The osmorespiratory compromise, the trade-off between the requirements for respiratory and ionoregulatory homeostasis at the gills, becomes more intense during environmental hypoxia. One aspect that has been previously overlooked is possible change in transepithelial potential (TEP) caused by hypoxia, which will influence branchial ionic fluxes. Using the euryhaline killifish, we show that acute hypoxia reduces the TEP across the gills by approximately 10 mV in animals acclimated to both freshwater (FW) and seawater (SW), with a higher PO2 threshold in the former. TEP becomes negative in FW, and less positive in SW. The effects are immediate, stable for at least 3 h, and reverse immediately upon return to normoxia. Hypoxia also blocks the normal increase in TEP that occurs upon transfer from FW to SW, but does not reduce the fall in TEP that occurs with transfer in the opposite direction. These effects may be beneficial in FW but not in SW.
The osmorespiratory compromise is a trade-off between the requirements of the gills to promote respiratory gas transfer, whilst restricting the passive fluxes of ions and water. In the traditional compromise, fish increase their effective gill permeability to O2 during exercise or hypoxia by changing blood perfusion patterns. This results in increases in branchial surface area and/or reductions in the mean blood-to-water diffusion distance, both of which favour gas exchange but also exacerbate unfavourable ion and water movements (Randall et al., 1972). The usual interpretation is that these effects result exclusively from changes in gill permeability. However, at least for ions, branchial permeability is only one of the three factors affecting both passive and active fluxes; the other two are concentration gradient and electrical gradient, which together dictate the electrochemical gradient on an electrolyte (Kirschner, 1970). However, the electrical gradient is often overlooked. While the impacts of ion concentration gradients on ionic fluxes have received considerable attention, to our knowledge the influence of electrical gradients has never been examined systematically in the context of the osmorespiratory compromise.
In the present study, we used the Atlantic killifish Fundulus heteroclitus, a model species which is both euryhaline and hypoxia tolerant (Burnett et al., 2007). The origin of the transepithelial potential (TEP) across the gills has been studied extensively in killifish and is relatively well understood in both freshwater (FW) and seawater (SW) (e.g. Pic, 1978; Potts, 1984; Wood and Grosell, 2008). To put the importance of this factor in perspective, using data from the FW-acclimated killifish and the Goldman–Hodgkin–Katz equation (Wood and Grosell, 2008), a modest ±10 mV change in the TEP could result from changing plasma [Na+] from the normal level of 140 mmol l−1 to 93 mmol l−1 or 221 mmol l−1 respectively, or altering the PCl/PNa permeability ratio over a 3-fold range.
We hypothesized that changes in blood perfusion patterns during hypoxia, or even hypoxia itself, could alter either electrogenic or diffusive components of the TEP, or both. Therefore, in the present study, we tested whether exposure to acute hypoxia alters the TEP across the gills of intact killifish in either FW or SW. Having found significant effects that persist for some time, we then examined the influence of hypoxia on the important immediate changes in TEP, which help killifish adjust quickly as they move between SW and FW.
RESULTS AND DISCUSSION
In the first set of experiments (series 1), FW- and SW-acclimated killifish were exposed acutely to various levels of hypoxia ranging from 10% saturation (PO2=2.06 kPa, 15.5 torr) to 40% saturation (8.22 kPa, 61.7 torr). The order was randomized to negate any cumulative effects of stress from handling (Pic, 1978; Wood and Grosell, 2008; see Materials and methods). While hypoxia affected TEP in both salinities, the response was clearly more sensitive in FW. The threshold in FW was between 40% and 30% saturation, with TEP (note: reference=water side as 0 mV) falling significantly from the control value of about +5 mV to hypoxic values of −5 to −7 mV, at 30%, 20% and 10% saturation, respectively (Fig. 1A). In SW, the control TEP in normoxia was about +13 mV, and this remained unchanged down to 20% saturation, but dropped significantly to about +8 mV at 10% saturation (Fig. 1B).
In the second set of experiments (series 2), FW- and SW-acclimated killifish were exposed to 10% saturation (PO2=2.06 kPa, 15.5 torr) for 3 h, together with a comparable control series, where the animals were handled identically but maintained at greater than 80% saturation (normoxia). In FW, the rapid decrease in TEP by about 8 mV to negative values was stable over 3 h, and immediately reversed back to the positive control value when the animals were returned to normoxic conditions (Fig. 2A). TEP in the controls did not change significantly. In SW, the higher normoxic TEP decreased significantly by about 10 mV immediately upon exposure to hypoxia, and remained stable at a lower, but still positive, value for 3 h (Fig. 2B). Upon return to normoxia, this decrement was quickly reversed, and again there were no significant changes in the controls.
A third set of experiments (series 3) examined whether hypoxia (10% saturation, PO2=2.06 kPa, 15.5 torr), either in the environment of origin, or in the environment of destination, affected the large changes in TEP that normally occur when killifish move acutely between FW and SW, and vice versa. When FW-acclimated killifish were transferred from normoxic FW to normoxic SW, TEP increased by about 11 mV (Fig. 2C). However, when the transfer was from normoxic FW to hypoxic SW, the increment was essentially abolished. The same lack of response occurred when the transfer was from hypoxic FW to hypoxic SW; thus hypoxia in the destination environment, and not hypoxia in the originating environment was responsible for this blockade (Fig. 2C). Indeed, transfer from hypoxic FW to normoxic SW was accompanied by an overshoot, an increase in TEP which was about 75% greater than in the normoxia-to-normoxia control, although this difference was not quite significant (P=0.07) (Fig. 2C). In contrast, when SW-acclimated killifish were transferred to FW, neither hypoxia in the destination environment, hypoxia in the originating environment, nor hypoxia in both had any significant effect on the approximate 45 mV drop in TEP that occurred (Fig. 2D).
The present TEP values measured under normoxic conditions are very similar to those reported in previous studies on similarly handled killifish in FW, SW and in those abruptly transferred between the two (Pic, 1978; Wood and Grosell, 2008). Current understanding of the TEP has been summarized in the latter study. In brief, in 100% SW, killifish gills exhibit high cation and low anion permeability. As a result, there is a positive Na+ diffusion potential, upon which is superimposed a positive electrogenic potential due to the secondary active excretion of Cl−; the positive TEP facilitates the passive extrusion of Na+ through the paracellular ‘shunt’ pathway. Upon acute exposure to FW, the Na+ diffusion potential immediately reverses to very negative values, and the electrogenic component is abolished. The resulting very negative TEP is beneficial in immediately limiting Na+ loss, and does not impact active Cl− uptake because there is none in this species in FW (Patrick et al., 1997). The shunt permeability stays high for a few hours, allowing the killifish to return to higher salinity and instantaneously re-activate NaCl excretion. However, after acclimation to FW (24 h+), the paracellular shunt pathway closes, anion permeability becomes slightly greater than cation permeability, and the TEP becomes slightly positive. The animals are now well acclimated for life in FW, but not to move rapidly back and forth between salinities.
In the present study, exposure to severe hypoxia (10% saturation) decreased the TEP by similar amounts in both FW and SW killifish (Fig. 2A,B), confirming that there is an important electrical element to the osmorespiratory compromise. This has been overlooked in previous analyses of the phenomenon, and as illustrated by the simulations below, it is a quantitatively significant factor in the trade-off. A fall in TEP of 10 mV is used in these calculations for the hypoxia effect in both FW and SW, representative of the data in Figs 1 and 2. This similarity may be happenstance, because the absolute TEP values in normoxia and hypoxia were very different in the two salinities, as were the threshold levels for the response (40–30% saturation in FW vs 20–10% in SW; Fig. 1) and perhaps also the mechanisms, as suggested subsequently.
In FW-acclimated killifish, there is no electrogenic component, so all acute effects were likely attributable to decreases in the PCl/PNa ratio caused by hypoxia, because the fall in TEP is immediate but stable for at least 3 h (Fig. 2A), whereas changes in plasma ion concentrations take some time to develop. This effect on the PCl/PNa ratio could result either from blood flow changes in the gills during hypoxia, or effects of low O2 itself on permeability; other environmental factors such as water pH and calcium concentration are well known to alter PCl/PNa (Potts, 1984; Wood and Grosell, 2008). The 10 mV drop in TEP in FW (Fig. 2A) would reflect a change in the PCl/PNa from about 1.30 to 0.82. In SW-acclimated killifish acutely transferred to FW, the PCl/PNa ratio, which is 0.36 under normoxia, would decrease further to 0.22 under hypoxia, based on the data in Fig. 2D. In both cases, the more negative potentials under hypoxia would favour the retention of Na+, reducing the electrochemical gradient driving Na+ loss by about 9% in FW-acclimated animals, and by about 12% in animals acutely transferred from SW to FW. Paradoxically, by changing the PCl/PNa ratio, hypoxia may actually help Na+ balance in FW, although from this analysis, we have no way of knowing whether hypoxia affects active Na+ uptake, which is electroneutral in FW fish (e.g. Kirschner, 1970; Potts, 1984).
In SW, the situation is more complicated because both electrogenic and diffusive components could be affected by hypoxia. Assuming a PCl/PNa ratio of 0.36 again for the SW-acclimated killifish, then this would predict a diffusion potential of +10 mV, about two-thirds of the total observed TEP, meaning that the electrogenic component is about +5 mV. However if PCl/PNa dropped to 0.22 again under hypoxia, the diffusion potential would actually increase to +16 mV, whereas a decrease in TEP was observed. Indeed when FW-acclimated animals were transferred to hypoxic SW, they were not able to increase TEP at all (Fig. 2C). The more likely explanation is that the inhibitory mechanism is different in SW, and the drop in TEP is mainly due to inhibition of the electrogenic component by hypoxia. Thus active Cl− extrusion would be reduced. To our knowledge, there is no published information on the effect of severe hypoxia on active ion transport rates in killifish in vivo, but in the isolated opercular epithelium of SW-acclimated killifish studied in vitro, complete anoxia reduces the active Cl− current by 75–83% (Karnaky et al., 1977; Barnes et al., 2014). This effect, together with the fact that the reduced TEP in the current study results in about a 70% increase in the electrochemical gradient driving Na+ entry, suggests that the electrical response to hypoxia would be deleterious to ionoregulatory homeostasis in SW-acclimated killifish. As SW fish also drink as part of their osmoregulatory strategy, there may also be direct or indirect effects on gut function.
Interestingly, a fish would occasionally escape during TEP measurement under hypoxia, reaching the better-oxygenated surface layer. In SW-acclimated killifish, TEP would rise immediately, but this did not happen in FW-acclimated animals, supporting the interpretation of a different etiology for the hypoxia-induced TEP reduction in the two media. In future studies, it will be of interest to evaluate these ideas by actually measuring the effects of acute severe hypoxia on unidirectional ion flux rates in vivo in FW- and SW-acclimated killifish. It would also be instructive to look at a freshwater fish with active Cl−-uptake mechanisms, as well as stenohaline species in FW as well as in SW.
MATERIALS AND METHODS
Adult killifish (Fundulus heteroclitus macrolepidotus Walbaum 1792) (5–12 g) were purchased from Aquatic Research Organisms (ARO) Ltd (Hampton, NH, USA), and acclimated for at least 1 month to either FW or 100% SW (24–26°C, composition as in Wood and Grosell, 2008) at the University of Miami. All procedures followed an approved University of Miami Animal Care Protocol (IACUC no. 13-225, renewal 02). The fish were fitted with saline-filled intraperitoneal catheters under anaesthesia as described by Wood and Grosell (2008) and allowed to recover for 24 h in shielded 800 ml beakers served with airlines and flowing SW or FW. TEP measurements were performed as in that paper, using 3 mol l−1 KCl–agar bridges connected via Ag/AgCl electrodes to a high-impedance voltmeter (Radiometer pHM84, Copenhagen, Denmark). TEP was expressed relative to the apical (water) side as 0 mV after correction for junction potential. Water O2 levels (as % air saturation) were monitored with a YSI Optical Probe and meter (Digital Professional Series, Yellow Springs, OH, USA), ensuring that measured values were within ±2% of nominal. Different water O2 levels were achieved by volumetric mixing of 100% and 0% (N2-equilibrated) water, and adjusted as necessary by air or N2 gassing.
In series 1, preliminary experiments showed that there was no effect of O2 on TEP down to 50% air saturation. Therefore each fish was acutely transferred, without air exposure, from normoxia (80–100%) to 40%, 30%, 20%, and 10% O2 in the same medium (FW or SW) in random order. Measurements were made after 2 min in the new environment. Because it was necessary to handle the fish during measurements, and TEP is known to be affected by handling stress (Pic, 1978; Wood and Grosell, 2008), the order was randomized. Between each measurement, the fish was returned to normoxia for 5–10 min, and the TEP was recorded again prior to the next transfer. Thus, the normoxic value for each fish represents the mean of 5 measurements, while the experimental values are single measurements.
Series 2 examined whether TEP changes under hypoxia (10% saturation) were stable over time, in both FW and SW fish. In the experimental series, TEP measurements were made under normoxia (P) and then at 2 min (0 h), 1 h, 2 h, and 3 h of hypoxia, and finally after 2 min recovery (R) in normoxia. A simultaneous control series of fish kept in normoxia throughout but handled identically was also performed. During the 3 h hypoxia exposure, a subsurface screen was necessary to prevent the fish performing aquatic surface respiration.
Series 3 examined whether hypoxia (10% saturation) in the originating environment, the destination environment, or both altered the acute TEP response to transfer between FW and SW, and vice versa. Fish were kept for 30 min in one medium under normoxia or hypoxia, and then transferred to the other medium under either normoxia or hypoxia (4 combinations for each acclimation salinity). TEP measurements were made before and 2 min after transfer.
Data were evaluated using Nernst and Goldman–Hodgkin–Katz (GHK) equations as in Wood and Grosell (2008), and the same plasma and water ion concentrations and activity coefficients were used (see supplementary material Table S1). The GHK equation was solved for PCl/PNa (permeability ratio) under various treatments. The net electrochemical gradient on Na+ was calculated as the difference between the Nernst potential and the TEP, signs considered (Kirschner, 1970). Data are shown as means±1 s.e.m. A one-way repeated-measures ANOVA followed by Dunnett's test was used to detect significant differences from the 100% value (series 1) or from the pre-exposure normoxia value (series 2). A one-way ANOVA followed by Tukey's test was used to differentiate treatment effects in series 3. Unpaired two-tailed Student's t-tests with the Bonferroni correction were used to evaluate simultaneous control versus experimental differences in series 2; P<0.05 was employed as the level of significance throughout.
We thank Dr W. S. Marshall for helpful advice and Kevin Schauer for aquatic system maintenance.
C.M.W. performed all experiments, analysed the data, and wrote the manuscript. C.M.W. and M.G. jointly contributed to conceptual design. M.G. provided advice and facilities, and edited the manuscript.
M.G. is Maytag Professor of Ichthyology and is funded by a National Science Foundation (USA) Grant [IOS 1146695] while C.M.W. held a Canada Research Chair and is funded by a Natural Sciences and Engineering Research Council (Canada) Discovery grant.
The authors declare no competing or financial interests.