In Vivo Whole-cell Patch-clamp Recording of Hypothalamic Peptidergic Neurons in Larval Zebrafish

As a popular vertebrate animal model, zebrafish (Danio rerio) are widely used in neuroscience research and have several advantages¹. Their complement of genes and brain architecture is highly conserved with mammals, and they have a rich behavioral repertoire, making them useful to study genetic and neuronal mechanisms that underlie behaviors such as sleep, anxiety, depression, and social interaction^2,3,4,5,6. Their small size and low maintenance costs make them ideal for high-throughput screening of genes or drugs aimed at treating neuropsychiatric diseases^7,8. Finally, the optical transparency of larval zebrafish, combined with advanced light-sheet microscope techniques, and the availability of multiple fluorescent reporters, makes them particularly well-suited for monitoring neuronal and astroglial activity at cellular resolution across the entire brain^9,10,11.

The zebrafish hypothalamus, located deep in the ventral diencephalon, is anatomically and molecularly conserved with the mammalian hypothalamus, but is 2-3 orders of magnitude smaller, providing a simpler system to study hypothalamus function¹². Hypothalamic neurons form extensive synaptic connections with multiple brain regions, including the thalamus, brainstem, pituitary gland, and telencephalon, through which they regulate physiological homeostasis, neuroendocrine signaling, and autonomic behaviors^13,14. Such modulation of behavior and physiology largely relies on neuropeptides, which are short-chain amino acids that act by binding to G-protein-coupled receptors (GPCRs)¹⁵. In contrast to fast-acting amino acid neurotransmitter release at the synapse, neuropeptides may diffuse over longer distances via volume transmission, and mediate slow modulatory effects on synaptic transmission and neuronal excitability¹⁶.

Recent studies in zebrafish have revealed diverse functions of hypothalamic neuropeptides through multidisciplinary approaches such as whole-brain calcium imaging, single-cell RNA sequencing, high-throughput behavioral analysis, and neuronal circuit mapping^4,12,17,18. These emerging techniques have transformed our ability to characterize cellular diversity at a large scale, bridge molecular mechanisms with circuit-level neuropeptide functions, and map neural circuits that underlie specific behaviors^19,20. However, these approaches remain limited in their ability to resolve the rapid physiological dynamics and precise synaptic mechanisms that underlie neuropeptide function.

The patch-clamp technique, developed by Neher and Sakmann, remains the gold standard for investigating neuronal physiology due to its unparalleled temporal resolution and biophysical precision, enabling direct measurement of membrane potential fluctuations, synaptic currents, and ion channel dynamics^21,22. This technique is particularly invaluable for studying neuropeptidergic cells, whose characteristic bursting patterns and neuromodulatory influences require direct electrophysiological measurement to fully understand their roles in behavior and homeostasis^15,16,23.

Here, we have developed an optimized protocol that enables reliable whole-cell patch-clamp recordings of larval zebrafish hypothalamic peptidergic neurons, despite the technical challenges posed by the deep anatomical location of these neurons. This methodology provides direct access to the electrical activity of these cells, permitting detailed investigation of their responses to visual and auditory stimulus (Figure 1). By integrating patch-clamp electrophysiology with optogenetics, we can precisely characterize functional connectivity between hypothalamic circuits and downstream neurons (Figure 2 and Figure 3), facilitating the study of neuropeptides in physiological homeostasis and behavioral control. This approach bridges a critical gap in neuroscience research, allowing mechanistic exploration of hypothalamic neuron function at the biophysical, cellular, and circuit levels simultaneously.

All experiments were approved by the Institutional Care and Use Committee (IACUC) at Shanghai University (animal protocol YS 2025-169) and IACUC at California Institute of Technology (animal protocol 1836). Zebrafish from 5 to 7 days post-fertilization (dpf) were fed with rotifers and used for experiments. At this stage of development, sex is not determined. Adult zebrafish on a nacre [mitfa(w2/w2)] background²⁴ were used for breeding.

1. Solutions and recipes

1. Prepare E3 medium containing 5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl₂, and 0.33 mM MgSO₄. Adjust the solution by NaOH to pH 7.2. Store at room temperature for use within 1 month.
2. Prepare the extracellular solution containing 134 mM NaCl, 2.9 mM KCl, 2.1 mM CaCl₂, 1.2 mM MgCl₂, 10 mM HEPES, and 10 mM glucose. The osmolality of the extracellular solution is ~290 mOsmol L^-1. Adjust the solution by NaOH to pH 7.8. Store at 4 °C for use within 1 month.
3. Prepare the intracellular solution containing 100 mM K-gluconate, 10 mM KCl, 2 mM CaCl₂, 2 mM Mg-ATP, 0.3 mM Na-GFP, 10 mM HEPES, and 10 mM EGTA. The osmolality of the intracellular solution is ~280 mOsmol L^-1. Adjust the solution by KOH to pH 7.4. Aliquot 100 µL of the internal solution into microfuge tubes and store at -20 °C for use within 1 year.
   NOTE: The addition of Mg-ATP and Na-GTP is recommended for prolonged and stable recordings. Intracellular solution needs to be filtered through a 0.22 µm membrane, which removes particulates and microbial contaminants that could compromise seal formation or lead to pipette clogging.
4. α-bungarotoxin: Dissolve 1 mg of α-bungarotoxin in 1 mL of E3 medium to a final concentration of 1 mg/mL. Aliquot 20 µL of the solution into microfuge tubes and store at -20 °C for use within 1 year.
5. Low-melting point agarose: Dissolve 0.75 g of agarose in 50 mL of E3 medium in a glass bottle to a final concentration of 1.5%. Heat the mixture in a microwave oven until the agarose is melted and the solution appears clear and homogeneous. Maintain the melted agarose at 42-45 °C in a water bath for use within 1 month.

2. Zebrafish preparation and dissection

1. Place single male and female adult fish containing a desired transgene, e.g., Tg(hcrt:RFP) fish²⁵, in a mating tank and separate them by a clear divider overnight. In the morning of the second day, remove the divider to allow the fish breed. Collect fertilized embryos at the bottom of the tank and return adult fish to their home tanks by early afternoon.
2. Raise the embryos in a 9 cm diameter Petri dish with E3 medium including 0.0005% methylene blue. Place the embryos in an incubator at 28.5 °C with 14:10 h light:dark cycles. Change the E3 solution every day.
   NOTE: Methylene blue in E3 medium is used to prevent fungal growth until 3 days post-fertilization (dpf), but is replaced with E3 medium alone at 3 dpf, since methylene blue may affect cellular metabolism and mitochondrial function at later stages of development²⁶.
3. Screen larval fish under an epi-fluorescence stereomicroscope with maximum magnification (~115x) starting at around 3 dpf. Place individual fish in water droplets (larger than their body size) to facilitate the sorting. For the fish older than 3 dpf, anesthetize the fish using 3 mL of E3 medium containing 0.01% tricaine to suppress spontaneous locomotor activity. Collect larvae with bright fluorophore expression in the hypothalamus.
   NOTE: Wear nitrile gloves, safety glasses, and a lab coat when handling tricaine. Excess tricaine must be disposed of as chemical waste according to animal protocols.
4. After the screening, wash out the tricaine using E3 medium immediately, and put these larvae back in the incubator. Add ~5 mL of feeding solution containing rotifers into the Petri dish starting at 5 dpf until the day of the experiments. Change the E3 solution and feed the fish with rotifers every day.
5. At 5-7 dpf, collect larval fish that appear healthy and show active locomotion. Transfer them to a small petri dish (3.5 cm in diameter) using a glass pipette.
6. Remove the water around the fish, and add 20 µL of 1 mg/mL α-bungarotoxin onto the fish for ~30 s. Then wash out the drug by adding ~3 mL of E3 medium into the Petri dish.
   NOTE: Wear nitrile gloves, safety glasses, and a lab coat when handling α-bungarotoxin. Excess α-bungarotoxin must be disposed of as chemical waste according to animal protocols.
7. Cover the Petri dish with aluminum foil to keep the fish in the dark and gradually become paralyzed.
   NOTE: α-bungarotoxin is an acetylcholine receptor blocker that specifically acts on the neuromuscular junction, thereby paralyzing the fish without affecting the nervous system. The fish can still sense environmental stimuli, and brain function is normal, after paralysis^9,27.
8. Check for paralysis after ~15 min using a stereomicroscope and select a paralyzed fish that shows a normal heartbeat and robust blood circulation in the brain and body.
9. Transfer the fish into a custom-made recording chamber, which is a 35 mm Petri dish with a 10 mm diameter hole in the center and a 20 x 20 mm glass coverslip glued at the bottom to cover the hole. Remove the water and add ~200 µL of 1.5% low-melting agarose.
10. Position the fish dorsal side up in the agarose using forceps (Dumont #5). When the agarose has fully solidified, add the extracellular solution to cover the agarose. It takes 5-10 min for the agarose to fully solidify.
11. Cut and remove the agarose above the fish's head using a micro knife. Then make a small incision in the skin above the brain ventricle between the optic tectum and cerebellum using a glass micropipette with ~1 µm tip opening for pipette advancing into the brain next step.
12. Remove tissue debris or blood clots around the dissection hole using a broken glass pipette.

3 Whole-cell patch-clamp recording

1. Prepare recording micropipettes with ~1 µm tip opening and ~20 MΩ resistance using borosilicate capillaries on the day of the experiment. Store pulled micropipettes in a container and use them on the same day to prevent contamination.
2. Turn on the computer, CCD camera, monitor, amplifier, digitizer, micromanipulator, and associated software. Amplifier and digitizer must be turned on prior to opening software.
3. Carefully move the fish to an electrophysiological rig where the patch clamp experiment will be performed.
4. Perfuse the fish with oxygenated extracellular solution at ~2 mL per min using a peristaltic pump to maintain fish health and cell viability.
5. Move the recording chamber using an X-Y translator (custom-made) to center the fish in the visual field of a 4x objective under an upright microscope.
6. Switch to a 60x objective and move the translator to visualize the target hypothalamus region. Adjust Nomarski differential interference contrast (DIC) with 900 nm infrared light to get the best visualization of the cells.
7. Turn on the fluorescence excitation light source (e.g., mercury bulb) and use it to visualize fluorescent neurons in the hypothalamus. Locate healthy neurons that show bright fluorescence and clear cell boundaries, which often indicate good cellular health and facilitate high-quality electrophysiological recordings.
8. Slowly back-fill a recording micropipette with intracellular solution using a fine microloader tip.
9. Affix the electrode to the headstage of the amplifier and tighten the knob to prevent air leakage. Apply a slight positive pressure (~150-200 mBar) to the inside of the micropipette through the tubing connected to the holder, which prevents contaminants at the air-solution interface from entering the electrode tip.
10. Lower the micropipette into the bath solution using a micromanipulator in fast mode and position the micropipette tip above the dissected hole under the 4x objective.
11. Switch the micromanipulator to slow mode, and change the objective to the 60x one. Advance the tip of the glass pipette slowly through the dissection hole, and along the ventricle between the two tectal hemispheres.
12. Bring the pipette tip orthogonally towards a target fluorescent cell viewed using both DIC and fluorescence. Once the pipette tip touches the cell membrane, a small indentation is formed.
    NOTE: To visualize deep tissues like the hypothalamus, 900 nm IR-DIC is used together with polarizer and analyzer filters that are finely adjusted for IR-DIC.
13. Release of the positive pressure quickly when the tip touches a cell enables the cell to form a seal with the pipette. Simultaneously, apply a brief negative pressure via the tubing connected to the holder.
    NOTE: The value of resistance on the software should quickly increase to > 1 GΩ, indicating a high-resistance seal is formed. When the fluorescent hypothalamic cell is recorded successfully, there should be fluorescent cell membrane within the tip of the recording micropipette.
14. Clamp the cell at -60 mV and wait for 1-2 min until the baseline and access resistance are stable. Apply a brief electrical pulse or gentle suction by mouth to rupture the cell membrane beneath the micropipette tip.
15. Compensate for the capacitive current and series resistance in an electrophysiologic software configuration. Discard the data if the series resistance is larger than 100 MΩ and varies > 20% in the recordings.
16. Apply a 1 s -10 mV voltage pulse in voltage mode to monitor intrinsic membrane resistance before and after the experiments. Discard the cell if the membrane resistance or membrane potential varies a lot.
17. First record in voltage-clamp (VC) mode when the cell is held at a negative value. After establishing a stable recording, switch to current-clamp (IC) mode if needed.
    NOTE: The giga-seal may not be stable shortly after the rupture, so clamping the cell at -60 mV under VC mode for a few minutes at the beginning of an experiment can provide better cellular conditions.
18. Process the analysis of electrophysiological data and spike detection offline using MATLAB after the experiment is complete.
19. After the experiments, euthanize the larval fish using 1.5% Sodium hypochlorite (NaClO) and dispose of them according to institutional animal care protocols.

In this manuscript, we present an improved in vivo whole-cell patch-clamp recording technique for investigating hypothalamic neurons in intact zebrafish, particularly focusing on hypocretin (Hcrt) neurons.

As presented in Figure 1, our methodology enables direct electrophysiological recordings from these neurons deep within the brain in an intact animal (Figure 1A), overcoming limitations of traditional in vitro slices or dissected brain preparations. Hcrt neurons are located deep in the ventral diencephalon, with extensive axonal projections into the forebrain, midbrain, and hindbrain (Figure 1B). Using the in vivo samples, we first measured several electrophysiological parameters for Hcrt neurons (Mean ± SEM for 5 neurons; resting membrane potential: -53.26 ± 3.64 mV; membrane resistance: 2.67 ± 0.22 GΩ; baseline frequency of action potential: 0.50 ± 0.17 Hz; peak amplitude of action potential: 33.61 ± 8.49 mV; half-width of action potential: 5.04 ± 0.84 ms). Next, we examined the responses of Hcrt neurons to visual and auditory stimuli. Figure 1C demonstrates that an Hcrt neuron exhibits robust physiological responses to visual stimuli, including membrane potential depolarization and increased action potential firing at either onset or offset of 2 s white LED flashes (in current clamp mode), along with excitatory postsynaptic currents (EPSCs) at both stimulus onset and offset when voltage-clamped at -60 mV, revealing this Hcrt neuron receives dual ON and OFF visual inputs. Furthermore, as shown in Figure 1D, this neuron also responds to brief (10 ms) auditory stimuli with enhanced spike activity and EPSCs, demonstrating multimodal sensory convergence at the single-cell level.

Optogenetic manipulation of Hcrt neurons was achieved using Tg(hcrt:ReaChR-YFP) transgenic zebrafish28, which selectively expresses the red-shifted opsin ReaChR in Hcrt cells (Figure 2A)29. Optogenetic stimulation was delivered through a custom optical system with color LEDs and a lens for light illumination and focusing. This optical device was incorporated into the light pathway of an upright microscope, so that the optogenetic light was projected onto the target region using a 60x objective of the microscope. Precise delivery of 590 nm laser light at 1 Hz evoked consistent membrane depolarization and reliable single action potential per stimulus cycle (Figure 2B). The neurons maintained faithful spike generation across a range of stimulation frequencies (Figure 2C), though intermittent failures occurred during high-frequency stimulation (Figure 2D), likely reflecting intrinsic limitations in the recovery kinetics of the ReaChR photocycle. This optogenetic approach demonstrates robust temporal control of Hcrt neuron activity while revealing their capacity for frequency-dependent encoding of neuromodulatory signals.

To investigate functional connectivity between Hcrt neurons and the locus coeruleus (LC), we combined optogenetic activation with simultaneous whole-cell patch-clamp recordings (Figure 3A,B). High-frequency (10 Hz) optical stimulation of Hcrt neurons reliably elicited robust action potential firing in postsynaptic LC neurons (Figure 3C), demonstrating strong excitatory drive from hypocretinergic to noradrenergic systems. These results provide direct electrophysiological evidence that LC neurons serve as downstream targets of hypothalamic Hcrt signaling, potentially mediating arousal-related functions of this neuromodulatory circuit.

Figure 1: Whole-cell patch-clamp recording of a Hcrt neuron. (A) Top: A recording chamber with a larval fish embedded in the center of the chamber. Bottom: The zoomed-in view of the fish with an incision at the skin above the ventricle between the tectal hemispheres and cerebellum. (B) The location and projections of Hcrt neurons in a 5-dpf Tg(hcrt:RFP) larval zebrafish. (C) Electrical responses of an Hcrt neuron to a 2 s white flash stimulus. The right traces are the zoomed-in view of the left boxed region. (D) Electrical responses of an Hcrt neuron to a 10 ms 500 Hz auditory stimulus with an intensity of 80 dB. The right traces are the zoomed-in view of the left boxed region. IC: current clamp; VC: voltage clamp. The gray lines indicate 10 individual trials, and the black lines indicate the average of individual trials (C, D). Please click here to view a larger version of this figure.

Figure 2: Characterization of optogenetic activation of Hcrt neurons by whole-cell patch recording. (A) Schematic showing the location of Hcrt neurons in a Tg(hcrt:ReaChR-YFP) fish. Both eyes were removed on the day before electrophysiological experiments to avoid the activation of Hcrt neurons via visual pathways by optogenetic light. (B) Five example traces showing activation of an Hcrt neuron by 1 Hz 50 ms optogenetic stimulation with 30 s between trials. (C) The responses of an Hcrt neuron to optogenetic stimulation at different frequencies. The 50 ms light pulses were used for 1 Hz, 2 Hz, and 5 Hz stimulation, whereas 10 ms light pulses were used for 10 Hz stimulation. The raster dots above each histogram indicate individual action potentials. The histogram is the average of all raster responses from 10 trials. (D) The relationship between stimulus frequency and evoked action potential frequency from the data shown in (C). Please click here to view a larger version of this figure.

Figure 3: Confirmation of Hcrt-LC connectivity by combining optogenetics and whole-cell patch recording. (A) The locations of Hcrt and LC neurons from a 5-dpf Tg(hcrt:RFP); Tg(dbh:GFP) fish. (B) Schematic showing the optogenetic activation of Hcrt neurons and simultaneous electrophysiological recording of an LC neuron using Tg(hcrt:ReaChR-YFP); Tg(dbh:GFP) fish. Both eyes were removed on the day before the electrophysiological experiments. (C) Raster and histogram plots showing the effect of 10 Hz optogenetic stimulation of Hcrt neurons on LC neuron action potentials. Please click here to view a larger version of this figure.

The protocol described here enables patch-clamp recordings of peptidergic neurons in the larval zebrafish hypothalamus, one of the deepest and most technically challenging brain regions to access. Due to the inherent difficulty of this preparation, successful patch-clamp recordings require meticulous attention to a few critical parameters, namely pipette quality, approach technique, solution purity, and tissue health. These factors collectively determine the likelihood of achieving and maintaining stable giga-seal recordings. Firstly, a clear visualization of target cells and the pipette tip is critical for successful pipette-cell touching and giga-seal formation. This requires that the optical pathway of the microscope be finely calibrated by adjusting the condenser, aperture diaphragm, and field diaphragm.

Due to the long distance between the dissection site and hypothalamic target neurons in our zebrafish preparation, application of optimal positive pressure during pipette advancement can not only blow away debris from the air-solution interface, but also clean the target cell surface prior to contact. Through systematic optimization, an ideal pressure range of 150-200 mBar was identified that balances effective cleaning without displacing target neurons. Insufficient pressure fails to prevent pipette contamination, while excessive pressure may cause cell displacement. A pressure meter was integrated into the headstage tubing for real-time monitoring of pressure. When the pipette tip contacts the cell membrane of hypothalamic neurons, positive pressure is released, and a brief negative pressure should be applied simultaneously. The negative pressure can be applied via mouth to precisely control the pressure strength and stop the suction immediately when the resistance reaches the giga ohm. Successful recordings of fluorescent neurons can be confirmed by visualization of the fluorescent cell membrane in the pipette tip. Although fluorescent dyes can be loaded into patched superficial cells, e.g., retinal cells, to visualize neuronal morphology^30,31, it is difficult to load fluorescent dyes into deep hypothalamic neurons, since the inclusion of dyes in the pipettes often interferes with a good giga-seal and decreases the probability of whole-cell recordings.

Achieving stable giga-seal formation depends on maintaining clean tissue conditions throughout the experiment, beginning with selecting healthy transparent animals with robust blood circulation. Paralysis duration should be limited to less than 30 min before the dissection, as prolonged paralysis may disrupt tissue integrity and increase the formation of cellular debris. Tissue viability needs to be carefully assessed under a high magnification objective. Good samples often have active microcirculation, clear tissue boundaries, and few cell degeneration. Throughout recordings, continuous perfusion with oxygenated extracellular solution ensures proper cellular metabolism, which is particularly important for older larval fish (8-14 dpf) when metabolic demands are higher.

This optimized methodology, incorporating rigorous animal selection criteria, minimized procedural stressors, and comprehensive physiological maintenance protocols, significantly improves the reliability of obtaining high-quality whole-cell patch-clamp recordings from deep hypothalamic structures. Due to technical demand, most patch-clamp experiments are often restricted to superficial brain regions or in vitro brain slices, which disrupts long-range connectivity and vascular supply, potentially altering physiology^32,33. Our in vivo methodology enables patch-clamp recording of neurons in their natural environment, with intact synaptic inputs and network activity, and may facilitate the investigation of hypothalamic neurons' response to physiological cues, e.g., sensory stimuli, hormones, glucose, pH, as well as interoceptive cues from peripheral body organs, including gut and heart. This protocol provides a powerful technical basis for investigating the electrophysiological properties of neuropeptidergic neurons and their roles in regulating physiological homeostasis and behaviors.

Compared to optical imaging techniques, patch-clamp has several limitations, such as low throughput and technical difficulty, that constrain its broader applicability, particularly in mapping behavior-related neural circuits. Whole-cell mode causes mechanical stress on the cell membrane, dialyzes cytoplasmic contents, and cannot provide spatial information about neuronal processes' activity. However, in the future, combining this technique with other methods, e.g., in vivo patch-seq or simultaneous patch plus calcium imaging, can expand its utility and provide a more complete picture of brain function^34,35.