Effects of Na⁺ channel isoforms and cellular environment on temperature tolerance of cardiac Na⁺ current in zebrafish (Danio rerio) and rainbow trout (Oncorhynchus mykiss)

Temperature is a major environmental factor that has exerted a strong selective pressure on animal life forms. Temperature-driven evolution has led to significant variation in thermal tolerance of ectothermic vertebrates, including fishes (Beitinger, 2000; Johnston and Bennett, 2008). Some fish species are adapted to a relatively narrow range of temperatures such as the stenothermic teleosts of the Antarctic Ocean and the polar cod (Boreogadus saida) of the Arctic Ocean (Somero and DeVries, 1967; Beers and Sidell, 2011; Drost et al., 2016; Abramochkin et al., 2019). In temperate climates, fishes usually have a wider thermal tolerance range, e.g. rainbow trout (Oncorhynchus mykiss) are cold-water fish that survive temperatures between 0 and 28°C (Hokanson et al., 1977; Beitinger, 2000). The zebrafish (Danio rerio), a teleost fish of the lakes and rivers of South-eastern Asia, also tolerates a wide temperature range (7–41°C) but occupies warmer habitats than rainbow trout (Cortemeglia and Beitinger, 2005; López-Olmeda and Sánchez-Vázquez, 2011).

Despite considerable research efforts, the physiological basis for the high temperature tolerance of fishes and other ectotherms remains unresolved. Practically all major processes and functions of the animal body (e.g. circulation, sensory/motor functions, behaviour, metabolism, digestion, immune defence, reproduction) are affected by temperature and therefore more or less adapted to the habitat temperature of the animal (Angilletta et al., 2002; Gracey et al., 2004; Podrabsky and Somero, 2004; Vornanen et al., 2005; MacMillan, 2019). Therefore, it is likely that when approaching the upper temperature tolerance limit of an animal, several processes simultaneously weaken in the animal's body, but the severity of their effect at the organismal level may vary and manifest with varying delays (Vornanen, 2020). Alternatively, different processes/functions may have slightly different thermal optima or failure temperatures, depending for example on the level of biological organization where the process/function appears (Lagerspetz, 1987; Clark et al., 2013). A fundamental question arises: is there a single underlying process or interaction principle that drives the thermal collapse of multiple organ systems? Given that similar thermal limitations appear to apply to unicellular organisms and Metazoa, it has been suggested that the limiting processes of animal life are found at the molecular level (Tattersall et al., 2012).

Electrical excitability is a common process for most tissues of the animal body including nerves, skeletal muscles, heart and smooth muscles. These tissues are responsible for almost all vital functions of the animal body like sensation, learning, behaviour, locomotion, blood circulation, digestion and homeostasis of the body (Hille, 2001). Recently, we have gathered evidence showing that high temperatures can impair electrical excitability and therefore potentially limit the upper thermal tolerance of both ectothermic and endothermic animals (Vornanen, 2020). Electrical excitability or generation of propagating action potentials (APs) is the result of interaction between inward and outward directed flow of ions across the plasma membrane. In studies on thermal tolerance of electrical excitability we have used fish ventricular myocytes as a model system, as they are well suited for patch-clamp experiments. Atrial myocytes are a less suitable model as they fail to maintain stable resting potential in the current-clamp mode of patch-clamp due to the tiny background inward rectifier current (I_K1) and the absence of acetylcholine-dependent inward rectifier current (I_KACh) in isolated cells (Vornanen et al., 2002; Molina et al., 2007). In a quiescent ventricular myocyte, there is a negative resting membrane potential (V_rest), which is maintained by K⁺ efflux via I_K1. For initiation and propagation of cardiac AP, V_rest must be depolarized to the threshold potential (V_th) of an AP. This is accomplished by Na⁺ influx through the voltage-gated Na⁺ channels, which generate the sodium current (I_Na). AP is initiated only when the charge transfer of the inward I_Na (the source current) exceeds the charge transfer of the outward I_K1 (the sink current or resting membrane leak). At high temperatures, electrical excitability of fish ventricular myocytes may fail due to mismatch between the source current (I_Na) and the sink current (I_K1) (Vornanen et al., 2014). Acute warming reduces charge transfer via the I_Na, while K⁺ leak via I_K1 increases: I_Na fails to depolarize V_rest to the threshold potential of AP. At the level of a working heart this appears as atrioventricular block and depression of ventricular beating rate (Haverinen and Vornanen, 2020), which may eventually result in a collapse of cardiac output and thermal death of the fish.

The present study aimed to test the source–sink mismatch hypothesis with respect to the source current, I_Na (Vornanen, 2020). Given the differences in temperature tolerances between rainbow trout (a cold-active temperate species) and zebrafish (a warm-dwelling subtropical species), it was hypothesized that thermal tolerance of the zebrafish I_Na is higher than that of the rainbow trout I_Na. This assumption was tested by comparing I_Na of zebrafish and rainbow trout ventricular myocytes at different temperatures. Another prediction of the source–sink hypothesis is that the channels that generate currents with slow gating kinetics can withstand high temperatures better than channels that produce currents with fast kinetics (Touska et al., 2018; Vornanen, 2020). This is based on the assumption that stiffer molecules have higher activation energy and slower kinetics, i.e. there is a trade-off between flexibility and thermal stability of the proteins (Somero, 1995; Zavodszky et al., 1998; Fields, 2001). Studies on mammalian Na⁺ channels have shown that the skeletal isoform, Na_V1.4, has faster inactivation kinetics than the cardiac isoform, Na_V1.5 (Wang et al., 1996). I_Na with slow inactivation kinetics is able to provide more depolarizing charge at high temperature than I_Na, which is rapidly inactivated. As zebrafish and trout ventricular I_Na are mainly generated by Na_V1.5 and Na_V1.4 channels, respectively (Haverinen et al., 2007; Haverinen et al., 2018), it was hypothesized that those channel isoforms are involved in adaptation of cardiac I_Na to high and low temperature, respectively. To this end, the inactivation kinetics and charge transfer of I_Na between rainbow trout and zebrafish ventricular myocytes were compared at different temperatures.

Finally, it was hypothesized that thermal stability and kinetics of fish I_Na depend in part on the cellular environment (e.g. biophysical properties of the lipid membrane and/or ancillary protein subunits) where they are expressed. To test this, the two main alpha subunits of zebrafish and rainbow trout Na⁺ channels were cloned and expressed in a mammalian cell line, the human embryonic kidney (HEK) cell. This allowed comparison of inactivation kinetics and thermal resistance of the Na⁺ channels in an identical cellular environment.

The wild-type zebrafish, Danio rerio (F. Hamilton 1822) (ab strain, kindly donated by Dr Maxim Lovat, Lomonosov Moscow State University), were raised and maintained at the animal facilities of Lomonosov Moscow State University according to common practices (Westerfield, 2007). The rearing temperature of the fish was 28°C. Fish of either sex, about 1.5 years old, were used for electrophysiological experiments (N=8). Zebrafish were killed by immersion in an ice-water bath and cutting of the spine. Rainbow trout, Oncorhynchus mykiss (Walbaum, 1792), were obtained from a local fish farm (Kontiolahti, Finland). Fish weighing 118.7±20.6 g (N=11) of either sex, acclimated at 12°C for more than 3 weeks, were used in electrophysiological experiments and gene cloning. Trout were stunned by a quick blow to the head and killed by cutting of the spine immediately behind the head. The experiments conform to the ‘European Convention for the Protection of Vertebrate Animals used for Experimental and other Scientific Purposes’ (Council of Europe No. 123, Strasbourg 1985) and were authorized by the national animal experimental board in Finland (permission ESAVI/8877/2019).

Total cardiac RNA was extracted by TriReagent (Thermo Fisher Scientific, Vilnius, Lithuania) and the quality and quantity of RNA were determined by agarose gel electrophoresis and NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA), respectively. RNA (2 µg) was treated with RNase free DNase (Thermo Fisher Scientific) and converted to cDNA using SuperScript IV reverse transcriptase (Invitrogen, Glasgow, UK) and oligo(dT) primers. cDNA (1 µl) was used as a template in 25 µl polymerase chain reaction (PCR) including a final concentration of 0.2 mmol l⁻¹ dNTP mix, 0.2 mmol l⁻¹ primers (synthesized by Invitrogen) as shown in Table 1, and 0.02 U µl⁻¹ Phusion High Fidelity DNA polymerase (Thermo Fisher Scientific). Cycling conditions for all PCR reactions were as follows: initial denaturation at 98°C for 1 min, 35 cycles at 98°C for 10 s, at 60°C for 30 s and at 72°C for 100–260 s (40 s per kb; for the length of the products see Table 1), followed by final extension at 72°C for 5 min. The PCR products were separated on a 0.8% agarose gel, and the nucleotide chains were extracted from the gel using a GeneJET Gel Extraction Kit (Thermo Fisher Scientific). Overhang adenines were added to the 3′-ends of PCR products using Dynazyme II DNA polymerase (Thermo Fisher Scientific) and products were ligated into the pGEM-T Easy vector (Promega, Madison, WI, USA). The inserts were digested from the pGEM-T Easy vector and directionally cloned into the expression vector. If the coding sequence of the SCN gene was amplified by PCR in two parts, the pieces were ligated into the expression vector to form the entire coding sequence. Coding sequences of the SCN5LA genes of rainbow trout (om for O.mykiss) and zebrafish (dr for D.rerio) were cloned into the expression vector pcDNA3.1/Zeo(+) (Invitrogen). Several attempts to clone the full-length om/drSCN4A into the pcDNA3.1 failed. To overcome this problem, the high-copy number vector pcDNA3.1/Zeo(+) was converted to a low-copy number vector that was better tolerated by bacterial cells. To this end, the ColE1 origin was replaced by the pMB1 origin via digesting pcDNA3.1/Zeo(+) with BsmI and PvuI and replacing this fragment (bases 2768–4462) with the digested fragment from pBR322 (bases 1353–3733). The coding sequences of dr/omSCN4A were successfully cloned into this modified vector. Plasmids were isolated using PureLink HiPure Plasmid Midiprep Kit (Thermo Fisher Scientific) and sequenced by GATC Biotech AB (Köln, Germany). Nucleotide sequences were converted to protein sequences using EMBOSS Transeq software (https://www.ebi.ac.uk/Tools/st/emboss_transeq/), and the protein sequences of the species-specific isoforms were aligned with EMBOSS Needle (https://www.ebi.ac.uk/Tools/psa/emboss_needle/). The fish SCN4A and SCN5LA genes are orthologous to the mammalian SCN4A and SCN5A genes. For simplicity, the protein names (Na_V1.4 and Na_V1.5) of the orthologous genes are used throughout the text.

Due to the whole genome duplications of the teleost lineage, the number of gene paralogues expressed in fish striated muscles is high (Alderman et al., 2012; Glasauer and Neuhauss, 2014). All eight teleost Na⁺ channel genes are expressed in the heart of both zebrafish and rainbow trout (Haverinen et al., 2018; Hassinen et al., 2021). Because striated muscle isoforms, Na_V1.4 and Na_V1.5, make up the great majority (>99%) of all Na⁺ channel transcripts in zebrafish and trout hearts, genes for heterologous expression were selected from these paralogues. For both Na_V1.5 and Na_V1.4, there are three paralogues in the trout genome and two paralogues in the zebrafish genome. For Na_V1.5 channels, drSCN5LAb and omSCN5LAba, the most abundant paralogues of ventricular myocytes, were expressed in HEK cells. For the expression of zebrafish Na_V1.4 channels, drSCN4Ab, the most abundantly expressed SCNA paralogue of ventricular myocytes, was inserted in HEK cells. In the case of trout Na_V1.4 channels, omSCN4Abb, the second most abundant SCN4A paralogue of ventricular myocytes (most abundant in atrial myocytes) was expressed in HEK cells. We failed to clone the most abundant paralogue of the ventricle, omSCN4Aba, and were therefore forced to use the major atrial isoform, omSCN4Abb. Human embryonic kidney (HEK293; ECACC) cells were grown at 37°C in DMEM (Biowest) supplemented with 10% fetal bovine serum (FBS, sterile-filtered and heat inactivated; Sigma-Aldrich) and 100 U ml⁻¹ penicillin and streptomycin (Sigma-Aldrich) in a 5% CO₂ environment. HEK cells were transiently co-transfected with pEGFP-N1 (Clontech), and either drSCN4Ab, drSCN5LAb, omSCN4Abb or omSCN5LAba construct using TurboFect transfection reagent (Thermo Fisher Scientific). Cells were transfected for 16–18 h at 37°C; plate medium was then refreshed, and the cells were further incubated at 28°C in a 5% CO₂ environment for at least 24 h. Incubation at lower temperature increased the expression level of Na⁺ channels. Whole-cell patch-clamp experiments were conducted 48–72 h after transfection.

The procedure of isolating zebrafish cardiomyocytes was essentially similar to the original method of isolating crucian carp (Carassius carassius) cardiac cells (Vornanen, 1997), but scaled down to the size of small zebrafish hearts and using slightly higher enzyme concentrations (1 mg ml⁻¹ collagenase Type IA and 0.67 mg ml⁻¹ Trypsin IV; both from Sigma). A blunt-ended syringe needle (34 gauge, TE734025; Adhesive Dispensing Ltd, Milton Keynes, UK) cannula was inserted via the bulbus arteriosus into the ventricle and secured in place with a fine thread. The heart was perfused first with Ca²⁺-free solution for 5 min and then with the same solution but with added hydrolytic enzymes together with fatty acid-free serum albumin (1 mg ml⁻¹; Sigma) for 25–30 min. Myocytes were stored at 5°C and used on the same day as they were isolated. The Ca²⁺-free solution contained (mmol l⁻¹): 100 NaCl, 10 KCl, 1.2 KH₂PO₄, 4 MgSO₄, 50 taurine, 20 glucose and 10 HEPES at pH 6.9 (adjusted with KOH at 20°C). From rainbow trout heart, myocytes were obtained using essentially the same procedure, but with lower enzyme concentrations (0.75 mg ml⁻¹ collagenase Type IA and 0.5 mg ml⁻¹ Trypsin IV) and shorter digestion time (10–12 min).

The whole-cell voltage-clamp recording of I_Na was performed using an Axopatch 200B or an Axopatch 1-D amplifier (Molecular Devices, San Jose, CA, USA) as previously described in detail (Haverinen et al., 2018). Cardiac myocytes or coverslips containing cultured HEK cells were placed in a small chamber with a continuous flow of K⁺-based external saline solution containing (mmol l⁻¹): 150 NaCl, 5.4 KCl, 1.8 CaCl₂, 1.2 MgCl₂, 10 glucose and 10 HEPES, with pH adjusted to 7.7 at 20°C with NaOH. The temperature of the saline solution was set using a Peltier device (CL-100, Warner Instruments, Hamden, CT, USA; or TC-10, Dagan, Minneapolis, MN, USA) and monitored continuously with thermistors placed close to the cells. Patch pipettes with a resistance of 1.5–2.5 MΩ were drawn from borosilicate glass (Hilgenberg GmbH, Malsfeld, Germany) and filled with Cs⁺-based electrode solution containing (mmol l⁻¹): 5 NaCl, 130 CsCl, 1 MgCl₂, 5 EGTA, 5 Mg₂ATP and 5 HEPES, with pH adjusted to 7.2 with CsOH. Current amplitudes were normalized to the capacitive cell size to obtain current density (pA pF⁻¹).

The time constant of I_Na inactivation at different membrane potentials (–30 to 0 mV) was derived by fitting the decay phase of I_Na using the double exponential function of the Chebyshev transformation procedure of the Clampfit 10.3 software package. The amplitude of the fast component (τ_f) was over 90% of the current. The slow component could only be reliably determined at the voltages at which I_Na was close to its peak amplitude. Therefore, only the results of τ_f are reported. Thermal coefficient (Q₁₀) values of I_Na inactivation rate were calculated using the equation: Q₁₀=(R₂/R₁)^{10°C/(T₂−T₁)}, where R₁ and R₂ are fast time constants of inactivation (τ_f) at temperatures T₁ and T₂.

All mentioned properties of I_Na were analysed in native trout or zebrafish myocytes and HEK cells expressing Na_V1.4 or Na_V1.5 channels at three different temperatures: 12, 20 and 28°C. However, additional experiments to estimate the dynamics of the temperature dependence of I_Na were done using the ‘heat ramp’ protocol. In these experiments I_Na was elicited by repetitive depolarizations to −20 mV and the temperature was steadily raised from 12 to 28°C (up to 39°C in the case of native zebrafish myocytes).

Statistical analyses were performed using SPSS (IBM; version 25). One-way ANOVA (with Tukey's or Dunnett's T3 post hoc test) and unpaired t-test were used to compare normally distributed data with homogenous variances. If the assumptions of parametric tests were not met, non-parametric Kruskal–Wallis and post hoc Mann–Whitney U-tests were used. P<0.05 was considered to show a statistically significant difference between means.

Temperature dependency of ventricular I_Na was markedly different between zebrafish and rainbow trout myocytes, as shown by the current–voltage (I–V) relationships (Fig. 1). In zebrafish myocytes, an acute rise of temperature from 12 to 20°C resulted in a significant increase in I_Na density at voltages between −50 and −30 mV (Fig. 1C). A similar difference was detected in I_Na density between 12 and 28°C (P<0.05). However, the current densities at 20 and 28°C were almost identical, suggesting plateauing of I_Na in this temperature range (Fig. 1C). In rainbow trout myocytes, warming from 12 to 20°C increased I_Na density at voltages from −30 to −10 mV, but further warming to 28°C caused a strong depression of I_Na (Fig. 1E). Although the density of the zebrafish I_Na was higher at 20 and 28°C than at 12°C, the charge transfer (integral of I_Na) was significantly smaller than at 12°C (Fig. 1D). In trout ventricular myocytes, the integral of I_Na decreased with increasing experimental temperature (Fig. 1F). Notably, charge transfer/I_Na density ratio, i.e. depolarizing power, was 4.5–22.2 times higher (at −20 mV) for zebrafish than rainbow trout I_Na at all experimental temperatures (0.6, 0.7 and 0.9 V A⁻¹ for rainbow trout and 2.7, 4.5 and 20.0 V A⁻¹ for zebrafish at 28, 20 and 12°C, respectively).

Temperature tolerance of peak density and charge transfer of I_Na (at −20 mV) were further studied using acute heat ramps. To this end, the cells were warmed from 12°C to their upper thermal tolerance limit while I_Na was elicited by depolarization from −100 to −20 mV for every second (Fig. 1G). Consistent with the I–V data, there was a dramatic interspecies difference in temperature dependence of the peak I_Na (Fig. 1G). Initially, when the temperature was raised above 12°C, the density of I_Na increased in both species. However, the breakpoint temperature (T_BP, the temperature above which peak I_Na started to decrease steadily) was much lower for rainbow trout I_Na (18.3±0.6°C, N=13) than for zebrafish I_Na (26.6±0.5°C, N=16) (P<0.05).

Taken together, these findings indicate that the ventricular I_Na of rainbow trout heart is much less tolerant of high temperatures than the zebrafish ventricular I_Na.

Temperature dependence of I_Na inactivation kinetics was measured in the voltage range −30 to 0 mV at 12, 20 and 28°C (Fig. 2). Kinetics of zebrafish or rainbow trout I_Na accelerated with increasing temperature and membrane depolarization. There was a striking difference in the rate of I_Na inactivation between trout and zebrafish ventricular myocytes (Fig. 2). At 12, 20 and 28°C, the time constant (τ_f) of I_Na inactivation (at 0 mV) was about 10.0, 6.3 and 4.0 times faster in rainbow trout ventricular myocytes in comparison with zebrafish ventricular myocytes (P<0.05).

Voltage dependence of steady-state (SS) activation and inactivation of I_Na was studied at 12, 20 and 28°C (Fig. 3). In zebrafish ventricular myocytes, warming from 12 to 20°C and further to 28°C decreased the slope factor of the SS inactivation curve (P<0.05), while V_0.5 remained unaffected (P>0.05; Fig. 3C; Table 2). Voltage dependence of SS activation of I_Na in zebrafish myocytes was more temperature sensitive, as warming to either 20 or 28°C shifted V_0.5 to the left by almost 10 mV and decreased the slope factor (P<0.05) (Table 2; Fig. 3C). In trout ventricular myocytes, warming from 12 to 20°C and further to 28°C failed to shift V_0.5 or change slope factor of both curves (Table 2; Fig. 3D).

Zebrafish and rainbow trout Na_V1.4 and Na_V1.5 Na⁺ channels were expressed in the same cellular environment for a direct comparison between temperature dependencies of the orthologous gene products (Fig. 4). Protein sequences of these Na⁺ channels suggest a relatively high degree of structural similarity between the species. Trout and zebrafish Na_V1.4 shared 64.5% identity and 76.0% similarity, whereas identity and similarity for Na_V1.5 were 79.9 and 87.1%, respectively (Figs S1 and S2).

I_Na generated by the zebrafish Na_V1.4 channels responded to acute temperature increases in a similar manner to native ventricular I_Na: the current density increased, and the charge transfer decreased with increasing temperature (Fig. 4A,B). I_Na generated by the trout Na_V1.4 channels did not show any statistically significant differences between the three temperatures for either current density or charge transfer (Fig. 4C,D). (The expression level of trout Na_V1.4 channels in HEK cells was much lower than zebrafish Na_V1.4 channels, which may have reduced the resolution of the analysis.) The I_Na generated by trout Na_V1.5 channels was, however, much more heat tolerant than the native trout ventricular I_Na: current density and charge transfer strongly increased with increasing temperature in HEK cells (Fig. 4G,H), while in ventricular myocytes both variables were strongly reduced at 28°C (Fig. 1E,F).

I_Na density of the zebrafish Na_V1.5 channels in HEK cells increased with increasing temperature (P<0.05) (Fig. 4E). Differently from ventricular I_Na, the heterologously expressed Na_V1.5 channels positively responded to warming (at −50 mV) up to 28°C, i.e. I_Na indicated a better heat tolerance in HEK cell membranes than in ventricular myocytes. No statistically significant differences were found in charge transfer by the heterologously expressed Na_V1.5 channels (Fig. 4F). I_Na density of the rainbow trout Na_V1.5 channels in HEK cells increased with warming from 12 to 28°C (Fig. 4G). This is a striking deviation from the thermal response of the ventricular I_Na, which was strongly depressed at 28°C. Charge transfer by rainbow trout Na_V1.5 channels was significantly increased by acute warming in HEK cells, in contrast to the observations of ventricular I_Na (Fig. 4H).

The inactivation rate of I_Na increased with increasing temperature and with membrane depolarization. There were marked gene-specific differences in the rate of I_Na inactivation (Fig. 5). The rate of inactivation of I_Na produced by Na_V1.5 channels was much slower than that produced by Na_V1.4 channels for both zebrafish and trout genes (P<0.05). In contrast to gene-specific differences, interspecies differences in inactivation rate of I_Na were non-existent or small. The time constant of I_Na inactivation of Na_V1.5 channels was almost identical for zebrafish and trout variants of the channel. In contrast, the I_Na generated by Na_V1.4 channels was faster for zebrafish than the trout channel variant at more negative voltages (at −30 mV at 20°C, and at −30 and −20 mV at 28°C) (Fig. 5D,F). Temperature dependence (Q₁₀ values) of I_Na inactivation varied between 1.8 and 6.0 with no differences between channel isoforms. However, the Q₁₀ value of I_Na inactivation rate was higher in zebrafish ventricular myocytes than in trout ventricular myocytes at voltages of −10 and 0 mV (P<0.05) (Table S1).

Interestingly, the rate of I_Na inactivation of heterologously expressed zebrafish Na_V1.4 and Na_V1.5 Na⁺ channels was much faster than the inactivation kinetics of the endogenous ventricular I_Na. Opposite to the findings in zebrafish, the I_Na inactivation rate of trout Na_V1.5 channels was much slower than that of trout ventricular myocytes. In contrast, the inactivation rates of I_Na for heterologously expressed trout Na_V1.4 channels and trout ventricular myocytes were similar.

Acute increases in temperature had only weak effects on the voltage dependence of SS activation and inactivation of I_Na generated by the heterologously expressed Na⁺ channels (Table 2; Fig. 6). The only change for the zebrafish I_Na produced by Na_V1.5 channels was the reduced slope of SS activation at 20 and 28°C relative to that at 12°C (P<0.05). Similar changes were observed for the rainbow trout I_Na generated by Na_V1.5 channels (P<0.05) (Fig. 6). In addition, the rainbow trout Na_V1.5 channels had a positive shift in the voltage dependence and a decrease in the slope factor of the I_Na SS inactivation (P<0.05) (Table 2; Fig. 6D). Elevated temperatures also reduced the slope factor of SS activation of the trout Na_V1.4 channels (P<0.05) (Table 2; Fig. 6B).

The present results can be summarized in three major findings. (1) The properties of the endogenous I_Na of zebrafish and rainbow trout ventricular myocytes differ markedly in terms of heat tolerance and inactivation kinetics, with the zebrafish I_Na being more heat tolerant and more slowly inactivating. (2) The major Na⁺ channel isoforms of zebrafish and rainbow trout ventricles, Na_V1.5 and Na_V1.4, respectively, show only minor interspecies differences, when expressed in HEK cells, i.e. the orthologous Na⁺ channel alpha subunits are functionally (heat tolerance, inactivation kinetics) similar in the same membrane matrix. (3) When expressed in HEK cells, I_Na generated by Na_V1.4 and Na_V1.5 isoforms of both species show large channel-specific differences in inactivation kinetics, with Na_V1.4 being fast and Na_V1.5 slow. Taken together, the species-specific properties of ventricular I_Na seem to be determined partly by the expressed Na⁺ channel alpha subunit – Na_V1.5 in zebrafish and Na_V1.4 in rainbow trout – and partly by the biophysical properties of the lipid matrix/ancillary subunits of the channel assembly. Notably, the better heat tolerance of I_Na seems to be related to the slower inactivation kinetics of Na⁺ channels.

The endogenous I_Na of zebrafish and rainbow trout ventricular myocytes have very different inactivation kinetics: the I_Na of zebrafish ventricular myocytes inactivates much more slowly than I_Na of rainbow trout ventricular myocytes. At 12°C, the inactivation time constant of the zebrafish I_Na was almost an order of magnitude bigger than that of the rainbow trout (5.0±0.7 versus 0.5±0.03 ms at −20 mV). This difference largely, but not completely, disappeared when the rate of I_Na inactivation was measured at the acclimation temperatures of the fish (0.8±0.09 ms for zebrafish at 28°C versus 0.5±0.03 ms for trout at 12°C). As thermal acclimation does not have any effect on the inactivation kinetics of the rainbow trout ventricular I_Na (Haverinen and Vornanen, 2004), the differences in the inactivation rate between the two species can be regarded as adaptations of I_Na to the respective habitat temperatures of the species. In general, electrical excitation in nerves and muscle tissues of ectothermic animals is adapted to work best at lower temperatures in comparison with the tissues of the endothermic animals. For instance, at the typical mammalian body temperatures (36–38°C), the AP conduction of ectothermic nerve fibres suffers from heat block (Hodgkin and Katz, 1949; Volgushev et al., 2000). The gating kinetics of the plasma membrane ion channels are probably responsible for the thermal adaptation of AP conduction, allowing APs to propagate at the proper rate and frequency at the typical habitat temperatures of the species. Because zebrafish live at warmer habitats than rainbow trout, the slow inactivation kinetics of its ventricular I_Na may provide better excitability at higher temperatures than the fast inactivating I_Na of the rainbow trout. However, when comparing the intrinsic heart rates of the two species at their acclimation temperatures (130 beats min^–1 for zebrafish versus 60 beats min^–1 for rainbow trout) (Aho and Vornanen, 2001; Vornanen and Hassinen, 2016), the rate of I_Na inactivation appears to be slow. Apparently, the rate of I_Na recovery from inactivation does not limit heart rate in zebrafish, although this was not experimentally confirmed.

The slow rate of I_Na inactivation in zebrafish ventricular myocytes may protect against heat-induced impairment of electrical excitability (Park et al., 2016; Touska et al., 2018). This is consistent with our hypothesis that Na⁺ channels with slow gating kinetics better maintain electrical excitability at high temperatures (Vornanen, 2020). Indeed, findings from the pain receptors of human skin strongly suggest that the slow gating kinetics of Na⁺ channels are needed for heat resistance of I_Na (Touska et al., 2018). At mammalian body temperatures, most Na⁺ channel isoforms operate close to their optimum and a slight increase in temperature above the typical body temperature limits the density and charge transfer of I_Na (Touska et al., 2018). However, I_Na of the mammalian pain receptors works well at temperatures much above body temperature (43–50°C), largely owing to the thermal properties of the Na_V1.9 Na⁺ channel isoform. Inactivation kinetics of this channel are more than an order of magnitude slower than those of Na_V1.1–Na_V1.8 isoforms (Balbi et al., 2017; Touska et al., 2018). Due to the slow inactivation of Na_V1.9, it provides enough charge for depolarization of the plasma membrane at high temperatures. Analogously, the slow inactivation rate of the zebrafish cardiac I_Na provides more inward charge for the same peak amplitude of I_Na or the same number of active Na⁺ channels than the fast inactivating trout cardiac I_Na (Fig. 4). Indeed, the charge transfer/peak current ratio of the zebrafish I_Na is much higher at 28°C than that of the rainbow trout I_Na at 12°C.

The species-specific difference in the inactivation kinetics of the ventricular I_Na is partly explained by the difference in the alpha subunit composition of the Na⁺ channels. In zebrafish ventricular myocytes, the main alpha subunit is Na_V1.5, which at the transcript level represents 83.1% of all ventricular Na⁺ channels. Transcripts of the Na_V1.4 comprise only 16.2% of the zebrafish ventricular Na⁺ channels (Haverinen et al., 2018). In the ventricle of the rainbow trout, the situation is opposite: Na_V1.4 channels form 80% of all Na⁺ channel transcripts, while Na_V1.5 channels represent only 20% of the total channel population (Haverinen et al., 2007). Thus, in zebrafish ventricle Na⁺ channels are mainly of the slow isoform, while in the ventricle of the rainbow trout they are mainly the fast isoform. Based on their tissue distribution in mammals, Na_V1.5 and Na_V1.4 are often called ‘cardiac’ and ‘skeletal’ isoforms, respectively (Zimmer et al., 2015). The mammalian Na_V1.4 is kinetically faster than Na_V1.5 and therefore functionally better at eliciting fast twitches at high frequencies in skeletal muscle fibres (Wang et al., 1996; Sheets and Hanck, 1999). Although the classification of Na_V1.5 and Na_V1.4 to cardiac and skeletal isoforms, respectively, is not valid for fish, the kinetic similarities between the orthologous mammalian and piscine Na⁺ channels persist. The fast inactivation kinetics of the rainbow trout Na_V1.4 can be regarded adaptive for heart function of this cold-dwelling fish.

The dominance of Na_V1.5 alpha subunits in the zebrafish heart only partially explains the slow inactivation kinetics of the ventricular I_Na. Unlike the rainbow trout, where the inactivation rate of Na_V1.4 channels in HEK cells relatively closely matches the inactivation rate of the endogenous ventricular I_Na, in zebrafish there is a large difference in the inactivation kinetics of the heterologously expressed Na_V1.5 and the endogenous ventricular I_Na. As Na_V1.4 and Na_V1.5 Na⁺ channel alpha subunits of zebrafish and rainbow trout share the inactivation kinetics in the heterologous environment of HEK cells, the slow inactivation rate of the zebrafish ventricular I_Na is probably partly due to the general biophysical properties of the bulk lipid bilayer of ventricular myocytes in the membrane domain where they are located (Lundbaek et al., 2004). Phospholipid bilayer composition of the plasma membrane affects function of the integral membrane proteins mainly in a non-specific manner by its biophysical properties (Lundbaek et al., 2004). Elasticity (stiffness) of the lipid bilayer, determined by the lipid composition, regulates the inactivation of the voltage-gated Na⁺ channels. Decrease in membrane stiffness induced by amphiphiles like Triton-X or β-octyl-glucoside accelerate the rate of I_Na inactivation, while increase in membrane stiffness due to elevated cholesterol content decreases the rate of I_Na inactivation (Lundbaek et al., 2004). Therefore, it is possible that the lipid composition of the sarcolemma in zebrafish ventricular myocytes differs from that of HEK cells and rainbow trout myocytes, and results in increased stiffness, which slows the rate of I_Na inactivation.

Although we do not have a complete answer about the slow inactivation rate of the zebrafish ventricular I_Na, some possible explanations (in addition to membrane stiffness) can be mentioned to guide future studies. Na⁺ channels are heteromultimers of large pore-forming alpha subunits and small accessory beta subunits. Beta subunits are needed for proper transportation of the alpha subunits into the plasma membrane and they may also affect kinetics of the I_Na. However, beta subunits are unlikely to cause the slow inactivation of the zebrafish I_Na as they tend to enhance the rate of Na⁺ channel inactivation and recovery from inactivation (Chen and Cannon, 1995; Goldin, 2003). A more likely contributing factor is the fibroblast growth factor orthologous factor 2 (FGF2), which binds to the inactivation domain of the C-terminus in the Na⁺ channel alpha subunit and strongly slows the rate of inactivation (Liu et al., 2003; Li et al., 2020). Interestingly, FGF2 knock-out mice are highly sensitive to temperature change and show more cardiac conduction defects when their core body temperature is elevated (Park et al., 2016). Future studies should examine the role of FGFs in temperature dependence of electrical excitability of the fish heart.

Acute heat challenge experiments indicated that the zebrafish ventricular I_Na is much more resistant to high temperatures than the rainbow trout I_Na. Thus, the heat tolerance of I_Na seems to positively correlate with the upper thermal tolerance of the fish and its heart rate, although in both species the optimum temperature of I_Na is lower than the critical thermal maximum temperature of the fish and the T_BP of the intrinsic heart rate (Beitinger, 2000; Aho and Vornanen, 2001; Cortemeglia and Beitinger, 2005; López-Olmeda and Sánchez-Vázquez, 2011; Vornanen and Hassinen, 2016). It is clear that I_Na of the trout ventricle would be almost non-functional at the acclimation temperature of the zebrafish, and it is likely that the kinetics of the zebrafish I_Na would be too slow for the trout heart at freezing temperatures.

Warming-induced decrease of heart rate is shown to be caused by atrioventricular block, probably due to the reduced excitability of the ventricle (Haverinen and Vornanen, 2020). Therefore, the thermal properties of I_Na are likely to affect the species-specific T_BP of heart rate. Warming-induced decrease of I_Na and simultaneous increase in membrane K⁺ leak via I_K1 results in source–sink mismatch, which may prevent AP generation (Vornanen, 2016; Haverinen and Vornanen, 2020; Vornanen, 2020). At low temperatures there is some excess of I_Na relative to the membrane leak via I_K1 (I_Na/I_K1≥1.0) called safety factor. When temperature rises the charge transfer of I_Na starts to decline (at the species-specific T_BP) due to increased rate of inactivation: the safety factor is lost (I_Na/I_K1<1.0) and excitation fails. In this respect, the difference in heat tolerance between the ventricular I_Na of rainbow trout and the I_Na generated by the heterologously expressed Na_V1.4 and Na_V1.5 channels of the trout may be important. Expression of trout I_Na in the HEK cell membrane makes it much more heat tolerant than it is in the native membrane surroundings. Notably, when zebrafish and trout Na_V1.4 and Na_V1.5 were expressed in HEK cells, inactivation rate and heat tolerance of I_Na were similar for the orthologous channels. It seems that the biophysical properties of the mammalian cell membrane shift the thermal tolerance window of the trout Na⁺ channels to higher temperatures which would, however, be suboptimal at the habitat temperature of the cold-dwelling fish.

Heat tolerance and inactivation kinetics of I_Na differ strongly between zebrafish and rainbow trout ventricular myocytes. In contrast, heat tolerance and inactivation kinetics of Na_V1.4 and Na_V1.5 channels of zebrafish and rainbow trout are similar when expressed in the same cellular environment of HEK cells. Species-specific thermal adaptation of the ventricular I_Na is largely achieved by expressing a specific alpha isoform subunit of Na⁺ channel: the slowly inactivating Na_V1.5 in zebrafish that tolerate higher temperatures, and the fast inactivating Na_V1.4 in rainbow trout that favour cold waters. Differences in elasticity (stiffness) of the lipid bilayer and/or accessory protein components may also be involved in the thermal adaptation of I_Na. These findings are consistent with the hypothesis that slow Na⁺ channel kinetics are associated with increased heat tolerance of cardiac excitation. Future studies should examine the extent to which these components are flexible under temperature acclimation and therefore able to accommodate the electrical excitability of the heart for seasonal temperature changes and peak summer temperatures. As electrical excitability is regulated in basically the same way in all excitable cells, studying the biophysical properties of neuronal and muscular I_Na could provide clues to its role in thermal homeostasis and death of ectotherms.

We would like to thank Kontiolahti fish farm for providing the trout. Anita Kervinen is acknowledged for taking care of the fish at the university and preparing solutions for experiments.