian pancreatic genes (Perillo et al., 2016). The second cell type is distributed across the whole midgut epithelium and is rich in mitochondria and vesicles at the basal region of the cells (Burke, 1981). Immunohistological analysis demonstrated Na⁺/K⁺-ATPase-rich cells scattered over the gut epithelium in sea urchin larva (Stumpp et al., 2013), suggesting that this type of cell is relevant for trans-epithelial ion transport. The recent cell model for pH and ion regulation additionally includes the V⁺-type ATPases (VHA) and secondary active sodium/hydrogen exchangers (NHEs) (Stumpp et al., 2015). Ocean acidification experiments in combination with expression analyses demonstrated an upregulation of NHE-2 (spslc9a2) under chronic hypercapnic conditions. This observation has been interpreted as a compensatory mechanism to defend midgut pH against environmental acidification (Stumpp et al., 2012). Pharmacological studies in combination with immunohistological investigations of NHEs suggest an important role for these transporters in larval midgut pH regulation (Stumpp et al., 2015). The ubiquitous appearance of Na⁺/H⁺ exchangers in all organisms and cell types and their primary function for intracellular pH regulation, cell volume and regulation of intracellular sodium concentration by removing protons in exchange for extracellular sodium ions, makes these transporters prime candidates for controlling acid–base conditions in sea urchin larvae (Dibrov and Fliegel, 1998; Saier et al., 1999; Xu et al., 2018). Also, in the intestines of mammals, marine teleosts and insect midguts, alkalization of luminal fluids involves NHEs to control [Na⁺] and pH (Onken et al., 2004, 2008; Grosell, 2011). For example, inhibitor studies revealed that in most mammalian species NHE-2 and NHE-3 contribute to neutral [Na⁺] absorption in the small intestine and colon coupled to Cl⁻/HCO₃⁻ exchange (Bachmann et al., 2004).

Previous studies addressing the mechanisms of midgut alkalization in the sea urchin larva pointed towards a central role of NHEs in the alkalization processes (Stumpp et al., 2013, 2015; Hu et al., 2017; Stumpp et al., 2020). However, the identification and contribution of specific NHE isoforms to midgut alkalization remained speculative. Here we investigate the role of NHEs in regulating pH and [Na⁺] in the alkaline fluids of the larval sea urchin midgut. Using a suite of methods ranging from ion-selective microelectrodes (ISMs) over immunocytochemistry, morpholino knockdown and gene expression analyses, this work provides a detailed analysis of the contribution of NHEs to maintain midgut [Na⁺] and pH in the sea urchin larva.

Adult purple sea urchins [Strongylocentrotus purpuratus (Stimpsin 1857)] were collected on the coast of California (USA) and shipped to the Helmholtz Center of Ocean Research (GEOMAR) in Kiel, Germany. The animals were kept in a re-circulating seawater system within a climate chamber at 11°C and a salinity of 33 practical salinity units (PSU). Spawning of males and females was induced by gentle shaking. Eggs were collected in a 500 ml beaker with filtered natural seawater (FSW). Dry sperm was collected with a pipette from the aboral side and diluted (1:1.000 in FSW) for fertilization. Eggs and sperm were transferred to 500 ml beakers filled with FSW. The fertilized eggs were washed several times with FSW and cultured in 2 liter Erlenmeyer flasks at a density of 25 eggs ml⁻¹ at 15°C. For each culture, one male and one female were used. For all experiments, larvae were fed from 3 days post-fertilization (dpf) onward, with Rhodomonas sp. at a concentration of 5×10³ cells ml⁻¹.

In addition to H⁺-selective microelectrodes, which were built as previously described (Stumpp et al., 2013, 2015), sodium selective microelectrodes were used to investigate the [Na⁺] of midgut fluids in the sea urchin larvae. Sodium-selective microelectrodes were backfilled with 100 mmol l⁻¹ NaCl electrolyte and front-loaded with a 150 µm column of the sodium-selective ionophore cocktail II [10% (w/w) sodium ionophore II, 89.5% (v/v) o-NPOE, 0.5% (v/v) sodium tetraphenylborate; all chemicals were obtained from Sigma-Aldrich, Germany]. To seal the ISM opening, microelectrodes were additionally front-loaded with PVC (300 µl tetrahydrofuran, 0.01 g PVC) containing 100 µmol l⁻¹ of the respective ionophore cocktail.

To ensure a rapid and selective response, the microelectrodes were calibrated by measuring the Nernstian property in a series of FSW adjusted to different pH values (7, 8 and 9; adjusted with NaOH/HCl) or different sodium concentrations including 10 mmol l⁻¹, 50 mmol l⁻¹, 100 mmol l⁻¹, 250 mmol l⁻¹, 300 mmol l⁻¹, 350 mmol l⁻¹, 400 mmol l⁻¹, 450 mmol l⁻¹, 500 mmol l⁻¹ and 1 mol l⁻¹. As the calibration for sodium-selective electrodes is not linear, the concentration was natural log-transformed to produce a linear relationship between the measured voltage (mV) and the concentration. An Ag–AgCl (RC2F; WPI, Germany) electrode was used as a reference cell and connected with a 1 mol l⁻¹ KCl-containing agar bridge to the perfusion bath.

To exclude measurement artefacts due to an epithelial potential, a microelectrode only filled with 1 mol l⁻¹ KCl was used, to confirm the absence of an electrical potential generated by the midgut epithelium as demonstrated earlier (Stumpp et al., 2013).

To measure the intracellular pH of midgut epithelial cells using the ratiometric, pH-sensitive dye 2′,7′-bis-(2-carboxyethyl)-5-(and-6)-carboxyfluorescein-acetoxymethyl ester (BCECF-AM; Invitrogen, Germany), epidermal cells were removed using the ‘bag isolation’ method described by Harkey and Whiteley (1980). Briefly, the larvae were carefully washed twice in Mg²⁺- and Ca²⁺-free artificial seawater (ASW) and afterwards washed twice in bag isolation medium as previously described (Harkey and Whiteley, 1980), followed by two washing steps in FSW. For dye loading, larvae were incubated in FSW containing 50 µmol l⁻¹ BCECF-AM (Sigma, Germany) for 60 min at 15°C.

Intracellular pH (pH_i) regulatory abilities of midgut cells were measured by real-time fluorimetry in combination with the ammonia pre-pulse technique (Stumpp et al., 2012; Hu et al., 2018). Fluorescence intensities were recorded using the Visitron imaging system and an inverted microscope (Observer 1; Zeiss, Germany) equipped with a CCD camera (CoolSNAP HQ2; Photometrics, Tucson, AZ, USA), and analysed using the imaging software Metaflour (Visitron, Puchheim, Germany). The perfusion set-up was equipped with a water jacket cooling system and all measurements were performed at 15°C. The fluorescence ratios were translated to actual pH_i values using ASW solution containing nigericin (Millipore, Darmstadt, Germany) and high potassium (150 mmol l⁻¹) as described previously (Stumpp et al., 2012). For this calibration, different pH steps were used (pH 5.5, 6.5, 7.0, 7.5, 8.0, 9.0) and 9–12 regions of interest were summarized and treated as one replicate (N=1). The ammonia pre-pulse method was used to determine the pH_i regulatory capacities of midgut epithelial cells. Here, the recovery rate (pH_i units min⁻¹) expressed as the slope of the compensation reaction after acidosis was used to determine pH_i regulatory capacities in the presence of different EIPA concentrations. Each measurement was conducted with 4–8 dpf larvae. To generate an inhibition curve and to determine the IC₅₀ value, 10, 1, 0.1 and 0.01 µmol l⁻¹ EIPA were added during the wash-out period after the ammonia pre-pulse.

The sea urchin larva has a tight midgut epithelium with very little influx of external fluids during swallowing of food particles (Stumpp et al., 2013), whereas the extracellular matrix that is surrounded by the ectoderm is highly permeable for molecules up to 4 kDa (Stumpp et al., 2012). As EIPA is ∼0.3 kDa in size, we expected the inhibitor to reach the basolateral cell side of the midgut by diffusion through the leaky ectoderm within 45 min (Stumpp et al., 2012). Therefore, inhibitors were applied by bathing larvae in the respective inhibitor solution to access the basolateral side or by direct injection into the midgut to reach luminal membranes. Within the experimentally relevant timeframe (0.5–1 h), we expected EIPA to affect only the target side via these different application methods for two reasons. First, because of its molecular structure, including charged amino and chloride groups, we expect EIPA to have a low membrane permeability. Second, even if low concentrations of EIPA, may reach the other cell side, the concentration of minimal inhibition will be strongly diluted, probably having no or minimal inhibitory effects. For all incubation experiments, 5–10 larvae (4–8 dpf) were picked from the culture, incubated in the inhibitor (FSW containing the respective EIPA concentration dissolved in DMSO), or control solution (FSW containing 0.1% DMSO) for 90 min at 15°C. After incubation, larvae were transferred to the microscope (Axiovert 135; Zeiss, Germany) equipped with a cooling stage (15°C) and midgut pH or Na⁺ concentration were measured using ISMs as previously described (Stumpp et al., 2013, 2015).

To apply the inhibitor to the luminal side, larvae were gastrocopically injected with FSW containing 1 mmol l⁻¹ FITC (fluorescein sodium salt; Sigma-Aldrich, Germany) and 0.1% DMSO for controls or FSW containing 1 mmol l⁻¹ FITC+10 µmol l⁻¹ EIPA for the inhibitor treatment. To avoid unnecessarily long time periods after injection, four to five larvae were taken from the culture and injected under an inverted fluorescence microscope (Axiovert). For the injections, larvae were held in place using a glass pipette (tip diameter 20 µm) to which a slight vacuum was applied. The injection pipette (3–4 µm tip diameter) was carefully inserted through the oesophagus into the stomach using an oil-hydraulic micromanipulator (Nikon; Narishige, Tokyo, Japan), without injuring the gut epithelium. Both pipettes were pulled from glass capillary tubes (GB150F-8P; Science Products, Hofheim, Germany) using a DMZ puller (Zeitz Instruments, Munich, Germany); 0.1 nl was injected into the midgut using a microinjection system (Picospritzer III; Parker, Hollis, NJ, USA). The fluorescence signal of FITC was used to observe successful injection into the larval midgut and to avoid double injections of the four to five larvae. Due to the spherical shape of the larval midgut, we calculated the increase in volume before (V₁) and after injection (V₂) using V=4/3×π×r³, and used this value to determine the injected volume by subtracting V₂–V₁. For the calculations, pictures of larvae before and after injection were taken and the midgut radius was determined using the free imaging software ImageJ. The average injected volume was calculated to be 0.1 nl and allowed the determination of a final concentration of 5 µmol l⁻¹ EIPA in the midgut. To determine the effect of the injection itself, an additional control group with non-injected larvae was measured.

In total, 15–18 larvae from three different fertilizations were measured for each treatment and each larva was treated as one biological replicate.

In addition to the inhibitor experiments, luminal midgut sodium and pH were detected over a time course ranging from 4 to 8 dpf. For every measurement, seven to 20 larvae from at least three different fertilizations were measured.

For the localization of the sea urchin NHE, SpSlc9a2, a polyclonal antibody raised against a synthetic peptide of the carboxyl-terminal region (CHGHHWIKEKWEEVNHK) of the SpSlc9a2 peptide was used (Hu et al., 2020).

For double staining, the Na⁺/K⁺-ATPase was localized using a monoclonal antibody designed against the avian alpha subunit of the protein (anti-Na⁺/K⁺-ATPase alpha 5, Hybridoma Bank, diluted 1:400 in PBST) that has been previously described to recognize the Na⁺/K⁺-ATPases in sea urchin gut (Stumpp et al., 2013, 2015).

Pluteus larvae were taken from cultures and fixed in 4% (w/v) PFA dissolved in FSW for 15 min at room temperature (RT), carefully rinsed in PBST, and subsequently used for immunofluorescence staining.

To block non-specific binding of the primary antibody, larvae were incubated in PBST containing 5% (v/w) bovine serum albumin (BSA) for 1 h. The samples were incubated with the primary antibodies overnight at 4°C. To remove non-bound antibodies, samples were washed four times in PBST. The secondary antibody (Alexa anti-rabbit Fluro 488 for anti-NHE-2; Santa Cruz Biotechnology; diluted 1:250 in PBST and Alexa anti-mouse Fluro 633 for Na⁺/K⁺-ATPases; Santa Cruz Biotechnology; diluted 1:250 in PBST) were incubated for 1 h at RT, followed by three washing steps. Samples were analysed with a confocal laser scanning microscope (LSM510; Zeiss).

For negative controls, a peptide compensation assay was performed as previously described (Hu et al., 2018). Therefore, the SpSlc9a2 antibody (diluted 1:150 in PBST) was pre-absorbed with the immunization peptide for 8 h at a peptide concentration of 0.1 mg ml⁻¹. Afterwards, immunofluorescence staining was performed as described previously.

Vivo Morpholinos (Vivo MO) are antisense oligonucleotide analogues containing a covalently linked delivery moiety consisting of an octa-guanidine dendrimer, which allows the Vivo MO to enter the cell membrane without further manipulation of the cells, and binds to complementary RNA sequences and blocks the translation of the target gene (Morcos et al., 2008; Ferguson et al., 2014). Here we used a Vivo MO specially designed against the spslc9a2 gene with the sequence 5′-AGCCCCCAAAAGTCATCTCTGTATT-3′complementary to the start codon region. The spslc9a2 Vivo MO was obtained from Gene Tools (Philomath, OR, USA). Previously published studies on sea urchin embryos and pluteus larvae treated with Vivo MOs demonstrated clear knockdown phenotypes in skeleton development and bone morphogenetic protein (BMP) signalling (Heyland et al., 2014; Luo and Su, 2012).

The MO was dissolved in H₂O at a concentration of 0.5 mmol l⁻¹ for the stock solution, which was further diluted to 5, 1 and 0.1 µmol l⁻¹ as the final working concentrations to which larvae were exposed. For the experiment, 25–30 larvae (4 dpf) were picked from the culture and transferred into 1 ml FSW containing the respective Vivo MO concentration. Negative controls only contained 0.1% H₂O deionized in FSW. Larvae were incubated in these solutions for 24 h and midgut [Na⁺] was determined by ISMs. Quantification of SpSlc9a2 protein abundance was determined by immunostaining (performed as described earlier) using the custom-made antibody for SpSlc9a2 (Hu et al., 2020), followed by analysis of the intensity of the fluorescence signal in control larvae and in larvae treated with 1 µmol l⁻¹ Vivo MO. Pictures were taken on an inverted fluorescence microscope equipped with a pco.panda sCMOS camera (PCO, Kelheim, Germany) in combination with the VisiScope Life Cell Imaging System (Visitron Systems, Germany) with excitation at 488 nm and emission at 520 nm. To avoid artefactual differences in immunostaining, control larvae and the Vivo MO-treated larvae were stained and treated in parallel using the same solutions and incubation conditions. Pictures were taken in an alternate mode from controls and the Vivo MO-treated larvae using the same exposure times and microscopy settings for every image. The fluorescence intensity analysis was performed using ImageJ. For every larva, 10 regions of interest were set into SpSlc9a2-positive cells of the midgut epithelium and the intensity was measured and the average value from one larva was used as one replicate.

To examine if acidic seawater influenced the pH or [Na⁺] of the midgut fluids, pluteus larvae (5 dpf) were exposed to acidified FSW (pH 6.5, buffered with 0.03 mol l⁻¹ MES and adjusted with NaOH) for a period of 10 h. Samples were taken directly before the start of the incubation (t₀) and after 1, 2, 4, 6 and 10 h (t₁, t₂, t₄, t₆ and t₁₀). The larval cultures were split directly before the experiment started and half of the culture was incubated in acidic FSW (pH 6.5, low pH treatment) while the other half remained in FSW (pH 8.1, controls). For midgut pH and [Na⁺] measurements, larvae were picked from the culture and midgut pH was measured using ISMs. At every time point, two to four larvae from three different cultures (N=3) were measured. To examine if exposure to acidic FSW influenced mRNA expression of putative NHEs in the midgut, quantitative PCR (qPCR) samples were taken from the acidification experiment at the same time points as used for microelectrode measurements and immediately flash-frozen in liquid nitrogen and stored at −80°C.

For RNA isolation, larvae were homogenized in TriZol (Invitrogen; Waltham, MA, USA) and mRNA was isolated by using Direct-zol RNA MiniPrep Kit (Zymo Research, Freiburg, Germany). Genomic contamination was removed by DNAase I treatment. cDNA was synthesized using a SuperScript IV cDNA synthesis kit (Invitrogen) and expression levels of three different NHE isoforms were quantified by real-time qPCR using LightCycler 480 SYBR Green I Master (Roche, Germany) as previously described (Hu et al., 2018).

Target genes were three members of the Slc9 family (SpSlc9a; NHEs) found in the genome of S. purpuratus. The sequences of these genes were obtained from EchinoBase (Cameron et al., 2009; Kudtarkar and Cameron, 2017; http://legacy.echinobase.org/shiny/quantdev/) and the coding regions were used for qPCR primer design (NCBI Primer-Blast; www.ncbi.nlm.nih.gov/tools/primerblast/; Table 1). Primers were tested by PCR to confirm a single amplification product. After calculation of the ΔCt value, expression was normalized to elf 1a (elongation factor 1a), which is a suitable reference gene during development and environmental perturbations in the sea urchin larva (Hu et al., 2018, 2020).

All experiments and measurements were repeated at least three times. For the statistical analysis and figures, every larva was treated as one biological replicate. All statistical analysis and figures were generated using GraphPad Prism 7. The EIPA incubation experiments and the fluorescence intensity analysis of the Vivo MO were statistically analysed with Student's t-test, while the results of EIPA injection experiments, real-time qPCR analyses, spslc9a2 Vivo MO experiments and the pH_i measurements were analysed using one-way ANOVA followed by Tukey’s multiple comparisons post hoc test. Correlation analyses from the acidification experiment were analysed by calculating linear regressions between midgut pH and mRNA expression of spslc9a2, spslc9a7 and spslc9a8.

Midgut fluid pH and sodium concentration of early pluteus larvae were measured daily from 4 to 8 dpf. The sodium concentration in the midgut was markedly below that of seawater, fluctuating between 68 and 233 mmol l⁻¹ with a total average (N=64) of 121±5.47 mmol l⁻¹ (Fig. 1A). Accordingly, the ECM contains on average a four times higher Na⁺ concentration than fluids of the larval midgut. Highest variability in midgut [Na⁺] was detected in 5 and 8 dpf larvae ranging from 70 to 230 mmol l⁻¹ with an average [Na⁺] of 156 mmol l⁻¹.

Midgut pH during daily measurements showed a slight fluctuation from the lowest pH 8.63 up to pH 9.28 with a total average of all larvae (N=55) at pH 8.79±0.03. In total, the pH measurement had low scattering, and the highest pH values were measured in 5 and 6 dpf larvae (average values: pH 8.90±0.04 and pH 8.99±0.04).

As molecules up to 4 kDa can diffuse through the ECM within several minutes (Stumpp et al., 2012), EIPA incubation was used to address the basolateral side of the larval midgut epithelial cells. The inhibitor incubation led to a partial but significant decrease in midgut pH by 0.13 pH units (P=0.0025) from pH 8.88±0.027 in DMSO-treated control larvae (N=17) to pH 8.75±0.024 in EIPA-treated larvae (N=18).

However, midgut [Na⁺] remained unaffected by exposure to 5 µmol l⁻¹ EIPA from the basolateral side (Fig. 1B). Control (N=22) and EIPA (N=21)-treated larvae had average sodium concentrations of 107±4.9 and 107±7.5 mmol l⁻¹, respectively.

EIPA injections into the midgut were used to address the function of NHEs on the luminal side of the midgut epithelium. Here, our pH measurements demonstrated no effect at a final concentration of 5 µmol l⁻¹ EIPA, directly injected into the larval midgut (Fig. 1C). Larvae injected with EIPA (N=14) had an average pH of 8.91±0.024, while those injected with FSW only containing 0.1% of the vehicle DMSO had an average pH of 8.89±0.029 (N=14). In addition, non-injected controls (N=14) had an average pH of 8.91±0.039 (Fig. 1B). Statistical analysis using one-way ANOVA showed no significant differences between all three groups.

Luminal sodium measurements in EIPA-injected larvae (N=18) revealed a significant increase in [Na⁺] (P<0.001, d.f.=46) by 135 mmol l⁻¹ compared with larvae injected with the control solution (N=16) only containing 0.1% DMSO and FITC. The average midgut sodium concentration of EIPA-injected larvae was 370±24 mmol l⁻¹, while larvae injected with the control solution had a total average sodium concentration of 235±17 mmol l⁻¹. Despite a significant difference between EIPA- and DMSO-injected larvae, both treatments had a significant (P=0.011) increase in luminal [Na⁺] compared with uninjected larvae (N=15) with an average sodium concentration of 166±13 mmol l⁻¹.

To test the overall contribution of NHEs on pH_i regulatory abilities in midgut epithelial cells, the ammonia pre-pulse technique was used in combination with the pH-sensitive dye BCECF-AM and the NHE specific blocker EIPA. The compensation reaction against the intracellular acidosis was determined by the rate of pH_i recovery expressed as pH_i units min⁻¹. Calibration of BCECF fluorescence ratios to actual pH values resulted in a sigmoidal calibration curve (Fig. 2D). Controls (N=15) recovered completely from the pre-pulse-induced acidosis within 20 min at a rate of 0.146±0.003 pH_i units min⁻¹. The ability to recover from the induced acidosis was only partially but significantly (P=0.0008) reduced to 0.067±0.007 and 0.047±0.006 pH_i units min⁻¹ in cells exposed to 1 µmol l⁻¹ (N=4) and 10 µmol l⁻¹ (N=12) EIPA, respectively. Different concentrations of EIPA demonstrated that the midgut epithelial cells reacted sensitively to the NHE inhibitor EIPA in a dose-dependent manner (Fig. 2C) with an IC₅₀ of 0.153 µmol l⁻¹. Although the recovery rate was affected by the treatment with EIPA, all cells were able to largely recover from the NH₃/NH₄⁺ pre-pulse induced acidosis.

SpSlc9a2 was localized using a custom-made antibody, designed against a synthetic peptide of the S. purpuratus SpSlc9a2 protein using confocal microscopy. This antibody was validated in a previous study, demonstrating specific immune reactivity with a 95 kDa protein (Hu et al., 2020). In the pluteus stage, immunofluorescence staining revealed a strong signal in a specific cell type of the midgut epithelium and at higher laser intensities positive immunoreactivity was also detected in primary mesenchyme cells (Fig. 3A–C). Positive immunoreactivity of the SpSlc9a2 antibody was located in luminal membranes of midgut epithelial cells. An overview image of the larval digestive organs (oesophagus, midgut, hindgut) demonstrated a regular distribution of the SpSlc9a2-positive cells scattered over the midgut epithelium with lower signal intensity in the ventral part of the organ (Fig. 3A). A lower density of SpSlc9a2-positive cells was detected close to the cardiac and the pyloric sphincter. The signal for the Na⁺/K⁺-ATPases was also detected at the luminal side of midgut cells in a similar pattern as observed for SpSlc9a2. Double staining of SpSlc9a2 and Na⁺/K⁺-ATPases demonstrated a co-localization of the fluorescence signals (Fig. 3C–E). The co-staining of SpSlc9a2 and mitochondria-rich cells using Mitotracker demonstrated a partial co-localization in midgut epithelial cells. Additionally, Mitotracker staining showed a cytosolic signal with different intensities among different cell types of the midgut epithelium (Fig. 3B). Negative controls, using the SpSlc9a2 specific antibody pre-absorbed with the immunization peptide, demonstrated no unspecific binding of the antibodies or auto-fluorescence (Fig. 3E–F).

Knockdown using Vivo MO specially designed against the coding region of SpSlc9a2 was used to determine the contribution of this protein to the maintenance of [Na⁺] in midgut fluids. The ability to maintain midgut [Na⁺] was significantly affected by the MO treatment in a dose-dependent manner (Fig. 4; P<0.0001). After exposure to 1 µmol l⁻¹ SpSlc9a2, MO larvae showed a significant increase in midgut [Na⁺] by 97 mmol l⁻¹ compared with control larvae that were only exposed to the vehicle. Larvae exposed to a concentration of 5 µmol l⁻¹ fully failed to regulate midgut [Na⁺] with an average concentration of 428 mmol l⁻¹. However, at the highest MO concentration of 5 µmol l⁻¹, larvae started to have morphological changes including reduced midgut volume and an increased presence of immune cells in the extracellular matrix. Because of these effects of the MO treatment on the overall physiology and morphology, 1 µmol l⁻¹ MO-treated larvae, where no visible side-effects were observed, were used to validate the SpSlc9a2 knockdown by immunofluorescence intensities. The quantification revealed a significant decrease in signal intensity of 21.05 arbitrary units (AU) from 100.51±7.98 AU in the SpSlc9a2 morphants to 79.46±5.23 AU in control larvae (control N=31, morphants N=40; P=0.027) (Fig. 4).

To stimulate the midgut pH regulatory machinery, larvae were exposed to acidic seawater (pH 6.5) over a period of 10 h and midgut pH and Na⁺ concentration, as well as the expression levels of the three identified Na⁺/H⁺ exchanger isoforms in the sea urchin larva, were examined.

The acidic seawater treatment did not affect the luminal sodium concentration, with average values of 103.6±3.7 and 103.9±7.6 mmol l⁻¹, in control and low pH-treated larvae, respectively (Fig. 5B). At the beginning of the low pH experiment, larvae had an average midgut pH of 9.13±0.04. Over the 10 h time course, the low pH-treated larvae reduced midgut pH in a biphasic manner and larvae were able to fully restore pH after 10 h of incubation (Fig. 5A). The first drop in midgut pH by 0.26 pH units occurred after 1 h of incubation, followed by a recovery within the next 2 h. The second drop in pH started after 4 h of incubation with the maximum decrease of 0.72 pH units measured after 6 h. Full pH recovery was measured after 10 h.

Real-time qPCR analysis was used to detect changes in the relative expression of spslc9a2, spslc9a7 and spslc9a8 between control (normal seawater pH 8.1) and larvae exposed to acidic seawater (pH 6.5) over the period of 10 h. qPCR results demonstrated that the luminal located SpSlc9a2 did not respond to the low pH treatment (Fig. 5C). In contrast, spslc9a7 constantly increased expression levels with a peak after 6 h of low pH exposure (P=0.045; Fig. 5C,D). The third NHE isoform, spslc9a8, progressively increased expression levels from 0.00141±0.000276 2^−ΔCt to 0.0048±0.001833 2^−ΔCt over the course of the low pH experiment (Fig. 5E). To relate changes in midgut pH to the expression of NHEs, midgut pH values were plotted against the relative expression of the three different NHE isoforms in control and low-pH treated larvae (Fig. 5F–H). Linear regression models demonstrated a significant (P=0.0208; Fig. 5G) relationship between midgut pH and the expression levels of spslca7 in low pH-treated larvae.

Marine larval stages of the superphylum Ambulacraria, including echinoderms and hemichordates, were demonstrated to strictly regulate extracellular pH in their digestive tract (Stumpp et al., 2013, 2015). This acid–base regulatory feature is a key factor associated with energy allocations during exposure to simulated ocean acidification (Stumpp et al., 2013; Hu et al., 2017). On the one hand, a larger fraction of energy is required to maintain pH (Hu et al., 2017) and on the other hand, slight reductions in midgut pH were demonstrated to reduce nutrient uptake by a shift in pH optima of key digestive enzymes (Stumpp et al., 2013). Furthermore, reductions in midgut pH were associated with a higher susceptibility of the sea urchin larva to bacterial infections of the gut (Stumpp et al., 2020). Thus, understanding the mechanistic basis of homeostatic regulation in the larval sea urchin gut will provide important insights into a largely unexplored physiological system and its response to changes in environmental pH. Accordingly, the present work used a suite of methods ranging from the whole organism to the molecular level to address the role of Na⁺/H⁺ exchangers in regulating pH and [Na⁺] in midgut fluids of the sea urchin larva.

In accordance with previous studies, the present work also confirmed the presence of a highly alkaline environment in the midgut of sea urchin larvae. In addition to this existing knowledge, the present work demonstrated that these organisms also strictly maintain low [Na⁺] in their midgut fluids. While midgut pH is typically maintained between pH 8.8 and 9.2, midgut [Na⁺] is far below that of seawater (ca 450–500 mmol l⁻¹), ranging from 70 to 230 mmol l⁻¹. These observations resemble the situation found in the intestine of marine teleost fish where alkaline conditions together with high rates of bicarbonate secretion support water uptake from the lumen by active NaCl absorption. In marine teleosts, two NHEs (NHE-2 and NHE-3) are localized in luminal membranes of intestinal cells, which are putatively involved in Na⁺ resorption from the luminal fluids. Another putative NHE (NHE-1), located at the basolateral side, is thought to be involved in transepithelial Na⁺ transport and cytosolic H⁺ extrusion (Grosell, 2011). Interestingly, this coordinated interaction between several NHE isoforms located on both sides of the teleost gut epithelium in marine teleosts also seems to be a characteristic of the midgut epithelium in the sea urchin larva. Here the luminal NHE SpSlc9a2, as well as putative NHEs located in basolateral membranes, differentially contribute to the regulation of pH and [Na⁺] in midgut fluids. Our experiments demonstrated that basolateral application of EIPA reduces midgut pH but has no effect on midgut [Na⁺]. In contrast, luminal injection of EIPA led to an increase in midgut [Na⁺] levels with no detectable changes in pH. Reaching the basolateral side is possible as large molecules up to 4 kDa can diffuse and distribute within the ECM within 60 min (Stumpp et al., 2012). At the same time, it is unlikely that large amounts of the inhibitor reach the luminal side as swallowing of microalgae is associated with negligible ingestion of external fluids (Stumpp et al., 2013). This is further supported by our observation that midgut pH or [Na⁺] showed negligible changes during swallowing, suggesting that no or only very small amounts of water are ingested into the gut. Our results demonstrated that NHEs located on the basolateral side are involved in pH regulatory processes of the midgut, whereas those located in luminal membranes contribute to the maintenance of midgut fluid [Na⁺].

Besides the example of the marine teleost gut, an involvement of basolateral NHEs in luminal alkalization has also been suggested in the highly alkaline ventriculus (up to pH 11) of the Aedes aegypti mosquito larvae. Pharmacological experiments using amiloride or Na⁺-free solutions together with immunocytochemistry suggested an involvement of basolateral NHEs to luminal alkalization (Onken et al., 2004; Pullikuth et al., 2006; Onken et al., 2008). Besides a putative Na⁺/2H⁺ exchanger at the luminal side, the recent cell model includes luminal HCO₃⁻ secretion via an HCO₃⁻/Cl⁻ exchanger and a Na⁺/2HCO₃⁻ co-transporter, probably supported by basolateral NHEs and V-ATPases (Onken et al., 2004; Pullikuth et al., 2006; Onken et al., 2008). We hypothesize that also in the sea urchin larva basolateral NHE activity acts in concert with luminal HCO₃⁻ secretion to maintain highly alkaline conditions in the midgut. Besides the sea urchin midgut, there are currently only two other known examples, including the anterior midgut of mosquito larvae and the retinal pigment epithelium of mammals, where the Na⁺/K⁺-ATPase is localized in luminal membranes (Okami et al., 1990; Patrick et al., 2006; Stumpp et al., 2015). Although it is striking that two of these systems are designed to generate a highly alkaline midgut, its biological significance remains a matter of future work.

Using teleost fish and insect larvae as a guide, we hypothesized that midgut alkalization in the sea urchin larva involves an as yet unidentified HCO₃⁻ transporter in combination with an intracellular carbonic anhydrase (Fig. 6). Protons generated by intracellular CO₂ hydration can be removed by the basolateral NHEs, and HCO₃⁻ can be secreted into the lumen to alkalize the larval midgut. This luminal bicarbonate secretion may involve an electrogenic anion exchanger, transporting HCO₃⁻/Cl⁻ in a 2/3:1 stoichiometry as described for the alkaline fish gut (Grosell, 2011). Such a HCO₃⁻ transport system could explain the near-zero transepithelial potential by using the lumen-positive potential generated by the Na⁺/K⁺-ATPases and the VHA located in luminal membranes (model; Fig. 6). Although previous studies found no effect of the carbonic anhydrase inhibitor acetazolamide and the anion exchange inhibitor 4,4′-diisothiocyano-2,2′-stilbenedisulfonic acid (DIDS) on midgut alkalization in the sea urchin larva (Stumpp et al., 2015), our experiments demonstrated that complex molecules are not necessarily able to reach the luminal side via bathing of larvae in the inhibitor solution. Inhibitors like DIDS or acetazolamide should be used by direct injection of the inhibitor into the midgut lumen. This would provide important pharmacological evidence about the potential involvement of HCO₃⁻ transport in the alkalization process, potentially leading to conserved mechanisms of luminal alkalization, as found in the guts of marine teleosts or several insect larvae (Ridgway and Moffett, 1986; Boudko et al., 2001a; del Pilar Corena et al., 2001; del Pilar Corena et al., 2004; Grosell et al., 2005; Grosell, 2011).

The lack of a transepithelial potential in the midgut of the sea urchin larva may also be explained by the presence of an electrogenic Na⁺/2H⁺ antiporter as located in insect epithelia and the distal nephron of humans (Kondapalli et al., 2017; Okech et al., 2008; Chintapalli et al., 2015). Alternatively, cation transport to counter a lumen positive potential may involve a putative K⁺/H⁺ antiporter with a stoichiometry of 1K⁺:2H⁺ as proposed in earlier studies using vesicles of highly purified apical membranes of goblet cells for measuring ATP-dependent proton transport and K⁺/H⁺ antiport (Wieczorek et al., 1991; Azuma et al., 1995). The presence of such an electrogenic cation antiporter at high expression levels would help to explain the near-zero transepithelial potential by counteracting the electrical potential generated by luminal VHA and NKA (Fig. 6). Immunohistological analyses of the sea urchin NHE isoform SpSlc9a2 using a custom-made antibody demonstrated localization of this protein in luminal membranes. Here it is partially co-localized with the Na⁺/K⁺-ATPase as well as with mitochondria-rich cells. These findings are following previous studies that demonstrated the luminal localization of an NHE isoform using an antibody designed against the teleost NHE-3 (Stumpp et al., 2015). Knockdown of spslc9a2 resulted in an increased midgut [Na⁺] compared with control animals, lending strong evidence for this NHE isoform to be responsible for the regulation of [Na⁺] in midgut fluids. The luminal localization of SpSlc9a2 leads to the export of Na⁺ from the lumen in exchange for H⁺. This import of protons into the midgut lumen seems to conflict the alkaline properties of the midgut. However, localization of NHEs in luminal membranes is associated with a range of digestive systems in animals (Zachos et al., 2005; Grosell, 2011; Xu et al., 2018). For example, in the teleost intestine, the putative NHE-2 and NHE-3 are suggested to excrete H⁺ into the alkaline intestinal lumen, where HCO₃⁻ is titrated near the apical surface of the cells to reduce luminal osmotic pressure (Grosell, 2011). In the sea urchin model, this acidified boundary layer may protect midgut epithelial cells from the alkaline digestive conditions. Furthermore, acidification of the intestinal mucosa by NHEs has been demonstrated to be an important feature of H⁺-coupled nutrient uptake in mammals (Thwaites et al., 2002; Zachos et al., 2005; Thwaites and Anderson, 2007). For instance, the short-chain fatty acid/OH exchanger in the rat colon as well as the H⁺/dipeptide co-transporter are driven by H⁺ gradients, which are both functionally linked to brush-border NHE-2 and NHE-3 (Zachos et al., 2005). According to these observations and the results of the present work, it can be concluded that NHE activity in the midgut epithelium of the sea urchin larva contributes to both acid–base and ionic regulation, potentially relevant for digestion and nutrient absorption.

Pharmacological characterization of pH_i regulatory capacities of midgut cells demonstrated that EIPA reduces the speed of the compensation reaction while pH_i is still fully restored. These results suggest that EIPA-sensitive mechanisms are not exclusively responsible for pH_i regulation, underlining the presence of other important acid–base regulatory mechanisms including HCO₃⁻ transport or VHA activity in midgut epithelial cells of the sea urchin larva. In fact, pH_i regulation in primary mesenchyme cells of the sea urchin larva has been demonstrated to strongly depend on HCO₃⁻ transport using knockdown and pharmacological approaches (Hu et al., 2018). Furthermore, the sensitivity of luminal alkalization in the sea urchin larva to the V-type H⁺-ATPase inhibitor Bafilomycin A1 provides additional pharmacological evidence for the VHA in midgut epithelial cells (Stumpp et al., 2015). As VHA activity has been demonstrated to be involved in pH_i regulatory processes in the Malpighian cells and goblet cells of Drosophila larvae and in the tobacco hornworm (Wieczorek et al., 1989, 2009; Bertram and Wessing, 1994) it can be hypothesized that the VHA may also contribute to pH_i regulation in midgut cells of the sea urchin larva. Here, future studies will address a detailed molecular characterization of other pH_i regulatory mechanisms including primary active H⁺ transport and the involvement of HCO₃⁻ transporters in midgut epithelial cells of the sea urchin larva.

To identify other Slc9 family members that are potentially involved in regulating the ionic composition of midgut fluids we challenged sea urchin larvae with acidic seawater conditions (pH 6.5). The low pH challenge did not affect midgut [Na⁺], indicating that maintenance of luminal pH is not directly linked to transepithelial Na⁺ transport rates. In addition, expression levels of spslc9a2 remained unaffected by the low pH treatment, indicating that this NHE isoform is not involved in defending intracellular or extracellular pH. These findings also corroborate our pharmacological, knockdown and immunohistochemical analyses on the function of SpSlc9a2, which is located in luminal membranes and is mainly responsible for maintaining the luminal Na⁺ concentration but not pH. Given the potential role of SpSlc9a2 in mediating boundary layer acidification and nutrient absorption, it is likely that these processes will not be significantly impacted by changes in environmental pH. However, extensive basolateral proton export by other NHE isoforms during more extreme environmental acidification may lead to an increase in intracellular [Na⁺] that could then affect Na⁺/H⁺ exchange in luminal membranes given the relatively low midgut [Na⁺]. Thus future studies will use different levels of acute acidification to determine a potential threshold where luminal NHE activity may also be affected.

In contrast to the [Na⁺], the pH of midgut fluids was dynamically impacted by the low pH challenge. The biphasic reduction in midgut pH at 1 and 6 h was fully restored both times. The lack of a transcriptional response during the acute compensation of midgut pH at the 1 h time point may be explained by acute post-translational modifications including phosphorylation or translocation of acid–base transporters in midgut epithelial cells. However, during prolonged low pH challenge increased expression of the midgut pH regulatory machinery (i.e. SpSlc9a7 and SpSlc9a8) is used to defend midgut pH. Using a similar approach, the importance of H⁺ transporters including NHEs and V-type H⁺-ATPases for midgut alkalization has been demonstrated previously (Stumpp et al., 2013). Acute acidification down to pH 7 by addition of CO₂ was accompanied by an initial reduction in midgut pH associated with an upregulation of NHE-2 (spslc9a2) and VHA (Stumpp et al., 2013). The contradicting pattern of spslc9a2 expression may result from the different experimental approaches using CO₂ or HCl to reduce seawater pH. During chronic exposure to environmental hypercapnia spslc9a2 expression levels were increased (Lee et al., 2019) accompanied by higher feeding rates (Stumpp et al., 2013) to meet the higher energetic demands under acidified conditions. Given the potential role of SpSlc9a2 in supporting nutrient uptake, its upregulation under ocean acidification simulations can be hypothesized to be a consequence of enhanced nutrient absorption rather than acid–base regulation. Finally, infection of pluteus larvae with Vibrio diazotrophicus led to an accumulation of bacteria in the gut (Ho et al., 2016; Buckley and Rast, 2019; Stumpp et al., 2020) accompanied by increased NHE-2 expression (Stumpp et al., 2020). Here it can be also suggested that a large fraction of ingested bacteria is digested and the free nutrients stimulate NHE-2 expression for the H⁺-coupled nutrient uptake.

In this study, two other NHE isoforms, SpSlc9a7 and SpSlc9a8, were stimulated by low pH challenge. Microelectrode measurements demonstrated that midgut pH was restored within 4 h from their lowest pH level (pH 8.4) at the 6 h time point to the 10 h time point (pH 9.1). This compensation reaction is reflected in the expression level of spslc9a7, with highest gene expression detected at the 6 h time point. The expression of spslc9a8 does not reflect the fluctuation of midgut pH but mRNA levels steadily increase towards the end of the experiments, where midgut pH is largely restored. Even though the cellular localization of these transporters is not known, the increased expression of spslc9a7 and spslc9a8 matches the pattern of midgut pH regulation, suggesting a potential link between these NHE isoforms and intra- or extracellular pH regulatory mechanisms in cells of the digestive tract. In mammals, Slc9a7 RNA dot blot analysis of human tissue revealed ubiquitous expression, with most dominant gene expression in regions of the brain, skeletal muscle and secretory tissue, including organs of digestive importance as the stomach, pancreas, small intestine and colon (Numata and Orlowski, 2001). It is suggested that mammalian Slc9a7 is involved in volume regulation of intracellular organelles via K⁺/H⁺ exchange, but further pharmacological investigation, especially regarding the function of the protein in digestive organs, is needed (Numata and Orlowski, 2001; Kagami et al., 2008; Xu et al., 2018). Slc9a8 mRNA was detected in the apical membrane in fundic and pyloric glands and the brush border of intestinal epithelial cells in mice (Kagami et al., 2008; Xu et al., 2008, 2018). In the mammalian intestine, Slc9a8 was suggested to be involved in sodium absorption from the intestinal brush border of mice and mucosal protection in the gut (Xu et al., 2018). Due to the current lack of histological and functional characterization of SpSlc9a7 and SpSlc9a8 in the sea urchin larva, their contribution to larval midgut pH regulation remains speculative. Thus future studies should be dedicated to a cellular localization and physiological characterization of these NHE isoforms in the sea urchin larva to better understand their potential contribution to acid–base regulatory processes in midgut epithelial cells.

This work presented novel insights into the role of Na⁺/H⁺ exchange mechanisms in the alkaline midgut of the sea urchin larva. Using pharmacological as well as molecular techniques, different NHEs were identified to contribute to the alkalization process and also to the maintenance of steep [Na⁺] gradients between the extracellular space and the larval midgut. While our results provide strong evidence for SpSlc9a2 in regulating midgut [Na⁺] and potentially also boundary layer acidification, its exact physiological function of luminal Na⁺/H⁺ exchange remains speculative. We hypothesized that H⁺ ions secreted into the luminal mucosa could be used to protect midgut cells from the highly alkaline digestive environment or may serve H⁺-coupled nutrient uptake in the sea urchin larva. However, future studies are needed to address the involvement of SpSlc9a2 in boundary layer acidification and the influence on nutrient uptake.

Determination of intracellular pH regulatory capacities underline an involvement of NHEs in pH_i regulatory abilities of midgut epithelial cells and also highlight the existence of other, as yet unidentified acid–base transport mechanisms. Thus, future studies will test the involvement of HCO₃⁻ transport pathways by direct injection of inhibitors like acetazolamide or DIDS into the midgut. The potential involvement of HCO₃⁻ transport in midgut alkalization in combination with Na⁺/H⁺ exchange mechanisms would highlight conserved alkalization mechanisms in digestive systems of different animal taxa including marine teleosts, insects and echinoderm larvae (Ridgway and Moffett, 1986; Boudko et al., 2001a,b; Corena et al., 2002; del Pilar Corena et al., 2004; Grosell et al., 2005; Grosell, 2011). Here the sea urchin larva can be used as a powerful model to study transmembrane transport mechanisms leading to deeper insights into ionic regulation in a basal deuterostome larva. This knowledge will broaden our understanding of the physiology and diversity of membrane transport systems and will help identify the mechanistic basis for the vulnerability of marine species towards changing environmental conditions.

The authors would like to thank M. Stumpp for fruitful discussions on this work. The authors further thank F. Thoben, J. Hildebrand and R. Lingg for maintenance of the sea urchin culture systems.