Experimental animals, O. vulgaris, were obtained from a local dealer in Keelung, Taiwan, and transported in well aerated containers to the Marine Research Station, Institute of Cellular and Organismic Biology, at the Academia Sinica (I-lan, Taiwan). Animals ranging from 300 to 600 g were held in closed circulatory seawater systems of ~600 l for several days before they were used for experimental procedures. Individuals were kept separately in cages at a constant 12/12 h light cycle at 26°C and low seawater ammonia levels <5.6 μmol l⁻¹ and were fed with live clams (Meretrix spp.). For experimental procedures including perfusion experiments and blood sampling, animals were anesthetized by cooling below 5°C until full depression of ventilation and were killed by decapitation. Best practices for handling cephalopods as experimental animals were followed, including currently discussed ethical standards for anesthesia and killing of cephalopods (Andrews et al., 2013). The experimental protocols were approved by the Academia Sinica Institutional Animal Care and Utilization Committee (approval no. RFIZOOHP220782).

Blood samples were collected from the vena cava via a gas-tight Hamilton syringe by dissecting the funnel and mantle from the ventral side. Determination of extracellular pH (pH_e) and NH4+ were conducted as previously described (also see Supplementary Information for Materials and Methods) using a WTW 340 pH meter (precision ± 0.01 units) equipped with a microelectrode (WTW Mic-D). The pH meter was calibrated with Radiometer precision buffers 7 and 10 (S11M44, S11 M007) and pH-values are provided using the NBS (National Bureau of Standards) scale.

For perfusion experiments gills were carefully dissected from the mantle along the branchial gland, and two cuts at the 1st order afferent and efferent vessels were made before the gill was carefully detached from the mantle and intestinal sac. All dissection procedures were conducted while the tissues were immersed in seawater. Five centimeters of long polyethylene tubes with an outer diameter of 1.52 mm and an inner diameter of 0.86 mm that were tapered toward the end, were inserted into the 1st order efferent and afferent vessels of the octopus gill. Synthetic woven threads of 0.2 mm diameter were used to tie branchial vessels to the PVC tubes. PVC tubes were connected to thicker (2 mm) silicon tubes and the afferent tube was connected to a peristaltic pump, pumping perfusion saline adjusted to 990 ± 10 mOsm Kg⁻¹ (see Supplemental Table 1 for saline composition) at a rate of 12 ml h⁻¹ through the gill. Furthermore, ionic composition of Octopus blood was analyzed to validate the composition of our artificial octopus blood (Supplemental Table 1). And the ionic composition of the saline is close to that of natural seawater as seawater has been demonstrated to be a suitable Ringer's solution for most invertebrates (Robertson, 1949). Before each experiment the pH of the saline was adjusted by addition of HCl or NaOH. The Na⁺, K⁺, and Ca²⁺ were measured using an atomic absorption spectrophotometer (Hitachi Z-8000, Tokyo, Japan). And Cl⁻ content was measured using a double-beam spectrophotometer (NanoDrop 2000/2000c UV-Vis Spectrophotometer, Thermo Scientific). For this 500 μl of diluted samples were mixed with 500 μL of solution containing Hg(SCN)₂ (0.3 g in 95% ethanol) and NH₄Fe(SO₄)₂ 12H₂O (30 g in 135 ml 6 N HNO₃). Standard solutions of major ions (Na⁺, K⁺, Ca²⁺, and Cl⁻; Merck, Darmstadt, Germany) were used to generate standard curves. In addition, HCO3− content was determined from total dissolved carbon measurements using a Corning 965 carbon dioxide analyzer (Olympic Analytical Service, England) as previously described (Hu et al., 2014b, 2016). At the efferent vessel of the gill perfusion saline was drained through a PVC tube into a collecting vessel. The entire gill was immersed in an aerated bathing solution (same as perfusion saline, without addition of ammonia) in a volume of 50 ml at 25°C that continuously irrigated the gill with oxygenated saline (see Figure S1 for a schematic illustration of the setup).

Before the start of experiments using perfusion salines with different NH₃/NH4+ concentrations (prepared by the addition of NH₄Cl to perfusion sline), pH (adjusted with HCL and NaOH) or pharmacological compounds (e.g., 10 μM KH7 and 75 μM cAMP) gills were perfused with the respective saline for 10 min to achieve full exchange of fluids within the tubes and gill vessels. 8-Bromoadenosine 3′,5′-cyclic monophosphate (cAMP) was dissolved in perfusion saline to a final concentration of 75 μmol l⁻¹ and a stock solution of the specific soluble adenylate cyclase (sAC) inhibitor KH7 (Bitterman et al., 2013) was dissolved in DMSO and was diluted to 10 μmol l⁻¹ with <0.1% DMSO in the final perfusion saline. The duration of perfusion experiments never exceeded 1.5 h to assure full vitality of the examined tissues.

Immunohistochemical analyses of acid-base transporters in O. vulgaris gills were conducted as previously described (Hu et al., 2010, 2014b). Gill tissues were fixed by direct immersion in Bouin's fixative followed by rinses in 75% ethanol. Samples were embedded in Paraplast (Paraplast Plus, Sigma, P3683) and sections of 4 μm were cut on a Leica RM2265 microtome. The slides were deparaffinized in Histoclear II® for 10 min and passed through a descending alcohol series and transferred to a PBS solution containing 5% bovine serum albumin (BSA) for 30 min to block non-specific binding. The primary antibodies, a rabbit polyclonal antibody H-300, raised against the human α subunit of the Na⁺/K⁺-ATPase (NKA) (Santa Cruz Biotechnology, INC) and custom made polyclonal antibodies raised against part of the carboxyl-terminal region (IYRVRKVGYDEQFIMSY) of Na⁺/H⁺-exchanger3 (NHE3), the subunit A region (SYSKYTRALDEFYDK) of the V-type-H⁺-ATPase (VHA), and the Rhesus protein (RhP) of squid, S. lessoniana (antibody designed against the synthetic peptide TRAGYQEFKW) were diluted in PBS (1:100) and placed in small droplets of 200 μl onto the sections, and incubated for 12 h at 4°C in a wet chamber. For the peptide competition assay (PCA), primary antibodies were pre-absorbed by their specific peptides at a concentration of 0.1 mg/ml for 8 h at 4°C. For detailed information regarding the antibodies used refer to Hu et al. (2010, 2014b). The sections were then washed in PBS and incubated for 1 h with small droplets (200 μl) of secondary antibody, anti-mouse Alexa Fluor 488 or anti- rabbit Alexa Fluor 568 (Invitrogen) (dilution 1:250). To allow double-color immunofluorescence staining, one of the polyclonal antibodies was directly labeled with Alexa Fluor dyes using the Zenon antibody labeling kit (Molecular Probes, Eugene, OR, USA). After rinses in PBS, sections were examined with a fluorescence microscope (Zeiss imager A1) equipped with an appropriate filter set.

Acid-base transporters' peptide sequences from cephalopods and other species (aquatic animals were given the highest priority) were used to BLAST the O. vulgaris expressed sequence tag (EST) and octopus genome databases in NCBI. Based on those putative sequences assembled from collected tags, gene specific primers (listed in Supplemental Table 2) were designed for the reverse transcription polymerase chain reaction (RT-PCR) analysis. Detailed cloning procedures and further phylogenetic analysis are presented in the Supplementary Information for Material and Methods.

The mRNA expressions of target genes were measured by qPCR with the Roche LightCycler® 480 System (Roche Applied Science, Mannheim, Germany). The sequence accession numbers and primers are depicted in Supplemental Table 3. qPCR assay procedures are described in the Supplementary Information for Materials and Methods.

Statistical analyses were performed using Sigma Stat 3.0 (Systat) software. Statistical differences between blood pH and NH4+ levels at different transit stations as well as the effects of cAMP, KH7 and pH on NH4+ excretion rates were analyzed by one-way ANOA followed by Tukeys post-hoc test. A Student's t-test was used to compare blood pH and NH4+ concentrations before and after gill passage in perfusion experiments. Two-way ANOVA was used for analyzing differences between H⁺ loss and NH4+ excretion rates with perfusion salines of different pH and NH4+ concentrations. Data sets were normally distributed (Kolmogorov-Smirnov test). Equal variance was tested using the Levene median test. The significance levels were set to p < 0.05 and 0.01.

The major excretory organs of O. vulgaris, such as gills and renal appendages, are connected by blood vessels that direct the blood flow from the vena cava through these organs to the systemic heart (Figure 1A). Blood returning from the body through the vena cava (station 1) enters the lateral vena cava and is distributed to the renal appendages from where the urine is secreted into the renal sac (station 4). After passage of the renal appendages the blood flow reaches the branchial heart by passing the branchial heart appendages. From here the blood is pumped through the gill via the 1st order afferent vessels (station 2) and leaves the gill via the efferent 1st order vessels (station 3) from where it enters the systemic heart. From here the oxygenated blood is pumped back into the body.

Sampling of blood from different blood transit stations before and after passage through major excretory organs, including gills and renal appendages, demonstrated changes in blood NH4+ concentrations and pH (Figures 1A,B). Blood pH-values in O. vulgaris ranged from pH 7.4 to 7.6 depending on the sampling site. Determination of blood pH at different transit stations demonstrated a decrease in pH by ~0.1 pH units between blood samples from the vena cava (station 1) and after gill passage (station 3). Blood NH4+ levels ranged from 240 to 300 μM, depending on the sampling site. NH4+ concentrations were significantly decreased by 60 μM after gill passage (station 3) compared to blood samples taken from the afferent vessel (station 2). The renal sac fluid (urine) was highly acidic with a pH-values being as low as 5.8 and is characterized by very high NH4+ concentrations of 3,192 ± 146 μM. Determination of pH and NH4+ concentrations before and after gill passage in perfused gills could demonstrate a drop in pH by 0.2 pH units associated with a drop in perfusion saline NH4+ levels of 40 μM (Figure 1C). The stability of this decrease in pH after gill passage served as a reliable indication for the viability of the perfused gill. During perfusion experiments exceeding 2 h the difference in pH between saline before and after gill passage decreased indicating a progressively reduced viability of gill tissues.

Perfusion experiments using artificial blood salines with different NH4+ concentrations ranging from 0 to 5,000 μM were used to test the NH4+ regulatory capacity of O. vulgaris gills. Surprisingly, an enrichment of ammonia was observed in the post branchial fluid at initial NH4+ concentrations below 300 μM, whereas at higher concentrations a loss of ammonia was detected in the post-branchial fluids (Figure 2A). This reduction in blood NH4+ was most evident at 5,000 μM with only 2,700 μM NH4+ remaining after a single gill passage leading to a non-linear flattened curve using blood salines with NH4+ concentrations >300 μM (Figure 2A). This bi-phasic regulation was also evident when looking at NH4+ excretion rates into the bathing saline where NH4+ excretion was maintained at a constant rate of ~1.5 μmol NH4+ h⁻¹ gFM−1 at blood NH4+ concentrations between 0 and 600 μM (Figure 2B). At blood (perfusion solution) NH4+ concentrations >600 μM an increased excretion rate was observed with 1.89 ± 0.49, 4.21 ± 1.48, and 5.14 ± 0.93 μmol NH4+ h⁻¹ gFM−1 at initial perfusion solution NH4+ concentrations of 1,200, 2,500, and 5,000 μM, respectively.

A gain of ammonia was observed in the perfusate, when gills were perfused with solutions initially containing NH4+ concentrations below 300 μM (Figures 2A,C). At 300 μM or higher NH4+ concentrations in the initial perfusion solution a positive rate of NH4+ loss from the blood was observed that increased in a linear fashion as a function of blood NH4+ levels up to 2,400 μM. At a blood NH4+ concentration of 5,000 μM the rate of NH4+ removal from the blood after gill passage was not further increased but was maintained at a rate of ~3 μmol NH4+ h⁻¹ gFM−1 similar to excretion rates determined for 2,400 μM NH4+. A linear regression analysis for blood NH4+ concentrations ranging from 0 to 600 μm shows the blood NH4+ concentration of 287 μM at which no net excretion of NH4+ was measured (Figure 2C, inset).

At a blood NH4+ concentration of 0 μM the metabolically produced NH4+ by the gill led to a total excretion of 4.55 ± 1.73 μmol NH4+ h⁻¹ gFM−1 of which 84% were excreted across the apical side into the bath and 16% being transported into the blood (Figure 3A). Addition of cAMP had no significant effect (p = 0.098) on relative NH4+ transport rates toward both, the apical and the basolateral side regardless whether initial perfusate NH4+ concentrations were set to 0 and 300 μM (Figures 3B,C). However, the specific soluble adenylyl cyclase (sAC) inhibitor KH7 significantly (One-way ANOVA df : 2.4; F: 19.565; p < 0.001, Post-hoc test p = 0.034) decreased apical NH4+ excretion rates into the bath and increased basolateral secretion of NH4+ into the blood at blood NH4+ concentrations of 0 μM (Figure 3B). No effects of KH7 were measured at an initial blood NH4+ concentration of 300 μM (Figure 3C).

In a second set of experiments we investigated the effects of an extracellular acidosis (pH 7.2) on the branchial pH regulatory capacities and epithelial NH4+ transport (Figure 4). The effects of an acidosis were combined with the initial blood NH4+ concentrations of 0 and 300 μM, respectively. The pH levels for the two NH4+ concentrations before gill passage were compared to those after gill passage. At an initial (before gill passage) blood pH of ~7.6 a significant decrease in blood pH after gill passage was observed for 0 μM NH4+ (Figure 4A). Under initial conditions of ~pH 7.2, an elevation of blood pH back to control levels ranging between pH 7.4 and 7.5 was measured for both blood NH4+ concentrations, 0 and 300 μM (Figure 4A). This compensation reaction toward an initially induced acidosis was also indicated by a significantly increased transport rate of protons from the blood during gill passage (Figure 4B). Here an increase in transport rate of H⁺ from the blood was measured for blood containing initially 300 μM NH4+ (9.52 ± 1.21 pmol h⁻¹ gFM−1) compared to blood containing 0 μM NH4+ (5.33 ± 0.72 pmol h⁻¹ gFM−1). A higher rate of proton removal from the blood under acidified conditions was accompanied by increased NH4+ excretion rates compared to control (pH 7.6) conditions (Figure 4C). Compared to control saline of pH 7.6 NH4+ excretion rates increased by 6-fold and by 16-fold when perfused with acidified saline of 7.2 initially containing either 0 or 300 μM NH4+.

Tissue expression levels of VHA demonstrated highest abundance of transcripts in brain (B) tissues (Figure 5A and Figure S2). Relatively similar mRNA expression levels were detected in gills, renal appendages (RA), branchial heart appendages (BA), and optical lobes (OL). Low expression levels of VHA were determined for systemic heart (SH), gill hearts (GH), and mantle (M) tissues. Statistical analyses demonstrated significant differences in the expression level between SH vs. GH; p = 0.04, SH vs. OL; p = 0.03 and between GH vs. M; p = 0.04). PCR as well as qPCR analyses demonstrated highest NKA expression levels in RA, followed by neurons (B + OL) and other excretory organs including gills (G) and (BA) (Figure 5B). Statistical analyses demonstrated significant differences in the expression level between SH vs. G; p = 0.04, SH vs. BHA; p = 0.04 and between BHA vs. M; p = 0.03). Expression levels of RhP are highest in neurons (B + OL) followed by tissues of the circulatory system including branchial gland (BG) and excretory organs including RA, BA, and G (Figure 5C). Lowest transcript levels of RhP were detected in SH and M tissues. Statistical analyses demonstrated significant differences in the expression level between RA vs. M; p = 0.04). Highest expression levels for NHE3 that belongs to the invertebrate clade of NHE3 (see Figure S3 for phylogenetic analysis) were detected in GH and G, whereas the remaining tissues had relatively similar expression levels (Figure 5D). Very low expression levels for this transporter were measured in the systemic heart (SH).

Immunohistochemical analyses were used to describe the morphology of the octopus gill and to identify epithelia that are rich in acid-base transporters relevant for pH and NH4+ regulation (Figure 6A). The morphology of the octopus gill differs from that of most cephalopods (e.g., squid and cuttlefish) by having a less organized hierarchy in folding pattern of 1st to 3rd order lamellae. Instead of continuous fan like folds as found in squid and cuttlefish the 2nd order lamellae of octopods are joined and the 3rd order lamellae are branching folds (Figure 6A). Immunohistochemical stainings demonstrated a predominant signal of the NKA in the blood vessels with a predominant basolateral localization in membranes of the outer epithelium. While some larger blood vessels face the seawater, others are located within the gill. All blood vessels branch out toward the periphery of the gill where the blood is merely separated by the thin branchial epithelium from the surrounding seawater. Additionally, branchial blood vessels show positive immunoreactivity for VHA. The VHA antibody shows weaker immunoreactivity in basolateral membranes compared to the NKA but shows a distinct localization of the VHA in the cytosol of endothelial cells of blood vessels. However, this antibody shows a strong signal in endothelial cells lining the inner side of the blood vessel (Figure 6A). Distinct NHE3 and RhP immunoreactivity was observed in apical membranes of blood vessel epithelial cells. Additionally, endothelial cells facing the lumen of the blood vessel also show a weak immunoreactivity of NHE3 and RhP in basolateral membranes. Besides expression of NHE3 and RhP in blood vessels, a distinct positive immuno-reactivity was also found in apical membranes of gill epithelial cells facing the seawater. Peptide compensation assays using antibody-specific peptides demonstrated full abolishment of the immunoreactivity on gill section (Figure 6B). Western blot analysis demonstrated distinct immunoreactivity with proteins in the predicted size range (Figure 6C).

In vivo determinations of O. vulgaris blood pH and NH₃/NH4+ are in general in accordance with earlier findings by Potts (1965), demonstrating that during passage through excretory organs, including renal appendages, branchial heart appendages and gills, the majority of NH4+ is excreted across gill epithelia. The passage of blood through excretory organs is accompanied by a drop in blood pH-values. In contrast, determinations of blood pH during gill passage using implemented catheters demonstrated an increase in blood pH after gill passage in the free swimming octopus (Houlihan et al., 1986). These different findings may be explained by the different sampling methods in anesthetized and free swimming animals and requires further clarification. In the anesthetized animals a slight but significant increase in blood NH4+ level was observed during blood passage through renal appendages. The renal appendages are enclosed by the renal sacs, filled with an acidic (~pH 6) urine that contains high concentrations of ammonia (~3.2 mM). Despite potential ammonia trapping in this acidic urine, a small amount of NH₃/NH4+ may diffuse back into the blood (at regular blood NH4+ levels below 300 μM) across epithelia of the renal appendages. This would explain the increase in blood NH4+ levels after passage of excretory organs within the renal sacs. Ex vivo measurements of pH and NH4+ concentrations within the artificial blood before and after passage through the isolated perfused gill were in accordance to observations made in the intact animal, demonstrating that the isolated gill behaves in a similar fashion compared to in vivo conditions. Furthermore, contractile movements of the gill supported perfusion of the gill and demonstrated the viability of the isolated perfused gill under ex vivo conditions for up to 2 h. In other aquatic species, isolated gill tissues showed viability and functionality over a time span of several hours. For instance, long-term recordings of oxygen consumption and ion transport capacities in isolated fish gill preparations demonstrated the viability of the perfused organ for at least 1–2 h and for the gills of the blue crab Callinectes sapidus 4 h, respectively (Perry et al., 1984; Burnett and Towle, 1990; Deigweiher et al., 2010).

In order to test the transport properties of ammonia across the gill epithelium we perfused gills with artificial blood containing different concentrations of NH4+, ranging from 0 to 5,000 μM. In response to an ammonia load exceeding in vivo blood NH4+ concentrations (≈300 μM) the octopus gill showed an outward directed net transport of NH4+. This finding is in general accordance with observations made in other aquatic organisms including fish and crustaceans (reviewed by Weihrauch et al., 2009). The magnitude of NH4+ excretion rates across gill epithelia are in accordance to findings in crustaceans and fish. For example, the green shore crab C. maenas has branchial NH4+ excretion rates of ca. 35 μmol h⁻¹ gFM−1 when gills were perfused with hemolymph-like salines containing 200 μM NH₄Cl and excretion rates of 10–20 μmol h⁻¹ gFM−1 when symmetrical bath and perfusion salines with NH4+ concentrations of 100 μM were applied (Weihrauch et al., 1998, 2002). Ammonia excretion rates of perfused gills of marine teleosts demonstrated excretion rates of 0.18–0.3 μmol h⁻¹ gFM−1 at relatively high blood NH4+ concentrations of 1 mM (Goldstein et al., 1982). Interestingly, our results on the octopus gill demonstrated that below in vivo blood NH4+ levels (<250 μM) the perfused gill responded with an accumulation of ammonia in the perfusate, indicating that the gill itself is capable of generating ammonia (ammoniagenesis) which is used to elevate blood ammonia levels to concentration close to 300 μM. For example, under conditions where the gill was perfused with artificial blood containing 0 mM NH4+ the gill generated and transported ammonia at a rate of 0.4 μmol h⁻¹ gFM−1 into the blood. Although ammonia is generally considered toxic to organisms, our results suggest that gill tissues of O. vulgaris regulate extracellular NH4+ homeostasis to maintain blood [NH4+] at levels of 250–300 μM. The physiological reasons for retaining a certain amount of ammonia in the blood remain speculative. Some mid-water cephalopod species accumulate NH4+ in exchange for Na⁺ in specialized tissues to improve buoyancy (Seibel et al., 2004). However, since these species accumulate NH4+ in the mM range in specialized vacuoles it is questionable in how far blood ammonium in the μM range may support buoyancy. In most animals NH4+ is generated through amino acid metabolism, wherein L-amino acids are first transaminated to form glutamate, which is then deaminated to NH4+ and α-ketoglutarate by glutamate dehydrogenase (GDH) (Wright, 1995; Nissim, 1999; Weiner and Verlander, 2013). The mammalian kidney is known to be capable of synthesizing NH4+ from glutamine in order to regulate pH homeostasis (Nissim, 1999; Weiner and Verlander, 2013). Here, a respiratory acidosis stimulates renal H⁺ secretion, accompanied by an increase in HCO3− accumulation and NH4+ excretion. Since cephalopods are powerful acid-base regulators that accumulate HCO3− in the mM range to compensate for an extracellular acidosis (Gutowska et al., 2010; Hu et al., 2014b) it is tempting to speculate that also here branchial ammonia production generates HCO3− that can be used to regulate blood pH.

Interestingly, many aquatic organisms including fish (Lin et al., 2012), crustaceans (Fehsenfeld and Weihrauch, 2013), echinoderms (Stumpp et al., 2012; Hu et al., 2014a), and molluscs (Thomsen and Melzner, 2010) respond with increased NH4+ excretion rates in response to acidified conditions as well. It has been hypothesized that NH4+ based proton secretion might represent a universal and evolutionary ancient pathway to counter an acidosis in many organisms (Wright, 1995; Hu et al., 2014b). Applying an acidosis by reducing the pH of the artificial octopus blood to pH 7.2 our results clearly showed a substantial capability of the octopus gill to regulate blood pH homeostasis to maintain pH-values around 7.4–7.5. Similar to the situation in teleost fish (Evans et al., 2005) and crustaceans (Henry et al., 2012), the cephalopod gill represents an important site for extracellular pH regulation equipped with an acid-base regulation machinery located in specialized epithelial cells (Hu et al., 2011, 2014b). The present work provides a direct evidence for the coupling of extracellular pH regulation and NH4+ excretion during an acidosis in a molluscan system.

The selective soluble adenylyl cyclase (sAC) inhibitor KH7 demonstrated that the excretion of NH4+ across brachial epithelia is cAMP-dependent. sAC is an evolutionary ancient enzyme that is involved in HCO3− sensing from cyanobacteria to humans and has been mainly associated with ion/pH regulatory epithelia (Tresguerres et al., 2011). sAC has been demonstrated to be an important regulator of primarily active ion pumps including NKA (Schmitz et al., 2014) and V-type H⁺-ATPase (Tresguerres et al., 2011). Accordingly, it can be hypothesized that also in molluscs like O. vulgasis sAC has inherited the evolutionary conserved role as a regulator of acid-base homeostasis. Recent advances in understanding branchial pH regulation and excretion in cephalopods have led to first models of proton and NH4+ secretion pathways. The cephalopod gill is equipped with acid-base transporters including NHE3, VHA, and NKA that are stimulated by a hypercapnia-induced acidosis (Hu et al., 2014b). Furthermore, a Rhesus protein (RhP) that is co-localized with NHE3 in apical membranes has been suggested to be involved in an acid-trapping mechanism of NH₃ by protons in the semi-tubular lamellar of the squid gill (Hu et al., 2014b). Similar to the situation in decapodiform cephalopods (e.g., squid and cuttlefish) the octopus gill also expresses acid-base transporters, including NHE3, VHA, NKA, and RhP. Predominant localization of NHE3 and RhP in apical membranes and NKA in basolateral membranes suggest that also in octopus branchial epithelia, NHE3, and RhP operate in concert to promote ammonia excretion. Furthermore, the predominant cytoplasmic localization of the VHA could indicate that this transporter underlies post-translational control through membrane trafficking mechanisms and/or may be involved in an alternative, vesicular NH4+ excretion pathway suggested for crustacean gills and the hypodermis of Caenorhabditis elegans (Weihrauch et al., 2002; Henry et al., 2012; Adlimoghaddam et al., 2015, 2016). Results of these studies suggested that NH4+ is trapped in VHA-rich, acidified vesicles. The vesicles are then thought to be transported along the microtubules network to the apical membrane of the gill epithelium, where vesicular content is excreted by exocytosis (Weihrauch et al., 2002). Also here the establishment of gill perfusion techniques in octopus provides a new and powerful model to study pH and ammonia regulatory mechanisms in complex invertebrate systems.

Comparisons of substrate preferences in different tissues of octopus indicated that gill, kidney (renal appendages), and liver tissues preferentially oxidize glutamate, which has been speculated to represent an important source of NH4+ (Hochachka and Fields, 1982). In vitro determinations of glutamate oxidation rates in gill and kidney tissues of Octopus macropus were 1.1 and 1.3 μmol h⁻¹ gFM−1, respectively (Hochachka and Fields, 1982). Furthermore, purine catabolism has been suggested to represent another source of gill ammoniagenesis supported by the evidence of high adenosine deaminase activity in the squid gill (Hoeger et al., 1987). These biochemical data corroborate with our functional results demonstrating that the isolated octopus gill is capable of generating NH₃/NH4+ at rates of 4.55 ± 1.73 μmol h⁻¹ gFM−1 under simulated in vivo conditions with no NH4+ added to the perfusion saline. Here it remains to be investigated through which pathways endogenous ammonia is exported across the basolateral membrane. In the thick ascending limb of the mammalian kidney NH4+ can exit the basolateral membrane via NHE4 and through a so far uncharacterized, presumably diffusive, mechanism (Weiner and Verlander, 2013). The expression data and immuno-histochemical localization of VHA, NKA, NHE3, and RhP underline the functional role of the gill in excretory processes. However, our mRNA expression data also demonstrate that among excretory organs highest expression levels for VHA, NKA, and RhP were detected in renal appendages. These organs are homologous to the vertebrate kidneys, and produce a highly acidic and NH4+ rich urine. Similar to gills, the renal appendages (kidneys) are a significant site of active ammonia synthesis and excretion into the renal sac (Potts, 1965; Potts and Todd, 1965; Hochachka and Fields, 1982). An acid-trapping mechanism of NH4+ in the highly acidified urine has been speculated, but the mechanisms of proton and proton equivalent secretion remain unknown. Furthermore, although the majority of NH4+ is removed after the passage of the gills, higher ammonia excretion rates during feeding (Boucher-Rodoni and Mangold, 1985) may require additional excretion capacities by the renal appendages (kidney). Thus, future studies addressing the regulatory mechanisms in different excretory organs of octopus will be important to improve our holistic understanding regarding ammonia and pH homeostasis in these highly developed molluscs.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.