Temperature and pH-dependent potassium currents of muscles of the stomatogastric nervous system of the crab, Cancer borealis

Summary

Marine crustaceans such as the crab, Cancer borealis, experience fluctuations in temperature and pH, yet their stomatogastric neuromuscular system must remain functional for feeding. We examined 16 of ∼40 stomach muscle pairs and found that warming consistently hyperpolarized muscle fibers (∼10 mV per 10°C) and reduced excitatory junctional potentials and currents. Muscle responses were also strongly influenced by extracellular pH, with an optimal range between pH 6.7 and 8.8; outside this window, abnormal activity emerged. Voltage-clamp analysis of gastric muscle gm5b revealed a temperature- and pH-sensitive conductance with a reversal potential near the potassium equilibrium potential and insensitivity to tetraethylammonium and barium, providing evidence against classical voltage-gated potassium channels. Quantitative RT-PCR detected the expression of two putative two-pore domain potassium (K2P) channels in these muscles. Together, these results suggest that muscle excitability in C. borealis is shaped by temperature- and pH-sensitive potassium currents consistent with contributions from K2P channels.

Graphical abstract

Highlights

• •Cancer borealis stomach muscles are sensitive to temperature and pH
• •Warming or alkalizing hyperpolarizes fibers and reduces synaptic response amplitude
• •RT-qPCR detects K2P gene transcripts CbKCNK1 and CbKCNK2 in muscles
• •CbKCNK1 and CbKCNK2 are candidates for the temperature and pH-dependent conductances

Biological sciences; neuroscience; zoology

Introduction

Many long-lived marine crustaceans, such as lobsters and crabs, face daily, monthly, and seasonal changes in water temperature and other environmental conditions, including pH and salinity. Specifically, in the New England Atlantic Ocean, the habitat of the crab, Cancer borealis (C. borealis), water temperatures vary from ∼3 °C to ∼25 °C.^1,2,3,4 Local pH also fluctuates depending on water depth, temperature, and tidal cycles.⁵ Nonetheless, the function of these animals is preserved, despite the strong influence both temperature and pH have on all biological processes, including electrical excitability and synaptic transmission. For this reason, it is interesting to understand how temperature and pH affect the neuronal and muscular components that must function for the animal’s behavior to be temperature and pH resilient. In this study, we examine the biophysical substrate of the effects of temperature and pH on muscle fiber membrane potential and synaptic transmission in the stomatogastric nervous system.

The crustacean foregut is a complex mechanical structure, involving the coordinated movements of more than 40 pairs of striated muscles, innervated by ∼20 excitatory motor neurons found in the stomatogastric ganglion (STG).⁶ These muscles are responsible for grinding food by moving teeth inside the stomach and filtering food through the pylorus. Many of the STG motor neurons provide the sole innervation of one or more muscles of the stomach, using either glutamate or acetylcholine to excite the muscles.^7,8,9 In this study, we take advantage of the simple innervation patterns of the stomatogastric neuromuscular system to study the effects of temperature and pH on these muscles.

Historically, it has long been known that crustacean muscles hyperpolarize when they are warmed, but this was attributed to effects on the Na⁺/K⁺ pump and/or changes in leak conductance.^10,11,12,13 We now argue instead that this hyperpolarization reflects the activation of two-pore K⁺ channels (K2P), which were discovered after the earlier work on the effects of temperature on crustacean muscle membrane potential.

Recent work has characterized a family of two-pore K⁺ channels, K2P channels, which are temperature and/or pH-sensitive.^14,15,16,17 In humans and mice, some of these channels are implicated in mechanosensation, temperature sensing, pain sensing, and are modulated by general anesthetics or bioactive molecules such as arachidonic acid.^18,19,20 In vertebrates, there are approximately 15 genes that encode this family of channels.²¹ In Drosophila, there are approximately 10 putative homologs,²² while in C. elegans, about 50 have been tentatively identified.²¹ Recent studies of the Homarus americanus²³ and C. borealis²⁴ genomes have identified only two candidate two-pore K⁺ channel candidates thus far, which we have studied here.

We now present electrophysiological and molecular data that identifies a temperature- and pH-sensitive K⁺ conductance whose reversal potential follows the Nernst prediction for K⁺. Then, we show its insensitivity to Barium and TEA, typical voltage-gated K⁺ channel blockers. We further show that two K2P channel genes are expressed in these muscles by using quantitative RT-PCR. Features of their sequences are consistent with pH sensitivity. Taken together, these findings support K2P channels as candidate contributors to the temperature- and pH-dependent modulation of resting membrane voltage in crustacean muscle.

Results

Figure 1A shows a diagram of the stomach muscles of C. borealis studied here. These include the cpv1ab muscles that are innervated by the cholinergic pyloric dilator (PD) motor neurons, and muscles that are innervated by glutamatergic motor neurons lateral gastric (LG) (gm5b; gm6; gm8a,b), dorsal gastric (DG) (gm4b; gm4c), inferior cardiac (IC) (cv2; gm5a), lateral pyloric (LP) (p1; cpv4; cpv6), pyloric (PY) (p2; p8), and lateral posterior gastric (LPG) (p7). Figure 1B is a schematic showing the recording configuration used to stimulate the innervating motor nerves with a suction electrode while recording intracellularly from the muscle fibers.

Figure 1.

Temperature effect on stomach muscles

(A) Schematic of the stomatogastric musculature based on the nomenclature from,⁶ including gastric muscles (gm), pyloric muscles (p), and cardio-pyloric (cpv) muscles—adapted from.²⁵ Only the left side of the stomach is shown, with the posterior directed toward the bottom.

(B) Schematic of the experimental setup, showing in vitro dissected muscles with the innervating nerves attached. The LG neuron projects its axon to the lateral ventricular nerve (lvn) and the lateral gastric nerve (lgn). The lgn is stimulated via a suction electrode with a 2-s train at 2 or 5 Hz. The EJPs are recorded via an intracellular electrode in the muscle fiber. .

(C) Effect of temperature on nerve-evoked EJPs and resting membrane voltage (V_rest) (dashed gray line).

(D) Effect of temperature on nerve-evoked EJCs. A gm5b and gm6 muscle fiber are clamped at −80 mV in a two-electrode voltage-clamp configuration. The lgn is stimulated at 5 Hz for 2 s at 11°C and 21°C.

(E) Quantification of the effect of temperature on V_rest for LG muscles (gm5b, gm6, gm8a), PD muscle (cpv1a), LP muscles (cpv4, cpv6), and DG muscle (gm4b). Horizontal and vertical lines represent the median and standard deviation, respectively. Individual points correspond to single muscle fibers. Paired t tests were performed between 11°C and 21°C (gm5b, n = 68, p < 0.0001; gm6, n = 50, p < 0.0001; gm8a, n = 11, p < 0.0001; cpv1a, n = 26, p < 0.0001; cpv4, n = 4, p = 0.0057; cpv6, n = 5, p = 0.0068; gm4b, n = 3, p = 0.0391).

Figure 1C illustrates the effect of increasing the saline temperature from 11°C to 21°C on the amplitude of the evoked excitatory junctional potentials (EJPs) and the resting membrane potential in 16 examples of the more than 40 pairs of stomach muscles. Low-frequency stimulation was used to avoid muscle contraction. LP muscles, which exhibited a large first EJP amplitude, were stimulated at 2 Hz. As the temperature was increased, EJP amplitude decreased, and the membrane potential hyperpolarized approximately 10 mV in all isolated muscle fibers. When gm5b and gm6 fibers were voltage-clamped at −80mV (Figure 1D), the resultant excitatory junctional currents (EJCs) also decreased in amplitude at 21°C, while retaining their characteristic facilitation. Figure 1E shows data from more than 167 muscle fibers from seven different muscles, including those innervated by the LG, PD, DG, and LP neurons, in response to temperature increase. In every muscle, warming produced a significant hyperpolarization of the membrane potential. This occurred regardless of the innervating neuron or its neurotransmitter. The generality of this result allowed us to further quantify it and study its mechanism in a smaller subset of muscles.

We voltage-clamped fibers from three muscles, gm5b, gm6, and cpv1a (Figure S1), to quantify the effect of temperature on the input resistance. In all three muscles, the median input resistance was significantly decreased when the temperature was raised from 11°C to 21°C (gm5b, −37%; gm6, -43%; cpv1a, −27%).

We further investigated the effect of altered temperature on gm5b by examining the evolution of the EJP to a 2-s 5Hz stimulus during a temperature ramp (Figure 2A). The EJP amplitude decreased monotonically with increasing temperature and recovered when the temperature was returned to control 11°C. The resting membrane voltage (V_rest) decreased monotonically with temperature (Figure 2B). To determine the probable conductance responsible for the temperature-dependent change in the resting membrane potential, we voltage-clamped fibers at different temperatures (Figure 2C). We plotted steady-state I–V curves obtained during a temperature ramp from 5°C to 21°C and overlaid the resulting current traces for comparison. Increasing temperature enhanced current activation. Current measurements at −120 mV and −65 mV further confirmed the outward and inward currents elicited by warming (Figures 2E and 2F).

Figure 2.

Effects of temperature on V_rest, EJPs, and voltage-clamp assessment of temperature-sensitive current in gm5b

(A) Current-clamp recording of the muscle fiber from 11°C to 25°C while the lgn is stimulated at 5 Hz for 2 s (gray dashed line is V_rest).

(B) The resting membrane voltage hyperpolarizes monotonically with warming.

(C) Example current traces acquired while stepping the membrane potential (from −130 mV to −60 mV over 2 s, left) at 11°C (center) and 21°C (right).

(D) Temperature-dependence of the current from 5°C to 21°C (from black to red).

(E and F) Currents at −120 mV and −65 mV show the inward and outward currents in warm solutions. The horizontal and vertical lines represent the median and the standard deviation in (B, E, and F).

We studied the effects of altered pH on gm5b by examining EJPs (Figure 3A) and EJCs (Figure 3B). While both EJP and EJC amplitudes decreased with increasing temperature, their responses to pH were different. We observed a functional pH range in which activity remained stable (approximately pH 6.7–8.8). Deviations outside this range led to reduced amplitudes (e.g., at pH 5.5) and, in some cases, abnormal activity such as spontaneous firing (e.g., at pH 10.8). A similar pH-dependent operating range has been described in stomatogastric ganglion neurons.^5,26 The resting membrane voltage hyperpolarized monotonically with increased pH (Figure 3C). To directly assess the pH sensitivity of this channel, we voltage-clamped muscle fibers in gm5b at different pH values (Figure 3D) and overlaid the resulting steady-state current traces for comparison (Figure 3E). Increasing pH, or alkalinizing, enhanced current activation. Current measurements at −120 mV and −60 mV further further showed the outward and inward currents elicited by basic solutions (Figures 3F and 3G).

Figure 3.

Effects of pH on V_rest, EJPs, EJCs, and voltage-clamp assessment of pH-sensitive current in gm5b

(A) Current-clamp recording of the muscle fiber while the lgn is stimulated at 5 Hz for 2 s in pH solutions ranging from 5.5 to 10.8. Both acidic and basic solutions affect the EJP amplitude.

(B) gm5b is clamped in a two-electrode voltage clamp at −80mV for the same stimulus.

(C) The resting membrane voltage at various pH.

(D) Example current traces acquired while stepping the membrane potential (from −130mV to −60mV over 2 s) in acid (left, pH 5.5), control (center, pH 7.8), and basic (right, pH 9.8) solutions.

(E) Voltage assessment of the pH sensitivity of the channel in gm5b. The IV curve of gm5b in different pH conditions shows that increasing pH activates the potassium channel.

(F and G) Currents at −120 mV and −60 mV demonst the outward and inward activation by basic solutions. The horizontal represent the median.

The IV curves obtained by altering temperature or pH intersected near −85mV, suggesting the presence of a potassium current (Figures 2D and 3E). Operationally, we define the reversal potential (E_rev) of the temperature-sensitive current as the voltage at which the I–V curves at 11°C and 21°C intersect (Figure 4A). At this potential, warming does not change the net current, so the temperature-dependent component must reverse. Other approximately linear, temperature-insensitive leak currents contribute similarly at both temperatures and therefore cancel in this comparison. For the recording shown in Figure 4A, in normal saline, E_rev was −86 mV, which suggests that an increase in K⁺ conductance could be responsible. To determine if this was the case, we reexamined E_rev in low (0.5K⁺) and high (2.0K⁺) to shift the K⁺ equilibrium potential (E_K). Figures 4A and 4B shows that, as predicted, E_rev of the temperature-sensitive current hyperpolarized in 0.5K⁺, and depolarized in 2.0K⁺. Pooled data from many fibers (Figure 4B) show statistically significant shifts in the reversal potential of the temperature-induced current across 0.5K⁺, 1.0K⁺, and 2.0K⁺, consistent with the activation of a K⁺ current.

Figure 4.

Characterization of a potassium-current in gm5b, insensitive to TEA and barium (Ba)

(A) IV curve of the muscle fiber at 11°C and 21°C in saline (top), marked with the reversal potential (E_rev, diamond), which is shifted in low K⁺ (0.5K, center) and high K⁺ (2.0K, bottom).

(B) The shift in E_rev follows the Nernst equation, revealing a potassium channel. The horizontal and vertical lines represent the median and standard deviation, respectively (Wilcoxon rank-sum test; 0.5K – 1.0K, p = 0.002; 1.0K - 2.0K, p < 0.0001; 0.5K - 2.0K, p < 0.0001, with p < 0.05/3, where three is the Bonferroni correction).

(C) IV curve at 11°C and 21°C in saline (top) and TEA (10⁻² M, bottom), marked with the reversal potential (E_rev).

(D) The reversal potential is not significantly different in saline and TEA. The horizontal and vertical lines represent the mean and standard deviation, respectively (Paired t test, n = 5, p = 0.8973, p > 0.05).

(E) The difference in current (ΔI) obtained at 21°C and 11°C at −65 mV, compared between saline and TEA, shows no significant difference. The horizontal and vertical lines represent the mean and standard deviation (Paired t test, n = 5, p = 0.9365).

(F) The resting membrane voltage (V_rest) is monitored in gm5b and gm6 at 11°C, 16°C, and 21°C in different solutions: saline (white background), barium 1 mM (Ba, teal), and barium 1 mM with TEA 10-2M (Ba + TEA, lilac). Ba with or without TEA has no effect on the resting membrane voltage (gm5b, n = 3; gm6, n = 4). In each image, each dot represents a single muscle fiber. The dots and the vertical lines represent the median and the SEM.

To estimate how much of the hyperpolarization could be explained by temperature dependence of the potassium equilibrium potential alone, we used the Nernst equation with typical ionic conditions for our experiments ([K⁺]_o ≈ 11 mM, [K⁺]ᵢ ≈ 330 mM). Over a 10°C change from 11°C to 21°C, E_K is predicted to shift by only ∼2.7 mV, whereas we observe ∼10 mV hyperpolarization of V_rest per 10°C. Thus, the temperature term in the Nernst equation acting on a fixed K⁺ leak is insufficient to account for the observed effect. This indicates that there must be a contribution of a K⁺ conductance to the temperature sensitivity of V_rest.

The behavior of this temperature-sensitive and pH-sensitive K⁺ current suggests it may be due to one or more of the K2P channels, such as TASK-1, TASK-3, TREK-1, TREK-2, TRAAK^{17,18,20,27,28,29,30,31} which have been previously described. This temperature-sensitive K⁺ current in stomach muscles is insensitive to the classical K⁺ channel antagonist, tetraethylammonium (TEA).^32,33 Bath application of 10⁻² M TEA did not significantly change the reversal potential between saline and TEA conditions, nor the difference in current between the two temperatures seen at −65mV (Figures 4C–4E). Insensitivity to TEA is commonly used to exclude classical voltage-gated K⁺ (K_v) channels.

We tested barium, which has been reported to block several K2P channels.³⁴ Figure 4F shows the evolution of resting membrane potential in the gm5b and gm6 muscles at different temperatures (11°C, 16°C, and 21°C) under three conditions: saline, 1 mM Ba²⁺, and 1 mM Ba²⁺ combined with 10⁻² M TEA. In all conditions, the muscle fibers hyperpolarized with increasing temperature, with no detectable effect of barium alone or in combination with TEA. These experiments indicate that neither Ba²⁺-sensitive K⁺ channels nor TEA-sensitive Kv channels can readily account for the temperature-induced hyperpolarization under our recording conditions.

Two candidate K2P channel genes were previously identified in C. borealis by sequence homology, referred to as CbKCNK1 (Accession #KU681438) and CbKCNK2 (Accession #KU681437). Comprehensive sequence analysis indicates that these two crab channels are members of the TASK family of K2P channels, with sequence homology clustering most closely with mouse KCNK3 and KCNK9, which encode TASK-1 and TASK-3, respectively (Figure 5). Of particular significance is the conserved positioning of a histidine residue (H98) adjacent to the pore selectivity filter in both crab K2P channel sequences (Figures 5A and 5B) that has been demonstrated to be responsible for pH-sensitive in the mouse homologs KCNK3 and KCNK9. Positive expression of both CbKCNK1 and CbKCNK2 was detected in all three muscle types via RT-qPCR (Figure 5D). Quantitative RT-PCR is a fluorescence-based assay that measures the relative abundance of specific mRNA transcripts, allowing us to confirm that both K2P genes are expressed in each muscle type. There were no significant quantitative differences detected among the expression levels of either gene across muscle types.

Figure 5.

Sequence analysis and expression of K2P channels in crab muscles

(A) Maximum likelihood phylogeny of mouse (Mm) and crab (Cb) K2P channels. Branch lengths are proportional to expected replacements per site. The alignment was constructed with the MAFFT model and phylogeny was inferred with the neighbor-joining method. Multiple sequence alignments are plotted for each protein where red indicates highly conserved positions and blue indicates lower conservation. The inset shows the alignment which spans the pore 1 domain. Indicated in bold red is the H98 residue known to confer pH sensitivity in mouse TASK-1 (MmKCNK3) and TASK-3 (MmKCNK9) channels and its conserved position in both crab channels.

(B) Amino acid alignment for the mouse and crab TASK channels. The four putative transmembrane domains (M1–M4) are highlighted in gray, while the two putative pore regions (P1 and P2) are highlighted in black. The K⁺ selectivity region of the pore is noted, and the red star indicates the pH sensitive H98 residue.

(C) Schematic diagram of TASK K2P channels. M1 to M4 are transmembrane segments and P1 and P2 are the pore-forming domains.

(D) Steady-state mRNA levels of CbKCNK1 and CbKCNK2 in crab muscle. Relative mRNA levels are shown for the putative two-pore K⁺ channels CbKCNK1 and CbKCNK2 from cpv1ab, gm5b, and gm6 muscle from individual animals. Each data point represents one tissue sample from one animal. For both genes, each sample is expressed as a fold difference relative to the median Cq value of the cpv1ab group. The horizontal lines represent the mean of the fold expression, and the vertical lines span the lower and upper limits of the standard error. There were no significant differences across muscle types for either gene.

Discussion

Many marine crustaceans, such as the lobsters and crabs in New England waters, encounter substantial temperature and pH fluctuations both daily and seasonally,^1,2,3,4 and yet their motor systems remain robust within a substantial range. Most of the previous work on the effects of temperature and pH on the stomatogastric nervous system has been done on the isolated nervous system.^{5,26,35,36,37,38,39,40,41} Nonetheless, the actual stomach movements produced by the animal depend both on the motor patterns and the properties of the neuromuscular synapses and the muscles themselves. Consequently, we now ask how temperature and pH influence the neuromuscular junctions made by the stomatogastric ganglion neurons.

All of the changes in EJP and EJC amplitude are unlikely to be accounted for solely by changes in input resistance, as in some muscles, the rate of facilitation and depression was altered, thus affecting the envelope of membrane depolarization. Moreover, hyperpolarization alone would increase EJP amplitude by increasing the driving force on transmitter-activated conductances.

The effects of temperature on membrane potential and EJP amplitude of the crab stomach muscles (this paper) are consistent with earlier reports on many other crustacean muscles and neuromuscular junctions.^{10,12,13,42,43} Given the substantial differences in ion channel expression and function across muscles and species, it is notable that temperature-associated hyperpolarization of the muscle membrane potential appears to be a general property. Moreover, despite the differences in the synaptic dynamics and transmitters across different stomach muscles,^7,8,9 it is interesting that all stomach muscles display consistent hyperpolarizing responses to increases in temperature (Figure 1). Because of this shared behavior, we examined the underlying mechanism in detail only in a subset of muscles, with the working hypothesis that similar biophysical processes operate across the full muscle set. While these effects were historically attributed to a generic leak conductance or Na⁺/K⁺ pump activity,^10,13,42,43 our results point to a more specific mechanism: the activation of a potassium conductance with K2P-like properties. At the same time, the resting membrane potential is necessarily shaped by multiple conductances, including other K⁺ and non-selective leak currents, Cl⁻ conductances, and the electrogenic Na⁺/K⁺ pump. We therefore view K2P channels as dynamic contributors to temperature- and pH-dependent modulation of V_rest, rather than sole determinants of the resting potential.

Voltage-clamp evidence (Figures 2 and 4) demonstrates that the reversal potential of the conductance affected by temperature is shifted by changes in K⁺ concentration, as would be expected for a K⁺ conductance. Transcriptomic data identified the expression of two putative K2P channel types in C. borealis,⁴⁴ and Figure 5 confirms the presence of both of these transcripts in stomach muscles. When they were first identified by sequence homology, these channels were designated CbKCNK1 and CbKCNK2, although these designations do not correspond to the same numbered homologs in mammals: the C. borealis K2P coding sequences share greatest homology with mammalian KCNK9/TASK-3 and KCNK3/TASK-1 channels, but form a distinct cluster within the sequence phylogeny where there is not a clear one-to-one crab homolog to a given mammalian channel (Figure 5).

Within the mammalian K2P channel family, TREK-1, TREK-2, and TRAAK (encoded by KCNK2, KCNK10, and KCNK4, respectively) are well-known polymodal sensors, activated by temperature, mechanical stretch, arachidonic acid, and volatile anesthetics. Our specific search of both the C. borealis genome²⁴ and neural transcriptome⁴⁴ to identify candidate homologs for KCNK2, KCNK10, and KCNK4 failed to detect such sequences in the crab. By contrast, TASK-1, TASK-3, and TWIK-1 (encoded by KCNK3, KCNK9, and KCNK1) are particularly sensitive to extracellular pH, and the identified crab KCNK channel sequences (CbKCNK1 and CbKCNK2) show clear homology to TASK-1 and TASK-3. Interestingly, both TREK-1 and TREK-2 display strong activation by temperature, while other K2P family members show more modest thermal sensitivity. An increase in pH activates TASK-1, TASK-3, TWIK-1, and TREK-1, whereas TREK-2 is instead activated by acidification.^17,31,45,46 The two crab channels identified share key sequence similarity to the TASK-1 and TASK-3 channels that confer sensitivity to pH, specifically the presence of a key histidine residue (H98) adjacent to the pore selectivity filter in pore 1.^47,48 These sequence similarities suggest that the proteins encoded by both CbKCNCK1 and CbKCNK2 likely are pH-sensitive. However, the source of temperature sensitivity of these currents in crab muscle remains unknown.

The crab genome appears to contain far fewer K2P channels than mammals, and we do not know whether the two crab sequences identified encode channels that have both temperature and pH sensitivity, whether the two functions may be separately conferred by the two genes, or some as of yet identified crab protein may be responsible for the thermal sensitivity we measured. A previously identified C-terminal motif that confers temperature sensitivity in TREK K2P channels⁴⁵ was not found in any protein of the crab genome or neural transcriptome through a tblastn search, thus this motif does not appear to be responsible for the temperature sensitivity we detected even in some as of yet identified channel.

Taken together, these observations raise the possibility that in C. borealis, at least two channel subtypes contribute to these muscle responses: the pH sensitive TASK channels (CbKCNK1 and CbKCNK2), and unidentified temperature-sensitive channel(s). We predict either a single temperature-sensitive channel, or multiple such channel types may share similar temperature activation profiles. Conversely, the two putative crab TASK channels may diverge in their pH sensitivity, providing a plausible explanation for the monotonic thermal response yet non-monotonic and process-dependent pH effects we observed.

Functionally, these channels provide a plausible mechanism for maintaining robust neuromuscular performance during environmental change. The pyloric and gastric rhythms accelerate 2- to 3-fold as temperature rises from 11°C to 21°C.^35,38,49 As the phase relationships and coupling across networks are preserved,³⁸ we see an increase in the spike frequency within the burst – acting as a direct input to the neuromuscular junction. Without compensation, a depolarized muscle receiving inputs at twice the spike frequency would contract more strongly or continuously, risking an abnormal movement pattern. By hyperpolarizing fibers as temperature increases, K2Ps reduce excitability and EJP amplitude, partially offsetting the enhanced synaptic drive. Thus, muscle fibers themselves could contribute to temperature compensation, complementing circuit-level robustness described previously in the STG.^35,36,50

In conclusion, we provide evidence that temperature- and pH-sensitive K2P channels contribute to the modulation of muscle excitability under varying environmental conditions. These findings revise our understanding of how crustacean muscles adapt physiologically to temperature and pH fluctuations. We propose that these channels play a critical role in maintaining the appropriate amplitude and timing of muscle contractions as motor pattern frequency changes with temperature. More broadly, this work opens new avenues for exploring temperature compensation at the neuromuscular junction and how multiple environmental variables may affect motor function in ectothermic animals.

Limitations of the study

Our study is limited by the lack of genetic tools in C. borealis. As a result, we cannot perform knock-down or knock-out experiments to causally link specific K2P genes to the temperature-sensitive K⁺ conductance. The K2P channels are known to be modulated by a wide range of general anesthetics and bioactive molecules, including arachidonic acid, bupivacaine, and volatile compounds.^19,20,32 Pharmacological modulators of K2P channels are not selective. As a result, pharmacological manipulations do not allow unambiguous identification of the channels underlying the temperature-sensitive current. Our measurements focused on a subset of identified stomach muscles, and we infered that similar mechanisms operate across the broader musculature based on shared phenomenology rather than direct recordings from every muscle pair.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Kathleen Jacquerie (kjacquerie@brandeis.edu).

Materials availability

This study did not generate any new unique reagents.

Data and code availability

• •Electrophysiology data and metadata are publicly available at Zenodo (https://doi.org/10.5281/zenodo.17279833). The two-pore K⁺ channel gene sequences referenced in this study are available in GenBank. Their accession codes are CbKCNK1 (KU681438) and CbKCNK2 (KU681437).
• •This study does not report original code. Custom MATLAB codes extracting the resting membrane voltage, the IV curves steady-state data, or the plotting functions are available on demand.
• •Any additional information required to reanalyze the data reported in this study is available from the lead contact upon request.

Acknowledgments

This work was supported by the United States National Institute of Health R35 NS 097343 (E.M.), Belgian American Education Foundation – BAEF, Wallonie-Belgium International – WBI Postdoctoral Fellowship (K.J.).

We thank Sonal Kedia and Kyra Schapiro for their support and valuable feedback on this project, Margaret Lee and Maria Ivanova for training assistance, Jackie Seddon for help with the installation of the muscle recording rig, and Gwen Harris for management of the animals and laboratory facility.

Author contributions

Conceptualization, K.J. and E.M.; methodology, K.J., D.S., and E.M.; investigation, K.J., A.P., and D.S.; writing – original draft, K.J. and E.M.; writing – review and editing, K.J., A.P., D.S., and E.M; funding acquisition K.J. and E.M; resources, E.M. and D.S.; supervision, E.M.

Declaration of interests

The authors declare no conflict of interest.

Declaration of generative AI and AI-assisted technologies in the writing process

During the preparation of this work, the authors used ChatGPT (OpenAI) to check spelling. The authors reviewed and edited the content and take full responsibility for the publication.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Chemicals, peptides, and recombinant proteins
Tetraethylammonium chloride (TEA-Cl)	Sigma-Aldrich	T2265
Deposited data
Electrophysiology data	Zenodo	https://doi.org/10.5281/zenodo.17279833
Cancer borealis sequencing	National Center for Biotechnology Information BioProject archive	https://www.ncbi.nlm.nih.gov/bioproject/PRJNA524309
Cancer borealis CbKCNK1	NCBI GenBank	GenBank: KU681438
Cancer borealis CbKCNK2	NCBI GenBank	GenBank: KU681437
Experimental models: Organisms/strains
Cancer borealis (Jonah crab), Adult male	Commercial Lobster Co., Boston, MA, USA	NCBI: txid 39395
Software and algorithms
MATLAB 2019b	MathWorks	https://www.mathworks.com/products/matlab.html - SCR_001622
pClamp/AxoClamp	Molecular Devices	SCR_011323
Adobe Illustrator 2024	Adobe	https://www.adobe.com/products/illustrator.html

Experimental model and study participant details

Adult male Jonah Crabs (Cancer borealis) were purchased from Commercial Lobster Company (Boston, MA) and held in filtered, circulating, aerated, artificial seawater (Instant Ocean Sea Salt Mix) at 11°C. Crabs were acquired between May 2024 and June 2025. Animals were typically held in the laboratory for one week before use.

Method details

Experimental preparation

Stomach musculature and the stomatogastric nervous system (STN) were dissected following the description and nomenclature of.⁶ Nerve dissection followed the procedures described by.⁵¹ Specific neuromuscular preparations were isolated from the foregut. Before starting experiments, the fat layer and connective tissue overlaying the muscles were carefully removed to improve access for recording from muscle fibers while keeping nerve terminals attached.^52,53

Reduced preparations of parts of the stomach and nerves were pinned into a silicone elastomer-lined (Sylgard 184: Dow Corning) dish. The preparation was maintained in C. borealis saline solution (440 mM NaCl, 11 mM KCl, 26 mM MgCl₂, 13 mM CaCl₂, 11 mM Trizma base, 5.4 mM maleic acid, pH 7.8 at 11°C) throughout the dissection and experiment. Saline was continuously superfused at 8–11 mL/min via a peristaltic pump (model Ismatec Ecoline). Temperature was adjusted from 5°C to 21°C using a Peltier system (Warner Instruments SC-20) with a temperature controller (CL-100, Warner Instruments) and a thermocouple probe in the bath near the recording site. Temperatures within ±0.3°C of the target value (e.g., 10.7°C–11.3°C) were considered as 11°C for analysis.

High K⁺ saline (2.0K⁺, 22 mM KCl) and low K⁺ saline (0.5K⁺, 5.5 mM) were prepared by adjusting KCl salt to the normal saline. TEA-containing saline was prepared by adding 10⁻² M of tetraethylammonium chloride (Sigma Aldrich). Barium-containing saline was prepared by adding 1 mM of BaCl₂ (Sigma Aldrich). Additional quantities of Trizma base or maleic acid were added to achieve solutions with pH 5.5, 6.5, 7.5, 8.5, 9.5 at room temperature and at 11°C pH 5.5, 6.7, 7.8, 8.8, 9.8, and 10.8 according to.⁵ For a more basic solution, NaOH was used to reach 10.5 at room temperature. Solutions were calibrated using a Mettler Tolledo pH meter. Values within ±0.2 pH units of the target pH (e.g., 6.4–6.6) were considered as pH 6.5 for analysis.

Electrophysiology

Electrophysiology experiments were performed as previously described.^52,53,54 Muscles were pinned to Sylgard dishes. Nerves were stimulated using suction electrodes attached to the cut nerve end, connected to an AM Systems isolated pulse stimulator (Model 2100). Intracellular recordings of excitatory junctional potentials (EJPs) in muscle fibers were obtained using sharp borosilicate glass microelectrodes (Sutter Instrument; ID 0.86 mm, OD 1.5 mm), pulled with a micropipette puller (Sutter Instrument, P-97) and backfilled with 400 mM K₂S0₄, 20 mM KCl, 33 mM NaCl. Electrode resistances were between 5 and 20 MΩ.

Electrodes were mounted on Leica Leitz mechanical micromanipulators with HS-2A-x1LU Axoclamp 2B and HS-9A-x1U Axoclamp 900A headstages (Molecular Devices). Signals were amplified using Axoclamp 2B or 900A amplifiers (Molecular Devices), digitized at 10 kHz using a Digidata 1440, and recorded in Clampex 10.7. The electrodes can have an offset between 0 and 3 mV for a 10°C increase from 11°C to 21°C. Pulse trains (2 s duration) were delivered at 2 or 5 Hz, frequencies chosen to avoid eliciting muscle contraction. Because LP muscles contracted at 5 Hz, they were stimulated at 2 Hz. Individual pulse durations ranged from 0.2 to 0.9 ms. Pulse amplitudes were calibrated to reliably trigger single EJPs at 11°C while keeping a 10–20% safety margin. Muscle fibers depolarized beyond −50 mV were excluded from recordings.⁵⁵

Excitatory junctional currents (EJCs) were recorded in two-electrode voltage clamp mode at −80 mV. Two electrodes were inserted in the same fiber along the longitudinal axis. The distance between the two electrodes was less than the diameter of the fiber. Current-voltage (I-V) curves were obtained by stepping the voltage from −130 to −60 mV (starting from −80 mV) for 2 s. Saline levels were minimized to reduce capacitive coupling. The headstage for recording currents was a Molecular Devices HS-2A-x10MU Axoclamp 2B, connected via an adapter to the Axoclamp 900A. Because the muscle fibers are large and electrically coupled, producing substantial currents, voltage steps were accepted only if they reached the intended values and the recorded currents remained stable.

Bioinformatics and phylogenetic analysis

Mouse K2P potassium channel protein sequences were first identified through searches of the NCBI and Ensembl databases. These sequences were used to search both the C. borealis transcriptome and draft genome assemblies for potential homologs via the tblastn BLAST algorithm.⁵⁶ Predicted protein sequences from top hits were then aligned using the NCBI Constraint-based Multiple Alignment Tool (COBALT) with default settings. These Multiple Sequence Alignments were plotted and used for sequence comparison to identify potential thermo- and pH sensitive motifs among mouse and crab protein sequences. For the phylogenetic analysis, we performed a MAFFT v7 multiple sequence alignment (MSA) to create a matrix of related sequences⁵⁷ and then used the neighbor-joining algorithm to construct a phylogenetic tree.⁵⁸

Reverse transcription and real-time quantitative polymerase chain reactions (qRT-PCR)

After electrophysiology experiments, whole muscles from the left and right sides of each animal were collected and stored at −80°C. The cpv1a and cpv1b muscles were pooled and analyzed together (hereafter, “cpv1ab”). Tissues were homogenized in 750 μL TRIzol (Invitrogen) for RNA extraction. Total RNA was isolated according to the manufacturer’s protocol (Invitrogen). 100 ng of total RNA from each sample was reverse transcribed using qScript reverse transcriptase (QuantaBio) primed with a mixture of oligo-dT and random hexamers. The final volume of the reverse transcription reaction was 20 μL and contained a final concentration of 2.5 ng/μL random hexamers, 2.5 μM oligo-dT, 40 U of RNase inhibitor, and 200 U of reverse transcriptase. Following heat inactivation of the enzyme, samples were diluted in ultrapure water to a final volume of 50 μL before this template was used in qRT-PCR analyses.

cDNA was used in qRT-PCR reactions for KCNK1 and KCNK2 genes with primer sets previously validated in this system.⁵⁹ qRT-PCR reactions consisted of primer pairs at a final concentration of 2.5 μM, cDNA template (equivalent to 3.33 ng of total cellular RNA), and SsoAdvanced SYBR mastermix (BioRad) according to the manufacturer’s instructions. Reactions were carried out on a CFXConnect (BioRad) machine with a three-step cycle of 95°C-15s, 58°C-20s, 72°C-20s, followed by a melt curve ramp from 65°C to 95°C. Data were acquired during the 72°C step and every 0.5°C of the melt curve. All reactions were run in triplicates of 10 μL, and the average Cq (quantification cycle) was used for analysis. Reactions were normalized to a fixed amount of total cellular RNA^60,61 and expression level for each individual sample was then calculated as a fold-expression level relative to the median Cq of the cpv1ab muscle group as follows: ${FoldExpression}_x=2^{-\left({Cq}_x-{Cq}_{medianCtrl}\right)}$.

Quantification and statistical analysis

We developed custom MATLAB scripts to visualize and analyze recordings obtained using Clampex software. These scripts were used to extract the resting membrane potential, the IV curve, the input resistance (R_in), and the reversal potential (E_rev).

Resting membrane potential values were obtained by averaging over 0.5 s prior to the onset of the pulse train. Input resistance (R_in) was calculated from current–voltage (IV) relationships obtained during voltage-clamp recordings. For each temperature condition, the steady-state current associated with each voltage step was averaged over the final 0.17 s of the 2-s step. R_in was then estimated by performing a linear regression on the IV data within a restricted voltage range (−92 mV to −73 mV), corresponding to the near-linear region of the IV curve. Only voltage steps falling within this range were included in the fit. The slope of the resulting linear fit (ΔI/ΔV) was used to compute input resistance as R_in = 1/slope. Goodness of fit was assessed using the coefficient of determination (R²). Raw traces were plotted and smoothed using 8–30 points (MATLAB smooth function), while voltage-clamp current traces were smoothed using ∼10 points. Reversal potential was determined by calculating the intersection of the IV curves at 11°C and 21°C.

For statistical analyses, data are reported as either the median or the mean ± standard deviation (SD), as indicated in the figure legends. Paired t-tests were used to compare the same preparation across different conditions. Bonferroni corrections were applied to adjust for multiple comparisons using the Wilcoxon rank-sum test. Significant differences are indicated as follows: ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001. Detailed values are also summarized in Table S1.

Published: March 5, 2026