Dendritic patch-clamp recording

Standard ACSF

125 mM NaCl, 2.5 mM KCl, 2 mM CaCl₂, 1 mM MgCl₂, 25 mM NaHCO₃, 1.25 mM NaH₂PO₄ and 25 mM D-glucose. Final osmolarity should be ~310 mmol/kg. At all points it is essential that the ACSF in contact with the brain tissue is saturated with carbogen (which provides both oxygenation and pH buffering).

Slicing solution

Depending on the brain area (i.e., prevalence of NMDA receptors and susceptibility to hypoxia), high quality brain slices can be cut in normal ACSF, saturated with carbogen. However, several procedures can be used to reduce activity levels during slicing in order to minimize excitotoxicity. These include thorough and rapid cooling of the preparation (ideally with cardiac perfusion of ice-cold slicing solution), replacing from 50% to 100% of the NaCl with iso-osmolar sucrose⁴⁴, changing the Ca:Mg ratio (e.g., 6 MgCl₂, 1 CaCl₂) or adding 2–10 mM kynurenic acid to the slicing solution, to eliminate ionotropic glutamatergic transmission.

Patch-pipette solution for whole-cell recording

Standard intracellular solution contains 130 mM K-methanesulfate, 10 mM HEPES, 7 mM KCl, 0.05 mM EGTA, 2 mM Na₂ATP, 2 mM MgATP and 0.5 mM Na₂GTP, titrated with KOH to pH 7.2. Alternative principal anions are gluconate and methanesulfonate; note that all internal solutions are associated with washout of intracellular factors and some may also have pharmacological effects⁴⁵ (also see Kaczorowski C.C., Disterhoft, J.F. & Spruston, N. Soc. Neurosci. Abst. 31, 737.17, 2005).

A fluorescent dye (e.g., 1–25 μM Alexa 594) can be included to visualize dendrites during the experiment, and biocytin (0.1–0.4%) can also be added for subsequent recovery of cell morphology. Final osmolarity should be ~285 mmol/kg.

Patch-pipette solution for cell-attached recording

As with somatic cell-attached recording, the patch-pipette solution is based on the bath ACSF. Often a HEPES-buffered version is used to avoid potential changes in pH in the absence of carbogenation. A typical recipe is the following: 125 mM NaCl, 10 mM HEPES, 3 mM KCl, 2 mM CaCl₂, pH 7.4 with NaOH. Changes in solution composition can be made to pharmacologically isolate conductances of interest¹², to make ionic substitutions, or to increase ion concentration gradients across the patch membrane in order to increase the size of currents recorded⁸.

Pipettes

Preferably thick-walled, borosilicate glass; can be filamented to ease filling with internal solution. The tip size depends on the size of the structure to be patched. Typical resistances range from 6 to 15 MΩ. Smaller tips can improve the chances of successful seal formation or recording stability, but can lower the chances of breaking in and will result in higher access resistance.

Experimental slice chamber

A glass-bottomed chamber must be used with IR-DIC optics. The use of thin coverslip glass allows appropriate focusing of condensers with high numerical aperture and short working distance.

Perfusion system

This consists of a reservoir of ACSF, bubbled to saturation with carbogen via a sintered glass bubbler (Robu), siphoned into the recording chamber via oxygen-impermeable tubing (e.g., Teflon), and removed into a collection reservoir via outflow tubing connected to suction. A ‘dripper’ interrupting the inflow tubing breaks the continuous flow of solution, reducing the electrical noise transmitted to the recording chamber via the ACSF column acting as an aerial. It also reduces the number of physically disruptive bubbles that flow into the chamber, and allows monitoring ofthe rate of ACSF inflow to the recording chamber.

Temperature controller

Solution can be heated prior to entry to the experimental slice chamber using a self-made or commercial temperature-regulated heating jacket around the inflow; a small temperature probe close to the slice can be used to accurately monitor recording temperature.

‘Harp’

AU-shaped piece of flattened platinum with nylon threads (pulled from nylon tights) glued across it, used to hold down the brain slice in the recording chamber.

Slice preparation ● TIMING Approximately 90–120 min

1. Prepare slicing and recording solutions by bubbling to saturation with carbogen (takes at least 0.5 h). Cool the slicing solution on ice. Prepare the incubation chamber; fill with carbogen-bubbled ACSF and warm to 35 1C. ◼ PAUSE POINT These materials can be prepared in advance.

2. Anesthetize the animal (e.g., using isoflurane). ! CAUTION Appropriate guidelines and regulations for animal experiments must be followed. Younger animals (e.g., <P25 rats) generally yield slices of higher quality.

3. Kill the animal by decapitation. With the head immersed in ice-cold slicing solution, lift off the skull and carefully remove the brain, transferring it rapidly to ice-cold slicing solution. ! CAUTION Appropriate guidelines and regulations for animal experiments must be followed. ▲ CRITICAL STEP There must be a balance between care and speed for this and the next step in order to minimize physical damage and hypoxia of the tissue. ? TROUBLESHOOTING

4. Place the brain on a suitable cutting surface (e.g., a Petri dish with a Sylgard bottom; the brain can be secured with dissecting needles if necessary) and submerge it in ice-cold, carbogen saturated slicing solution. If required, gently remove the meninges above the area of interest. Deftly cut a block of the tissue of interest. Cut a base parallel to the plane of the dendrites to be patched. (See also Table 1.) ▲ CRITICAL STEP The orientation of the slice should be parallel to the plane of the dendritic tree under investigation. This may involve using an angled slicing stage to ensure the correct orientation, as well as optimizing the orientation of the tissue block prior to slicing. Table 1 details optimal cut orientations used to prepare brain tissue blocks from several commonly studied brain areas. ? TROUBLESHOOTING

5. Slide the base of the tissue block onto a thin film of cyanoacrylate glue on the cooled slicing stage. Fill the chamber with ice-cold carbogen-saturated ACSF.? TROUBLESHOOTING

6. Cut thin slices (150–300 mm) of tissue, advancing the slicer blade at a speed, vibration amplitude and frequency that ensures the tissue is cleanly cut, without being compressed or displaced. Between slices, raise the blade/lower the tissue by 100–200 mm in order to avoid dragging the blade across the surface of the tissue as it is moved back to the start of the slicing cycle. Keep the slicing chamber cool throughout the entire procedure (using ice or a Peltier element). ▲ CRITICAL STEP High slicing speeds will squash the tissue rather than cut cleanly through it, while overly slow speeds increase the duration of time the tissue remains as a block, and thus unexposed to oxygenated ACSF. Overly high vibration frequencies or amplitudes are associated with larger z-displacements of the slicing blade, while overly low frequencies prevent clean slicing of the tissue. Optimal settings on different slicers differ, but the following parameters might be used as an initial guide: blade advance speed, 0.1–0.25 mm/s; blade vibration frequency, 45–85 Hz; blade vibration amplitude, >1 mm. ? TROUBLESHOOTING

7. Place slices in incubation chamber filled with ACSF, store at 35 1C for 30–60 min, then keep at room temperature until use. Saturate the ACSF with carbogen before slicing, and maintain carbogen bubbling throughout. Slices will exhibit the best optics and most healthy cells in the first hour following the warm incubation period. They are generally usable for 3–5 h after incubation; after this, cell death and swelling usually make it progressively more difficult to visualize healthy dendrites, thus making the chance of successful dendritic patching very poor.

Slice transfer and visualization ● TIMING Approximately 10–30 min

• (8)

  Transfer a brain slice to the recording chamber and place the ‘harp’ on top of it to secure it. Make sure that the harp holds the slice firmly in place, as movement of the slice will make dendritic recording impossible. Perfuse the chamber at a fairly high rate (typically 1–2 ml/min, although 4–6 ml/min can be used in some cases; see Hajos, N. et al. FENS Abstr. 2, A009.13, 2004) with carbogen-saturated (and heated, if relevant) ACSF to ensure adequate oxygenation. ▲ CRITICAL STEP Transfer of the slice is best accomplished using a broken-off glass Pasteur pipette with the broken end inserted into a small rubber bulb. Do not touch the surface of the slice with a brush or any other instrument when transferring slices. The harp should be oriented so that the fine nylon threads are parallel to the orientation of the dendrites to be patched. Note that adequate oxygenation of the slice is essential. This can be aided by improving solution flow in the chamber (e.g., by positioning the open end of the u-shaped harp so that it faces toward the ACSF inflow). ? TROUBLESHOOTING

• (9)

  Lower the water-immersion objective into the bath solution and, at high magnification, focus on the surface of the slice. ▲ CRITICAL STEP At high magnification, it often becomes apparent that vibrations (often too ‘minor’ to have been observed at lower magnification) are being transmitted to the set-up. Find and remove these sources of instability. ? TROUBLESHOOTING

• (10)

  Adjust the optics for the best resolution (see Box 1). ▲ CRITICAL STEP

• (11)

  Scan the slice for healthy looking somata; inspect these candidate cells for suitable (connected) dendrites. Center the chosen dendrite in the middle of the video monitor screen. (See also Fig. 2.) ▲ CRITICAL STEP Aim for a dendrite with a smooth, well-defined membrane surface (Fig.2). For neurons with extensive branching, draw the dendrite-to-soma trajectory on the screen with a marker pen, for later distance measurement. Avoid dendrites where the membrane appears either dark, crisp and jagged, or faint and swollen (Fig. 2e,g). (Unfortunately, these dead dendrites/cells often are the structures most clearly visible in the slice. They can, however, serve the useful function of allowing one to check that the main axis of the dendrites of interest run roughly parallel to the cut surface of the slice, which is essential when aiming for distal dendrites).

  Dendrites deeper than about 50 mm in the slice are very difficult to visualize and/or to approach while maintaining clean pipette tips. It is possible, however, to follow a dendrite deep into the slice from the soma to a distal, more superfical location. Even if the dendrite seems to disappear at points it is possible to spot the same dendrite again at a more distallocation. The diameter and general appearance and trajectory of the dendrite are good clues to its identity. ? TROUBLESHOOTING

Patch-clamp recording from dendrite ● TIMING Approximately 10 min (until break-in)

• (12)

  Fill a patch-pipette with enough solution to just cover the electrode wire and insert it securely into the pipette holder. ▲ CRITICAL STEP The solution should be filtered through a 0.2 μm pore filter, to remove debris that can interfere with sealing and block the pipette tip.

• (13)

  Apply positive pressure to the pipette holder (via tubing connected to pipette holder). Use a switchable valve or three-way tap to hold the pressure behind the pipette.

  Pressure application is required before entering the bath solution in order to deflect any dirt floating on the surface of the bath, which might otherwise block or dirty the pipette tip. Increasing the pressure too much for extended periods of time, on the other hand, tends to drive any particles present in the intracellular solution into the pipette tip, thus clogging it. This is more critical during dendritic recording due to the use of small pipettes. It is useful (but not essential) to be able to monitor the amount of pressure in the pipette via a manometer. This allows the pressure used for patching to be adjusted, and will highlight any loss in pressure due to air leaks. The amount of pressure used will depend on tip size; typically 30 mbar is used.

• (14)

  Lower the pipette into the ACSF, and move it into position, in the center of the video monitor. Check that the tip of the pipette is not blocked.

  Searching for cells and inserting pipettes is faster under low magnification (1× on the magnifier); switch to 2× or 4× to target dendrites.

• (15)

  In voltage clamp, apply a ‘test pulse’ (typically a 10 mV, 10 ms square pulse of voltage, applied at >10 Hz), and monitor the current required for this voltage deflection on an oscilloscope (or computer monitor). The resistance of the pipette tip can be calculated using Ohm’s law (Resistance ¼ Voltage/Current).

  The tip resistance gives a useful indicator of the tip size (6-15 MΩ is typical for dendritic patching), and to verify that the pipette is unblocked (with block identifiable when the pipette resistance jumps to a higher value).

• (16)

  Lower the pipette to ~50 μm above the surface of the slice (Fig. 3a).

  Note that the time spent with the pipette close to the neuron should be minimized when using physiological internal solution, as the stream of high potassium solution can lead to continuous excitation of neurons in the slice, potentially triggering plasticity and/or excitotoxicity. However, when making paired somatic and dendritic recordings, this excitation, seen at the soma, can be a helpful indication that the dendritic pipette is approaching a dendrite belonging to the same cell (this is especially useful when trying to record distally from long dendrites, where the path of the dendrite may be obscured or ambiguous). Once this excitation has been observed, it can be helpful to hyperpolarize the soma (or hold it in voltage clamp) while establishing the dendritic recording.? TROUBLESHOOTING

• (17)

  On the patch-clamp amplifier, zero any voltage offset between the electrode and earth.

• (18)

  For the final approach towards the dendrite of interest (Fig. 3b), increase the positive pressure and move the pipette towards the dendrite of interest. The flow of intracellular solution from the pipette tip should clear away extracellular debris as you move towards the dendrite. This can also increase the visibility of the dendrite (see Supplementary Videos 1 and 2 online). The amount of positive pressure used will depend on pipette tip size and the type of dendrite being patched; e.g., for pyramidal neurons 120–180 mbar is typical (using ~8–10 MΩ pipettes), while for Purkinje neurons 30–50 mbar is typical (using ~7–8 MΩ pipettes).

  You should endeavor to move the pipette cleanly through the slice towards the dendrite, without squashing the tissue (this can subsequently cause problems maintaining the contact between the pipette and the dendrite as the tissue slowly ‘relaxes’ back to its original position). One way of achieving this is by moving the pipette mainly in the diagonal ‘in and out’ axis. ? TROUBLESHOOTING

• (19)

  Aiming towards slightly off center (to the side closest to the pipette) of the dendrite, move the pipette diagonally onto, then slightly down into the dendrite (see Supplementary Videos 1 and 2). While doing this, focus continuously on the tip of the electrode.

• ▲ CRITICAL STEP The aim is to solidly connect with the dendrite, but not to deform it by more than a few micrometers. Actual movement of the dendrite by the pipette by a few micrometers is desirable, as it verifies that the pipette is in direct contact with the dendrite. ? TROUBLESHOOTING

• (20)

  When good contact is made, a dimple should start to form in the dendritic membrane (Fig. 3c, Supplementary Videos 1 and 2). When the dimple starts to appear, immediately release the pressure on the pipette, then apply light negative pressure to the pipette via gentle suction (<10 mbar), and finally hyperpolarize the voltage clamp command potential. ▲ CRITICAL STEP Hyperpolarization helps a giga-ohm seal to form between the pipette and cell membrane; holding potentials down to ~-75 mV can be used, but this should be readjusted to approximately the cell’s resting potential before breaking into the cell. ? TROUBLESHOOTING

• (21)

  Continue light suction while monitoring the seal resistance by observing the size of the test-pulse current. If successful, a giga-ohm seal should form (Fig. 3d). The quality of the subsequent recording is normally directly correlated with the quality of the giga-ohm seal (at least 5 GΩ is desirable). In cases where the dendrite moves upon release of pressure (because the tissue expands again), it can be helpful to gently withdraw the pipette tip during seal formation, helping to insure continuous solid contact. If a giga-ohm seal doesn’t form, reject the seal, change the patch pipette and try again, on another cell if necessary. ? TROUBLESHOOTING

• (22)

  Once a stable giga-seal has formed, withdraw the electrode slightly, so that it is not pressing into the dendrite. Cell-attached, perforated or inside-out patch recordings can be made starting at this point: continue on to options A, B or C, respectively. For making whole-cell or outside-out recordings, continue on to the next steps.

  A second dendritic recording could also be made at this point (before visibility of the structure has deteriorated with time or dialysis of the cell); alternatively it can be made after breaking into the first dendrite (as break-in can be difficult after a long period spent cell-attached).

  • (A)
    Cell-attached patch recording
    Cell-attached patch recordings can be made using conventional techniques^4,46 as soon as the dendritic giga-ohm seal has formed. If current densities are being measured, care should be taken to keep the patch pipette tip size constant, reducing one possible source of variation of the membrane patch size. It is often useful to measure the dendritic resting potential at the end of the experiment in order to determine the transmembrane potential. This can be achieved by rupturing the patch and immediately recording the resting membrane potential (before significant wash-in of the patch pipette solution).
  • (B)
    Perforated patch recording
    If a perforating agent has been included in the patch-pipette solution, such as nystatin or gramicidin⁴⁰, then a perforated-patch recording can be made simply by waiting for the perforating agent to form channels. Electrodes should be tip-filled with solution not containing the perforating agent to permit high-quality seals to form. Electrode stability is of particular importance when making perforated patch recordings from dendrites given the long time usually required for perforation (typically 10–20 min).
  • (C)
    Inside-out patch recording
    Inside-out patches can be made using standard techniques⁴⁶ from the starting point of a cell-attached recording by slowly and gently removing the pipette from the dendrite. Care should be taken to detect vesicle formation (which can be rectified by briefly exposing the patch to air).
• (23)

  To enter the whole-cell recording mode, break into the cell by rupturing the membrane beneath the pipette tip using strong, brief suction (Fig. 3e). ▲ CRITICAL STEP Start gently; if ‘break-in’ isn’t achieved, gradually increase the strength (but not duration) of the suction pulse. An alternative approach is to use a gentle ‘‘ramp’’ of suction, terminating immediately on break-in. If break-in using suction is unsuccessful, another method is to use strong hyperpolarization (e.g. below –120 mV), to cause dielectric breakdown of the membrane below the pipette. In this case, switch to current clamp immediately upon break in; otherwise the entire cell membrane will break down. ? TROUBLESHOOTING

• (24)

  After breaking into the cell, monitor the amplitude of the test-pulse current for a few minutes; if it starts to decrease (indicating an increase in the access resistance) gently try to clear the electrode tip using brief suction. A stable, low access resistance achieved at this point is more likely to remain low for a longer time.

• (25)
  At this point, you can make a whole-cell patch-clamp recording (A), or proceed with outside-out patch formation (B).

  • (A)
    Whole-cell patch recording
    During whole-cell recordings, continuously monitor the position of the pipette relative to the dendrite, as well as the access resistance. It is possible to accommodate slight movements of the dendrite (e.g. due to swelling of the slice) by using the micromanipulator to gently maintain the position of the pipette on the dendrite.
    Monitor and compensate for the access resistance of the pipette by applying a test pulse and adjusting your amplifier’s bridge balance and capacitance compensation circuitry. Deteriorating access resistance can sometimes be improved by using brief, gentle suction or positive pressure (although this can also destroy the recording and even the dendrite). Note that the relatively high access resistance associated with dendritic recordings (typically >20 MΩ) means that bridge balance and capacitance compensation (in current clamp) must be scrupulously applied. In voltage clamp, expected series resistances are high, and likely beyond accurate compensation. This, together with the space clamp errors associated with recordings from neurons with dendrites, makes whole-cell dendritic voltage clamp imprudent for most circumstances, although offline compensation can help overcome many of these limitations¹¹.
    If access resistance is sufficiently high (e.g., >50 MΩ) and cannot be improved, the recording should be abandoned. Recordings should also be terminated if seal resistance deteriorates (as indicated by depolarization of the membrane potential towards 0 mV, together with a large decrease in apparent input resistance).
  • (B)
    Outside-out patch recording
    An outside-out dendritic patch can be made using conventional techniques^4,46. Briefly, this involves gently withdrawing the pipette from the dendrite, and monitoring the gradually increasing access resistance using a test pulse.
    Continued access to the dendrite can also be detected by monitoring dendritic action currents (for dendritic locations with substantial backpropagating action potentials).
    A sudden increase in access resistance while maintaining a high patch resistance indicates that the elongating tube of membrane connecting the pipette to the dendrite has broken and sealed over to form an outside-out patch. ? TROUBLESHOOTING
• (26)

  Simultaneous double dendritic (or double somatic-dendritic) recordings can be carried out from the same neuron by repeating the above process with a second electrode. Care must be taken to minimize vibrations in the setup when mounting the second electrode and inserting it into the tissue. One way this can be achieved is by lowering both electrodes at the same time into the chamber until they are both above the neuron of interest, and then making seals in quick succession (this also minimizes problems of dendritic membranes becoming less visible with time spent whole cell, possibly due to dialysis of the cell with internal solution). Alternatively, the setup should be optimized to minimize vibrations when changing/inserting electrodes (e.g., transfer of vibrations during activation of mechanical manipulators).

● TIMING

Step 1, preparing equipment: 30–60 min

Steps 2–6, slice preparation: 20–30 min

Step 7, slice incubation: 30–60 min

Steps 8–11, slice transfer and dendrite visualization: 10–30 min

Steps 12–26, establishing giga-seal and relevant patch-clamp configuration: ca. 10 min (10–20 min for perforated patch-clamp) Step 22/25, recording: 10–40 min (depending on conditions)

Repeat Steps 9–26 (and 8 as necessary) while healthy dendrites can be seen in the slices (3–5 h)