Patch clamp

The patch clamp is a technique for recording the electric current through ion channels, the protein pores in cell membranes. A glass micropipette is pressed against a cell, gentle suction seals the glass to the membrane, and an amplifier holds the voltage across the sealed patch at a chosen value while it measures the current that flows. When the seal has a resistance of a gigaohm or more, the recording is quiet enough to show the current through a single channel, about a picoampere, as the channel flips between open and shut.^[1][2]

Erwin Neher and Bert Sakmann reported the first such recordings in 1976, from frog muscle fibers, and shared the 1991 Nobel Prize in Physiology or Medicine "for their discoveries concerning the function of single ion channels in cells".^[3][2] The amplifiers built for the method measure picoampere currents at bandwidths of kilohertz, and the same hardware records the current through a nanopore. The experiment of 1996 that put single strands of DNA into a protein pore, and so began nanopore sequencing, was recorded on an amplifier sold for physiology.^[4] Molecular Reality Corporation streams the current of its laboratory bench from a four-channel patch-clamp amplifier, the EDR4.

Until the 1970s ion channels were an inference. Alan Hodgkin and Andrew Huxley held the membrane of the squid giant axon at set voltages with a feedback amplifier, a voltage clamp, and in 1952 described the sodium and potassium currents it passed in a set of equations, the Hodgkin-Huxley model.^[5] The equations described conductances and said nothing about what carried them. Estimates of the current through one channel came from indirect experiments. In 1972 Katz and Miledi analyzed the noise of the response to acetylcholine at the frog neuromuscular junction and inferred an elementary current pulse lasting about 1 ms at 20 °C and carrying about 50,000 ions.^[6] Antibiotics and proteins inserted into artificial lipid membranes gave single-channel steps of a few picoamperes. The background noise of a living cell, however, was thought to be at least ten times larger than that.^[7]

Neher and Sakmann, at the Max Planck Institute for Biophysical Chemistry in Göttingen, isolated a small patch of membrane with a pipette. Their pipettes had an opening of 3 to 5 µm and a resistance of 2 to 5 MΩ, and they pressed them against muscle fibers that had been cleaned with enzymes. The seal between glass and membrane reached only 50 to 100 MΩ, yet the patch current stepped between two levels as acetylcholine receptor channels opened and closed.^[7] Neher also spent parts of 1975 and 1976 in Charles Stevens's laboratory at Yale, where by the same history he obtained similar results.^[7]

A seal is a resistor, and it generates thermal current noise of $\sqrt{4k_{B}TB/R}$ over a bandwidth B. A seal of 100 MΩ gives about 0.40 pA rms in 1 kHz, a large fraction of the current being sought.^[7][a] The group tried different pipette glass, cleaner surfaces and enzyme treatments, with little effect. By one account they were close to abandoning the attempt when they noticed that gentle suction on the pipette produced seals of 10 to 100 GΩ. The new seals were called giga-seals, to tell them from the earlier megaohm seals.^[7]

The 1981 paper by Hamill, Marty, Neher, Sakmann and Sigworth set down the method. It described how to make the pipettes, how to obtain giga-seals with resistances of 10⁹ to 10¹¹ ohms, the recording circuit, and the ways of isolating patches from cells. Detached patches allowed single-channel recording with defined solutions on both sides of the membrane.^[1]

The 1981 paper described the cell-attached, inside-out, outside-out and whole-cell recordings, and noted that single-channel resolution is possible in the whole-cell mode when the cell is smaller than 20 µm across.^[1] The perforated patch came in 1988, when Horn and Marty used nystatin in the pipette to give electrical access to a cell without the loss of the cytoplasmic compounds that its response to acetylcholine needed.^[8]

A patch-clamp amplifier is a transimpedance amplifier in a small headstage next to the pipette, followed by gain, filtering and a digitizer. The element in its feedback loop sets the noise. A resistor adds thermal noise, which falls as the resistance rises, and the very large resistors that picoampere work calls for are imperfect parts. A capacitor adds almost none.^[9]

The Axopatch 200B, made by Axon Instruments and now sold by Molecular Devices, is built on that argument. In patch mode its feedback element is a 1 pF capacitor, so the headstage measures current as the rate at which voltage rises across it. The voltage cannot rise forever, so the circuit resets when it reaches the supply limit, in 50 µs, and cooling the critical components to about −15 °C lowers their thermal noise. Whole-cell mode adds a resistor of 500 MΩ, or 50 MΩ for large currents, in parallel with the capacitor, because the cell and its surroundings make more noise than the amplifier does.^[9] The data sheet gives the maximum noise of the headstage in patch mode, with the input open and an 8-pole Bessel filter:

With a pipette holder attached the figure at 10 kHz is 0.145 pA.^[9] A 1 GΩ resistor alone would add 0.41 pA in 10 kHz, three times as much, and a 1 MΩ resistor 12.9 pA.^[b] The measured noise rises almost in proportion to the bandwidth, ten times the bandwidth giving 8.7 times the noise. That is the behavior of dielectric and capacitive noise, and thermal noise alone would rise only as the square root of the bandwidth.^[10] The sources are set out in Nanopore noise.

The amplifier also corrects for the voltage drop across the pipette (series-resistance compensation), cancels the capacitance of the pipette and of the cell, with a whole-cell range of 100 pF or 1,000 pF, and filters with a 4-pole Bessel filter set between 1 and 100 kHz, the limit of its recording bandwidth. It can apply a pulse of +1.3 V to rupture a patch on the way to the whole-cell mode.^[9]

The Nobel Assembly's announcement of 1991 said that the new knowledge had helped explain the cellular mechanisms behind several diseases, among them diabetes and cystic fibrosis, and that drugs commonly used to stimulate insulin secretion act directly on ATP-controlled ion channels.^[2]

The method is used to measure what a drug does to an ion channel. The safety guideline ICH S7B (2005) recommends an in vitro IKr assay, which evaluates a compound's effect on the ionic current through a native or expressed channel such as the one encoded by hERG, because inhibition of that current is the most common way a drug delays ventricular repolarization.^[11] Such currents are recorded by patch clamp.

Manual patch clamping takes a trained experimenter and a micromanipulator. In 2002 Fertig, Blick and Behrends reported whole-cell recording on a planar glass chip, which replaced the pipette with a hole in the chip; automated patch-clamp systems use chips of this kind.^[12] Nanion Technologies sells instruments of both the planar patch-clamp and the bilayer kind, and supplies the bilayer platform with the software that Elements developed for its amplifiers.^[13]

A nanopore experiment is a patch-clamp experiment with a different membrane. Two chambers of salt solution are separated by a membrane with one pore, each chamber holds a silver chloride electrode (Ag/AgCl electrode), an amplifier holds the voltage and reads the current, a low-pass filter and a digitizer follow, and a metal box shields the headstage from interference.^[10] The channel of a patch recording is replaced by a protein channel in an artificial bilayer or by a hole in a chip, and the open-channel current becomes the open-pore current. Recording single channels in artificial bilayers is older than the patch clamp, and nanopore laboratories use the Montal-Mueller bilayer method.^[7][10]

In the paper of 1996 the current "was converted to voltage and amplified by either an Axopatch 200A (Axon Instruments, Foster City, CA) or a Dagan 3900A patch clamp amplifier", then filtered with a Bessel filter and sampled at typically 64,000 samples a second, and the bilayer apparatus sat in a μ-metal box to shield it from "electric and magnetic noise sources".^[4] The Axopatch 200B data sheet has a section for bilayers, which says the headstage is stable with an input capacitance of 1,000 pF and that its integrating design lets the bilayer voltage be stepped quickly.^[9] In a 2020 comparison of noise in biological and solid-state nanopores the example traces of both kinds were recorded with an Axopatch 200B, and the one solid-state recording at 500 kHz needed a dedicated high-bandwidth amplifier, the VC100 of Chimera Instruments.^[10]

Where nanopore work differs is the load on the input. A membrane chip adds capacitance, over 300 pF for standard silicon chips and under 1 pF for the best glass ones, and that capacitance largely sets the noise above 10 kHz.^[10]

The company's guide describes its amplifier in the terms of the field: "the instrument that applies the voltage and hears the current." It samples at 200 kHz, one reading every five microseconds.^[14] In a post of March 2026 the founder said that the stream page should take data from any patch-clamp amplifier or nanopore device, and from other kinds of molecular sensor.^[15]

The first recordings in the stream's archive are of air, a standard first check: with no liquid there is no ionic path, and the amplifier reports its own floor of leakage and pickup.^[16] In its data sheet the Axopatch states its noise with the headstage input open. On the stream the pore is a laser-drilled hole 200 µm across,^[16] more than 100,000 times wider than the narrowest part of a protein nanopore, which is about 1 nm.

• Ion channel
• Hodgkin-Huxley model
• Lipid bilayer
• Picoampere
• Bessel filter

• Sakmann, B.; Neher, E. (1995). Single-Channel Recording. Plenum Press.