Pulse-mode scanning ion conductance microscopy—a method to investigate cultured hippocampal cells

https://doi.org/10.1016/S0165-0270(02)00023-7Get rights and content
Full text access

Abstract

Scanning ion conductance microscopy (SICM) takes advantage of the increase in the resistance which occurs if a glass microelectrode is closely approached to a poorly conducting membrane (Science 243 (1989) 641) and has been shown to be a promising technique to study membranes of living cells (Biophys J 73 (1997a) 653; J Microsc 188 (1997b) 17). Based on a newly designed set-up on top of an inverted light microscope in combination with a speed optimized low noise intracellular amplifier, a novel mode for control of the distance between the probe and surface has been developed. By application of current pulses, the change in the resistance is monitored independently from electrode drift and parasitic DC currents. We demonstrate the applicability by showing first high-resolution images of neural cells produced with the pulse-mode operated SICM.

Keywords

Scanning ion conductance microscopy (SICM)
Cell culture
Neurons

1. Introduction

The scanning ion conductance microscope (SICM) was first described by Hansma et al. (1989). The SICM belongs to the family of scanning-probe-microscopes (SPM) with the common feature that a fine probe is scanned over the sample surface, while the distance between probe and surface is held constant. The interactions between the probe and the sample surface depend on the method used, for example a mechanical force is recorded in atomic force microscopy (AFM) or a current in scanning tunneling microscopy (STM). In SICM, an electrolyte filled glass microelectrode is approached towards a surface and the resistance changes that occur when the electrode approaches a high resistance surface in the bath electrolyte are measured. Distance control depends on the actual value of the resistance in comparison to the resistance at infinite distance between tip and sample. The electrode is arrested at the point in z-direction were the resistance changes by a given value. Recording sequentially the positions of z-arrest at every point of a given x–y plane results in a topographical image. Positioning of the probe is usually achieved by means of piezoelectric elements or step motors.
Recently, it was also shown that the SICM can be combined with whole cell patch clamp recording to measure ion currents through topographically resolved membrane pores in heart cells (Korchev et al., 2000). The main advantage of the SICM with respect to other scanning probe methods, such as AFM, is the ease of use combined with the ability to scan living cells within growth medium, since no special preparation of the scanned cells is necessary (Korchev et al., 1997a, Korchev et al., 1997b). Another essential aspect is the ability of the SICM to scan a sample without having physical contact between probe and sample surface. This promises that this technique will allow one to measure dynamic processes, such as cell growth and differentiation, over long periods of time without damaging or disturbing the cell.
Here we established a pulse mode of height control in order to obtain more stable measurements over long time periods. The basic feature of this mode consists in the application of a test pulse that eliminates the effects of electrode drift over time. We found that this drift, which cannot be totally eliminated because of temperature shifts during scanning sessions in the magnitude of hours, can often cause a premature arrest of the scanning procedure due to erroneous resistance readings. Further, the SICM technique can be combined with other electrophysiological investigations which are coupled to the application of currents or drugs that can cause additional steady state potential shifts on the electrodes which could disturb the control of the z-position. In the following, we describe a new version of an SICM using a repetitive application of test pulses to measure the access resistance of the electrode. This new approach is highly robust and suitable for long term experiments.

2. Methods

To test the function of the microscope on living cells, hippocampal neurons were cultured as described by Lessmann et al. (1994) and Potthoff and Dietzel (1997). Hippocampi were removed from postnatal Wistar rats (P0–P5) and cut into 1 mm thick tissue blocks. They were collected in ice-cold phosphate-buffered saline (PBS) supplemented with 10 mM HEPES, 1 mM pyruvate, 10 mM glucose and 2 mM glutamine. Some 100 ml of this solution were supplemented with 0.1 ml of 6 mg/ml DNAse stock, 100 mg bovine serum albumin, 250 μl of a stock solution of 104 U/ml penicillin and 10 mg/ml streptomycin, 5 mg Phenol red, pH was adjusted to 7.3 with NaOH (MPBS). MPBS+ contained, in addition, 0.9 mM CaCl2 and 0.5 mM MgCl2. After washing twice in ice-cold MPBS+ and once in MPBS, the tissue was incubated for 8–15 min at 37 °C in MPBS containing 0.25% trypsine. Cells were dissociated by trituration using plastic pipette tips and suspended in 5 ml DMEM (Gibco) supplemented with 10% fetal calf serum (FCS). The suspension was centrifuged at 160×g for 5–8 min at 4 °C. After resuspension of the pellet in 2 ml ice-cold DMEM supplemented with 10% FCS, cells were plated at densities of 50 000 cells per dish in d,l-polyornithine-coated (1 mg/ml in borate buffer) 3.5 cm diameter plastic culture dishes in 1 ml DMEM plus 10% FCS and incubated at 37 °C in a humidified atmosphere containing 5% CO2. When cells had attached to the dishes after 24 h, the medium was exchanged to serum free B18 medium (Brewer and Cotman, 1989).

3. Results

The SICM was constructed on the basis of an inverted Zeiss microscope with a holder for 35 mm cell culture dishes, standing on a solid granite platform. This assembly was positioned on a polystyrene foam socket placed in a Faraday cage (Dietzel et al., 1992).
The microelectrode holder was mounted at an angle of 90° to the surface of the culture dish in order to allow for a maximal exposure of the cell surface to the electrode tip. Hence, the condensor of the microscope was replaced by a glass fiber optics illuminating the culture dish at an angle of ≈45° from the side. Although the image did not reach the clarity of the original phase contrast optics, it was sufficient to manually position the microelectrode within the scanning region.
The electrode position was coarsely adjusted by a three-axis micromanipulator (OWIS, Staufen, Germany) and the fine position was achieved using computer controlled 60 μm piezoelectric elements (PSt 150/5/60 Gh10, Amplifier SQV3/150, both Piezomechanik GmbH, München), assembled as described by Hengstenberg et al. (2000). The configuration of manipulators and head stage is illustrated in Fig. 1.
  1. Download: Download full-size image

Fig. 1. Schematic illustration of the experimental setup. The electrode holder is mounted directly above the scanning region on a solid metal plate. This plate is connected to a micromanipulator, which is driven by piezoelectric elements. Illumination is achieved by means of a light conducting fiber bundle, aiming at the scanning region at an angle of ≈45°.

The electrode plus access resistances were measured with a current clamp amplifier, as routinely used in electrophysiological recordings (Jens Meyer Technical Developments, Bochum, Germany). The current clamp feature of the amplifier permitted the generation of currents of predetermined amplitude and duration through the microelectrode. Glass electrodes were pulled in two steps from borosilicate glass tubing (GB 150-8P, o.d. 1.5 mm, i.d. 0.86 mm, Science Products, Hofheim Germany) using a Narishige PP-830 puller. The opening diameter of the electrode tips was estimated to be in the range between 200 and 500 nm. The electrodes were filled with a solution containing 145 mM NaCl, 3 mM CaCl2, 4 mM KCl, 10 mM glucose, 10 mM HEPES and 2 mM MgCl2 with a pH of 7.3, mimicking the ion concentration in the extracellular space.
For the operation of the SICM a Visual Basic program was developed. This software controls the positioning of the electrode, application of the test pulses and measurement of the access resistance of the electrode. Measured data points are stored in a file and then processed by a self-developed program. Processed data were displayed using the freeware raytracing program PovRay.
The distance control mechanism of the microscope consists in the application of constant current pulses through the electrode and measurement of the resulting voltage drop across the electrode and access resistance (Fig. 2A). The application of short test current pulses eliminates potential errors due to slowly developing DC potentials at the electrode tip. For the determination of each z-position, the electrode is advanced in steps of a predetermined size, that is adjusted to larger values if electrodes with larger opening diameter and smaller resistance are used. Useful step sizes were in the range of 10–100 nm. After each step, a current pulse is applied and the resistance is calculated. The microscope can also be operated in an alternative mode, using a patch-clamp amplifier, applying voltage pulses and advancing the electrode tip until the current decreases by a given factor. Although this mode reduces the settling time of the electrodes and thus enhances scanning speed, it cannot be used in all applications, for instance, if additional ion activity measurements with double-barreled electrodes are performed while scanning the surface. Thus, we implemented an algorithm to dynamically calculate the optimal current-pulse length to speed up the scanning procedure in these cases.
  1. Download: Download full-size image

Fig. 2. (A,B) Schematic illustration of the principle of the pulse-mode SICM. Pulses of constant current amplitude are passed through an electrolyte filled electrode. The approximation of the electrode tip to the membrane surface is sensed by an increase in the voltage drop between amplifier and a distant reference electrode in the bath solution. (C,D) Influence of the time constant of the electrode on the voltage signal after application of a constant current pulse. The trace in (C) shows the behavior of an electrode with a resistance of 106 Ω, while the trace in (D) shows the behavior of an electrode with a resistance of 108 Ω. Ten voltage readings are digitized and then averaged to minimize noise (symbolized by the gray grid on the trace). Because a voltage measurement before fully charging the electrode capacitance will result in erroneously low voltage readings (time points of too early voltage readings symbolized by gray grid), the pulse duration was automatically adjusted to perform the voltage recording after the electrode had been charged (see text and gray trace in (D)). In practice, pulses in the range from 500 to 1000 μs are needed.

The minimal time necessary for the reliable resistance measurement depends on the time constant of the electrode. Hence, electrodes with larger tip diameters and smaller resistances show a shorter time constant compared to smaller electrodes with higher resistances which need a longer time for recharging after application of a current step (Fig. 2C,D). To keep the scanning time as short as possible, the resistance is optimally measured as soon as the electrode capacity has been recharged and this time point depends on the tip diameter of the scanning electrode. We hence designed software to find the optimal test pulse length by calculating regression lines through the voltage traces during the rising phase of the voltage drop across the electrode. Recharging is complete when the slope of the regression lines approaches zero:(1)B=∑inti−t̄Ui−Ū∑inti−t̄2⇒0where B is the slope of trace; t is the time basis; t̄ is the average of time basis; U is the voltage difference; and Ū is the average of voltage differences.
As soon as a constant voltage is reached, several successive voltage values are measured and averaged to reduce noise. The number of voltage readings averaged can be set by the experimenter depending on the digitization rate and frequency of the voltage noise. At this time point, the constant current pulse is terminated and following the decline of the voltage transient, controlled by regression line calculations, ten more points are averaged to determine the baseline voltage. The difference between the test and baseline voltages, which is proportional to the access resistance at a given point in space is compared to the access resistance at a long (virtually infinite) distance from the surface of interest. If the resistance increases by a value set by the experimenter (in practice 3–10% is a useful setting) the electrode movement in z-position is stopped and the corresponding z-position is stored. The electrode is now elevated and then translated laterally, the size of the ‘backstep’ depending on the size of the lateral steps. We found that setting the backstep twice the size of the lateral steps leads to good results. Choosing higher values increases scanning time of the cells without increasing the safety factor for not touching the membrane. The test pulses are repeated and the electrode is approached to the membrane at the new point of the x–y-plane.
Images of living cells were taken from cells grown in dissociated culture from the hippocampal region of postnatal rats (Fig. 3). The scanning electrode used for the images shown in Fig. 3 had a resistance of 6 M. The x–y plane was scanned using increments of 500 nm between lateral spots in Fig. 3(A). The surface of the cell shown in Fig. 3(B) was scanned using the same electrode, the increment between two points was 250 nm. In both cases, scanning time was ≈30 min.
  1. Download: Download full-size image

Fig. 3. Images of hippocampal cells from postnatal rat brain. (A) The scanned region is 30×30 μm, lateral step size 500 nm, scanning time 30 min, pulse height 2 nA, electrode resistance 5×106 Ohm. The electrode was arrested at a resistance change of 3%. The cell body is clearly shown and neurites start to become visible at this resolution. (B) The scanned region is 15×15 μm, lateral step size 250 nm, scanning time 30 min, pulse height 2 nA. Cellular processes are clearly visible with this resolution.

4. Discussion

Here we present a variant mode of scanning ion conductance microscopy using constant current pulses to control the distance between cell and electrode tip. By backstepping the electrode tip twice the distance of the lateral steps, we succeeded in scanning the contour of an entire hippocampal neuron on a culture dish.
Previous applications of SICM used a feedback system which keeps the current flow between bath and electrode constant. The present mode extends the applicability of the SICM-method in several respects: first, it allows one to investigate cells even if the soma or processes do not smoothly fit to the culture dish surface. In this case, the ‘constant distance mode’ inevitably leads to situations where the electrode shaft runs into the overhanging membrane. In addition, the pulse application method used in our system could significantly improve the stability of the scanning procedure. Finally, the new method allows one to obtain relatively fast low resolution preview scans of a cell of interest by using step sizes of ≈500 nm. In contrast to the constant distance mode, where smaller tip sizes demand smaller scanning steps, the backstep-mode allows a low resolution scan even with smaller electrode tips.
In the present experiments, the electrode was arrested at a 5% increase in resistance. This change in resistance occurs at a distance of ≈100 nm from the bottom of the dish, as estimated from the approximation curves using the same electrodes. This distance avoids contact with surface charges of the membrane and adhesion of glycosylated protein residues to the electrode tip. In the present scanning mode, positive voltage changes were applied to the electrode, avoiding a possible stimulation of the neurons.
Since, in contrast to AFM-techniques, close contact with the membrane can be avoided, we presently succeeded in repeating a line scan over a cell several times without visibly damaging its surface. If one is interested in the perfect visualization of the membrane topography, a shortcoming is that this procedure displays an ‘equi-resistance’ surface, rather than the true topology of the cell. Hence, the membrane grooves seen in Fig. 3 could be due to slow movements of the membrane, reflecting interactions of the membrane with the cytoskeleton as well as areas of decreased membrane resistance, due for instance to active transporters or ion channels.
In its present mode of operation, this tool permits one to monitor the surface of living cells during a time course of minutes to hours. In a next step, we will increase the resolution to investigate finer details not resolved with the electrodes used in the present paper, such as processes or synapses. Repeated scans of the same membrane area with a higher resolution will also provide more information about the nature of the membrane ripples seen in Fig. 3.

Acknowledgements

We are grateful to R. Heumann, U. Dodt, W. Zieglgänsberger and J. Meyer for helpful discussions and support with equipment. The project was supported by the German ‘Hochschulsonderprogramm II’, a graduate student fellowship of Nordrhein-Westfalen to G.H. and the DFG grant Schu-929/5-1.

References

View Abstract
Copyright © 2002 Elsevier Science B.V. All rights reserved.