Multiple functions of ion transport by the nuchal organ in embryos and neonates of the freshwater branchiopod crustacean Daphnia magna

The gill is the predominant organ for ionoregulation in euryhaline crustaceans, and the structures of the transporting cells and mechanisms involved have been well characterized in large species such as the blue crab, Callinectes sapidus, and the Chinese mitten crab, Eriochier sinensis (Henry et al., 2012; Larsen et al., 2014). Smaller crustaceans such as branchiopods pose technical challenges and have been less well studied, but in adult brine shrimp, the branchiae are thought to be the main sites of ion uptake. Classic work involving staining with AgNO₃ and use of KMnO₄ to oxidize the transporting cells indicates that the first 10 pairs of branchiae are the sites of ion excretion in adult Artemia salina in hypertonic conditions, and are probably the site of ion uptake in hypotonic media (Conte, 1984; Croghan, 1958). Hyperosmotic regulation in embryonic ostracods (seed shrimp) is proposed to reflect both salt reserves in the yolk of the egg and the absorption of salts by special cells located in the non-calcified zone of the inner shell layer (Aladin and Potts, 1996). In the freshwater copepod Eurytemora affinis, structures termed Crusalis organs are thought to function as osmoregulatory organs (Johnson et al., 2014); the organs consist of ionocytes on the penultimate leg segment of each swimming leg and are enriched in ion transport enzymes.

Another ionoregulatory structure, termed the nuchal organ (also referred to as the neck organ or dorsal gland), is present in a wide variety of crustaceans, both larval and adult, including branchiopods, copepods and malacostracans (Martin and Laverack, 1992). In branchiopods, the nuchal organ is the preferred term to describe structures that contain mitochondria-rich ion transporting cells that are probably involved in salt uptake in freshwater and salt excretion in saline water (Aladin and Potts, 1995). In the nauplius larva of the brine shrimp Artemia salina, there is direct evidence for salt excretion by the nuchal organ in larvae preloaded with ²²Na (Russler and Mangos, 1978).

Although a role for the nuchal organ in ion uptake by freshwater species has been inferred from ultrastructural studies, direct evidence for uptake is lacking. In the freshwater cladoceran Daphnia magna, the nuchal organ appears as an expanded portion of a dorsal ridge which runs from the front to the back of the head in first instar juveniles. In electron micrographs, the perimeter of the nuchal organ is delineated by a densely staining portion of cuticle which separates the thin cuticle covering of the nuchal organ from the thicker and less densely staining surrounding cuticle. The cells that form the nuchal organ fill much of the haemocoelic space between the cuticle and the gut, and are differentiated from surrounding squamous epidermal cells by greater apical–basal depth, extensive amplification of plasma membranes through apical microvilli and basal infoldings, and abundant mitochondria (Halcrow, 1982).

It has been suggested that the nuchal organ is most useful whilst juveniles remain in the brood chamber, particularly in the earlier stages of embryonic development when the thoracic appendages move very little (Halcrow, 1982). Cladocerans incubate their eggs in brood chambers formed by the carapace, except when laying resting eggs (Aladin and Potts, 1995), and the brood chamber remains open to the environment in most brackish and freshwater genera, including Daphnia. The egg membrane must therefore be impermeable until the larval organs of osmoregulation have developed. These include both the nuchal organ and the maxillary gland for water excretion. In the free-swimming neonate, the nuchal organ probably functions for about 12 h only (Halcrow, 1982). The nuchal organ disappears at the first embryonic moult when the animal begins to feed and the epipodites of the thoracic appendages probably assume a role in ion uptake (Aladin and Potts, 1995).

In this study, we used the scanning ion-selective electrode technique (SIET) to provide the first direct measurements of the transport of Na⁺, K⁺, Cl⁻, H⁺, Ca²⁺ and NH₄⁺ across the nuchal organ and across regions of the body surface away from the organ in both embryonic and neonate Daphnia magna. Because development of the embryo is associated with formation of the heart and the haemocoel, changes in ion transport may be related not just to the need to replenish ions lost to the environment in freshwater but also to the conversion of the intracellular volume, typically with a Na⁺/K⁺ ratio much less than 1, into extracellular space with a Na⁺/K⁺ ratio much greater than 1. There may also be changes in ion transport associated with metabolism of yolk proteins into amino acids and subsequent synthesis of new proteins during the formation of tissues in the neonate such as the thoracic appendages and cuticle. Metabolism during development may thus lead to the formation of organic ions that may alter total cation and anion levels in the newly formed extracellular compartment. Lastly, in view of evidence that isolated embryos do not equilibrate with calcium in the environment and that calcium is transferred to the embryo from the mother (Giardini et al., 2015), we used ion-selective microelectrodes to determine whether embryos are exposed to concentrations of Ca²⁺ and other ions in the brood chamber that differ from the concentrations in the surrounding water.

A starter culture of Daphnia magna Straus was obtained from a commercial supplier and maintained at room temperature (23°C) in aerated 20 l tanks of dechlorinated Hamilton tap water (DHTW). The water was sourced from Lake Ontario water, containing (in mmol l⁻¹): 1 Ca, 0.6 Na, 0.70 Cl, 0.3 Mg and 0.05 K, with titration alkalinity of 2.1 mequiv l⁻¹, hardness of ∼140 mg l⁻¹ as CaCO₃ equivalents, and pH ∼8.0 (Hollis et al., 2001; Leonard et al., 2014). Daphnia were fed a 2:2:1 mixture of Spirulina powder:Chlorella powder:yeast 3 times per week.

SIET measurements were made with hardware from Applicable Electronics (Forestdale, MA, USA) and Automated Scanning Electrode Technique (ASET) software (ASET-LV4, Science Wares, Falmouth, MA, USA). Micropipettes were pulled from 1.5 mm borosilicate glass (World Precision Instruments Inc., Sarasota, FL, USA) to tip diameters of ∼3 μm on a P-97 Flaming-Brown pipette puller (Sutter Instruments Co., Novato, CA, USA). Na⁺-selective microelectrodes were backfilled with 150 mmol l⁻¹ NaCl and tip filled with a cocktail consisting of 3.5% Na ionophore X, 0.6% potassium tetrakis (4-chlorophenyl) borate and 95.9% 2-nitrophenyl octyl ether (Jayakannan et al., 2011). Na⁺ ionophore X has a high selectivity for Na⁺ over Ca²⁺ (>3000-fold) and for Na⁺ over K⁺ (∼400-fold). Ion-selective microelectrodes for the other ions were constructed with the following ionophores (Sigma-Aldrich, St Louis, MO, USA), with backfill and calibration solutions (in mmol l⁻¹) indicated in parentheses: K⁺ ionophore I, cocktail B (150 KCl backfill, 0.5/5 KCl calibration); Ca²⁺ ionophore I, cocktail A (100 CaCl₂ backfill, 0.1/1/10 CaCl₂ calibration); H⁺ ionophore I, cocktail B (100 NaCl/100 sodium citrate at pH 6 backfill, 1 Hepes, 0.6 NaHCO₃, 1 CaCl₂ at pH 6.5, pH 8.3 calibration); NH₄⁺ ionophore I, cocktail A (100 NH₄Cl backfill, 0.1/1 NH₄Cl calibration); Cl⁻ ionophore I, cocktail A (150 KCl backfill, 0.5/5 NaCl calibration). Because Cl⁻-selective microelectrodes based on chloride ionophores are known to be sensitive to organic anions that may be released from tissues (Chao and Armstrong, 1987; Del Duca et al., 2011; Kondo et al., 1989; Messerli et al., 2008), SIET measurements of Cl⁻ flux were also made with a solid-state Cl⁻ microelectrode (Donini and O'Donnell, 2005) that is insensitive to organic anions such as bicarbonate and acetate (Saunders and Brown, 1977). The solid-state Cl⁻ microelectrode consisted of the fine tip (∼10 μm diameter) of a chlorided silver wire glued into the barrel of a glass micropipette with hot melt glue so that a piece of wire approximately 50 μm long and 10 μm in diameter protruded from the micropipette tip. To further reduce the exposed surface area of silver at the tip, the solid-state microelectrode was coated with a layer of petroleum jelly (∼5 μm thick), which was then partially removed at the tip by wiping with a small piece of tissue paper so that the exposed chlorided silver wire was reduced to approximately 10 μm in diameter and 5–10 μm in length, thus allowing finer spatial resolution for measurement of Cl⁻ concentration.

Measurements of Na⁺, K⁺, Cl⁻ and Ca²⁺ flux were made in DHTW. Measurements of K⁺ flux were also made in DHTW containing 1 mmol l⁻¹ KCl. NH₄⁺ flux was measured in DHTW containing 0.1 mmol l⁻¹ NH₄Cl. Preliminary measurements of H⁺ flux were also made in DHTW. However, because protons may diffuse freely or in association with buffers in the saline, proton transport rates must be corrected for buffering using equations described in Messerli et al. (2006). For these experiments, a synthetic Hamilton tap water containing similar levels of Na⁺, Cl⁻ and Ca²⁺ and known buffer concentrations was made using (in mmol l⁻¹): 0.6 NaHCO₃, 1 CaCl₂ and 1 Hepes, adjusted to pH 8. Measurements of Na⁺ transport kinetics were done in water containing six concentrations of NaCl from 0.07 to 2.62 mmol l⁻¹, and 0.5 mmol l⁻¹ CaCl₂, and Michaelis–Menten curves were fitted to the mean flux at each concentration. Measurements of Ca²⁺ transport kinetics were done in water containing 0.04–1.56 mmol l⁻¹ CaCl₂ and 1 mmol l⁻¹ NaCl. We began with 0.04 mmol l⁻¹ and added CaCl₂ from a stock solution to approximately double the concentration for each step increase. Five concentration steps were sufficient to reach plateau values for the flux for some animals, whereas six or seven concentration steps were required for others. We therefore determined the Michaelis–Menten parameters for each animal and the mean values presented in the Results are thus the means of the K_m and V_max values for each animal.

Ion flux was measured in embryos and neonates, corresponding to developmental stages 5 and 6, respectively (Kast-Hutcheson et al., 2001). Stage 5 is late in embryonic maturation; the second embryonic membrane has ruptured, and the second antennae are partially extended. The antennal setae are poorly developed and the tail spine is folded against the carapace. This stage occurs 45–50 h after deposition into the brood chamber. Stage 6 corresponds to a fully developed neonate, >48 h after deposition into the brood chamber. The organism is free swimming (i.e. emerged from the brood chamber), the setae on the second antennae setae are developed and the tail spine is fully extended from the carapace.

Flux at each of the three sites was averaged, and then a mean value for the three sites was calculated. The typical interval between removal of the embryo from the brood chamber and the first flux measurement was 2–3 min. To determine whether there were changes over time associated with the securing of the embryos and neonates in the Petri dish, five sets of measurements at 3 min intervals were made for each preparation. To determine whether ion flux occurred at sites other than the nuchal organ, control measurements were made along the postero-lateral surface of the carapace, >100 μm from the nuchal organ.

Ion-selective microelectrodes fabricated as described above were used to measure the concentrations of Na⁺, K⁺, NH₄⁺, Ca²⁺, H⁺ and Cl⁻ in the brood chamber of Daphnia in three different states: without eggs in the brood chamber, with eggs, and with embryos. Adult Daphnia were secured with petroleum jelly to the bottom of a Petri dish filled with DHTW and a micromanipulator was used to position the microelectrode tip in the brood chamber. A second type of liquid membrane Cl⁻ selective microelectrode based on 2% Cl⁻ ionophore II, 0.03% tridodecylmethylammonium chloride and 97.97% 2-nitrophenyl octyl ether (Messerli et al., 2008) was used in some of the measurements of the brood chamber water, as discussed below. Potential differences between the ion-selective microelectrode and a reference electrode consisting of a Ag/AgCl pellet connected to the bath through an agar bridge containing 150 mmol l⁻¹ KCl in 4% agar were measured using a high-impedance electrometer (pH AMP, ADInstruments, Bella Vista, NSW, Australia) connected to a data acquisition system (Powerlab) running LabChart software.

Graphing and statistical tests of significance were done in GraphPad Prism 6 (San Diego, CA, USA). Changes in ion flux at the nuchal organ over time and between embryos and neonates at the same time points were assessed with two-way ANOVA followed by Šidák's multiple comparisons test. Differences between the magnitude of ion flux at sites away from the nuchal organ and zero were assessed with a one-sample t-test. Differences were considered significant if P<0.05.

The nuchal organ in embryos and neonates of D. magna is located near the base of the second antennae, opposite the rostrum and overlapping the anterior portion of the heart (Fig. 1A,B). Neonates were readily distinguishable from embryos by the polygonal patterning of the cuticle, well-developed setae on the second antennae, a more pronounced dorsal ridge and a prominent tail spine, which extended away from the carapace (Fig. 1). Na⁺ influx was localized to the nuchal organ, declining to near-zero values at the junction of the nuchal organ and the surrounding cuticle (Fig. 1C). The influx of Na⁺ at the nuchal organ was sustained for five sets of measurements made at 3 min intervals in both embryos and neonates (Fig. 2). There were no significant differences in magnitude of the flux between embryos and neonates at each time point, and there were no significant changes over time (two-way repeated measures ANOVA followed by Šidák's multiple comparisons test). One-sample t-tests indicated that Na⁺ flux at sites away from the nuchal organ was not significantly different from zero in embryos (4.4±4.0 pmol cm⁻² s⁻¹, N=6) or neonates (−22.4±13.8 pmol cm⁻² s⁻¹, N=6).

Analysis of transport kinetics revealed that the maximum rate of Na⁺ influx (V_max) across the nuchal organ of embryos was 518.1±25 pmol cm⁻² s⁻¹ and the bath Na⁺ concentration at which transport was half-maximal (K_m) was 0.433±0.06 mmol l⁻¹ (Fig. 3). The latter value is below the measured Na⁺ concentration in the DHTW used in these experiments (0.74 mmol l⁻¹). Comparison of Figs 2 and 3 indicated that flux in neonates measured with microelectrodes based on Na⁺ ionophore X in DHTW water containing 0.74 mmol l⁻¹ Na⁺ was ∼−320 pmol cm⁻² s⁻¹, similar to the value of −315 pmol cm⁻² s⁻¹ predicted using 0.74 mmol l⁻¹ and the Michaelis–Menten parameters derived from measurements in water containing 0.07–2.62 mmol l⁻¹ NaCl and 0.5 mmol l⁻¹ CaCl₂ (Fig. 3).

There were pronounced changes in K⁺ flux at the nuchal organ during development. In DHTW containing 1 mmol l⁻¹ KCl, K⁺ influx at the nuchal organ of neonates (∼−50 pmol cm⁻² s⁻¹; Fig. 4A) was approximately one-quarter of the magnitude of Na⁺ influx. By contrast, there was an efflux of K⁺ from the nuchal organ of embryos of ∼80 pmol cm⁻² s⁻¹. One-sample t-tests indicated that K⁺ flux at sites away from the nuchal organ was not significantly different from zero in embryos (4.1±2.3 pmol cm⁻² s⁻¹, N=6) or neonates (−0.5±1.5 pmol cm⁻² s⁻¹, N=6). K⁺ flux was also measured at the nuchal organ of neonates bathed in DHTW without any added K⁺. This water contained 0.04 mmol K⁺ and there was an influx of K⁺ of −10±2.2 pmol cm⁻² s⁻¹ (N=5), approximately 20% of the influx seen in water containing 1 mmol l⁻¹ K⁺. For five embryos bathed in DHTW without any added K⁺, the water near the nuchal organ contained 0.078 mmol K⁺ and there was a K⁺ efflux of 59.4±10.2 pmol cm⁻² s⁻¹ across the nuchal organ, approximately 75% of the efflux seen in water containing 1 mmol l⁻¹ K⁺.

Measurements of Cl⁻ flux at the nuchal organ with microelectrodes based on Cl⁻ ionophore I, cocktail A, indicated a sustained efflux of Cl⁻ in both embryos and neonates (Fig. 4B). There were no significant differences in the magnitude of Cl⁻ flux between embryos and neonates at each time point, and there were no significant changes over time (two-way repeated measures ANOVA followed by Šidák's multiple comparisons test). Cl⁻-selective microelectrodes based on Cl⁻ ionophore I are known to be sensitive to organic anions, and we therefore measured Cl⁻ flux at the nuchal organ with solid-state Cl⁻ microelectrodes (Fig. 4C). These measurements also revealed a sustained efflux of Cl⁻ at the nuchal organ; there were no significant differences in the magnitude of Cl⁻ flux between embryos and neonates at each time point, and there were no significant changes over time (two-way repeated measures ANOVA followed by Šidák's multiple comparisons test). Moreover, there were no significant differences between the magnitude of Cl⁻ flux measured with solid-state Cl⁻ microelectrodes relative to that based on Cl⁻ ionophore I in embryos or neonates (two-way repeated measures ANOVA followed by Šidák's multiple comparisons test). This last result indicates that Cl⁻ flux at the nuchal organ was not due to interference by organic anions on the Cl⁻-selective microelectrodes based on Cl⁻ ionophore I. One-sample t-tests indicated that Cl⁻ flux at sites away from the nuchal organ was not significantly different from zero in embryos (3.2±5.7 pmol cm⁻² s⁻¹, N=6) or neonates (−2.0±9.4 pmol cm⁻² s⁻¹, N=6).

There was an efflux of H⁺ from the nuchal organ of embryos of 3.2±1.1 pmol cm⁻² s⁻¹ (N=7) after 15 min in DHTW. One-sample t-tests indicated that H⁺ flux at sites away from the nuchal organ was ∼3% of that at the nuchal organ, but was significantly different from zero (0.08±0.02 pmol cm⁻² s⁻¹, N=6). Although most protons diffuse in association with a buffer in relatively hard water such as DHTW, it was not possible to correct for buffer effects because of uncertainties regarding the precise concentrations of carbonate, bicarbonate and dissolved organic matter. H⁺ flux was therefore measured in a synthetic Hamilton tap water of known ionic and buffer composition and the raw flux was corrected for buffering using the equations of Messerli et al. (2006). There were no significant changes in corrected H⁺ flux over time (Fig. 4D), but the larger mean H⁺ efflux in neonates relative to embryos was close to significance (P=0.06; two-way repeated measures ANOVA followed by Šidák's multiple comparisons test). One-sample t-tests indicated that H⁺ flux at sites away from the nuchal organ of embryos in synthetic Hamilton tap water was less than 1% of that at the nuchal organ, but was significantly different from zero (8.1±2.0 pmol cm⁻² s⁻¹, N=6). H⁺ flux at sites away from the nuchal organ of neonates was not significantly different from zero (−4.7±6.3 pmol cm⁻² s⁻¹, N=6).

Preliminary measurements indicated that there were dramatic changes in the magnitude of NH₄⁺ efflux from the nuchal organ during development. Measurements were therefore made in neonates within 1 h of emergence from the adult and at >2 h after emergence.

There were no significant changes in NH₄⁺ efflux from the nuchal organ of embryos or >2 h neonates over time. However, NH₄⁺ efflux from neonates within 1 h of emergence from the brood chamber was 4- to 5-fold greater than that in embryos or in neonates >2 h after emergence (Fig. 4E) and there was a significant increase in the efflux at 15 min relative to that at 3 min in neonates <1 h after emergence (two-way repeated measures ANOVA followed by Šidák's multiple comparisons test). One-sample t-tests indicated that NH₄⁺ efflux at sites away from the nuchal organ of neonates <1 h after emergence was less than 1% of that at the nuchal organ, but was significantly different from zero (1.0±0.4 pmol cm⁻² s⁻¹, N=6). NH₄⁺ flux at sites away from the nuchal organ of embryos was not significantly different from zero (1.5±0.9 pmol cm⁻² s⁻¹, N=6).

In contrast to transport of other ions, Ca²⁺ transport was not confined to the nuchal organ. In embryos, there was a small influx of Ca²⁺ at the nuchal organ at 3 min (Fig. 4F), but the value at the nuchal organ (−2.4±1.3 pmol cm⁻² s⁻¹, N=7) was not significantly different from that away from the nuchal organ (−5.2±1.5 pmol cm⁻² s⁻¹, N=7). Ca²⁺ influx across the nuchal organ in embryos was not sustained, and was not significantly different from zero after the first measurement at 3 min (Fig. 4F). Ca²⁺ influx increased dramatically in neonates relative to embryos. However, the influx over the nuchal organ at 3 min (−43.6±5.2 pmol cm⁻² s⁻¹, N=8) was not significantly larger than that over the body surface at sites away from the nuchal organ (−27.7±8.7 pmol cm⁻² s⁻¹, N=6), consistent with Ca²⁺ uptake over the entire exoskeleton. Flux at sites away from the nuchal organ for all ions measured is summarized in Table S1.

Neonates were exposed to five to seven Ca²⁺ concentrations between 0.04 and 1.56 mmol l⁻¹ for analysis of transport kinetics (Table S2). The Michaelis–Menten parameters calculated by non-linear regression for each neonate (N=6) were: K_m=0.146±0.040 mmol l⁻¹ and V_max=−68.5±15.3 pmol cm⁻² s⁻¹. The mean R² value for the non-linear regression equations was 0.93±0.02 (range 0.84 to 1.00).

The concentrations of K⁺, Na⁺, NH₄⁺ and Ca²⁺ in the brood chamber were 2- to 4-fold higher than those in the bathing water for Daphnia with or without eggs (Fig. 5A–D). For Daphnia with embryos in the brood chamber, the concentrations of K⁺ and NH₄⁺ were 24% and 126%, respectively, above those in the water, whereas the concentrations of Na⁺ and Ca²⁺ were within 5% of those in the bath water. Chloride concentrations in the brood chamber were 3- to 4-fold higher than those in the bathing water for Daphnia with or without eggs (Fig. 5E). The larger size of the tip of the solid-state Cl⁻ electrode relative to the liquid membrane ion-selective microelectrodes precluded measurement of Cl⁻ concentration within the brood chamber of Daphnia containing embryos. Attempts to measure brood chamber Cl⁻ concentration with a liquid membrane Cl⁻ microelectrode based on Cl⁻ ionophore I, cocktail A were discontinued because there was evidence that some component within the brood chamber interfered with the microelectrode. The time required to respond to a change in Cl⁻ concentration increased from a few seconds to >15 min after the microelectrode tip had been positioned within the brood chamber (N=17; data not shown). Measurements with microelectrodes based on chloride ionophore II (N=8, data not shown) were also unsuccessful; the response time increased and the slope of the microelectrode decreased after sampling of the brood chamber. The pH within the brood chamber did not differ significantly from that in the bath in Daphnia with or without eggs or embryos (Fig. 5F).

Our results provide direct evidence for a role of the nuchal organ in Na⁺ uptake, pH regulation and ammonia excretion (Fig. 6). We suggest below that transport of K⁺ and Cl⁻ across the nuchal organ may be related to the formation and expansion of the haemocoel as the circulatory system develops. Our results also show influx of Ca²⁺ across the cuticle of neonates but not embryos, consistent with calcification of the exoskeleton through deposition of calcium salts.

Our results indicate a sustained influx of Na⁺ at the nuchal organ of both embryos and neonates. Kinetic analysis of the influx suggests that the K_m (0.433 mmol l⁻¹) was slightly below the level of Na⁺ in the water in which the animals were reared. The significance of this influx to ionoregulation in the embryos and neonates can be appreciated through estimates of haemolymph volume and nuchal organ surface area. Approximating the embryo shape as an ellipsoid with major and minor axes of 0.700 and 0.430 mm (Fig. 1), respectively, gives a volume of 4/3π×0.350×0.215×0.215=0.068 mm³. It has been estimated that haemolymph volume in adult D. magna corresponds to 61% of animal volume (Kobayashi and Nezu, 1986). An upper limit of haemolymph volume in the embryo is thus 0.61×0.068=0.041 μl. Based on a diameter of the nuchal organ (Fig. 1) of 80 μm, its area is 5×10⁻⁵ cm². Transport across this area at rate of ∼−320 pmol cm⁻² s⁻¹ (Fig. 2) is equivalent to 0.057 nmol h⁻¹. Our measurements of Na⁺ concentration in adult D. magna (C.M. and M.J.O., unpublished data) indicate a value of 51 mmol l⁻¹ for animals reared in DHTW; the estimated haemocoel Na⁺ content is thus: 0.051×0.041=2.1 nmol. Complete replacement of Na⁺ in the haemocoel of the embryo could be achieved by ∼37 h of transport (2.1/0.057) across the nuchal organ, a similar magnitude to the time required (∼24 h) for development from the time of appearance of the nuchal organ (coincident with appearance of the antennae; Kast-Hutcheson et al., 2001; Mittmann et al., 2014) through to the emergence of a free-swimming neonate. A smaller haemolymph volume as a proportion of animal volume in embryos and neonates relative to adults will decrease the time required for transport of the entire haemocoel content of Na⁺. Clearly, haemolymph volume is negligible in the early embryo before the haemocoel is formed, and there may be significant Na⁺ content of the egg, in which case the nuchal organ's ionoregulatory role may be partly to replenish passive losses of Na⁺ across other body surfaces. Later in development, there may be additional uptake of Na⁺ across the epipodites (gills) of the thoracopods (thoracic appendages), which are present in embryos by the time the second antennae have elongated (Mittmann et al., 2014). It has been suggested that the nuchal organ is most useful during the juvenile's stay in the brood chamber, and that it functions for <<12 h in the neonate, before the nuchal organ switches from ion transport to cuticle secretion (Halcrow, 1982).

A previous study of Na⁺ uptake by Daphnia investigated ion exchange across the whole animal (Bianchini and Wood, 2008) and therefore multiple sites (gut, gill) and mechanisms may have contributed. These authors proposed that a vacuolar-type H⁺-ATPase sensitive to bafilomycin in the apical membrane of cells involved in uptake from the water by neonates creates an electrical gradient favouring Na⁺ uptake through channels sensitive to the drug phenamil. Such a proposal is consistent with our findings of outwardly directed H⁺ flux and inwardly directed Na⁺ flux at the nuchal organ. The K_m for Na⁺ uptake by whole neonates in that study (0.351 mmol l⁻¹; Bianchini and Wood, 2008) is similar to the K_m for Na⁺ uptake at the nuchal organ (0.433 mmol l⁻¹). Further studies of the nuchal organ using SIET will allow the role of specific transporters in Na⁺ influx across a single epithelium to be assessed through the application of transport inhibitors and toxins. Silver, for example, causes mortality at extremely low concentrations through inhibition of sodium uptake pathways (Bianchini and Wood, 2003), and it will be of interest in future studies to examine the influence of silver on Na⁺ influx across the nuchal organ. Whole-animal studies have also shown that both the epithelial Na⁺ channel blocker phenamil and the vacuolar H⁺-ATPase inhibitor bafilomycin A1 inhibit Na⁺ uptake in Daphnia neonates (Bianchini and Wood, 2008), and that the Na⁺ channel and Na⁺:H⁺ exchange inhibitor amiloride blocks Na⁺ uptake in adults (Glover and Wood, 2005). The latter study also revealed complex relationships between ambient Ca²⁺ levels and Na⁺ uptake, with Ca²⁺ inhibiting Na⁺ uptake at low Na⁺ levels, but stimulating Na⁺ uptake at high Na⁺ levels. Acidic pH severely inhibits sodium influx in adults when calcium concentration is high (Glover and Wood, 2005). Additional studies using SIET will allow analysis of the interrelationships of water pH and hardness on Na⁺ uptake by the nuchal organ.

The magnitude of H⁺ efflux from the nuchal organ of embryos and neonates bathed in synthetic Hamilton tap water was approximately 300-fold larger than the uncorrected flux calculated from the measured H⁺ concentration values at two points within the unstirred layer. This difference between corrected and uncorrected H⁺ flux is a common finding in media containing significant levels of buffers. In a study of mammalian gastric oxyntic cells, buffers enhanced the diffusion of protons by a factor of 2249 (i.e. 1374 by 1 mmol l⁻¹ Hepes and 875 by 5 mmol l⁻¹ HCO₃⁻; Demarest and Morgan, 1995). The large flux of H⁺ across the nuchal organ suggests a significant role in acid–base balance, particularly as H⁺ flux across the body surface was only 1% of that at the nuchal organ. An efflux of H⁺ could be used to drive Na⁺ uptake through a Na⁺–H⁺ exchanger. Alternatively, efflux of H⁺ could indicate the activity of a vacuolar H⁺-ATPase known to be implicated in Na⁺ uptake, or hydration of metabolic CO₂ passing out through the nuchal organ, followed by hydration of CO₂ and dissociation of carbonic acid into H⁺ and HCO₃⁻ through the actions of carbonic anhydrase.

The efflux of NH₄⁺ across the nuchal organ of both embryos and neonates may be a consequence of catabolism of protein from yolk granules into amino acids for energy production in embryos and neonates. Given the large H⁺ efflux across the nuchal organ, the NH₄⁺ gradient measured with SIET could be a consequence of diffusion trapping of NH₃ that has diffused across the nuchal organ. It is worth noting in this context that ammonia excretion in adult Daphnia is enhanced at low environmental pH relative to the rate of excretion at circumneutral pH (Al-Reasi et al., 2013).

Although our measurements indicated efflux of ammonia across the nuchal organ, it must be noted that ammonium ionophore I, cocktail A is only 4 times more selective for NH₄⁺ than for K⁺. Efflux of K⁺ from the nuchal organ of embryos may thus lead to an overestimate of apparent NH₄⁺ efflux, and influx of K⁺ across the nuchal organ of neonates will lead to an underestimate of apparent NH₄⁺ efflux. A corrected flux can be estimated by accounting for the effects of K⁺ on the NH₄⁺-selective microelectrode during SIET measurements. For embryos, the corrected NH₄⁺ flux is 55% of the uncorrected value, whereas interference from K⁺ at the nuchal organ of neonates results in a small underestimate (2%) of the NH₄⁺ flux (see Appendix).

Most freshwater animals require uptake of both Na⁺ and Cl⁻ to replace passive loss of these ions. The efflux of Cl⁻ across the nuchal organ was, therefore, an unexpected finding. We suggest that Cl⁻ efflux reflects displacement of extracellular Cl⁻ by production of other anions such as bicarbonate. Daphnia pulex is known to have both elevated levels of bicarbonate in the haemolymph (20.9 mmol l⁻¹) and an elevated extracellular pH of 8.33 (Weber and Pirow, 2009). If similar conditions apply to D. magna, then both bicarbonate and negative charges on circulating amino acids, peptides and proteins could lead to an anion surplus, favouring efflux of Cl⁻ across the nuchal organ. Efflux of K⁺ from embryos bathed in water containing 1 mm K⁺ may also be a consequence of developmental processes. Development of an egg into an embryo requires the formation of extracellular space (typically with Na⁺/K⁺>>1). If there were no change in the volume of cytoplasm (with Na⁺/K⁺<<1), formation of extracellular fluid would require uptake of both Na⁺ and K⁺. We suggest that intracellular volume is converted into extracellular volume, and that release of cytoplasmic K⁺ into the extracellular environment may thus lead to excess K⁺ in the extracellular space during early development and expansion of the haemocoel. Later in development, there is an influx of K⁺ across the nuchal organ of neonates. This influx is coincident with tissue development (e.g. gills, gut, epidermal cells) that re-expands total intracellular volume, necessitating uptake of K⁺. Although we have no data indicating changes in total intracellular volume during development, the smaller number of yolk granules in neonates relative to embryos is consistent with breakdown of the yolk and release of ions.

Our SIET measurements indicating negligible Ca²⁺ transport across the nuchal organ or body surface of embryos is consistent with an earlier study which used radioactively labelled calcium (⁴⁵Ca) to trace calcium from mothers to embryos (Giardini et al., 2015). That study demonstrated that calcium is transferred to the embryo from the mother, and that isolated embryos do not equilibrate with calcium in the environment.

SIET measurements indicated Ca²⁺ influx in neonates, but there was no difference in the magnitude of the Ca²⁺ flux across the nuchal organ relative to sites away from the nuchal organ, when the Ca²⁺-selective microelectrode tip was positioned over the posterior regions of the carapace. Daphnia require dissolved calcium to harden the new carapace post-moult, and previous studies have shown that the necessary Ca²⁺ is acquired by uptake from the environment (Tan and Wang, 2009). Our measurements of Ca²⁺ kinetics derived a K_m of 0.146 mmol l⁻¹, considerably below the levels in the relatively hard water (dechlorinated Hamilton tap water) used for rearing D. magna in this study. The efficiency of Ca²⁺ uptake presumably aids rapid calcification of the cuticle.

Further studies using SIET will aid analysis of Ca²⁺ transport across the body surface of Daphnia in low-Ca²⁺ waters, particularly in neonates, as juveniles are more sensitive to calcium deficiency than adults (Hessen et al., 2000). Such studies could examine the influence of water chemistry (pH, HCO₃⁻, hardness) and temperature on Ca²⁺ uptake, and the possible impacts of freshwater acidification from anthropogenic increases in atmospheric CO₂.

Increases in the concentration of K⁺, Na⁺ and Cl⁻ in the empty brood chamber of D. magna above the corresponding values in the surrounding water indicate that these ions are released from the female. Care was taken to avoid touching the surface of the brood chamber when positioning the microelectrode tip, and the finding of lower Ca²⁺ concentrations in the brood chamber of D. magna without eggs suggests that elevated concentrations of the other ions are not simply the result of leakage following damage to the wall of the brood chamber. One caveat is that for measurements with the solid-state Cl⁻ microelectrode, the larger tip size relative to that of the liquid membrane ion-selective microelectrodes made it more likely that the surface of the brood chamber was contacted by the microelectrode tip during measurements of Cl⁻ concentration. The concentration of each ion species within the brood chamber will reflect the rate of ion release from the female, uptake or release by the egg or embryo, and convective and/or diffusive exchange of the brood chamber water with the water outside the female. When eggs are present in the brood chamber, the increases in concentration of Na⁺, K⁺ and Cl⁻ in the brood chamber could result from release from either the female or the egg, but irrespective of the source of ions in the brood chamber, these increases would tend to reduce any passive loss of ions from the developing eggs. The increased concentration of NH₄⁺ in the brood chamber above that in the bathing water, by contrast, creates a larger gradient opposing efflux of NH₄⁺ out of the egg or embryo if ammonia is transported as the ion (but not if excretion is occurring as the gas NH₃, which is then trapped as the ion by combining with H⁺ to form NH₄⁺). Measurements of Ca²⁺ concentration in the brood chamber revealed a complex pattern of changes. Although the concentration of Ca²⁺ in the brood chamber of D. magna was slightly lower than the bathing water around D. magna with no eggs in the brood chamber, the increase in Ca²⁺ concentration above that in the bathing water when eggs were present is consistent with a previous suggestion of maternal provisioning that was based on the flux of Ca⁴⁵ (Giardini et al., 2015). However, this raises the question of how Ca²⁺ is taken up through the egg membranes, which are assumed to be impermeable to ions to minimize ion loss by the eggs before development of ion-transporting organs. Although we saw only transient uptake of Ca²⁺ by isolated embryos, it is conceivable that uptake is sustained in the ionic and hormonal milieu within the brood chamber. It will be of interest in future studies to determine whether maternal provisioning of Ca²⁺ through release of Ca²⁺ into the brood chamber is of greater significance for Daphnia reared in soft water, given that effects of low Ca on growth rate are most apparent during the first days after hatching, reflecting the higher Ca demands of the early juveniles (Hessen et al., 2000). It is worth noting that eggs of land isopods (suborder Oniscidea) are brooded in a fluid-filled maternal marsupium until a few days following the second embryonic moult and that there is evidence for maternal control of the marsupial environment (Surbida and Wright, 2001). Eggs of Armadillidium vulgare possess a well-developed dorsal organ underlying a broad silver-staining saddle on the vitelline membrane. Like the nuchal organ of Daphnia, the dorsal organ has been implicated in ion regulation and acid excretion, but it also plays a role in calcium provisioning (Wright and O'Donnell, 2010).

Correction for this interference requires estimation of the concentration of K⁺ at the inner and outer limits of microelectrode excursion. By re-arranging Fick's equation (see Eqn 2 in Materials and Methods) to solve for ΔC, a K⁺ efflux in embryos in DHTW of 59.4 pmol cm⁻² s⁻¹ corresponds to ΔC of 0.0155 mmol l⁻¹. The mean K⁺ concentration in the unstirred layer near the nuchal organ in DHTW was 0.078 mmol l⁻¹, so the K⁺ concentration at the inner and outer limits of microelectrode excursion can thus be estimated as 0.078+(0.0155/2)=0.86 mmol l⁻¹ and 0.078−(0.0155/2)=0.70 mmol l⁻¹, respectively. For embryos in K⁺ in DHTW containing 0.1 mmol l⁻¹ NH₄⁺, the concentration of NH₄⁺ near the nuchal organ was 0.14 mmol l⁻¹ and the uncorrected NH₄⁺ efflux was 36 pmol cm⁻² s⁻¹, corresponding to ΔC of 0.0086 mmol l⁻¹ NH₄⁺, and uncorrected NH₄⁺ concentrations at the inner and outer excursion limits of 0.144 and 0.136 mmol l⁻¹. The selectivity coefficient for NH₄⁺ microelectrodes based on ammonium ionophore I is 0.25. The corrected NH₄⁺ concentration at the inner and outer limits of electrode excursion is thus 0.144−(0.25×0.086)=0.123 mmol l⁻¹ and 0.136−(0.25×0.070)=0.118 mmol l⁻¹. The corrected ΔC is therefore 0.0047 mmol l⁻¹ and the corrected NH₄⁺ efflux is 19.8 pmol cm⁻² s⁻¹, approximately 55% of the uncorrected value. Corresponding calculations for neonates within 1 h of emergence indicate that interference from K⁺ produces only a small underestimate (2%) of the NH₄⁺ efflux.