Neuroscienceand Behavioral Physiology, VoL 27, No.

6, 1997

USE OF ARTIFICIAL ION CHANNELS FOR

QUASI-INTRACELLULAR RECORDING OF
CEREBRAL CORTEX NEURON ACTIVITY

M. Yu. Inyushin, V. Yu. Tsytsarev, A. Yu. Ignashchenkova,

and D. N. Lenkov

Action potentials and synaptic potentials were recorded in vivo from cortical neurons in baby rats aged 20-25
days using a new method based on the ionophore nystatin. Nystatin solution was used to fill a standard glass
extracellular microelectrode, and became inserted into membranes. Spikes winch were initially recorded as
extracellular spikes showed increases in amplitude and were transformed into unipolar quasi-intracellular
spikes at 0.5-5.0 min after formation of the high-conductance contact. This method allows stable recording of
neuronal activity from cells for at least I h, and provides a good signal-to-noise ratio. The electrode does not
puncture the neuron membrane, with the result that experiments do not require any isolation ,from vibration.
Thus, the results obtained demonstrate that the method is highly efficient for recording the in vivo activity of
small nerve cellsn.
Key words: Ion channels, neuronal activity, action potentials, rats.

Recording of neuronal activity in mammalian cortical cells, especially in early ontogenesis, is usually made difficult
by the small sizes of the neurons and the fragility of their membranes. Intracellular positioning of the electrode usually results
in rapid death of the cell, and extracellular recording is usually of low information value [4]. The use of quasi-intracellular
recording can significantly counteract methodological difficulties in a variety of experiments where other recording methods
are not possible or are difficult. In the present studies, we used a new method for recording the intracellular potential, using
artificial ion channels formed by polyene antibiotics, of which the best known are nystatin and amphotericin B. The polyene
chains of these agents insert into the lipid bilayer, forming artificial channels in membranes.
Nystatin is a mixture of antibiotics of the so-called nystatin complex. Nystatin does not form pores in artificial
membranes lacking ergosterol [16], which has been shown [17] to be necessary for maintaining the integrity of the nystatin
channel. Starting with the studies of Horn and Marty [10] in 1988, numerous reports have appeared in which nystatin has been
used in a variety of clamping methods. Horn and Marty [10] used a whole-cell configuration and the main argument in favor
of the new method was that it forms an electrical contact without loss of cytoplasmic components due to diffusion of the
intraelectrode medium. This allowed recording of responses to acetylcholine over a prolonged period of time; in the standard
method, effects of losses were seen within 10 min. In recent years this method has also been used for recording ion currents
in cortical neurons [13].
Further studies reported in 1993 [15] used clamping with amphotericin B instead of nystatin; this was found to produce
channels with lower resistance (up to 3-4 M~) in a variety of cell types, though most investigators preferred to use the more
extensively studied nystatin.
We started to use artificial ion channels based on amphotericin for recording spike and synaptic activity of nerve and
muscle cells in mollusks [1, 2]. The method allowed recording of synaptic potentials in small (8-10 /zm) muscle cells in
experiments involving intracellular stimulation of motoneurons in central ganglia. With some modifications, we have used a
basically similar technique for recording cortical neuron activity in developing rats.
A. A. Ukhtomskii Physiological Science Research Institute, St. Petersburg State University. Translated from
Fiziologicheskii Zhurnal imeni I. M. Sechenova, Vol. 82, No. 1, pp. 18-24, January, 1996. Original article submitted April
11, 1995.
702

0097-0549/97/2706-0702518.00 9

Plenum Publishing Corporation

Nystatin channel

I~Ul'on

Fig. 1. Use of nystatin channels for recording neuron

activity. A) Diagram showing the formation of an ion
channel by eight nystatin subunits (the number of subunits
may not be constant and may depend on the nystatin
concentration in the solution); B) diagram showing quasiintraceUular recording of neuron activity with a glass
mieroelectrode via artificial ion channels.

METHODS
Acute experiments were carried out on white mongrel male rats aged 10-25 days and weighing 17-40 g. Animals were
anesthetized with urethane, given i.p. at a dose of 1 g/kg; the surface of the neocortex running rostral from the bregma was
then exposed. Rats were then placed in a screened chamber and clamped by the parietal bone using "Akriloksid ~ self-hardening
plastic. Electrodes were prepared from "Pyrex" glass blanks using an Mt~-2 apparatus, and had tip diameters of 3-5 gm.
Neuron activity was recorded using an amplifier with integral amplitude and signal duration calibrators and an input
resistance of 1 Gfl.
Electrodes were filled with a nystatin solution; this was prepared immediately before use, as its channel-forming ability
lasts 1.5-2 h [10]. Solutions were prepared using nystatin obtained from E. R. Squibb & Sons Ltd. Aliquots of 5 mg were
dissolved in 0.5 ml of dimethylsulfoxide (DMSO), and the resulting solution was supplemented with 0.5 ml of 0.2 M potassium
chloride solution. Thus, the potassium ion concentration in the electrode was essentially the same as the potassium content in
the intracellular environment.
Instability of polyene antibiotic solutions can pose difficulties; these solutions retain activity for 1.5-2 h. Nystatin
solutions can be stabilized with antioxidants, hydroxyquinoline providing the greatest level of stability [15]. However, since
any given experiment with one electrode lasted no more than 2 h, no stabilizing additives were used here.
Prepared electrodes had an active resistance of 4-6 Mf~; the stability of this measure served as an indicator of electrode
integrity during experiments. Microelectrodes were moved using a manual micromanipulator. Electrode insertion was also
performed using a manual microscrew, without any additional mechanical stabilization, as required in standard extracellular
recordings. Oscilloscope monitoring of spontaneous and evoked activity was performed simultaneously.

703

Fig. 2. Neuron activity in the cerebral cortex of neonatal rats during quasiintracellular recording. A) An action potential from a corticalneuron after
mechanical contact of the electrode and membrane for 1 min; B) contact for
10 rain; 6") spontaneous discharges of a cortical neuron recorded using an
extracellular electrode containing nystatin solution; D) AP from a cortical
neuron of the primary motor projection of the forelimb evoked by sensory
stimulation. AP are inverted. The arrow shows the moment of stimulation.
Calibration: 5 mV, 2 msec.
Electrodes were inserted into the cortex until normal extracellular action potentials (AP) of characteristic amplitude
(0.5-2.0 mV) were detected from a neuron.
At the zone of contact between the microelectrode tip and the cell membrane, nystatin subunits then formed highconductivity ion channels on the cell surface (Fig. 1), the result of which was that the AP amplitude in most cases gradually
increased. During experiment, activity was monitored on the screen of an SI-83 oscilloscope, with recording on tape using an
FOR-2 light recording device; activity was also entered into a PC/AT-286 computer. Evoked neuronal activity was produced
by stimulation of the distal regions of the forelimbs using single square-wave impulses of duration 0.3-0.5 msec, passed through
metal electrodes of diameter 0.3 mm, inserted subcutaneously.

RESULTS
Eight experiments were performed on white mongrel male rats aged 10-25 days. Recordings were made from a zone
in the frontal cortex with coordinates of 1.0-1.5 mm rostral and 0.8-1.2 mm lateral to the bregma, at penetration depths of 0.2704

0.8 mm. Channel formation probably took from 30 sec to 5 min, since dynamic increases in AP were recorded over this time
period (Fig. 2, A, B); the amplitude of the recorded signal increased several-fold. Spike size increased smoothly from the
moment at which recording was started. In most cases, when the electrode contacted a cell, standard discharges of different
polarities and amplitudes of 0.5-2.0 mV were recorded; these were of extracellular origin, and were subsequently converted
into unipolar spikes.
Apart from AP, synaptic potentials of differing polarity were clearly seen. The maximum amplitude of AP was never
greater than 20-25 mV, which was some 30% of the membrane potential. However, all experiments showed notable increases
in the signal-to-noise ratio after formation of stable "nystatin" co ntact~ The spontaneous activity of a number of neurons
consisted of trains interrupted by long pauses lasting 15-30 msee (Fig. 2, 6").
..
The shape and duration of spikes remained constant throughout
" "tim' enttre
:- period of stable recording. Only the amplitude
underwent some variation within series of spikes in train-emitting neurons, with increases towards the ends of trains. Inverted
spikes were seen in some cases (Fig. 2, D). Discharge amplitude in these cases was no greater than 6 mV. AP had a nonstandard shape, with a pronounced pre-potential (Fig. 2) of 0.5-1.0 inV. The shape of neuron discharges was independent of
whether activity was spontaneous or evoked. Trace hyperpolarization was seen in all recordings; this was also inverted when
AP were inverted.
The period of stable activity recording from a given neuron using this method lasted 30-60 rain, after which the
duration of AP started to increase and the amplitude to fall; spike activity then ceased after 5-6 min. It was of note that despite
movement due to respiration and the pulse, the instability of the manual manipulator with its poor protection from vibration,
and the absence of any measures for isolation from vibration, the period of stable recording was quite prolonged.
Additional experiments showed that upward and downward movements of the head of the manipulator through distances
of 10-15/~m, produced using the microserew during the period of stable recording, had no negative effects on recordings.

DISCUSSION
The results presented here show that this method, based on the use of nystatin for recording neuron activity, showed
it to be a quite efficient method for performing microelectrode experiments on the neocortex of baby rats aged 10-25 days.
There is no doubt that nystatin acts as a channel-forming agent by interaction with the neuron membrane. In most of our
experiments, the amplitude of AP significantly and quickly increased and then stabilized at a defined level. The time required
for the increase in the signal varied quite considerably, though this can apparently be explained by differences in the rate of
diffusion of nystatin from the electrode tip, the chance of forming air locks in the electrode, and formation of inadequate contact
with the neuron membrane. Of note is the point that when the electrode resistance was relatively high (greater than 8 Mfl) and
this was not due to an air lock, membrane penetration had frequently occurred and the affected cell rapidly died. In this
situation, the signal amplitude increased to 60-70 mV for a few seconds and then dropped sharply to zero. This was probably
due to mechanical penetration of the cell rather than a side effect of the antibiotic.
Published data on membrane puncturing by nystatin are contradictory; no problems were encountered in the present
work in determining the molecular and biophysical aspects of the method. For example, in one report [6] it was maintained
that the clamping method using nystatin sometimes transforms into the classical method, because of entry of organic stains of
molecular weight greater than 800, including fluorescein, into cells. The report did not make it clear whether nystatin was
responsible for the penetration. However, another report [12] showed that nystatin provokes the transport of fluorescein into
ceils by a mechanism other than simple osmotic transfer. In [18], the proposal to use a mixture of'nystatin and sodium
fluorescein was made, using a molar ratio of 1:10, since nystatin is well soluble in fluorescein, which can be used as a
substitute for DMSO as the polar solvent. A resistance of 20 Mfl at the contact zone in clamping experiments was achieved
with this method. It was noted that fluorescein also penetrated the cell in this case, producing cell staining.
At this point it is relevant to note that most reports of clamping experiments assume that the nystatin channel is only
accessible to cations. In [9], mammalian tracheal epithelial cells were loaded with chloride ions via these channels. Most studies
of osmoregulation involve filling of cells via nystatin channels with tetramethylammonium, as in [7], and it is supposed that
water molecules cannot pass along the channel. It is not clear how this relates to data [11, 14] on the transport of large
molecules by nystatin. Additionally, in [8] it was demonstrated that addition of nystatin or amphotericin B solutions to washed
yeast results in sequential liberation of low-molecular-weight substances such as potassium ions, followed by higher-molecularweight substances such as phosphate; only then are proteins and other large molecules released. The possibility of osmotic
705

shock should be considered as a criticism of this work. However, the question of the diameter of the nystatin pore remains
unanswered. In [3] it was proposed that polyene antibiotics form a penetrating artificial ion channel though the membrane, of
diameter 40 nm and consisting of eight antibiotic molecules, and that channels do not form until a defined threshold
concentration of the antibiotic is reached.
Despite these contradictory data, there is no doubt of the fact that in~acellular potentials are being recorded in vivo
through electrodes filledwith nystatin solution. Placement of such an electrode against a cell,without penetration of the outer
membrane, evidently produces less trauma to the neuron than an intracellularelectrode, a point of particular importance in
studies of small ceUs in the mammalian cortex [4] at early ontogenetic time points. The amplitudes of the spikes recorded were
smaller than those obtained by normal intracellularrecording with neuron penetration [4]. W e suggest that this is a result of
simple shunting of the siLnml because of incompletely tight contact of the electrode tip with the cell surface.
It is probable that in some cases) recordings represented a combination of potentials from the external and internal
surfaces of the membrane. This is theoreticallypossible in conditions of inadequately tightcontact of the electrode and the cell
membrane. This type of state should be very unstable, and would probably be characterized by a nonstandard recorded signal
shape which would be difficultto interpret. W e believe the inverted signals in Fig. 2, D to represent this situation.
Another problem could result from the intrinsicpolarity of synaptic potentials,with regard to the possibilitythat these
can be inverted. However, judging from the polarity of inhibitory post-synaptic potentials (PSP) separating the trains recorded
from a train-producing neuron (Fig. 2, C), the polarityof PSP also corresponds to the polarity obtained in standard intracellular
recordings, at least in the case of noninverted AP. Further testing is needed for other cases.
However, the significantincrease in the signal-to-noiseratio without increases in the complexity of the electronic part
of the apparatus and the ability to make stable recordings of both spike and synaptic activity of neurons in vivo are clear
advantages of the new method.
The method of quasi-intracellularrecordings could probably be used in a variety of elcctrophysiological investigations
into the activityof small cellular elements, including glialcells,though the channel-forming effect itselfof nystatin and other
polyenes does require further study.

REFERENCES
.

3.
4.
5.
6.
7.
8.

9.
10.
11.

706

I. Yu. Artem'ev and M. Yu. Inyushin, "The use of artificial ion channels for recording the electrophysiological
parameters of nerve and muscle cells," in: Proceedings of the 10th All-Union Conference on Evolutionary Physiology
[in Russian], Leningrad (1990).
M. Yu. Inyushin and I. Yu. Artem'ev, "Recording of electrical potentials from cells through artificial ion channels,"
in: Ion Channels and Biological Membranes [in Russian], Moscow (1990).
Kh. M. Kasmov, The Molecular Mechanism of the Interaction of Polyene Antibiotics with Lipid Membranes [in
Russian], Baku (1988).
D. N. Lenkov and L. A. Vasireva, "The formation of synaptic processes in the sensorimotor cortex in kittens," in:
Neuronal Mechanisms of the Developing Brain [in Russian], Moscow (1979).
Yu. D. Shenin, V. V. Belakhov, T. N. Kontelova, et al,, "Stabilization of standard samples of Levorin and nystatin
using antioxidants," Antibiot. Klfimioterap., 36, No. 8, 20-24 (1991).
I. Chung and L. C. Schlichter, "Criteria for perforated-patch recordings: ion currents versus due permeation in human
T lymphocytes," Pfliiger's Arch., 424, No. 5-6, 511-525 (1993).
C. U. Cotton, A. M. Weinstein, and L. Reuss, "Osmotic water permeability of Nectorus gallbladder epithelium," J.
Gen. Physiol., 93, No. 4, 649-679 (1989).
M. A. de Resende ad F. Alterthum, "Effect of nystatin, amphotericin B and amphotericin B methyl ester on
Saccharomyces cerevisiae with different lipid composition," Mycopatologia, 112, No. 3, 165-172 (1990).
M. Haas and D. G. McBrayer, " N a - K - C I cotransport in nystatin-treated tracheal cells: regulation by isoproterenol
apical UTR and CLi," Am. J. Physiol., 266, No. 5, Part 1, 1440-1450 (1994).
R. Horn and A. Marty, "Muscarinic activation of ionic currents measured by a new whole-ceil recording method,"
J. Gen. Physiol., 92, No. 2, 145-159 (1988).
R. Negry, G. Costanzo, M. Buttinelli, et al., "Effects of DNA topology in the interaction with histone octamers and
DNA topoisomerase I," Biophys. Chem., 50, No. 1-2, 169-181 (1994).

12.
13.
14.
15.
16.
17.
18.

A. A. Nikiforov and I. B. Ostresova, "Stimulation of weak organic acid uptake in rat renal tubules by cadmium and
nystatin," Biochem. Pharmacol., 47, No. 5, 815-820 (1994).
T. Omura, M. Munakara, and N. Akaike, "Nystatin-perforated patch recordings disc lose KA-operated outward
currents in rat cortical neurons," Brain Res., 627, No. 2, 345-348 (1993).
M. Opekarova and W. Tanner, "Nystatin changes the properties of transporters for arginine and sugars. An in vitro
study," FEBS Lett., 350, No. 1, 46-50 (1994).
J. Rae, K. Cooper, P. Gates, and M. Watsky, "Low access resistance perforated patch recordings using amphotericin
B," J. Neurosei. Methods, 37, No. 1, 15-26 (1991).
B. S. Whyte, R. P. Peterson, and S. C. Hartsel, "Amphotericin B and nystatin show different activities on sterol-free
vesicles," Bioehem. Biophys. Res. Commun., 164, No. 2, 609-614 (1989).
D. J. Woodbury and C. Miller, "Nystatin-induced liposome fusion. A versatile approach to ion channel reconstitution
into planar bilayers," Biophys. J., 58, No. 4, 833-839 (1990).
H. Yawo and N. Chuhma, "An improved method for perforated patch recordings using nystatin-fluorescent mixture,"
Jpn. J. Physiol., 43, No. 2, 267-273 (1993).

707