Sodium channel NaV1.3 is important for enterochromaffin cell excitability and serotonin release

In the gastrointestinal (GI) epithelium, enterochromaffin (EC) cells are enteroendocrine cells responsible for producing >90% of the body’s serotonin (5-hydroxytryptamine, 5-HT). However, the molecular mechanisms of EC cell function are poorly understood. Here, we found that EC cells in mouse primary cultures fired spontaneous bursts of action potentials. We examined the repertoire of voltage-gated sodium channels (NaV) in fluorescence-sorted mouse EC cells and found that Scn3a was highly expressed. Scn3a-encoded NaV1.3 was specifically and densely expressed at the basal side of both human and mouse EC cells. Using electrophysiology, we found that EC cells expressed robust NaV1.3 currents, as determined by their biophysical and pharmacologic properties. NaV1.3 was not only critical for generating action potentials in EC cells, but it was also important for regulating 5-HT release by these cells. Therefore, EC cells use Scn3a-encoded voltage-gated sodium channel NaV1.3 for electrical excitability and 5-HT release. NaV1.3-dependent electrical excitability and its contribution to 5-HT release is a novel mechanism of EC cell function.

Human and mouse colon and small bowel EC cells express voltage-gated sodium channels (Na V ). We used immunofluorescence to determine whether Na V 1.3 protein is present in EC cells of human and mouse colon and small bowel (Fig. 3A). We found that Na V 1.3 is not only present in both mouse and human, but it appears to be localized highly asymmetrically -almost exclusively at the basal side (Fig. 3A). In the mouse and human GI epithelium, we found that Na V 1.3 was present in most EC cells (mouse Tph1-CFP+ and human 5-HT+ cells) in both small bowel and colon (Fig. 3B). We quantified the frequency of CFP+/Na V 1.3+ cells and found co-localization in 89.4 ± 2.0% of small bowel EC cells (N = 3 animals, n = 71 ± 5 cells/animal) and 88.4 ± 4.4% of colon EC cells (N = 3 animals, n = 73 ± 5 cells/animal) (Fig. 3B). Similarly, in the human GI epithelium, we found that Na V 1.3 and 5-HT co-localized in 89.8 ± 1.1% of small bowel EC cells (N = 3 patients, n = 70 ± 3 cells/patient) and 92.8 ± 2.0% of colon EC cells (N = 3 patients, n = 68 ± 5 cells/patient) (Fig. 3B). Altogether, our data from the human and mouse small bowel and colon show that ~90% of EC cells express the voltage-gated sodium channel Na V 1.3.

Primary colon and small bowel EC cells express Scn3a and have robust Na V 1.3 currents.
To directly confirm Scn3a expression in EC cells, we used single cell RT-qPCR in Tph1-CFP mouse small bowel and colon primary cultures. We found that Scn3a and Tph1 mRNA were present in CFP+ EC cells but not CFP-cells or bath medium from both mouse small bowel (N = 3) and colon (N = 3) primary cell cultures (Fig. 4A, full size gel in Supplementary Figure 1).
To evaluate the functional role of Na V 1.3 in EC cell excitability, we performed whole cell voltage-clamp experiments on CFP+ EC cells from mouse small bowel and colon and compared their electrophysiologic properties (Table 1). We found fast voltage-dependent inward currents in 128 of 154 small bowel EC cells (81.3 ± 4.0%, N = 44 cultures) and 18 of 29 colon EC cells (64.1 ± 9.2%, N = 29 cultures) (Fig. 4B). CFP-cells in the same preparations did not have voltage-dependent inward currents (Supplementary Figure 2). The peak currents were −63.4 ± 5.6 pA/pF for small bowel and −68.0 ± 11.5 pA/pF for colon (n = 114 and 18, p > 0.05 by a nonparametric two-tailed t-test). Voltage-dependent channel activation was well fit by a Boltzmann function with the following parameters: half-activation voltage (V 1/2A ) −23.4 ± 0.9 mV and slope (δV A ) 6.7 ± 0.2 mV for small bowel, V 1/2A −25.3 ± 2.2 mV and δV A 6.5 ± 0.3 mV for colon EC cells (Fig. 4B inset, Table 1). The time constant of activation (τ A ) was 0.26 ± 0.02 ms for small bowel and 0.23 ± 0.03 ms for colon EC cells. Inactivation was best fit by a two-exponential function, with τ F 1.00 ± 0.04 ms and τ S 11.4 ± 0.6 ms for small bowel, τ F 0.68 ± 0.04 ms and τ S 11.2 ± 1.1 ms for colon EC cells (p < 0.05, τ F of colon vs. small bowel EC cells). The voltage-dependence of inactivation, or steady-state availability, was also well-fit by a two-state Boltzmann function. The half-inactivation voltage (V 1/2I ) and slope (δV I ) were −48.2 ± 1.1 mV and 7.1 ± 0.1 mV for small bowel, −53.3 ± 3.8 mV and 8.9 ± 0.9 mV for colon EC cells (p > 0.05) (Fig. 4B, inset). Overall, we found robust voltage-gated fast inward currents in colon and small bowel EC cells. These currents had similar biophysical properties in both small bowel and colon EC cells except that colonic EC cells had smaller capacitances and had faster fast inactivation time constants (Table 1, Supplementary Figure 3).
Although RNA expression experiments suggest Na V 1.3 as the primary sodium channel, the fast-inactivated, voltage-dependent inward currents that we observed could be carried, in principle, by either a Na + or Ca 2+ channel. Thus, we used ion substitution to determine which type of voltage-gated channel carries the current (Na + or Ca 2+ ), and pharmacologic blockade to identify the specific subtype (Nav1.3 or potentially others). The fast voltage-dependent inward currents in small bowel EC cells were eliminated by substituting Na + with N-methyl D-glucamine (NMDG + ) (  (Fig. 4C). Current-voltage (I-V) relationships showed decreases in peak currents for 100 mM Na + (Fig. 4C & D) without a change to voltage-dependent parameters when fit to a two-state Boltzmann function (data not shown). Full Na + replacement by 150 mM NMDG + eliminated the fast-inward currents (Fig. 4C &  D). We next used the selective Na V 1.3 inhibitor ICA-121431 23 (Fig. 4E). At the + 20 mV test pulse, ICA-121431 dose-dependently inhibited the fast-inward currents with an IC 50 of 204 ± 54 nM and block of 86.3 ± 0.1% at 10 µM, the highest concentration tested here (n = 4) (Fig. 4E & F). Together, the ionic substitution and pharmacological blockade experiments establish that the fast voltage-dependent inward current in EC cell is carried by sodium and specifically by the Scn3a-encoded voltage-gated sodium channel Na V 1.3. If other channels contribute as well, their currents were not measurable. Na V 1.3 channels are responsible for elicited action potentials (APs) in EC cells. If Na V 1.3 is the sole source of voltage-dependent inward current in EC cells, it should then be the logical candidate for AP generation. We examined whether Na V 1.3 channels were involved in EC cell APs. We current-clamped the EC cells from both small bowel and colon and used short pulses of current injection to elicit electrical activity. We found that for EC cells (CFP+) from both small bowel (Fig. 5A) and colon (Fig. 5B), a current stimulus of 20-50 ms resulted in action potentials in 48 of 57 cells (84.2%) in the small bowel and 5 of 7 cells (71.4%) in the colon. In the EC cells that lacked fast inward current from both colon (n = 5) and small bowel (n = 4), we were unable to elicit APs. For EC cells in which APs could be elicited, the threshold potential was not different between small intestine and colonic EC cells (ECJ, −47.8 ± 1.4 mV, n = 48; ECC, −57.7 ± 7.0 mV, n = 5; P > 0.05 t-test for unequal variance). The action potential peaks were +55.3 ± 3.1 mV and +39.1 ± 8.2 mV for small bowel and colon, respectively. To determine whether Na V 1.3 was responsible for the EC cell action potentials, we used Na + substitution (Fig. 5C) and Na V 1.3 blockade (Fig. 5D). Both approaches resulted in action potential elimination ( Fig. 5C & D).

EC cells exhibit bursts of spontaneous electrical activity.
Having found spontaneous electrical activity in EC cells ( Fig. 1) and having established the role of Na V 1.3 in the generation of action potentials, we examined whether Na V 1.3 was involved in EC cell spontaneous electrical activity. Current-clamped EC cells fired spontaneous but irregular bursts of action potentials from plateau potentials (Fig. 6A), often for minutes at a time during a single experiment (Supplementary Figure 4). The EC membrane potential oscillated between a resting potential of −72 ± 4 mV and a plateau potential of −56 ± 4 mV (Fig. 6B, Supplementary Figure 5). The action potential amplitudes showed a decay along the distribution of their baselines (Fig. 6C, black). Overall, EC cells had hyperpolarized resting potentials, discrete plateau potentials, and from plateau potentials fired bursts of action potentials that had properties consistent with the voltage-dependent function of Na V 1.3 channels examined in previous sections (Fig. 6D).

Na V 1.3 contributes to the regulation of 5-HT release by EC cells.
To determine whether Na V 1.3 contributes to 5-HT release, we used ELISA to measure 5-HT release from primary colon cultures (Fig. 7). Using KCl (50 mM) as a cell depolarization agent, we found that cell depolarization resulted in a large release of 5-HT from colonic EC cells (from t 0 of 0.24 ± 0.25 ng/mL to t 20,KCl of 1.87 ± 0.95 ng/mL, for Δ5-HT KCl of 1.54 ± 0.30 ng/mL, n = 7). This response was significantly decreased by the addition of ICA-121431 to block Na V 1.3 (from t 0 of 0.40 ± 0.36 ng/mL to t 20,KCl,ICA of 1.14 ± 0.54 ng/mL for Δ5-HT KCl,ICA of 0.74 ± 0.11 ng/mL, n = 5, p < 0.05 between Δ5-HT KCl and Δ5-HT KCl,ICA by Student's t-test). BDS-1, a known Na V 1.3 agonist 24 , also induced 5-HT release from primary cultures (from t 0 of 0.34 ± 0.33 ng/mL to t 20,BDS of 0.91 ± 0.38 ng/mL, for a Δ5-HT BDS of 0.57 ± 0.08 ng/mL, n = 5). Addition of ICA-121431 significantly inhibited the response to BDS-1 (from t 0 of 0.35 ± 0.30 ng/mL to t 20,BDS of 0.66 ± 0.20 ng/mL, for a Δ5-HT BDS,ICA of 0.30 ± 0.07 ng/mL, n = 5, p < 0.05 between Δ5-HT BDS and Δ5-HT BDS,ICA by Student's t-test), suggesting that Na V 1.3 contributes to 5-HT release by EC cells. We next examined whether NaV1.3 is involved in EC cell response to luminal stimulants, such as the short chain fatty acid (SCFA) butyrate. We found that while butyrate resulted in 5-HT release (from t 0 of 0.36 ± 0.12 ng/mL to t 20,butyrate of 1.21 ± 0.13 ng/mL, for Δ5-HT butyrate of 0.86 ± 0.21 ng/mL, n = 3), this response did not appear to rely on Na V 1.3 channels since ICA failed to inhibit 5-HT release (from t 0 of 0.51 ± 0.13 ng/ mL to t 20,butyrate,ICA of 1.32 ± 0.60 ng/mL, for a Δ5-HT butyrate,ICA of 0.81 ± 0.58 ng/mL, n = 3, p > 0.05 between
Δ5-HT butyrate and Δ5-HT butyrate,ICA by Student's t-test). Our data show that while Na V 1.3 channels contribute to 5-HT release by colonic EC cells, the mechanism upstream of Na V 1.3 is currently undetermined.

Discussion
The enterochromaffin (EC) cell is the single most important source of systemic serotonin (5-HT) 1 , and EC cell 5-HT plays critical physiologic roles within the GI tract and systemically 2 . However, the molecular mechanisms of EC cell function and serotonin release are poorly understood because only a few studies have examined EC cells in isolation from the rest of the GI epithelium 13,25,26 . In isolated guinea pig EC cells, there was an inward current consistent with voltage-gated calcium channels but no fast sodium current 15 , while in a different study TTX-sensitive sodium current was present in murine EC cells 13 . It is unclear whether species differences are responsible for these results.
In this study, we examined purified EC cells as a group and as single cells. We discovered that EC cells are not only electrically excitable, exhibiting spontaneous bursting electrical activity, but that their electrical excitability depends on a specific voltage-gated sodium-selective ion channel, Scn3a-encoded Na V 1.3, and that Na V 1.3 contributes to 5-HT release.
We found that Scn3a mRNA is a single highly expressed voltage-gated sodium channel in dissociated and FACS-sorted small bowel Tph1-CFP cells, and it was expressed in single Tph1-CFP EC cells from both small bowel and colon primary cultures. Our data also show that the Na V 1.3 protein is present in ~90% of small bowel and colon EC cells in both human and mouse. Previous studies that examined gene expression in the GI epithelium suggested that Scn3a is expressed in enteroendocrine cells. Scn3a is expressed in intestinal neurogenin 3    (Ngn3) 27 positive cells, which are epithelial cells of a secretory phenotype 28 , and chromogranin A (CgA) 13 positive cells, which are enteroendocrine cells that include EE and EC cells 29 . More specifically, in small bowel enteroids Tph1+ EC cell single cell expression profiles showed that Scn3a was one of the most abundantly expressed ion channels 30 . The L-cell, a different type of enteroendocrine cell that produces glucagon-like peptides (GLP) and peptide YY (PYY), also expresses Scn3a 18 . In contrast, Scn3a was not found in the enteroendocrine K cells that produce and secrete glucose-dependent insulinotropic polypeptide (GIP) 27 . Overall, our results align with a number of studies that showed Scn3a-encoded Na V 1.3 is densely expressed in multiple, but not all, types of enteroendocrine cells. Intriguingly, Scn3a was previously found in endocrine and neuroendocrine cells outside the enteroendocrine system, such as neuroendocrine adrenal chromaffin cells 17 and pancreatic αand β-cells 16 . In addition to Scn3a (Na V 1.3), these endocrine cells express other Na V isoforms: Na V 1.7 for mouse αand β-cells 16,31 , Na V 1.6 and Na V 1.7 for human β-cells 32 , and Na V 1.9 for L-cells 18 . In EC cells, in addition to the highly expressed Na V 1.3, we found only one other Na V isoform, Na V 1.6, but at much smaller expression levels. With regard to the EC cell, it is unclear if the Scn8a-encoded Na V 1.6 contributes to electrophysiology. However, our data suggest that Na V 1.3 is the most functionally significant voltage-gated sodium channel isoform, since Na V 1.3 blockade nearly eliminated the EC cell voltage-dependent inward current and abolished action potentials.
We established and used EC cell primary cultures to determine whether EC cells from both small bowel and colon have voltage-gated inward currents. We found that as with immunohistochemistry, voltage-dependent inward currents were present in most but not all EC cells. In both small bowel and colon EC cells, Na V currents were inhibited by Na + substitution or by the Na V 1.3/Na V 1.1-selective blocker ICA-121431 23 . Since EC cells showed no Na V 1.1 expression, we concluded that the fast voltage-dependent inward current is carried by Na V 1.3. The biophysical properties of the inward currents were nearly identical between small bowel and colon EC cells and were similar to other endocrine cells that use Na V 1.3 for excitability, namely adrenal chromaffin cells 17 , αand β-cells 16 , and enteroendocrine L-cells 18 .
We were intrigued when we found that EC cells fire bursts of spontaneous action potentials. When we pursued further, we found that elicited EC cell action potentials rely on Na V 1.3, since Na + substitution and ICA-121431 eliminated EC cell excitability. Primary EC cells had a dynamic cell membrane resting potential that fluctuated between two dominant potentials. The likelihood of firing an action potential from the EC cell resting membrane potential (−72 ± 4 mV) was exceedingly low, but when the EC cell membrane potential reached a plateau (−56 ± 4 mV), a burst of action potentials quickly resulted. The bursts were self-terminating and action-potential amplitudes matched well with the voltage-dependent properties of EC cell Na V 1.3 channels. We currently do not know the mechanism downstream of Na V 1.3. There exist several possibilities. First, bursting could allow the EC cells to maintain a depolarized membrane potential that activates calcium channels that are critical for 5-HT release by EC cells [13][14][15] . While previous studies have agreed that calcium channel activation is important for 5-HT release, future studies will need to determine whether and how Na V 1.3 activation is coupled to calcium channel activation. Second, there is a possibility of Ca 2+ -independent exocytosis that relies on cytoplasmic Na + in the process of hormone secretion 33 . Third, Na V 1.3 may be involved in cellular signaling between EC cells and extrinsic afferents or intrinsic primary afferent neurons (IPANs) 34 . Further work is required to elucidate the molecular determinants of EC cell resting and plateau potentials, the relationships between the membrane potential and intracellular calcium dynamics, and the contribution to cell-cell communication.
Given the electrical excitability of EC cells, we were interested to find out whether Na V 1.3-mediated EC cell electrical excitability plays a role in EC cell 5-HT release. In colonic EC cells, we found that primary EC cell depolarization by KCl increased 5-HT release and KCl-induced 5-HT release was inhibited by Na V 1.3 blocker ICA-121431. We then used an established Na V 1.3 agonist -a sea anemone peptide toxin blood depressing substance-1 (BDS-1) 24 . BDS-1 stimulated 5-HT release from EC cells, and similar to KCl, the addition of Na V 1.3 inhibitor ICA-121431 inhibited 5-HT release, suggesting the involvement of Na V 1.3 channels in 5-HT release. However, 5-HT release by colonic EC cells in response to SCFA butyrate did not involve Na V 1.3, which is consistent with the notion that SCFA chemosensation requires G-protein coupled receptors 13,35 . Our data not only fit with the existing body of literature, where GI tissues were used to examine 5-HT release, but also provide intriguing novel possibilities. Depolarization of porcine mucosa by KCl produced a steady outflow of 5-HT, and submucosally-applied tetrodotoxin (TTX) reduced 5-HT release 36 . Interestingly, TTX applied to the luminal side did not inhibit 5-HT  37 . Similar to TTX, the Na V channel activator veratridine applied to the luminal side did not affect 5-HT release 36 . Therefore, manipulation of Na V channels from the luminal side by cell-impermeable substances does not inhibit EC cell 5-HT release. On the other hand, EC cell 5-HT release was effectively blocked by intraluminal application of lidocaine, a well-known lipid permeable blocker of sodium channels 37 . In our study, we found that Na V 1.3 was always localized to the basal side and in some instances within EC cell basal extensions, recently termed "neuropods" 38 . We suspect that the divergent data on EC cell voltage-gated sodium channels affecting 5-HT release may be due to the localization of Na V 1.3 channels at the EC cell basal surface, where hydrophilic drugs do not have access. Consistent with the idea of EC cell functional polarization, recent work shows that EC cells respond differently to luminal versus systemic glucose exposure 25 . Therefore, through localization of Scn3a to the basal side of EC cells, the amplification machinery of these cells is protected from luminal exposure, where there is a rich variety of potential chemical stimulants.
EC cell electrical excitability transforms the EC cell from a sensory receptacle, driven by receptor currents to activate 5-HT exocytosis, to a cell that can participate in complex bidirectional communication with the enteric and extrinsic nervous systems. In this respect, the EC cell joins the taste cells, which are also electrically excitable. In fact, Scn3a was also found specifically expressed at the basal side of sweet, bitter, and umami taste cells, where it is proposed to use electrical excitability to amplify currents generated by TRPM5 in response to tastant stimulation 39 , which is critical for the taste cells' communication with afferent neurons. Interestingly, the EC cells also express taste receptors 13,40,41 and multimodal chemosensor TRPA1 13,42 . In addition, recent evidence suggests functional connections between enteroendocrine cells with afferent neurons 13,34 . Therefore, stimulation of bursting electrical activity in EC cells may allow for a direct communication with afferent neurons. A complementary possibility is that EC cell Na V 1.3 is required for execution of efferent neuronal control of EC cells, as vagal stimulation was previously shown to affect 5-HT release from EC cells 13,43 . Such communication between an efferent neuron and a neuroendocrine cell is exemplified by splanchnic nerve activation of chromaffin cells in the adrenal medulla, where sodium channels are critical for defining the bursting properties of chromaffin cells 17 .
Our findings that EC cell electrical excitability and bursting behavior contribute to 5-HT release not only provide a novel mechanism of EC cell function, but may also have novel therapeutic possibilities. Future studies need to examine Na V 1.3 function in human EC cells, because if Na V 1.3 contributes to human EC cell function, novel therapeutic possibilities would exist. For example, in diabetes, studies have established not only that manipulation of islet cell electrical excitability affects hormone secretion 32 but that blockade of pancreatic islet sodium channels may be effective in combating diabetes 44 . An alternative therapeutic possibility is Na V 1.3 blockade in carcinoid syndrome, as human carcinoid cells were found to highly express SCN3A 45 . Since EC cell 5-HT is critical to many GI and systemic functions, blockade of EC cell Na V 1.3 may provide novel approaches for targeting GI and systemic pathophysiologic conditions.
In summary, we show that EC cells in human and mouse densely express the Scn3a-encoded voltage-gated sodium channel Na V 1.3, which is responsible for the electrical excitability and contributes to 5-HT release. The discovery of Na V 1.3-dependent EC cell excitability may have significant potential implications for EC cell physiology, pathophysiology, and therapy.

Methods
All experimental procedures were approved by the Institutional Review Board (IRB) and the Institutional Animal Primary EC cell culture 5-HT release. Mouse colon primary EC cells cultures were grown on 12 well dishes and 5-HT release was measured in response to stimuli using enzyme linked immunoassay (ELISA, BA E-5900, Rocky Mountain Diagnostics, Colorado Springs, CO). After 24 hours of incubation, media was removed, NaCl Ringer's solution was added and a sample was collected from each well for a baseline 5-HT reading. was added to each sample. Samples were spun for 5 minutes at 5000 rpm and stored at −80 °C. 5-HT release was measured by a serotonin enzyme immunoassay (BA E-5900, Rocky Mountain Diagnostics, Colorado Springs, CO) according to manufacturer's instructions. Absorbance was measured at 450 nm and concentrations were determined using a standard curve.
Intestinal epithelial single cell isolation and cell sorting. Mouse intestinal epithelial cells were isolated as previously described with modification 22 . The intestinal samples (~10 cm long pieces of duodenum/jejunum) were collected from Tph1-CFP BAC transgenic mice and cut into 2-cm small pieces. The mucosa suspensions were obtained by incubating the intestinal pieces in dissociation buffer (30 mM EDTA, 1.5 mM DTT in PBS) at room temperature for 30 minutes. Single cell suspension was generated from intestinal mucosa by digestion in protease buffer containing 50 µg pronase (Sigma-Aldrich) per 100 mL Basal Medium Eagle (Gibco). Cells were sorted by flow cytometry using a BD FACSAria II cell sorter. CFP-positive single cells were collected by gating for CFP fluorescence combined with side and forward scatter to select single cells. Non-virable cells were excluded by gating for 7-amino actinomycin D (7-AAD, BD Biosciences) shortly after treatment with 7-AAD for 5 minutes before analysis. Molecular Biology. Sorted TPH1-CFP qPCR. We employed standard qPCR techniques on sorted EC cells.
Total RNA from FACS sorted Tph1-CFP cells was extracted using Qiagen RNeasy Micro kit (Valencia, CA, USA) according to the manufacturer's protocol. cDNA was reversed transcribed using Maxima cDNA Synthesis kit (Thermofisher). qPCR reaction was prepared with iTaq Universal SYBR green supermix (BioRad) and amplified for 40 cycles using CFX96 Real Time System (BioRad). Any sample that did not generate melting temperature was considered as no target gene expression, and its Ct value was assigned to 40. Target gene expression levels were normalized to that of reference gene Actb, resulting in a Ct difference (ΔCt) that was used to determine relative expression level of each target gene. Primers for the sodium channel α-subunits and Tph1 are listed in Supplementary Table 1.
Single cell RT-qPCR. Single cell RT-qPCR was performed with an Ambion Single Cell-to-C t kit (4458236, ThermoFisher, Grand Island, NY). Scn3a and Tph1 gene expression was analyzed in mixed enterochromaffin cell primary cultures using single cell RT-qPCR. Media was removed from cultures after 24 h and replaced with NaCl Ringer solution. Individual cells were aspirated into a glass pipette by negative pressure from a syringe, and then ejected by positive pressure into a PCR tube, subsequently placed on dry ice. Cells were collected to a total of 2-6 CFP + cells, 2-6 CFP − cells, or 0 cells (buffer medium only) per tube. First, samples were lysed using Single Cell DNase I/Single Cell Lysis Solution. Then, cDNA was synthesized using Single Cell SuperScript RT and a thermal cycler protocol that consisted of 10 min at 25 °C, 60 min at 42 °C, and 5 min at 85 °C. Next, a pre-amplification step using a 0.2x mixture of Scn3a and Tph1 primers (Supplementary Table 1) was added to each sample and run for 10 min at 95 °C for enzyme activation, then 14 cycles of 15 s at 95 °C and 4 min at 55 °C for amplification, and finally 10 min at 99 °C for enzyme deactivation. After pre-amplification products were diluted 1:10, a RT-qPCR reaction was run on a LightCycler 480 (Roche, Indianapolis, Indiana) using 480 sybr green (Roche 04707516001, Indianapolis, IN) and Scn3a and Tph1 primers for 2 min at 50 °C, 10 min at 95 °C, and 40 cycles of 5 s at 95 °C and 1 min at 55 °C. Finally, PCR products were separated by 2% agarose gel electrophoresis. All bands were sequenced to confirm identity.
Immunohistochemistry. Tissue labeling. Tph1-CFP mouse tissues were cut into 1 cm × 0.5 cm flat sheets or 0.5 cm length tubes from colon and small bowel. Human colon and small bowel tissue samples were collected from surgical waste. All tissues were fixed in 4% paraformaldehyde phosphate buffer (PFA) for 4 hours, rinsed in phosphate buffered saline (PBS), incubated overnight in 30% sucrose in PBS at 4 °C, frozen in 2-methylbutane over dry ice in optimum cutting temperature (OCT) embedding compound (Sakura Finetek, Torrance, CA), and stored at −80 °C until sectioned. Blocks were then sectioned into 12 µm thickness sections and thawed onto slides. Human tissue slides underwent antigen retrieval using target retrieval solution, as per manufacturer's protocol (Dako, Carpinteria, CA), and background reduction process with 3% hydrogen peroxide for 5 min. All slides of mouse and human tissue were then rinsed with PBS twice for 5 min and blocked with 200 µL/slide of 1% bovine serum albumin (BSA)/PBS/0.3% Triton in a humidity chamber for 1 hour. Primary antibodies (Supplementary Table 2) were added at 200 µL/slide of BSA/PBS/0.3% Triton and were incubated at 4 °C overnight in a humidity chamber. Slides were then rinsed 5 times for 3 min with PBS. Secondary antibodies (Supplementary Table 2) at 200 µL/slide of BSA/PBS/0.3% Triton were incubated for 1 h in the dark. Slides were again rinsed 5 times for 3 min with PBS and mounted in slowfade gold with 4' ,6-diamidino-2-phenylindole (DAPI, Life Technologies, Grand Island, NY) mounting buffer. Control slides were also prepared, on which no primary antibody was used. Primary and secondardy antibodies are listed in Supplementary Tables 1 and 2, respectively.
Cell counting. To quantify Na V 1.3 localization in EC cells, Tph1-CFP and/or Na V 1.3 positive cells in mouse tissue and 5-HT and/or Na V 1.3 positive cells in human tissue were counted in epithelium only, as defined by the DAPI positive outer layer of the mucosa facing the lumen. Images were taken on Olympus BX51W1 epifluorescent (40x) and Olympus FV1000 confocal (20x, 0.95 NA and 60x, 1.2 NA, z-res 0.91 µm) microscopes (Olympus Corporation, Tokyo, Japan). We counted EC cells from the small bowel and colon of 3 mice and 3 humans.
Voltage-clamp protocols. The signals were filtered at 4 kHz and sampled at 20 kHz. Data were recorded from cells held at −120 mV, stepped for 50 ms to −80 through +15 mV in 5 mV intervals to measure steady state activation, then stepped to 0 mV for 50 ms to measure steady state inactivation. The start-to-start time was 250 ms per sweep and 6 s per run for up to 10 runs.
Current-clamp protocols. To examine elicited activity, EC cells were injected with −10 pA current to hyperpolarize the baseline membrane potential to −60 to −80 mV, then action potentials were elicited every second by 50-ms injections of depolarizing current in 2 pA intervals. To examine spontaneous activity, cells were injected with zero current, and the membrane potential was recorded in gap-free mode.