All electrophysiological experiments in the current study were conducted using a patch-clamp experimental system in the Department of Physiology and Biophysics, University of Illinois at Chicago from 2005 to 2006.

Animals used in this study were treated in strict accordance with the American Physiological Society’s Guiding Principles in the Care and Use of Animals and the U.S. National Institutes for Health Guide for the Care and Use of Laboratory Animals. The protocol for all experimental methods was approved by the Institutional Animal Care Committee of the University of Illinois at Chicago. Fisher rats (F-344; 14 - 18 days old; male and female) were decapitated and the brains quickly removed. The brains were placed in an ice-cold cutting solution (in mM: 220 sucrose, 2.5 KCl, 2.4 CaCl₂, 1.3 MgSO₄, 1.24 NaH₂PO₄, 26 NaHCO₃ and 11_D-glucose), which was constantly bubbled with 95% O₂ and 5% CO₂. Transverse brain slices (400 μm thick) were made on a Vibratome (Series 1000 plus, St. Louis, MO). Brain slices were incubated for 30 min in artificial cerebrospinal fluid (ACSF) (in mM: 126 NaCl, 2.5 KCl, 2.4 CaCl₂, 1.3 MgSO₄ 1.24 NaH₂PO₄, 26 NaHCO₃ and 11 _D-glucose; osmolarity 300 mOsm), which was constantly bubbled with 95% O₂ and 5% CO₂ at room temperature (23˚C - 25˚C). The brain slices were then incubated in a HEPES-buffered external solution (see the next section) containing papain (15 - 18 U/ml) at 32˚C - 36˚C for 20 - 26 min. After papain treatment, the brain slices were further incubated in ACSF for 20 - 40 min. VTA neurons were dissociated with a vibrating stylus apparatus for dispersing cells from the brain slices as previously described [9] . Once the cell dissociation procedure was completed (4 - 7 min), the brain slice was removed from the culture dish, and the dissociated neurons usually settled and adhered to the bottom of the dish within 20 min.

In current-clamp recording, nystatin-perforated patch recording was employed as previously described [8] . Microelectrodes were fabricated on a P-87 puller (Sutter Instrument Company, Novato, CA) from glass capillaries (LE16; Dagan, Minneapolis, MN) and heat-polished on a microforge (Narishige, Tokyo, Japan). The tip resistances of the electrodes were 1.5 - 3 MΩ when filled with pipette solution (in mM: 60 K-acetate, 60 KCl, 1 CaCl₂, 2 MgCl₂ and 40 HEPES; pH adjusted to 7.2 with KOH, final [K⁺]_i = 131 mM; 290 mOsm). Nystatin was dissolved in methanol at a concentration of 10 mg/ml. This nystatin stock solution was diluted with pipette solution to a final concentration of 100 - 200 μg/ml and the electrodes were backfilled with this solution. A HEPES- buffered extracellular solution contained (in mM) 145 NaCl, 2.5 KCl, 2 CaCl₂, 1 MgCl₂, 10 HEPES and 11_D- glucose. The pH was adjusted to 7.4 with NaOH and the osmolarity was 310 mOsm. The HEPES-buffered solution was constantly bubbled with 100% O₂. In some experiments, recording was done in a Ca²⁺-free external solution, in which Ca²⁺ was replaced by equimolar Mg²⁺ to block Ca²⁺ current. The liquid junction potentials between the pipette and extracellular solutions were estimated to be 5 mV [49] and the results have been corrected by this amount. The osmolarity of the solutions was measured by a Vaprovapor pressure osmometer (Wescor Inc., Logan, UT). Electrophysiological measurements were performed using an Axopatch-1B patch clamp amplifier (Axon Instruments, Union City, CA). Membrane currents and voltage were filtered at 1 kHz and acquired at a sampling frequency of 10 kHz. Data acquisition was performed by a DigiData 1322A interface and pClamp software version 9.0 (Axon Instruments). The dissociated VTA neurons were visualized under phase-contrast on an inverted microscope (Diaphot 300, Nikon, Tokyo, Japan). All experiments were performed at room temperature (23˚C - 25˚C).

In voltage-clamp recording of persistent Na⁺ currents, conventional whole-cell recording was employed with the pipette solution (in mM: 60 Cs-methanesulfonate, 80 CsCl, 5 tetraethylammonium (TEA)-Cl, 4 MgATP, 0.3 Na₂GTP, 5 EGTA and 20 HEPES; pH adjusted to 7.2 with Tris-OH; 310 mOsm). The extracellular solution to record persistent Na⁺ current contained (in mM) 100 NaCl, 40 TEA-Cl, 5 KCl, 2 CaCl₂, 1 MgCl₂, 1 BaCl₂, 1 CsCl, 0.2 CdCl₂, 0.1 NiCl₂, 10 HEPES and 11_D-glucose. The pH was adjusted to 7.4 with Tris-OH and the osmolarity was adjusted to 320 mOsm with sucrose. The extracellular solution was constantly bubbled with 100% O₂. In some experiments, recording was done in a Ca²⁺-free external solution, in which Ca²⁺ was replaced by an equimolar Mg²⁺. The liquid junction potentials between the pipette and extracellular solutions were estimated to be 8 mV and the results have been corrected by this amount.

Neurons were continuously bathed in the external solution and drugs were dissolved at their final concentration in the same solution. Drug solutions were applied via a multiple channel manifold (MLF-4; ALA Scientific Instruments, Westbury, NY). Each channel of the manifold was connected to a gravity-fed reservoir with tubing (860 μm, i.d.). The output of the manifold was connected to an outflow tube (500 μm, i.d.), the tip of which was placed within 200 μm of the soma of the recorded neuron. Solution flowed continuously through one manifold channel. Application of drug solutions was controlled by opening or closing valves connected to the reservoirs.

The following drugs and chemical agents were used in this study, apamin, BaCl₂, CsCl, CdCl₂, EGTA, HEPES, nystatin, riluzole, TEA-Cl and tetrodotoxin (TTX) were purchased from Sigma?Aldrich (Saint Louis, MO). Papain was purchased from Worthington (Lakewood, NJ). Riluzole was dissolved in dimethyl sulphoxide (DMSO) and made as a stock solution at a concentration of 100 mM. The stock solution for riluzole was diluted in the HEPES-buffered external solution. The final concentration of DMSO in the HEPES-buffered external solution was 0.01%.

Action potentials (APs) were analyzed offline with pClamp 9.0 software (Axon Instruments Inc.). To assess the change of spontaneous firing with drugs, a 60-s-long segment during the peak of the drug response was analyzed as previously described [50] . Inter-spike interval (ISI) histograms were created as previously described [51] . The number of bins was equal to the square root of the number of ISIs. Bin width was obtained by dividing the ISI range (maximum ISI minus minimum ISI) by the number of bins. The coefficient of variation (CV) of the events was obtained by dividing the standard deviation of the event intervals by the mean event intervals. Burst duration was measured between the first spike and the last spike of a spike train. The ISI during burst-like firing was obtained by dividing the burst duration by the number of spike intervals in the burst spike train. The plateau potential, on which the trains of spikes were generated, was measured from the bottom of membrane potential to the threshold of the first spike. The data from cells with AP amplitudes less than 50 mV were discarded. All average values are expressed as mean ± standard error of the mean (S.E.M.). Statistical comparison to assess significant differences was done by the paired or unpaired Student’s t-test. To analyze persistent Na⁺ current activation, the current was normalized to the maximal current and fitted by the Boltzmann equation, using Origin software version 7 (Origin Lab, Northampton, MA):

where y is the current at the test voltage of x, A₁ is the minimal current and set to be 0, A₂ is the maximal current and set to be 1, x₀ is the membrane potential for half-activation and dx is the slope factor.

Figure 1(A) shows typical spontaneous membrane activity before, during, and after the application of 1 μM TTX in DA VTA neurons. Before TTX treatment, the DA VTA neuron exhibited spontaneous firing (Figure 1(A)-a). TTX hyperpolarized the membrane potential and abolished spontaneous firing (Figure 1(A)-b). After the washout of TTX, AP generation reappeared (Figure 1(A)-c). Figure 1(B) shows typical spontaneous membrane activity before, during, and after the application of 10 μM riluzole, a voltage-dependent Na⁺ channel antagonist, in DA VTA neurons. Before riluzole treatment, the DA VTA neuron exhibited spontaneous firing (Figure 1(B)-a). Riluzole hyperpolarized the membrane potential and abolished spontaneous firing (Figure 1(B)-b). After the washout of riluzole, AP generation reappeared (Figure 1(B)-c).

To ascertain the presence of persistent Na⁺ currents in DA VTA neurons, we used a ramp-voltage protocol with potassium-based pipette solutions and physiological extracellular solutions Figure 2(A₁) shows typical currents before and after treatment with 1 μM TTX and TTX-sensitive currents were obtained by the digital subtraction of TTX-insensitive currents from control currents. Figure 2(B₁) shows typical currents before and after treatment with 10 μM riluzole and riluzole-sensitive currents were obtained by the digital subtraction of the riluzole-insensitive currents from control currents. The riluzole-sensitive inward current provided a similar current-

voltage relationship to the TTX-sensitive current. Figure 2(C) shows the current density-voltage relationship of riluzole-sensitive currents in DA VTA neurons.

We next examined the current-voltage relationship of persistent Na⁺ current by a voltage-step protocol using cesium-based pipette solution. To identify DA VTA neurons, cell-attached recording was performed before making conventional whole-cell configuration. In the 6 DA VTA neurons examined, average spontaneous firing frequency was 1.5 ± 0.2 Hz. Figure 3(A₁) shows currents induced by the voltage-step protocol before and after application of 2 μM TTX in the control extracellular solution. Figure 3(A₂) shows currents induced by the voltage-step protocol before and after application of 2 μM TTX in the Ca²⁺-free extracellular solution. The graph in

Figure 3(B) shows the current-voltage relationship of TTX-sensitive persistent Na⁺ currents in the control extracellular solution and the Ca²⁺-free extracellular solution. Figure 3(C) shows the normalized persistent Na⁺ current plotted as a function of test voltage before and after the elimination of Ca²⁺ from the extracellular solutions. In the control extracellular solution, average half-activating voltage (V_{act, 0.5}) was −53.6 ± 0.7 mV and average slope factor was 3.4 ± 0.2 (n = 6). In the Ca²⁺-free extracellular solution, average V_{act, 0.5} was −57.1 ± 0.7 mV and average slope factor was 3.7 ± 0.2 (n = 6). The elimination of Ca²⁺ from the extracellular solution significantly shifted the V_{act, 0.5} in a negative direction (P < 0.001) without changing the slope.

We examined the effect of Ca²⁺-free extracellular solution, which potentiated persistent Na⁺ current, on the firing of DA VTA neurons. Figure 4(A₁) shows typical membrane activity before, during, and after the application of Ca²⁺-free extracellular solution with or without continuous hyperpolarizing current injection. In the control extracellular solution, the spontaneous firing frequency was 1.6 Hz (Figure 4(A₂), a). The Ca²⁺-free extracellular solution depolarized the membrane potential and increased the firing frequency to 7.6 Hz (Figure 4(A₂), b). When a continuous hyperpolarizing current was injected to counter the depolarizing driving force, burst-like firing was generated in the Ca²⁺-free extracellular solution (Figure 4(A₂), c). When the continuous hyperpolarizing current injection was removed and the Ca²⁺-free extracellular solution was switched to the control extra-

cellular solution, the firing frequency became 1.7 Hz (Figure 4(A₂), d). Figure 4(B) shows APs in control and burst-like firing in a faster time scale. In the burst-like firing, a train of spikes rode on a plateau potential and the spike amplitude gradually decreased during the spike train (Figure 4(B)). As shown in Figure 4(C), the firing patterns before and after burst-like firing were analyzed by ISI histograms. The ISI is normally distributed with a peak at 590 ms in control (Figure 4(C₁), while the ISI histogram skewed leftward with a wider ISI distribution during the burst-like firing (Figure 4(C₂). In 11 neurons out of 16 DA VTA neurons (69%), burst-like firing was generated in the Ca²⁺-free extracellular solution, when continuous hyperpolarizing current was injected (−48.5 ± 5.6 pA); the average CV of firing frequency was 1.62 ± 0.17. The burst-like firing parameters were as follows: burst duration, 826.5 ± 105.2 ms; spike number in the train, 4.9 ± 0.2; ISI, 221.2 ± 18.9 ms; plateau potential, 10.7 ± 0.6 mV (n = 11). As illustrated in Figure 4(D), the burst-like firing induced by the continuous hyperpolarizing current injection in the Ca²⁺-free extracellular solution was abolished after treatment with 10 μM riluzole; average membrane potential after treatment with riluzole was −62.3 ± 1.4 mV (n = 4).

In Ca²⁺-free extracellular solution, reduced Ca²⁺ influx through voltage-dependent Ca²⁺ channels decreases Ca²⁺- activated K⁺ currents. If this occurred in DA VTA neurons, the counter effect of SK currents would oppose decreases in the depolarizing driving force and burst firing may be induced as previously described in midbrain dopamine neurons [28] [29] . To represent the condition of blocking both voltage-dependent Ca²⁺ current and SK current in Ca²⁺-free extracellular solution, we used nickel to produce concurrent blockade of T-type Ca²⁺ channels and SK channels, since these two ion channels are strongly coupled in midbrain dopamine neurons [30] . We ascertain that blockade of SK channels by 200 n Mapamin reduced AHP in DA VTA neurons (Figure 5(A)). In

a similar manner, Ni²⁺ (100 μM) reduced AHP, which is attributable to SK currents (Figure 5(B)). Before nickel treatment, the spontaneous firing frequency was 1.7 Hz (Figure 5(C₂), a). The firing frequency was decreased and firing pattern became irregular in the presence of 100 μM Ni²⁺ (Figure 5(C₂), b). When a continuous hyperpolarizing current was injected during nickel treatment, the firing frequency was decreased and abolished without inducing burst firing (Figure 5(C₂), c). After the continuous hyperpolarizing current injection was removed and Ni²⁺ was washed out, the spontaneous firing frequency became 1.6 Hz (Figure 5(C₂), d). In 5 DA VTA neurons, continuous hyperpolarizing current injection with nickel treatment failed to generate burst firing; the average continuous hyperpolarizing current was −20.0 ± 3.5 pA.