Bench philosophy: Patch Clamp 2.0

The second generation of patch clamp recording
by Jose Guzman & Peter Jonas, Labtimes 05/2017

41 years after Erwin Neher's and Bert Sakmann's seminal Nature paper, describing the first patch clamp recordings of ion channels, new patch clamp techniques open the doors to the functional dissection of synaptic circuits.

Within the first generation of patch clamp, we encounter the classical applications that include the excised patch (inside-out and outside-out) and the whole-cell modes. An excised patch requires the detachment of a small piece of the membrane for its electrophysiological analysis. It is the tool of choice for the analysis of single-channel properties. In whole-cell recordings, the pipette gains access to the intracellular medium and can deliver dyes, small molecules or genes. It is the approach used to study voltage and current fluctuations across the membrane of an entire neuron.

Simultaneous patching of several neurons in the visual cortex. Photo: Peter Jonas

The second generation of patch clamp techniques covered three major applications: subcellular patch clamp techniques, multiple simultaneous patch clamp recordings and in vivo recordings.

Subcellular patch clamp techniques are the minimalistic refinement of classic patch clamp experiments. They have opened the door for the study of new functional compartments, such as axons and dendrites, thanks to technical developments in micromanipulators, microscopes and brain slicing methods. Micromanipulators of outstanding mechanical stability and minimal pipette drift (in the micrometre range) facilitate stable recordings from small structures. Microscopy methods based on transmitted sources of illumination (infrared and Dodt microscopy) and fluorescence (e.g. confocal and two-photon), allowed users to move from a blind approach to a visually-guided patching experience. Finally, the new generation of brain slicing techniques is better-suited for delicate neuronal structures.

Applications of subcellular techniques are suitable for the analysis of neuronal subcompartments and the study of synaptic transmission. In the first group, we find the excised patch mode, a subcellular variant of the outside-out configuration that allowed the mapping of channel densities along the dendrites. This breakthrough paved the way for later refinements in subcellular patch clamp, involving the direct measurement of dendritic voltage changes with two or more microelectrodes. These experiments changed the view of the dendrite as a merely passive structure and opened a new field focussed on dendritic integration.

Similarly, recordings from axons have demonstrated that the axon’s initial segment (AIS) contains a high density of sodium channels. Hence, the AIS has been implicated in the generation of action potentials and pinpointed as the central processing unit of the neuron. Subcellular patching permits access to small postsynaptic boutons (1 µm diameter) that impinge on sensory hair cells and has shed light on the working mode of sensory ribbon synapses.

Paradigm change

When subcellular patching arrived in the field of synaptic transmission in the mid-1990s, the experimental paradigm to study synapses changed from the classic frog neuromuscular junction to mammalian central synapses. The calyx of Held, for example, offers a model with many advantages. It is a giant auditory synapse, with a large presynaptic terminal connected to a single postsynaptic cell body. The experimenter can patch both elements simultaneously and introduce drugs into their cytoplasm to test their effects on neurotransmission. Calcium chelators are available to spatially and temporally restrict the spread of calcium ions in the presynapse. It is also possible to homogeneously raise the concentration of calcium using caged compounds, such as DM-nitrophen. This molecule binds to calcium and releases it during a flash of ultraviolet light.

The excitatory synapse between the mossy fibre and the CA3 pyramidal cell in the hippocampus is probably the most representative cortical synapse that has been recently studied with subcellular techniques. Still considered a large synapse, it allows the simultaneous presynaptic patch-clamp recording from a bouton (3-5 µm), while postsynaptically accessing a CA3 pyramidal neuron. The mossy fibre-CA3 pyramidal neuron synapse represents the canonical form of the presynaptic expression of long-term potentiation (LTP).

Dual recordings at this synapse have unveiled an endogenous buffer that controls calcium between its entry point and the vesicle fusion sites (Science 343, 665-70). More importantly, this synapse is, to date, the only known cortical synapse with a one-to-one communication. It permits the discharge of an action potential with a single presynaptic stimulation, several seconds after a high-frequency activity has occurred (conditional detonator). Concerning its detonation properties, the mossy fibre–CA3 synapse is also called the “teacher synapse”.

Subcellular techniques have contributed to the most intriguing findings in neuroscience. For example, against Cajal’s principle of information flow from dendrites to soma, an action potential can travel back from the axon to the dendrites. The generation of action potentials is not an exclusive property of axons, since dendritic spikes are present in almost all neuronal types. Axonal patch clamp has permitted the molecular identification of ion channels with very fast activation kinetics and optimal energy consumption (Nav1.2, Kv1.1 and Kv3.3). Moreover, owing to this technique, new ways of encoding information have been suggested (analogue versus digital encoding) that may enrich the repertoire of possible synaptic communication between cells.

When the study of synaptic transmission extended its formal biophysical approach to include neuronal networks, it created a new research field. Functional connectomics aims at investigating the neuron's activity (i.e. function) inside its wiring diagram (i.e. connectome). Patch clamp is well-suited for studying this relationship. Because synapses bridge the information contained in the action potentials fired by a cell (presynaptic side) to a voltage response in the target neuron (postsynaptic side), simultaneous recordings from both neurons (i.e. paired recordings) offer a logical approach to investigate functional connections.

By evoking one presynaptic action potential, we can characterise many features of unitary synaptic responses at the biophysical level (i.e. response size, the number of functional contacts, release probability, etc.). When we elicit a train of action potentials, we can further evaluate the time variation of the response (i.e. short-term dynamics), which is a synaptic signature of a functional connection between two neurons.

Paired recordings could also precisely control the timing of action potentials occurring in two neurons. They provide the ultimate way to address the Hebbian postulate directly, which suggests that synaptic plasticity occurs when pairs of neurons fire in association. For instance, in pairs of layer 5 pyramidal neurons of the rat somatosensory cortex, it is possible to induce plasticity changes: the strength of the connection between neurons can be modified by just varying the timing between the action potentials of pre- and postsynaptic cells (Science 275, 213-15).

Inducing changes in plasticity

In connected pairs of neurons, the natural order of activation (first presynapse, second postsynapse) enhances synaptic responses, whereas the reversal of the sequence (i.e. postsynapse before presynapse) produces the opposite effect. Thus, the rules of spike timing-dependent synaptic plasticity (STDP), the canonical form of synaptic plasticity between excitatory neurons, were discovered with the improvement of patch clamp methods.

An extension of the paired recording technique has been achieved by heroically increasing the number of simultaneous patch clamp pipettes added to a given preparation. By simultaneously patching several neurons in the visual cortex (Science 353, 1117-23), researchers demonstrated functional connectivity principles (motifs), provided for the first time (PLoS Biol 3, e68).

Increasing the number of simultaneous paired recordings offers a distinct advantage; in a quadruple simultaneous recording configuration, six paired configurations can be tested, resulting in a total of 12 putative synaptic connections. However, if we use the same number of electrodes (four) in a standard paired-recording configuration, only two paired recordings and four possible synaptic connections are possible. To date, 12 is the largest number of cells recorded by simultaneous patch clamp recording.

Mouse genetic meets patch clamp

A modern twist on this technique is recording from one or several neurons of choice by exploiting genetically labelled neuronal sub-classes in transgenic mice lines. We live in exciting times where the symbiosis between mouse genetics, fluorescence imaging and advanced patch clamp techniques will unravel the functional microcircuits in neuronal networks.

Classically, blind patch clamp recordings have been used to record neurons in brain slice preparations, in the absence of any visual guidance. Just when this method was in decline, it experienced a sudden renaissance in the in vivo recording technique. In the in vivo patch clamp technique, a small brain craniotomy enables blind neuronal access to different parts of the brain following stereotactic coordinates.

This technique has the edge over its extracellular counterparts, such as local field, silicon probes or tetrode recordings. The whole-cell mode permits the unequivocal attribution of firing patterns to identified neurons. It also allows the detection of subthreshold fluctuations (e.g. postsynaptic potentials and currents) and an adequate labelling of patched cells. At present, it comes at the expense of relatively short recording times, limited possibilities for pharmacological intervention and the evaluation of a restricted repertoire of behaviours. Despite these limitations, neuroscientists have devised some ingenious implementations. Whole-cell recordings in fully anaesthetised animals are now in bloom. The low resistance access serves as an ideal intracellular conduit for delivering genes or dyes to neurons in their natural context, making possible detailed reconstructions of neurons. In vivo recordings combined with subcellular patching techniques allowed unprecedented access to the dendrites in the superficial layers of the visual cortex and the axons of the cerebellar mossy fibres. By using patch clamp in vivo, together with local field recordings, we can track the local interactions between a single neuron and the circuits that surround it.

Nowadays, recording in fully awake animals with access to deep brain structures is no longer fiction. It is possible to obtain whole-cell recordings from cortical neurons in awake rodents, while they move on a spherical treadmill or run on a linear belt, using either real sensory cues or virtual environments.

Despite being in their infancy, on vivo whole-cell recording techniques have contributed to the discovery of several new principles in neuroscience. We now know that the dendrites of pyramidal neurons from layer 2/3 in the mouse visual cortex are tuned to specific visual features and that dendritic spikes selectively match the preferred stimulus orientation (Nature 503, 115-20). Growing evidence supports that CA1 pyramidal cells in the hippocampus are the cells that fire when an animal enters a particular location (place cells).

Inhibitory hippocampal neurons, in contrast, show little spatial tuning, although there is evidence that parvalbumin-positive interneurons in the neocortex respond to sensory stimulation. Since its origins, this technique has stirred the scientific debate. One of its most intriguing findings is the sparse activity in granule cells, the input layer of the hippocampal formation. These cells can fire bursts of up to 100 spikes per second, leading to the proposal that the granule cell network could provide a way of compressing spatial information. This mechanism can potentially represent one of the most robust encoding strategies of the brain.

With the precise control of gene expression, a careful targeting of neurons and the assessment of an enriched palette of behaviours, we will soon be able to decipher the single-cell correlates of specific brain functions. This combined approach will hopefully reconcile single-cell computations with the general principles of population coding.

Last Changed: 01.10.2017