Differences in ionic currents between canine myocardial and Purkinje cells

An electrophysiological analysis of canine single ventricular myocardial (VM) and Purkinje (P) cells was carried out by means of whole cell voltage clamp method. The following results in VM versus P cells were obtained. INa3 was present, had a threshold negative to the fast activating–inactivating INa1, its slow inactivation was cut off by INa1, and contributed to Na+ influx at INa1 threshold. INa1 was smaller and had a less negative threshold. There was no comparable slowly inactivating INa2, accounting for the shorter action potential. Slope conductance at resting potential was about double and decreased to a minimum value at the larger and less negative IK1 peak. The negative slope region of I-V relation was smaller during fast ramps and larger during slow ramps than in P cells, occurred in the voltage range of IK1 block by Mg2+, was not affected by a lower Vh and TTX and was eliminated by Ba2+, in contrast to P cells. ICa was larger, peaked at positive potentials and was eliminated by Ni2+. Ito was much smaller, began at more positive values, was abolished by less negative Vh and by 4-aminopyridine, included a sustained current that 4-aminopyridine decreased but did not eliminate. Steeper ramps increased IK1 peak as well as the fall in outward current during repolarization, consistent with a time-dependent block and unblock of IK1 by polyamines. During repolarization, the positive slope region was consistently present and was similar in amplitude to IK1 peak, whereas it was small or altogether missing in P cells. The total outward current at positive potentials comprised a larger IK1 component whereas it included a larger Ito and sustained current in P cells. These and other results provide a better understanding of the mechanisms underlying the action potential of VM and P cells under normal and some abnormal (arrhythmias) conditions.


Introduction
The different functions of Purkinje (P) and ventricular myocardial (VM) cells are associated with several electrophysiological and mechanical differences (e.g., see Lin and Vassalle 1978;Cordeiro et al. 1998). Thus, the action potential (AP) of canine P fibers is longer (+71%, Lin and Vassalle 1978), their plateau is more negative (e.g., Bal ati et al. 1998) and their twitch is shorter (À40%) and smaller (À79%) (Lin and Vassalle 1978) than in ventricular myocardial fibers.
These findings suggest that sodium influx during the action potential may be greater in Purkinje fibers because it also includes I Na2 , which slowly inactivates at plateau potentials Bocchi and Vassalle 2008). In addition, in P cells the slowly inactivating sodium current I Na3 is activated at potential negative to that of I Na1 threshold (Rota and Vassalle 2003). Whether I Na2 and I Na3 are also present in VM cells or whether Na + currents have identical features in P and VM cells have not been determined. Furthermore, it is not known whether there are differences in negative slope (NS) and positive slope (PS) regions of the I-V relation between the two tissues. In P cells, I Na3 and I Na2 are involved in the NS region (Rota and Vassalle 2003), but the role of the block and unblock of inward rectifying I K1 channels (Ishihara 1997;Ishihara and Ehara 1998) in the NS and PS regions, respectively, is undefined. Furthermore, whether the mechanisms underlying NS and PS region are similar or differ in P and VM cells is unknown.
There are differences in electrophysiological features of other currents as rabbit P cells express smaller I K1 , a larger transient outward current I to than VM cells (Cordeiro et al. 1998) and a greater I to sensitivity to TEA (Han et al. 2000). Whether voltage-and time-dependent features of I K1 , I to , sustained current, and I Ca differ in P and VM cells have not been determined.
The general aim of the present experiments was to investigate several ionic currents by means of a whole cell patch clamp method in canine P and VM cells isolated with the same technique to determine their features (e.g., presence or absence, threshold potential, magnitude, time-and voltage-dependent characteristics).
The specific aims included the determination of differences in the following features in P and VM cells: (1) I K1 inward rectification and its characteristics; (2) slope conductance over the voltage range of the action potential; (3) presence of I Na3 and I Na2 and their characteristics; (4) I Na1 amplitude; (5) threshold potential for different Na + currents and their voltage-and time-dependent inactivation; (6) contribution of I Na3 to peak I Na1 ; (7) presence and magnitude of NS and PS regions and their underlying mechanisms; (8) I to peak and sustained outward current; (9) magnitude and voltage range of the inward component related to I Ca ; and (10) identification of the various currents by different means including different V h , different ramps slopes, and channel blockers.
It was found that the differences in ionic currents between P and VM cells are numerous and substantial and provide insights in the different mechanisms that shape the action potentials, in their modification by some physiological and pharmacological factors and in the mechanisms of induction of some ventricular arrhythmias.

Material and Methods
Institutional and national guide for the care and use of laboratory animals was followed. The protocols for the experiments were reviewed and approved by the local Animal Care and Use Committee.
The details of the methods have been published (Rota and Vassalle 2003;Vassalle et al. 2007;Bocchi and Vassalle 2008). In brief outline, adult dogs (beagle, n = 25) of either sex were euthanized by intravenous injection of sodium pentobarbital (60 mg kg À1 ). Once the respiration had stopped, the hearts were removed and rinsed in physiological saline solution. Purkinje fiber bundles and thin papillary muscles or trabeculae (diameter ≤ 1 mm) were cut from both ventricles and were driven at 60/min for 30 min while being superfused in a tissue bath at 37°C.
The composition of physiological saline solution in mmol L À1 was NaCl 140, KCl 5.4, CaCl 2 1.8, MgCl 2 1, HE-PES 5.0, and glucose 5.5. The solution was gassed with 100% O 2 and adjusted to pH 7.4 with NaOH. The P and VM fibers were then rinsed with Ca-free solution with added 25 mmol L À1 taurine, 5 mmol L À1 beta-hydroxybutyric acid and 5 mmol L À1 Na pyruvate for 5 min in the same tissue bath and washed in a test tube three times with the same Ca-free solution. Ca-free solution contained in mmol L À1 : NaCl 140, KCl 5.4, KH 2 PO 4 1.2, MgCl 2 1.5, HE-PES 5.0, and glucose 5.5 (pH adjusted to 7.2 with NaOH). P and VM tissues were separately digested at 37.5°C in Ca-free physiological saline solution to which collagenase (1 mg/mL, type VIII, Sigma, St. Louis, MO), elastase (0.6 mg/mL, type II-A, Sigma), and essentially fat-free bovine serum albumin (2 mg/mL) had been added ("enzyme solution"). The cells were separated from the digested fibers by agitation by means of a mechanical "triturator" (Datyner et al. 1985). The cells were suspended in Kraftbr€ uhe (KB) solution and samples of the cell suspension were perfused with physiological saline solution at 37°C in a chamber located on the stage of an inverted microscope (Nikon Diaphot, Nikon, Tokyo, Japan).
We elected to study ionic current profiles under physiological conditions (intact intracellular and extracellular ionic concentration and absence of channel blockers). Although this approach does not allow to fully isolate single currents, it preserves ionic balances and electrochemical gradients during the acquisition. Therefore, the currents in P and VM cells were studied in the absence of any channel blocker (such as Ba 2+ , Ni 2+ , tetrodotoxin, 4-Aminopyridine, etc.) to compare and contrast the currents under physiological conditions and to avoid the multiple effects of channel blockers on currents and ionic gradients. Later on, we identified the current under study and their role on different parameters in different ways, including different V h , different ramps slopes, and channel blockers.
Successive command steps of the same protocol were applied at intervals of at least 5 sec and different protocols were separated by intervals of 3-5 min to allow the effects of each procedure to fully subside.
The data were analyzed by means pCLAMP program (Axon Instruments Inc.). Steps from different holding potentials (V h ) were applied to activate voltage-and time-dependent currents and depolarizing and repolarizing ramps with different slopes were used to study the currents under different conditions. On step depolarization from V h À80 mV, I Na1 was often cut off at À10 nA by the saturation of the amplifier. As no differences were detected in the results obtained from male and female dog cells, the results were pooled together.
The amplitude of the slowly decaying component of I Na2 was measured as the difference between the current at the beginning and the end of the step. The beginning was taken as the value at the intersection between the rapidly inactivating I Na1 and the backward extrapolation of I Na2 , also checked by fitting the slowly inactivating I Na2 with a double exponential function.
Unless otherwise specified, the current traces were fitted with two term standard exponential function using the Chebyshev technique with Clampfit software according to equation (1): where A 1 and A 2 are the amplitudes, and ss and sf are the time constants and C is the offset constant. Data were analyzed by mean of the Clampfit (pCLAMP 10.2) and Microsoft Excel programs. The results of tests carried out for each procedure are shown in the tables as means AE SEM (standard error of the mean) together with the number (n) of cells studied. Student's paired t test between two terms of comparison and one-way ANOVA (analyses of variance) between a data group were applied and a P < 0.05 was considered significant and was marked by an asterisk (*) in the tables and in text.

Results
I Na3 and its relation to I Na1 In P cells, I Na3 appears at a potential (À57.8 mV) which is negative to I Na1 threshold (À52 mV) (Rota and Vassalle 2003). I Na1 suppresses the slow inactivation of I Na3, as at its threshold I Na1 is not followed by time-dependent current (Rota and Vassalle 2003;Vassalle et al. 2007;Bocchi and Vassalle 2008). Whether I Na3 also is present in VM cells or whether its slow inactivation is suppressed by I Na1 is not known.
In Figure 1A, in a VM cell during the step from V h À80 mV to À50 mV, an inward current appeared that decayed bi-exponentially. In Figure 1B, the step to À40 mV elicited I Na1 (partially shown), which (as in P cells) was not followed by a slowly inactivating component. In Figure 1 inset 1, the shaded area emphasizes the fact that I Na3 slow inactivation was present at À50 mV and absent at À40 mV.
The absence of I Na3 during the step at I Na1 threshold potential could be due to either the suppression of I Na3 slow inactivation by I Na1 (as in P cells) or to the less negative voltage. To clarify this point, a two-step protocol was applied, as a suitable conditioning step may reduce I Na1 channel availability just enough to shift its threshold to a less negative value. In Figure 1C, the conditioning step to À50 mV induced I Na3 and the test step to À40 mV failed to activate I Na1 and induced a smaller I Na3 . The finding suggests that in Figure 1B I Na3 inactivation was not present because it was suppressed by I Na1 and not because it could not occur at À40 mV.
In Figure 1D, the À30 mV test step initiated an inward transient (À3037 pA) which was followed by a small shallow tail (À93 pA), suggesting the induction of I Na2 with a small slow inactivation component.
In n = 10, with the two steps protocol in VM cells during the step from À80 mV to À50 mV I Na3 amplitude was À143.2 AE 54.9 pA and during the test step to À40 mV it was À72.4 AE 9.9 pA (not significantly different). During the test step to À30 mV, the inward transient was À3045 AE 576 pA and was followed by a decaying tail of 63.7 AE 14.8 pA. Therefore, in VM cells I Na3 was present during the À40 mV test step in the absence of I Na1 . During the À30 mV test step, I Na2 was followed by a small and quickly inactivating component (see below). Similarly, in P cells (n = 10) I Na3 could be activated at the I Na1 threshold if the activation of I Na1 was prevented by the conditioning step. One difference with the VM cells was that in P cells the slowly inactivating I Na2 was much larger (+673.1%*; see below). I Na3 was studied in VM and P cells by applying single steps (Fig. 1A) from V h À80 mV (Table 1). With respect to P cells, in VM cells during depolarizing steps I Na3 was consistently present, had a less negative threshold (*), and similar amplitude as well as time constants of inactivation.
The finding that the À50 mV conditioning step prevented the appearance of I Na1 but not of I Na3 during the À40 mV test step suggests that I Na3 might be less sensitive to voltage-dependent inactivation than I Na1 . This was tested by applying depolarizing steps from gradually less negative V h . As shown in Table 1, in P and VM cells with gradually less negative V h , the amplitude of I Na3 decreased very little until V h was À40 mV and the threshold remained less negative (*) in VM cells. At all V h , in both P and VM cells the inactivation of I Na3 was slow in the Figure 1. I Na3 and its relation to I Na1 . In a VM cell, a depolarizing step from V h À80 mV to À50 mV (lower trace in A) and to À40 mV (B) elicited the currents shown in the upper traces. The current traces have been superimposed in inset 1 and the shaded area emphasizes the suppression of the slow inactivation of I Na3 by I Na1 . In C, a conditioning step was applied to À50 mV and a test step to À40 mV and in D to À50 and À30 mV, respectively. In E, depolarizing steps were applied from V h À80 to À40 mV and increased by 1 mV to À30 mV: the horizontal arrow points to inactivating I Na1 . In F, in a P cell, ramps with the same slope and different duration were applied. Downward vertical arrow points to I K1 peak and downward oblique arrow to the negative slope region. In inset 2, part of the F traces are shown at greater gain. Dashed lines emphasize the different slopes of inward currents prior to and at beginning of I Na1 . In F and G, the shaded areas show that at the end of I Na1 inactivation the magnitude of the outward current approached that of I K1 peak. The short dash in each panel indicates zero current in this and subsequent figures. In G, ramps with different slopes (260 mV sec À1 , gray trace; 520 mV sec À1 , black trace) were applied. I Na3 voltage-dependent increase and sudden suppression of its slow inactivation by I Na1 Gradually increasing depolarizing steps might lead to a progressive increase in I Na3 . In Figure 1E, in a VM cell with depolarizing steps increasing by 1 mV between the I Na3 and I Na1 thresholds, I Na3 magnitude increased progressively and inactivated relatively more quickly up to À29 mV. At À30 mV (I Na1 threshold), the inactivation of I Na1 (arrow) suppressed I Na3 slow decay. Similar results with steps increasing at intervals of 1 mV were obtained in VM cells (n = 4) and in P cells (n = 16). The above results suggest that during depolarizing ramps I Na3 might precede I Na1 , as the continuous decline in voltage would initiate and increase I Na3 at potentials negative to the I Na1 threshold. Furthermore, the current at the end of I Na1 inactivation would be expected to be more outward than in its absence due to the suppression of I Na3 slow inactivation. In some cells, applying ramps of different duration with a borderline slope (150 mV sec À1 ) for I Na1 activation led to a nonuniform induction of I Na1 , so that the events in the presence and absence of I Na1 could be compared in the same cell as shown in Figure 1F.
In a P cell, during the ramps, the outward current gradually increased before peaking (I K1 peak, vertical arrow). The slope of the ramps being borderline for I Na1 activation, during the shorter ramp, I K1 peak was followed by a negative slope (NS) region (oblique arrow). No I Na1 was present and the current during the NS region (I NS ) was followed by a reincreasing outward current. During the longer ramp, I NS more quickly turned inward and its steeper slope merged into that of the activating I Na1 , as emphasized by the gray lines in Figure 1 inset 2.
In Figure 1F and inset 2, the shaded area shows that the current at the end of I Na1 inactivation was more outward than in the I Na1 absence, as expected from the suppression of the slow inactivation of I Na3 by I Na1 . The inactivation of I Na1 was still followed by NS region with smaller amplitude and less steep slope, consistent with the slow inactivation of I Na2 (see below). The patterns illustrated in Figure 1F were present in n = 3, the consistent presence of I Na1 with its inactivation approaching I K1 peak in n = 4 and the absence of I Na1 with the consistent presence of I NS in n = 19.
In another approach, ramps with different slopes (Fig. 1G) were applied, as in several instances in P cells no I Na1 was initiated during slower ramps. In Figure 1G, the 260 mV sec À1 ramp (lighter trace) did not activate I Na1 , whereas the superimposed 520 mV sec À1 ramp (darker trace) did. During the slower ramp, I K1 peak was followed by I NS but not by I Na1 . Instead, with the steeper ramp, the end of I Na1 inactivation approached the I K1 peak and was followed by I NS .
The asymmetry between slower activation and faster inactivation of the overall Na + current (Fig. 1 F and G) was a consistent finding that might be expected from I Na3 preceding I Na1 and its slow inactivation being cut off by it.
In VM cells, I Na1 was less frequently absent with slower ramps. With 260 mV sec À1 ramps, I Na1 was absent in 9/17 P cells and in 3/17 VM cells whereas with the 520 mV sec À1 ramp, I Na1 was absent only in 3/17 P cells and in none of 17 VM cells (the NS region being present with or without I Na1 ).
Therefore, in VM cells I Na3 : (1) was present with a similar magnitude and rate constants of inactivation; (2) had a threshold less negative than in P cells and negative to that of I Na1 ; (3) increased progressively at potentials between its threshold and that of I Na1 ; (4) its slow inactivation was consistently eliminated by I Na1 ; (5) could appear and inactivate slowly at voltages less negative than I Na1 threshold if I Na1 activation was prevented by a conditioning step; (6) contributed to the beginning of I NS during depolarizing ramps, and (7) was less sensitive than I Na1 to voltage-and time-dependent inactivation. Th (mV) À21.0 AE 1.0 À27.1 AE 1.8* I Na3 (pA) À29.9 AE 18.3 (3/10) À84 AE 35 (6/9) V h (mV), holding potential in mV; Param, parameters measured; VM cells, data from ventricular myocardial cells; P cells, data from Purkinje cells; Th (mV), threshold potential in mV of I Na3 ; I Na3 (pA), amplitude in pA of I Na3 measured as the difference between its peak and the end of the step; s f (msec) and s s (msec), fast and slow time constants, respectively, of I Na3 inactivation; Numbers in parenthesis (e.g., 18/18), number of cells in which I Na3 was present over the total number of cells studied; *statistically significant difference between P and VM cells data. Currents during larger depolarizing steps in Purkinje and myocardial cells Ionic currents in P and VM cells were investigated also by applying 500 msec depolarizing steps from V h À80 mV to +40 mV in increments of 10 mV (Fig. 2, protocol in c).
In Figure 2, in the P cell (A) the range of inward and outward currents was larger than in the VM cell (Fig. 2B). In P cell, I Na1 was truncated by the saturation of the amplifier at À10,000 pA whereas in VM cell the largest I Na1 was À8428 pA. In the P cell (Fig. 2c), the À50 mV step initiated I Na1 whose inactivation was followed by a steady current that overlapped the current trace at À60 mV (where no I Na1 was present; see also Vassalle et al. 2007;Bocchi and Vassalle 2008), thus suggesting no loss of voltage control. The slowly inactivating I Na2 appeared at À40 mV and reached its largest value during À20 mV step (shaded area in Fig. 2c). Typically, I Na2 was still decreasing by the end of the 500 msec step Bocchi and Vassalle 2008).
In the VM cell, the threshold for I Na1 activation was less negative (À40 mV, Fig. 2d) than in the P cell (À50 mV). During the depolarizing À20 mV step, the inactivation of I Na1 was not followed by a decaying I Na2 , in contrast to the P cell. The comparison of Figure 2a and b indicates that in VM cell the large and slowly decaying I Na2 was absent at other potentials as well. The Figure 2. Larger range of inward and outward currents in P than in VM cells. The protocol is shown at the bottom of c. In the P cell, the currents flowing during the steps from V h À80 to À60, À50, and À20 mV are shown in c, where the shaded area emphasizes the slow decay of I Na2 . In the VM cell, the currents recorded with steps from V h À80 to À40 and À20 mV are shown in d. In inset 1, the current traces were displaced by 200 pA for a better visualization. The light-shaded areas emphasize the time-dependent component of I K1 and the dark-shaded area I Na3 . In inset 2, the current traces during the À60 mV step are shown in control (Contr.) in the presence of 4-aminopyridine (4-AP), of 4-AP plus Ba 2+ (+Ba 2+ ), and during recovery (Rec.). The upward arrow indicates the initial decline in the outward current to a steady value and the downward arrow its suppression by Ba 2+ . sustained current at the end of À50 mV step (measured as the difference from the holding current I h ) was larger in the VM cell than in the P cell. In the VM cell, the sustained current at À40 mV (I Na1 threshold) was similar to that at À50 mV (not shown). In both cells, the sustained current at À20 mV was less outward than at À50 mV (P cell) and at À40 mV (VM cell), as expected from the onset of I NS . With more positive steps, the outward currents were far larger in P cell than in VM cell (cf. Fig. 2a and b).
In Table 2, with V h À80 mV, in VM cells I Na1 threshold was less negative in P cells (*), I Na1 was smaller (*) in VM cells, as it was truncated in all P cells, but only in 13/18 VM cells. In both VM and P cells, I Na1 inactivated exponentially with s~1.5 msec. In VM cells, I Na1 was smaller (*) than in P cells also with V h À70 mV when I Na1 was less often truncated.
In Table 3, at À20 mV in VM cells an inactivating I Na2 tail was present in 6/18 cells, was small and decayed quickly. In contrast, in P cells the slowly inactivating I Na2 was much larger (*) and inactivated more slowly (*). With gradually less negative V h , I Na1 (Table 2) and I Na2 (Table 3) gradually decreased.
The slope conductance was measured by superimposing small hyperpolarizing pulses on the parent steps in VM cells (n = 9). At the I Na1 threshold, after the I Na1 inactivation, the slope conductance was minimal and did not vary with time, as shown in P cells by Bocchi and Vassalle (2008). This finding also is consistent with no loss of voltage control.

Time-dependent decay of I K1 on depolarization in Purkinje and myocardial cells
In Figure 2, in the same VM cell, the currents during the À70, À60, À50, and À40 mV steps are shown in inset 1. At the beginning of the À70 and À60 mV steps, the outward current quickly declined (light-shaded areas) to a steady value. At À50 mV, I Na3 appeared and declined slowly (dark-shaded area), and at À40 mV I Na1 quickly activated and inactivated to a steady value.
One possible explanation for the initial decline of the outward current at À60 and À70 mV might be a noninstantaneous block of I K1 by polyamines during the depolarizing steps (Ishihara 1997;Ishihara and Ehara 1998), a block which is eliminated by Ba 2+ (Ishihara and Ehara 1998). In Figure 2 inset 2, the traces during the step to À60 mV were recorded in control (Contr.), in the presence of 4-aminopyridine (4-AP), of Ba 2+ (+Ba 2+ ), and during recovery (Rec.). The upward arrow indicates the initial decline in the outward current to a steady value, a decline that was little affected by 4-AP. Instead, Ba 2+ suppressed both the initial decay and the steady current during the step (downward arrow).
In Table 4, with V h À80 mV, at the voltages indicated in VM cells the outward current decreased more (*) in VM cells than in P cells. The time constant of the exponential decline was similar (~6 msec).
If the decline of the current during the step is indeed due to a time-dependent block of I K1 on depolarization, then I Na1 (pA), amplitude of I Na1 from its beginning to its peak in pA; s (msec), time constant of the exponential inactivation in ms of I Na1 ; VM cells, data from ventricular myocardial cells; P cells, data from Purkinje cells; Th (mV), threshold potential in mV of I Na1 ; Numbers in parenthesis (e.g., 18/18), number of cells in which I Na1 was present over the total number of cells studied; *statistically significant difference between P and VM cells data. decreasing V h should reduce the declining current, as increasing degrees of I K1 block would occur during the less negative V h , prior to the depolarizing step. To test this point, V h was reduced in 10 mV increment to À40 mV. As shown in Table 4, in both P and VM cells with gradually less negative V h , the initial decay of the current became gradually smaller and less frequent, and disappeared altogether with V h À40 mV. These findings indicate a timeand voltage-dependent block of I K1 , which was significantly larger in VM cells at V h À70 and À60 mV.
Contribution of I Na3 to Na + inflow due to I Na1 As I Na3 occurs also at the I Na1 threshold, at that potential I Na3 would be expected to precede I Na1 and therefore con-tribute to the peak Na + current. To verify such a possibility, the current traces at the beginning of depolarizing steps were displayed at suitably greater time base. In Figure 3, in a P cell (A) and in a VM cell (B), at the usual time base only I Na1 was visible. However, when the traces were displayed at much greater time base, a slower inward component (comprised between the downward and horizontal arrows) preceded I Na1 both in P (Fig. 3C) and in VM cell (Fig. 3D). In Figure 3C (P cell), the trace recorded at À60 mV show the initial decay of the outward current to a steady value (time-dependent block of I K1 ). Instead, the trace recorded at À50 mV departed from the exponential decay (vertical arrow), crossed in an inward direction the À60 mV trace (as expected from the activation of I Na3 ). After a delay, it was followed by the fast initiation of I Na1 (sudden beginning of the steeper  Peak (mV), voltage at which the largest slowly inactivating I Na2 in P cells or in VM cell was measured; I Na2 slow (pA), amplitude of slowly inactivating I Na2 in pA, measured from its beginning to the end of the step; VM cells, data from ventricular myocardial cells; P cells, data from Purkinje cells; s f (msec) and s s (msec), fast and slow time constants, respectively, of I Na2 inactivation; Numbers in parenthesis (e.g., 6/18), number of cells in which I Na2 was present over the total number of cells studied; *statistically significant difference between P and VM cells data. slope, horizontal arrow). Similar events occurred in the VM cell (Fig. 3D).
In Figure 3E (P cell), the current traces at the beginning of À50, À40, À30, and À20 mV steps were superimposed and show the progressively earlier onset of I Na1 with larger depolarizing steps until no distinct I Na3 component was apparent. In Figure 3F, similar events occurred in the VM cell in that larger depolarizing steps elicited an earlier I Na1 with the eventual disappearance of I Na3 .
In n = 18, in VM cells, I Na3 (measured between the departure of the trace from exponential decay and the sudden onset of I Na1 ) had a magnitude of À2677 AE 201 pA and a duration of 2.2 AE 0.3 msec; I Na1 (measured from the sudden increase in steepness to its peak) had a threshold of -36.6 AE 1.1 mV) and an amplitude of À8857 AE 461 pA. In P cells, I Na3 had a amplitude of À2330 AE 235* pA and a duration of 2.6 AE 0.8 msec; I Na1 had a more negative threshold (À48.3 AE 1.4* mV) and an amplitude >À10,000 AE 0.0* pA.
Thus, I Na3 was a substantial fraction of the inward current flowing at the threshold for I Na1 . With steps to less negative values, I Na3 consistently decreased and I Na1 was activated sooner.

Differences in I-V relation of the sustained current in Purkinje and myocardial cells
To investigate the quasi-steady state I-V relations in VM and P cells, the sustained current at the end of 500 msec depolarizing steps from V h À80 mV to +40 mV was mea-sured as the difference from I h . The same procedure was applied with less negative V h to determine how the I-V relation would be affected in P versus VM cells.
In Figure 4, in A, with V h À80 mV, in VM cells the sustained current was more outward at negative potentials and less outward at positive potential than in P cells. In both types of cells, past I K1 peak, the NS region was present, but in VM cells I NS was larger and peaked at a less negative value (~0 mV), whereas in P cells the smaller I NS peaked at~À20 mV. The outward current began to increase at +10 mV in VM cells and at À10 mV in P cells, suggesting a different I to threshold. I NS was differently affected by lower V h in P and VM cells. With V h À70 (Fig. 4B), overall the sustained current was much less outward in both VM and P cells, as expected from the inward rectification of I K1 channel at less negative V h . However, I NS was still large in VM cells whereas it was diminished in P cells, suggesting that in P cells the decrease in I NS might be related to a partial inactivation of I Na2 .
This interpretation is supported by the findings with still lower V h . In VM cells, the current became inward and I NS persisted unaltered up to V h À50 mV ( Fig. 4B-D). Only with V h À40 mV did I NS decrease (Fig. 4E) as apparently the channel contributing to I NS was partially blocked prior to the depolarizing step. Instead, in P cells I NS markedly decreased with V h À60 to disappear altogether with V h À40 mV. In Figure 4 inset 1, the graph shows the difference between the outward current peak prior to the NS region (corresponding to I K1 peak) and the smallest current value of I-V relation prior to the reincrease in outward Figure 3. The activation of I Na3 precedes that of I Na1 at the latter's threshold. In A (P cell) and B (VM cell), the traces are shown at the usual time base whereas in C, D, E, and F the traces are shown at the greater time base indicated. In C and D, the downward vertical arrows point to the beginning of I Na3 and the leftward horizontal arrows point to the beginning of I Na1 . The traces recorded at the beginning of the À50, À40, À30, and À20 mV steps have been superimposed in E (P cell) and those at the beginning of the steps to À30, À20, À10, 0, and +10 mV have been superimposed in F (VM cell). The arrows in E and the dots in F indicate the beginning of I Na1 . The numbers next to the traces indicate the voltage of the respective steps. current (a measure of I NS peak). The graph shows how differently I NS amplitude varied in VM and P cells as a function of V h , the decrease of I NS in P cells being statistically significant (ANOVA < 0.0001).
The increase in outward current with the larger depolarizing steps was much greater in P than in VM cells (helped in this by the inward shift of the current in VM cells), and it was little affected by less negative V h . At each V h , the outward current in P cells increased past À20 mV, as expected for I to . As the protocol applied at different V h was the same, with the gradually less negative V h the depolarizing steps attained gradually more positive values. Hence, the sustained current with the largest depolarizations increased to similar values in spite of the decreasing V h .
These results raise the possibility that in P cells Na + currents might mainly contribute to I NS (with this protocol, I Na2 ) whereas the voltage-dependent block of I K1 channel may predominantly determine I NS in VM cells.

I Na2 and the I Ca component in Purkinje and myocardial cells
As at plateau voltages the slowly inactivating I Na2 prevails in P cells and presumably I Ca prevails in VM cells, the different amplitude, voltage range, voltage-and time-dependent inactivation of I Na2 in P cells and of the I Ca component in VM cells were investigated as shown in Figure 5.
In the P cell (Fig. 5A), a step from V h À80 to À50 mV elicited I Na1 which was not followed by time-dependent Figure 5. Different amplitude, voltage range, kinetics, and time-dependent inactivation of I Na2 in P cells and of I Ca in VM cells. Steps were applied from V h À80 to À50, À30 mV +10 mV in the P cell (A) and in the VM cell (C) where the shaded area emphasizes I Na3 and the small arrow points to the beginning of a large I Ca . In the P cell (B), progressively longer conditioning steps to À50 mV were followed by test steps to À30 mV. In the VM cell (D), conditioning steps to À50 mV were followed by test steps to +10 mV. In E, the oblique upward arrow points to inactivating I Na2 during the test step to À20 mV in the P cell. In F, the filled circle labels I Na3 and the vertical upward arrow points to the absence of comparable slowly inactivating I Na2 at À20 mV in the VM cell. currents, as usual. During the step to À30 mV, I Na1 was followed by the slowly decaying I Na2 (s s 297 msec). During the +10 mV step, a small inward component was superimposed on a small outward current. In the VM cell ( Fig. 5C), the step from V h À80 to À50 mV elicited I Na3 (shaded area). During the step to À30 mV, I Na1 was followed by a small outward component, but not by decaying I Na2 . During the +10 mV step, I Na1 was followed by a large I Ca component (À1699 pA) which inactivated with a s s of 106 msec and whose beginning during I Na1 inactivation is indicated by the arrow in the magnified trace. In the P cell ( Fig. 5B), after progressively longer conditioning steps at À50 mV, the test steps to À30 mV elicited a gradually smaller inactivating I Na2 (with the last test step, À47%) (see Bocchi and Vassalle 2008). In the VM cell ( Fig. 5D), the conditioning step was the same, but (as there was no I Na2 at À30 mV) test steps were applied to the voltage where I Ca component was largest (+10 mV, protocol in Fig. 5D). The test step elicited an inward current, whose amplitude was not decreased by progressively longer conditioning steps (with the last test step, +3.7%). In the VM cell, I Ca decayed more quickly than I Na2 did in the P cell. The findings point to a different voltage range, kinetics, and voltage-and time-dependent inactivation of I Na2 in P cells and of I Ca in VM cells.
In order to separate I Na1 from I Na2 , a double step protocol was applied from V h À80 mV to À50 and to À20 mV in a P cell (Fig. 5E) and in a VM cell (Fig. 5F). In the P cell, at À50 mV I Na1 was followed by a steady current, and at À20 mV I Na2 activated rapidly and decayed slowly (oblique arrow). In the VM cell, at À50 I Na3 was present as usual at that voltage (filled circle; see Fig. 1) and at À20 mV an inward transient was followed by a faint and brief tail at a potential where the slowly decaying I Na2 was large in the P cell.
With V h À80 mV, in VM cells at +20 mV the I Ca component was À436.7 AE 105.2 pA (17/17 cells) with s f 19.1 AE 8.4 msec and s s 112 AE 15.4 msec, whereas in P cells at +18.8 mV the I Ca component was À96.8 AE 44.5* pA (present in 6/17 cells) with s f 11.4 AE 6.3 msec and s s 157.6 AE 61.2* msec.

I-V relation during slow depolarizing and repolarizing ramps in myocardial and Purkinje cells
Because in VM cells the sustained current at the end of depolarizing steps was larger at potentials negative to I K1 peak and smaller at positive potentials than in P cells (Fig. 4), the steady state I-V relation was studied during slowly depolarizing and repolarizing ramps in the two tissues.
In Figure 6A, in the P cell during 6.5 mV sec À1 depolarizing ramp, the outward current increased gradually less to stop increasing altogether between point 1 (À58 mV, I K1 peak) and point 2 (À30.7 mV). On further depolarization between points 2 and 3, the outward current increased markedly and before the ramp peak, underwent an enhancement (shaded area, the "bulge"; Du and Vassalle 1999). During the repolarizing ramp, the outward current initially decreased more rapidly, but 16 sec after the ramp peak it was similar to 16 sec before (at I K1 peak) (292 and 275 pA, respectively).
In Figure 6B, in the VM cell, the outward current also increased gradually less as a function of depolarization, but at point 1 (À48 mV, I K1 peak) the outward current was 179% larger than in the P cell. Also, a distinct NS region began at point 1 and was followed (change in slope) by an I Ca component that peaked at +2 mV (point 2). Between points 2 and 3, the current reincreased in an outward direction, but it was much smaller than in the P cell. Furthermore, there was no enhancement of the outward current (no "bulge") and at the ramp peak (point 3) the current was less outward than at point 1, in sharp contrast with the P cell.
During the repolarizing ramp, the outward current decreased much less than in the P cell and it was less inward (star) than at point 2. The outward current reincreased in the positive slope (PS) region to a peak value (786 pA) similar to that of I K1 peak (769 pA). Past the peak of I PS , the outward current underwent a progressively quicker decrease as a function of repolarization.
In Table 5, with respect to P cells, in VM cells I K1 was larger (*) and peaked at a less negative potential (*). I NS was larger (*), was more frequently present and peaked at less negative value (*). The I Ca component was larger when measured from its beginning to its peak and when the I Ca peak was compared to the symmetrical peak during repolarizing ramp. The outward current began to reincrease ("I to start") at less negative potential (*) to reach a value at the ramp peak ("I to ") which was smaller (*) (although it was similar when measured with respect to I h ).
During the repolarizing ramp, the decrease of outward current (I repol ) was smaller (*) and peaked at less negative potential (*). I PS (consistently present in VM but not in P cells) began at a more positive potential (*) and it was much larger (*) as it was (*) when measured with respect to I h although less so. In VM cells, I K1 peak was similar to I PS peak as it was in P cells at I PS peak (or at the value at which the outward current began to decrease rapidly). This suggests that in neither tissue the Na + currents contributed to I K1 or I PS peaks.
Thus, with slow ramps, with respect to P cells, in VM cells: (1) I K1 peak was much larger and peaked at a less negative potential; (2) I NS was larger and peaked at less negative potentials; (3) I Ca component was larger; (4) the outward current enhancement prior to ramp peak (the "bulge") was absent; (5) the outward current at ramp 2013 | Vol. 1 | Iss. 3 | e00036 Page 12 peak was much smaller (but less so when compared to I h ); (6) during repolarizing ramps, the smaller I repol declined much less; (7) I PS was present more frequently, was larger and with a more positive beginning, and (8) I PS and I K1 peaks were similar, both being larger than the ramp peak current.
Slope conductance changes during slow ramps in Purkinje and myocardial cells If, during depolarization, the gradually smaller increase in outward current is due to the inward rectification of I K1 channel, the slope conductance should decrease accord-ingly. To find out, the slope conductance was measured by superimposing small hyperpolarizing voltage pulses on the parent ramp (protocol in Fig. 6C).
In the P cell (Fig. 6D), the amplitude of the pulse current at V h À80 mV decreased gradually on depolarization to reach a minimal value at À44 mV, just prior the beginning of a small NS region. During I NS , the pulse current reversed polarity and reincreased. With further depolarization, the pulse current decreased again, became once more negative and reincreased in amplitude. During the repolarizing ramp, similar events occurred in reverse order, including a smaller increase in slope conductance during a rather small I PS . Figure 6. I-V relation and changes in slope conductance during slow ramps in P and VM cells. A 6.5 mV sec À1 depolarizing and repolarizing ramp was applied to a P cell (A) and to a VM cell (B). Point 1 indicates the I K1 peak, point 2 the beginning of increasing outward current, and point 3 the current at ramp peak. In A, the shaded area indicates the enhancement of the outward current ("the bulge"). In B, the asterisk indicates the transition between the decreasing outward current and the beginning of I PS . Hyperpolarizing voltage pulses (amplitude 7 mV, duration 200 msec, rate 90 min À1 ; C) were superimposed on the parent ramp to measure slope conductance. In D (P cell) and E (VM cell) (same heart but different from that for A and B), the horizontal lines indicate the sections of current records shown at higher gain underneath. In D, the upward vertical arrow points to an inward transient after the reversed pulse current and the triangle points to its absence. In the enlarged sections of VM trace, some of the capacity spikes have been deleted for a better visualization of the increase and reversal of pulse current in the NS and PS regions. The sections of the traces labeled with a horizontal line are shown underneath at higher gain for a better visualization of pulse current changes. The arrow under the magnified trace points to a small inward component that followed the outward pulse current: such an inward component was not present during the repolarizing ramp (triangle) or in the VM cells (see below), suggesting that the brief hyperpolarizing step allowed an increased availability of sodium channels.
In the VM cell (Fig. 6E), at V h the amplitude of the pulse current was larger (+72%) than in the P cell. The pulse current decreased on depolarization to become minimal at À31 mV, reversed polarity during I NS and increased to a maximum at À14 mV, decreased again and then underwent a much smaller increase than in the P cell during the remainder of the ramp. During the repolarizing ramp, similar events occurred in reverse order, including an increase in slope conductance during I PS .
In P cells (n = 21), the pulse current amplitude varied as follows. It was À190.8 AE 20.1* pA at V h À88.5 mV and 0 pA at I K1 peak (À47.1 AE 1.5* mV). I NS was 7.2 AE 4.4* pA and was present only in 3/21 cells. The pulse current amplitude was À110.0 AE 11.7* pA at the ramp peak. During the repolarizing ramp, the pulse current amplitude was À71.4 AE 6.8 pA at +17.7 mV (* with respect to À119 pA at +20.1 mV during depolarization) and 2.5 AE 2.2 pA at the potential where the final faster depolarization began (I PS was present in 2/21 cells). n 15, number of cells studied; D (mV or %), difference in mV or percent of VM cells data with respect to P cells data; I K1 peak (pA), amplitude of I K1 peak in pA, measured as the difference from I h ; I K1 peak (mV), voltage in mV at which I K1 peaked; I NS (pA), current amplitude in pA during the negative slope region; I NS , peak (mV), voltage in mV of I NS peak; I Ca start (mV), beginning of I Ca component in mV determined as the departure of current trace from I NS peak; I Ca (pA), amplitude of I Ca component in pA as the difference between its beginning and its peak; I Ca peak (mV), peak in mV of I Ca ; D (pA), difference in pA between I Ca peak during depolarization and minimum outward current preceding the beginning of I PS on repolarization; I to , start (mV), voltage in mV at which the increasing outward current started at I NS or I Ca peaks; I to (pA), amplitude of outward current in pA measured between its beginning and ramp peak; I ramp peak ÀI h (pA), outward current at ramp peak measured as difference from I h ; I repol (pA), amplitude in pA of the outward current between ramp peak and its smallest value prior to the beginning of I PS ; I repol peak (mV), voltage in mV at which the outward current was smallest prior to I NS beginning; I PS start (mV), voltage in mV at which I PS began; I PS peak (pA), current in pA at I PS peak, measured as difference between its beginning and its peak; I PS peak (mV), voltage in mV at which PS region peaked; I PS peakÀI h (pA), current in pA measured as difference between I NS peak and I h ; V h (mV), holding potential in mV; Param, parameters measured; VM cells, data from ventricular myocardial cells; P cells, data from Purkinje cells; *statistically significant difference between P and VM cells data. In the cells from the same three hearts with the same À83.3 mV V h , similar results were obtained in that in VM cells the pulse current amplitude at V h was larger by +88.8%, the pulse current fell to 0 pA at a less negative potential, I NS was +923% larger, the pulse current at ramp peak smaller by À87.5%, the reversed pulse current during I PS was 73 pA (there was no I PS in P cells).
Thus, with respect to P cells, in VM cells the pulse current: (1) was larger at V h (+100.0%*); (2) fell to a minimum at the 7.9 mV* less negative I K1 peak; (3) consistently reversed and reincreased during I NS ; (4) fell again by I NS end and reincreased but much less (À84.8% * than P cells at the ramp peak; and (5) during the repolarizing ramp, the smaller VM conductance underwent the converse changes, reincreasing during I PS .

Current during fast ramps in myocardial and Purkinje cells
As with the 6.5 mV sec À1 ramps, the Na + channels would be inactivated, depolarizing and repolarizing ramps with progressively steeper slopes were applied to VM and P cells.
In Figure 7, 260 mV sec À1 ramp, with respect to the P cell (Fig. 7A), in the VM cell (Fig. 7B) at point 1 I K1 peak was much larger and less negative. In both P and VM cells, immediately after I K1 peak, a slowly increasing I Na3 (shaded areas labeled by downward arrows) preceded the activation of I Na1 , as expected from the more negative threshold of I Na3 (see Figs. 1 and 3). In the P cell, I NS (empty circle) peaked at point 2 and was followed by an . I-V relation during steep ramps in P and VM cells. The 260, 520, and 1040 mV sec À1 ramps were applied from V h À80 to +50 mV (G) to a P cell (A, C, and E, respectively) and a VM cell (B, D, and F, respectively). The numbers 1, 2, and 3 label the peaks of I K1 , of I NS , and of the ramp, respectively. I NS is labeled by empty circles in the P cell and by filled circles in the VM cell. The inward component attributable to I Ca is labeled by empty squares and small shaded areas in the P cell, and by filled squares in the VM cell. In VM cells, I Ca was measured from its beginning (taken as the point at which the slope of I NS met the backward extrapolation of I Ca ) and I Ca peak. The downward vertical arrows point to the slowly increasing inward current (shaded areas) preceding the activation of I Na1 . In inset 1, the traces from C and D were superimposed by the end of I Na1 inactivation. In both cells, I Na1 was cut off by the saturation of the amplifier at À10 nA. increasing outward current. A very shallow inward component, peaking at +12 mV, appeared as an "indentation" on the increasing outward current (empty square and small shaded area). In the VM cell, I NS (filled circle) was smaller and was followed by a large inward component (filled square) which peaked at +17 mV, as expected from a larger I Ca component. Similar but not identical results were obtained during the 520 mV sec À1 (Fig. 7C and D) and 1040 mV sec À1 (Fig. 7E and F) ramps. I K1 peak increased in magnitude with the steeper ramps both in P and VM cell, still being much smaller in the P cell. In both cells, the steeper ramp slope caused I Na1 to inactivate closer to the end of I NS . As usual, the outward current at ramp peak (point 3) was larger than I K1 peak (point 1) in the P cell ( Fig. 7C and E) whereas it was smaller in the VM cell ( Fig. 7D and F). I NS of the P cell ( Fig. 7C) and that of the VM cell ( Fig. 7D) were superimposed by the end of I Na1 inactivation in Figure 7 inset 1. While I NS was larger in the P cell, the subsequent I Ca component was much larger (filled square) in the VM cell.
On repolarization, in the P cell the outward current decreased more and peaked at more negative values than in the VM cell. In both cells, on repolarization with steeper ramps the decrease in outward current was faster initially and it was larger. Also, I PS started from a less outward value and, on that account, I PS peak became smaller than the I K1 peak.
As shown in Table 6, during the 260 mV sec À1 ramp with respect to P cells, in VM cells I K1 peak was larger (*) and less negative (*), I NS was smaller and its peak was less negative (*). I Na3 amplitude was somewhat greater, I Na1 threshold was less negative (*), I Na1 amplitude was similar (there was I Na1 in 14/17 VM cells and in 8/17 P cells).
The I Ca component was larger (*) (there was a measurable I Ca component in15/17 VM cells and in 3/17 P cells). The outward current between I Ca and ramp peaks ("I to ") began at a more positive potential (*) and was smaller (*). When the ramp peak current was measured as the difference from I h , I to was smaller (*) but less so, due to the larger I K1 upon which the ramp peak current was superimposed. End I Na1 (pA), Amplitude in pA of the current at the end of I Na1 inactivation; End I Na1 (mV), voltage in mV of the current at the end of I Na1 inactivation; V h (mV), holding potential in mV; Param, parameters measured; VM cells, data from ventricular myocardial cells; P cells, data from Purkinje cells; I Na1 Th (mV), threshold potential in mV of I Na1 ; I Na3 (pA), amplitude in pA of I Na3 measured as the difference between its peak and the end of the step; I Na1 (pA), amplitude in pA of I Na1 ; Numbers in parenthesis (e.g., 14/17), number of cells in which the parameter was present over the total number of cells studied; *statistically significant difference between P and VM cells data. Other explanations as in the legend of The I K1 peak was similar to the current at the end of I Na1 inactivation both in the VM cells (0.003%) and in the eight P cells in which I Na1 was present (I K1 peak 389.1 pA and at the end of I Na1 324.6 pA, difference not statistically significant). These results are consistent with the elimination of the slow inactivation of I Na3 by I Na1 in both tissues. At the end of I Na1 , the voltage was less negative in VM cells (*), reflecting the less negative I Na1 threshold (*).
Thus, in both VM and P cells the faster ramps increased I K1 peak, I Na3 , I Na1 , I NS, and I ramp peak when measured either from its beginning during the ramp or from I h . The increase in I Na3 and in I NS was larger in P cells, consistent with I Na3 role in I NS . n, number of cells studied; D (mV or%), difference in mV or percent of VM cells data with respect to P cells data; I K1 peak (pA), amplitude of I K1 peak in pA, measured as the difference from I h ; I K1 peak (mV), voltage in mV at which I K1 peaked; I NS (pA), current amplitude in pA during the negative slope region; I NS , peak (mV), voltage in mV of I NS peak; I Na1 Th (mV), voltage at which I Na1 began; I Na1 (pA), amplitude in pA of I Na1 ; I Ca start (mV), beginning of I Ca component in mV determined as the departure of current trace from I NS peak; I Ca (pA), amplitude of I Ca component in pA as the difference between its beginning and its peak; I Ca peak (mV), peak in mV of I Ca ; I to , start (mV), voltage in mV at which the increasing outward current started at I NS or I Ca peaks; I to (pA), amplitude of outward current in pA measured between its beginning and ramp peak; I ramp peak ÀI h (pA), outward current at ramp peak measured as difference from I h ; I repol (pA), amplitude in pA of the outward current between ramp peak and its smallest value prior to the beginning of I PS ; I repol peak (mV), voltage in mV at which the outward current was smallest prior to I NS beginning; I PS start (mV), voltage in mV at which I PS began; I PS (pA), current in pA at I PS peak, measured as difference between its beginning and its peak; I PS peak (mV), voltage in mV at which PS region peaked; I PS peakÀI h (pA), current in pA measured as difference between I NS peak and I h ; End I Na1 (pA), amplitude in pA of the current at the end of I Na1 inactivation; End I Na1 (mV), voltage in mV of the current at the end of I Na1 inactivation; V h (mV), holding potential in mV; Param, parameters measured; VM cells, data from ventricular myocardial cells; P cells, data from Purkinje cells; I Na3 (pA), amplitude in pA of I Na3 measured as the difference between its peak and the end of the step; Numbers in parenthesis (e.g., 17/17), number of cells with the parameter present over the total number of cells studied; *statistically significant difference between P and VM cells data.
As for the repolarizing ramps, with respect to P cells, in VM cells during the 260 mV sec À1 ramp (Table 6), the decreasing outward current was smaller (*) and peaked at a more positive potential (*). I PS began at a less negative potential (*) and was larger (*). It was larger also when measured as the difference from I h (*).
In VM cells, with respect to the 260 mV sec À1 repolarizing ramp, the 520 (Table 7) and 1040 mV sec À1 (Table 8) repolarizing ramps induced the following changes, respectively: I repol larger by +60.3% and by +220%*, the difference in initiation of I PS 3.1 and 2.1 mV, difference in voltage of I PS peak 0.1 and 2 mV, amplitude of I PS peak +0.2% and +4.2% and, when compared to I h , À5.0% and À18.1%. Therefore, on repolari-zation the decrease of I repol (but not I PS ) was sensitive to the repolarizing ramp slope.
Therefore, in P cells I Na3 and I NS became greater with faster depolarizing ramps, suggesting that sodium currents play a larger role in NS region of P cells than in that of VM cells. Also, during repolarizing ramp, the outward n, number of cells studied; D (mV or%), difference in mV or percent of VM cells data with respect to P cells data; I K1 peak (pA), amplitude of I K1 peak in pA, measured as the difference from I h ; I K1 peak (mV), voltage in mV at which I K1 peaked; I NS (pA), current amplitude in pA during the negative slope region; I NS , peak (mV), voltage in mV of I NS peak; I Na1 Th (mV), voltage at which I Na1 began; I Na1 (pA), amplitude in pA of I Na1 ; I Ca start (mV), beginning of I Ca component in mV determined as the departure of current trace from I NS peak; I Ca (pA), amplitude of I Ca component in pA as the difference between its beginning and its peak; I Ca peak (mV), peak in mV of I Ca ; I to , start (mV), voltage in mV at which the increasing outward current started at I NS or I Ca peaks; I to (pA), amplitude of outward current in pA measured between its beginning and ramp peak; I ramp peak ÀI h (pA), outward current at ramp peak measured as difference from I h ; I repol (pA), amplitude in pA of the outward current between ramp peak and its smallest value prior to the beginning of I PS ; I repol peak (mV), voltage in mV at which the outward current was smallest prior to I NS beginning; I PS start (mV), voltage in mV at which I PS began; I PS (pA), current in pA at I PS peak, measured as difference between its beginning and its peak; I PS peak (mV), voltage in mV at which PS region peaked; I PS peakÀI h (pA), current in pA measured as difference between I PS peak and I h ; End I Na1 (pA), Amplitude in pA of the current at the end of I Na1 inactivation; End I Na1 (mV), voltage in mV of the current at the end of I Na1 inactivation; V h (mV), holding potential in mV; Param, parameters measured; VM cells, data from ventricular myocardial cells; P cells, data from Purkinje cells; I Na3 (pA), amplitude in pA of I Na3 measured as the difference between its peak and the end of the step; Numbers in parenthesis (e.g., 17/17), number of cells in which the parameter was present over the total number of cells studied; *statistically significant difference between P and VM cells data. Other explanations as in the legend of Table 7.
current decreased more with faster ramps in both P and VM cells. With faster ramps, starting from a lower value, I PS increased in P cells, but decreased when measured as the difference from I h .

Differences in I NS in Purkinje versus myocardial cells
A decrease of V h from À90 mV to À60 mV markedly decreased the amplitude of I NS in P cells (Rota and Vassalle 2003;Vassalle et al. 2007; present results), but not in VM cells as measured using the sustained current at the end of depolarizing steps (Fig. 4). The finding suggests that the Na + current may play a predominant role in the mechanisms underlying I NS in P but not in VM cells. This was tested by applying ramps from different V h . In Figure 8A, in a VM cells, V h was progressively decreased by 1 mV and 520 mV sec À1 ramps were applied from V h indicated above the I h trace (À60 to À43 mV). At V h À60 mV (top A trace), I Na1 was truncated at À10 nA and I NS was followed by I Ca component (À151 pA, as indicated above the trace) peaking at +22 mV. With V h À54 mV, I Na1 was still larger than À10 nA, but beginning with V h -53 mV, I Na1 gradually Figure 8. Persistence of NS region and of I Ca with ramps from lower holding potentials in VM cells. In A, 520 mV sec À1 ramps were applied from the V h indicated above I h (À60 to À43 mV in 1 mV decrements; not all traces shown). I Na1 amplitude in nA is indicated next to the tip of I Na1 traces. In inset 1, the traces with V h À60 and À48 mV have been superimposed. The amplitude in pA of I Ca component (measured from its beginning to its peak) is indicated by the number above the traces. In inset 2, the superimposed traces show that I Ca component was not affected by gradually smaller V h . In B (from a different heart), depolarizing steps were applied from gradually lower V h and the currents were superimposed at the indicated voltages. The amplitudes of I Na1 and of I Ca (shaded area) are indicated in nA at the right hand of the traces. In inset 3, the voltage step was applied from V h À40 to +10 mV in a VM and P cell. decreased as indicated in nA next to the tip of I Na1 traces. Also, as I Na1 became much smaller, the activation and inactivation of the inward transient became slower, as illustrated in Figure 8 inset 1 by the superimposed V h À60 and À48 mV traces. At V h À43 mV (bottom trace), there was no apparent I Na1 and I NS had an amplitude of À462 pA, a duration of 36 msec and a voltage range between À40 and À12 mV. Over the range of V h tested, the I Ca component peaked at~+20 mV and its amplitude remained at~À150 pA (see numbers above the I Ca component traces and the superimposed traces in Figure 8 inset 2). That the inward component positive to I NS was due to I Ca is supported by the results obtained with depolarizing steps applied from V h À80, À70, À60, À50, and À40 mV (Fig. 8B). The amplitudes of I Na1 and of that of I Ca (shaded areas) are indicated, respectively, in nA at the right hand of traces. I Na1 decreased gradually with the less negative V h and disappeared with V h À50 and À40 mV. Instead, the slowly inactivating current persisted even with V h of À40 mV (albeit somewhat decreased) as expected for I Ca (Isenberg and Kl€ ockner 1982). In Figure 8 inset 3, depolarizing steps were applied from V h À40 to +10 mV and show that I Ca was far larger in the VM than in the P cell, as usual.
The results are consistent with Na + currents playing little role in I NS of VM cells and also with I Ca underlying the inward component that peaked at~+20 mV (as well as the small indentation over a similar voltage range in P cells).
In Table 9, in VM cells fast ramps were applied from V h À80 and À50 mV with the following changes. I K1 peak decreased (*) as the K1 channel rectified inwardly at lower V h before the ramp was applied. I NS was not affected whereas I Na1 markedly decreased (*). The I Ca component was similar. The ramp peak current (I to ) increased (the ramp peak voltage being more positive), although it was smaller when measured as I ramp peak ÀI h (*) due to the inward shift of the current with less negative V h . I PS and its peak voltage were similar. However, I PS peak measured relative to I h was 820 pA in control n, number of cells studied; I K1 peak (pA), amplitude of I K1 peak in pA, measured as the difference from I h ; I K1 peak (mV), voltage in mV at which I K1 peaked; I NS (pA), current amplitude in pA during the negative slope region; I NS , peak (mV), voltage in mV of I NS peak; I Na1 (pA), amplitude in pA of I Na1 ; I Ca start (mV), beginning of I Ca component in mV determined as the departure of current trace from I NS peak; I Ca (pA), amplitude of I Ca component in pA as the difference between its beginning and its peak; I Ca peak (mV), voltage in mV at which I Ca peaked; I to , start (mV), voltage in mV at which the increasing outward current started at I NS or I Ca peaks; I to peak (pA), amplitude of outward current in pA measured between its beginning and ramp peak; I ramp peak ÀI h (pA), outward current at ramp peak measured as difference from I h ; I repol (pA), amplitude in pA of the outward current between ramp peak and its smallest value prior to the beginning of I PS ; I repol peak (mV), voltage in mV at which the outward current was smallest prior to I NS beginning; I PS start (mV), voltage in mV at which I PS began; I PS peak (pA), current in pA at I PS peak, measured as difference between its beginning and its peak; I PS peak (mV), voltage in mV at which PS region peaked; I PS peakÀI h (pA), current in pA measured as difference between I PS peak and I h ; V h (mV), holding potential in mV; Param, parameters measured; VM cells, data from ventricular myocardial cells; P cells, data from Purkinje cells; *statistically significant difference between the data at the two V h values either in VM or P cells. and À5.0 pA* with the less negative V h , reflecting the inward shift of I K1 with V h À50 mV. In 6 of 7 of these experiments, V h was also decreased by 1-2 mV with the results similar to those illustrated in Figure 8A. In Table 9, in P cells, the same procedure decreased I K1 peak (*), I NS (*), and I Na1 (*). I Ca was not measurable (only small indentations) either in control or at less negative V h . "I to " and I ramp peak -I h increased but not significantly. Furthermore, I PS was smaller than in VM cells by 83.1% with V h À80 mV and was not present with V h À50 mV.
Thus, with less negative V h , I Na1 markedly decreased in both tissues during both ramps and depolarizing steps. In VM cells, I NS and I PS regions persisted even at V h À50 mV. Instead, in P cells with V h À80 mV, I PS was much smaller than I NS and, with V h À50 mV, I NS was markedly diminished and I PS abolished. Furthermore, in VM cells I Ca was consistently present with V h À80 mV and did not decrease with V h À50 mV. In P cells, I Ca was either not apparent or it was a small inward indentation on the outward current.

The transient outward current I to in Purkinje and myocardial cells
Some of differences between currents in P and VM cells in the À20 to +40 mV range (see Fig. 2) were analyzed by applying the depolarizing steps from V h À80 and V h À40 mV in the absence and presence of I to blocker 4-aminopyridine.
In Figure 9A, in P cell, during the step from V h À80 to À20 mV, I Na2 slowly decayed (arrow). A small outward peak appeared at the beginning of the step to 0 mV and progressively increased in size with more positive steps. In Figure 9B, in VM cell from a different heart, no decaying inward I Na2 was apparent and, instead, a smaller outward current grew in amplitude over the range À20 to +40 mV. The sustained current at the end of the steps was smaller than in P cell (note the different current calibration).
In the presence of 4-AP in the P cell (Fig. 9C) the peak I to was abolished, the decaying I Na2 was larger at negative values, and the sustained current was reduced. In the VM cell (Fig. 9D), 4-AP abolished the peak I to and reduced the sustained current. An inward transient was present at À20 mV which increased at +20 mV. Thus, at À20 mV in P cells the inward current was much larger, inactivated more slowly than the current in VM cell, and decreased markedly at + 20 mV.
In Figure 9 inset 1, V h was decreased to À40 mV in order to inactivate I Na2 Bocchi and Vassalle 2008; present results) but not I Ca (Isenberg and Kl€ ockner 1982). At V h À40 mV, I to channel is partially inactivated, mid-inactivation voltage being~À35 mV (Du- I Ca peak (mV), voltage in mV at which the I Ca was largest; I Ca (pA), amplitude of I Ca in pA measured from its peak to the end of the step; V h (mV), holding potential in mV; Param, parameters measured; VM cells, data from ventricular myocardial cells; P cells, data from Purkinje cells; s f (msec) and s s (msec), fast and slow time constants, respectively, of I Ca inactivation; Numbers in parenthesis (e.g., 18/18), number of cells in which I Ca was present over the total number of cells studied; *statistically significant difference between P and VM cells data. maine and Cordeiro 2007). In the P cell (Fig. 9E), the inward transients were markedly reduced and peak I to was decreased. In the VM cell (Fig. 9F), inward transients were present during the +20 (arrow), +30, and +40 mV steps and I to was present during the +60, +70, and +80 mV steps.
In the presence of 4-AP, in the P cell (Fig. 9G), I to was eliminated and the small slowly decaying inward component was largest at 0 mV. In the VM cell (Fig. 9H), in the presence of 4-AP, I to peak was eliminated and the sustained current was smaller. I Ca (arrow) was largest at +30 mV and reversed between +60 and +70 mV. V h to test step (mV), depolarizing steps from holding potential to voltage indicated; I to (pA), amplitude of I to in pA as the difference between I to peak and the end of 500 msec steps; I to ÀI h (pA), amplitude of the current measured as the difference between I to peak and I h ; Param, parameters measured; VM cells, data from ventricular myocardial cells; P cells, data from Purkinje cells; s f (msec) and s s (msec), fast and slow time constants, respectively, of I to inactivation; Numbers in parenthesis (e.g., 18/18), number of cells in which I to was present over the total number of cells studied; *statistically significant difference between P and VM cells data. Figure 9. I to in P and VM cells. In A-D, steps were applied from V h À80 to the À20/+40 mV range. In P cell, the arrow points to slowly decaying I Na2 during the À20 mV step (A). The currents in a VM cell taken from a different heart are shown in B. The procedure was repeated in the presence of 4-AP in P (C) and VM cell (D). In inset 1, steps were applied from V h À40 to +80 mV in P (E) and VM cell (F), where the arrow points to the current recorded at +20 mV. The procedure was repeated in the presence of 4-AP in P (G) and VM cell (H), where the arrow points the current recorded at +20 mV. VM cells from the same heart as in Fig. 2 Thus, I to patterns were distinctly different in that in the P cell I to peak was larger and declined more by the end of the step. Also, I to block by 4-AP unmasked inward currents with voltage range, magnitude, and speed of inactivation consistent with I Na2 in P cell and with I Ca in the VM cell. With V h À40 mV, in the P cell I Na2 was small and in the presence of 4-AP the current decreased slowly. In the VM cell, with V h À40 mV, I Ca was larger in the presence of 4-AP and inactivated quickly.
In Table 11, with V h À80 mV and test steps to +40 mV, in P cells I to when measured as the difference between its peak and sustained current at end of 500 msec steps was larger (*) as it was when measured as the difference between I to peak and I h (*). I to decreased from its peak to the end of the step by À27.1% in VM and by À45.1% in P cells, the slow time constant of inactivation being smaller in VM cells (*).
In VM cells, with V h À70 and À60, I to decreased by À40.3% and À93.2%, respectively. With V h À50 and À40 mV there was no apparent I to on depolarization to +40 mV. In P cells, with V h À70, À60, À50, and À40 mV, at +40 mV I to decreased by À9.7%, À27.4%, À41.7%, and À51.9%, respectively. With V h À40 mV, at +80 mV (past the reversal potential of I Ca ) in VM cells I to was smaller (*) than in P cells. The difference between peak I to and I h (I to ÀI h ) was larger than I to in both tissues and was much larger (*) in P cells at all V h . Thus, with respect to P cells, in VM cells I to peak: (1) was much smaller; (2) decreased less to a smaller sustained current; (3) was reduced more or altogether eliminated by less negative V h ; and (4) could be made to reappear with steps from À40 to +80 mV, still being much smaller than in P cells. Figure 10. TTX does not eliminate I NS and Ni 2+ abolishes I Ca component. In A, during the 260 mV sec À1 ramp, I Na1 was superimposed on I NS which was followed by the I Ca component (shaded area). TTX (30 lmol L À1 , B) eliminated I Na1 (short arrow) but not I NS as confirmed by the presence of I PS . Ni 2+ (2 mmol L À1 , C) abolished the I Ca component (long arrow). Recovery is shown in D. In inset 1, depolarizing steps from V h À40 mV to +20 mV elicited I Ca (control, E) which was not suppressed by TTX (TTX, F) but was eliminated by Ni 2+ (TTX + Ni 2+ , G). During the recovery from TTX and Ni 2+ , I Ca returned to control value (recovery, H).

Effect of tetrodotoxin on I NS and of nickel on I Ca component
In P cells, the sodium channel blocker tetrodotoxin (TTX) markedly reduces I NS ), including the fraction caused by I Na3 (Rota and Vassalle 2003). As in VM cells, sodium currents do not seem to play a dominant role in I NS , TTX would not be expected to suppress the NS region. Also, if I Ca is responsible for the inward component positive to I NS , the Ca 2+ channel blocker Ni 2+ should eliminated it.
In Figure 10A, control, in the VM cell during the 260 mV sec À1 ramp I Na1 was superimposed on NS region and the I Ca component was emphasized by shaded area. In Figure 10B, TTX (30 lmol L À1 ) eliminated I Na1 (short downward arrow) and only I NS was left. That indeed I NS was not abolished by TTX is confirmed by the fact during the repolarizing ramp a distinct I PS was present as in control.
In the presence of TTX, I Ca (shaded area) was greater than in control. A possible reason for this increase might be that TTX decreases intracellular Na + activity (Abete and Vassalle 1988;Iacono and Vassalle1990), thereby increasing the transmembrane Na + gradient. In turn, a larger Na + gradient increases the extrusion of Ca 2+ through an enhanced Na + -Ca 2+ exchange: this would increase the transmembrane Ca 2+ gradient and therefore I Ca .
In Figure 10C, Ni 2+ (2 mmol L À1 ) altogether abolished I Ca , as emphasized by the long downward arrow. Although 2 mmol L À1 Ni 2+ blocks also the Na + -Ca 2+ exchange, such a block seems unlikely to account for the observed phenomenon, as in the voltage range of I Ca , the Na + -Ca 2+ exchange would be operating in the reverse mode generating an outward current.
During the recovery in physiological saline solution (Fig. 10D), I Na1 and I Ca reappeared.
In another approach (Fig. 10, inset 1), depolarizing steps were applied from a V h of À40 mV so that Na + channels (but not I Ca channel) were inactivated. In Figure 10E, control, a step from V h À40 mV to À30 mV did not elicit time-dependent currents, whereas depolarization to +20 mV elicited I Ca . In Figure 10F, TTX (30 lmol L À1 ) did not suppress I Ca at +20 mV. In Figure 10G, adding Ni 2+ (2 mmol L À1 ) to TTX solution eliminated I Ca and allowed a small decaying current to appear (presumably an unmasked small I to ). In Figure 10H, during the recovery from TTX and Ni 2+ , I Ca returned to the control value. Similar results were obtained in n = 2.

Effects of 4-AP and Ba 2+ on fast and slow ramps in myocardial cells
In Purkinje cells, Cs + and Ba 2+ (blockers of I K1 ) markedly decreased I K1 as well as the slope conductance, but only slightly reduced the outward current at positive potentials (Du and Vassalle 1999). Furthermore, Ba 2+ eliminated I K1 peak but not I NS , whereas low V h and TTX eliminated I NS (Rota and Vassalle 2003;Vassalle et al. 2007). As neither a low V h nor TTX abolished I NS in VM cells (present results), in VM cells I NS might be mostly related to Mg 2+ Figure 11. NS and PS regions are abolished by Ba 2+ . In A, 260 mV sec À1 ramps were applied in control, in the presence of 4-AP, of 4-AP plus Ba 2+ , and during recovery. In B, the same procedures were applied during 6.5 mV sec À1 ramp. The larger I Ca component in the presence of 4-AP is labeled with a rhombus. In the two panels, the left hand double headed arrows emphasize the disappearance of I NS and the right hand double headed arrows that of I PS . In inset 1, the current traces recorded between I Ca and ramp peaks in B have been superimposed: C, control; 4-AP, in the presence of 4-aminopyridine; +Ba 2+ , in the presence of 4ÀAP and Ba 2+ ; R, recovery in physiological saline solution. block of I K1 . To verify this point, in one experiment fast and slow ramps were applied to test whether Ba 2+ (by blocking I K1 and therefore preventing the block and unblock by Mg 2+ in the NS and PS regions, respectively) abolished I NS and I PS .
In Figure 11A, in a VM cell, during a 260 mV sec À1 ramp (bottom trace), in control the current exhibited the usual features. In the presence of 4-AP, I NS was little changed whereas I Ca component became larger (+129%, rhombus). While overall the current was less outward, I K1 and I PS peaks were somewhat larger. The current between I Ca and ramp peaks was unchanged (+1.6%) as in P cells (Du and Vassalle 1999). In the presence of Ba 2+ , not only the I K1 peak, but also I NS and I PS were abolished (as pointed out by double headed arrows). Neither the outward current between I Ca and ramp peaks nor the initial fast decrease in current at beginning of repolarization was suppressed. I NS and I PS reappeared during recovery (bottom current trace).
In Figure 11B, during a 6.5 mV sec À1 ramp (bottom trace), in the presence of 4-AP, I K1 peak was also somewhat larger and so was I Ca . Instead, the current between I Ca and ramp peaks was not affected (À0.6%, see 4-AP trace in Figure 11 inset 1). Ba 2+ abolished I K1 peak and I NS , as pointed out by the left hand double headed arrow, leaving only the slowly increasing outward current. As in P cells (Du and Vassalle 1999), the current during the last part of the depolarizing ramp was unaffected (see Ba 2+ trace in Figure 11 inset 1). During the repolarizing ramp, the current decreased continuously (no I PS , as indicated by the right hand double headed arrow). The recovery from the procedures is shown by the bottom current trace and by trace R in Figure 11 inset 1).
Thus, while in P cells Ba 2+ abolished I K1 peak, but not I NS (Rota and Vassalle 2003), in VM cells it abolished I K1 peak, I NS and I PS . However, neither 4-AP nor Ba 2+ abolished the increase in outward current between I Ca and ramp peaks.

Discussion
The present results show numerous dissimilarities between VM and P cell currents. The features involved concern: (1) the slope conductance at the resting potential and its changes during depolarization and repolarization; (2) the sodium currents I Na1 , I Na2 , and I Na3 (presence or absence, amplitude, threshold, voltage and time dependence, contributions to peak sodium current); (3) outward and inward current during ramps with different slopes; (4) time and voltage dependence of I K1 blocking and unblocking; (5) NS and PS regions (voltage range, voltage and time dependence, slope conductance, depolarization vs. repolarization); (6) prevalent role of slowly inactivating sodium currents in the mechanisms underlying NS region in P cells; (7) prevalence of block and unblock of I K1 in NS and PS regions, respectively, in VM cells; (8) characteristics of I Ca as well as I to and sustained current; (9) different contributions of I K1 and I to to the I-V relation; (10) voltage and time dependence of currents during repolarizing ramps, and (11) response to channels blockers.
We conclude that the differences between VM and P cells involved all the ionic currents studied and account for several electrophysiological differences in resting and action potentials.

Membrane conductance as a function of voltage and time in VM and P cells
As at negative potentials I K1 predominates in determining the I-V relation (Shah et al. 1987), at potentials negative to I K1 peak the I-V relation was taken to essentially reflect that of I K1 . Near the resting potential, the larger slope conductance in VM cells (+100%) is consistent with the findings that I K1 in rabbit myocardial ventricular cells is larger than in P cells (Cordeiro et al. 1998). Indeed, the expression of transcripts underlying I K1 channel (Kir2.1, Kir2.2, Kir2.3) in human Purkinje fibers is about half that in ventricular myocardium (Gaborit et al. 2007).
The larger slope conductance in VM cells at resting potential could be related to the fact that I K1 channels are located also in T-tubules (e.g., Lopatin and Nichols 2001), which are absent in P cells (see Vassalle et al. 1995 for references). In VM cells, the larger I K1 peak could possibly be due to the larger resting conductance and/or to lower polyamines level in myocardial cells.
The smaller conductance in P cells would seem at odds with the fact that in 2.7 mmol L À1 [K + ] o their maximum diastolic potential (~À95 mV; e.g., Vassalle 1965) is more negative than the resting potential in VM cells (~À80 mV). This discrepancy is due to the presence of the pacemaker current I Kdd , (the potassium current underlying diastolic depolarization; Vassalle 1966;Vassalle et al. 1995;Vassalle 2007Vassalle , 2013, which is present in P cells, but not in VM cells. Thus, in quiescent P cells in 5.4 mmol L À1 [K + ] o the resting potential is~À80 mV and does not increase when [K + ] o is lowered to 2.7 mmol L À1 (Vassalle 1965(Vassalle , 1966 due to K + -dependent fall in K 1 channel conductance. However, in cells active in 2.7 mmol L À1 [K + ] o , the activation of I Kdd during the AP (Vassalle 1966;Vassalle et al. 1995) is responsible for the voltage undershoot to the maximum diastolic potential. As I Kdd deactivates as a function of time in the diastolic potential range, the pacemaker potential and slope conductance decrease toward the resting potential value (Weidmann 1951 1965,1966). Therefore, in P cells diastolic conductance at resting potential is mostly a function of I K1 (as in VM cells) and at the maximum diastolic potential of both I K1 and I Kdd .
During ramp depolarization, the progressive decline of slope conductance to a minimum at I K1 peak shows that the gradually smaller increase in outward current is due to I K1 inward rectification, rather than a progressive increase of an inward current. At I K1 peak, I K1 is matched by the inward current, as the outward current stopped increasing and the slope conductance became minimal (Fig. 6). Soon after, in both tissues, the activation of I Na3 (together with the decreasing I K1 ) accounts for the initiation of the NS region.
During I NS , the reversal and reincrease of pulse current indicate that I K1 became smaller than the inward current. In fact, I K1 reaches its minimum at the peak of I NS (À20 mV in P cells) as shown by I K1 I-V relation (control minus I K1 blockers) (Shah et al. 1987;Cordeiro et al. 1998). In P cells, the reversal of pulse current occurred both during the part of I NS due to I Na3 (Rota and Vassalle 2003) and that due to the inactivation of I Na2 (Bocchi and Vassalle 2008). The relation of the reversed pulse current to the activation and decay of I Na2 is demonstrated by the findings that during depolarizing steps the amplitude of pulse current became larger when I Na2 appeared, decreased gradually during I Na2 slow inactivation, and was markedly reduced by lidocaine (Bocchi and Vassalle 2008).
However, as pulse current reversal in NS range occurred also during the 6.5 mV sec À1 ramp (Fig. 6E) when the Na + currents are inactivated (and in VM cells the inactivating I Na2 would play little role), the reversal appears to involve also the block of I K1 channel. In both cases (increase in Na + currents and decrease in I K1 ) the net current would become inward (I NS ) and therefore the pulse current would reverse. The role of I K1 change in pulse current reversal is supported by the occurrence of the reversal also during I PS (Fig. 6D and E), when any possible residual Na + currents would be inactivated.
Contribution of I Na3 , I Na2, and rectification of I K1 to NS region in P and VM cells Our findings show that the mechanisms underlying I NS in P cells differ qualitatively and quantitatively from those in VM cells and account for some previously reported results.
In the presence of TTX and of Ca 2+ blocker nicardipine in rabbit, the NS region was present in VM cells but not in P cells (Cordeiro et al. 1998). This difference is accounted by the finding that I Na3 (Rota and Vassalle 2003) and I Na2 Bocchi and Vassalle 2008) are blocked by TTX. However, when measured as the difference current (control minus Ba 2+ ), a small NS region was present also in P cells (Cordeiro et al. 1998), suggesting that, in the absence of Na + and Ca 2+ currents, the residual I NS was due to a block of I K1 .
Similarly, in P cells with steps from a V h of À50 mV, the NS region was not always found, but it was present in I K1 I-V relation (control minus current in Ba 2+ or Cs + ) (Shah et al. 1987). In retrospect, at V h À50 mV the Na + currents would have been inactivated or markedly reduced (Rota and Vassalle 2003;Vassalle et al. 2007;present results). Therefore, the findings of Shah et al. (1987) can also be accounted for by the inactivation of Na + currents at À50 mV and a small contribution of I K1 block to I NS in P cells. That in P cells, I K1 rectification contributes to I NS is also indicated by a net decrease of radioactive K + efflux in the À60 to À40 mV range (e.g., Vereecke et al. 1980).
In our experiments, in the absence of blockers, with respect to VM cells, in P cells the increase in I NS with the steeper ramps and its decrease or absence with 6.5 mV sec À1 ramp are consistent with time-dependent inactivation of I Na3 and I Na2 . In P cells, with V h À80 mV, the smaller I NS in I-V relation of the sustained current is consistent with inactivated I Na3 and substantially reduced I Na2 by the end of 500 msec steps. In addition, the gradual decrease of I NS with less negative V h in P but not in VM cells (Fig. 4, inset 1) indicates a voltage-dependent inactivation of Na + currents.
Conversely, with 6.5 mV sec À1 ramps the much smaller I NS in P cells is explained by I Na3 and I Na2 being reduced or absent, in agreement with the little effects of TTX on I NS and by the similar values of I NS and I PS . Furthermore, with V h À50 mV, during fast ramps the near abolition of I NS in P cells but not in VM cells also points to a greater role of Na + currents.
In P cells, TTX, lidocaine, and low V h markedly reduced I Na3 (Rota and Vassalle 2003) as well as I Na2 ) and so did low [Na + ] o (Bocchi and Vassalle 2008). Yet, I NS persisted in VM cells with some of these procedures (present results). Furthermore, in P cells, Ba 2+ abolished I K1 peak but not I NS (Rota and Vassalle 2003) whereas it abolished I K1 peak, I NS , and I PS in VM cells.
These results indicate a predominant role of Na + currents in P cell I NS and of I K1 inward rectification in VM cell I NS . Indeed, in guinea pig ventricular myocytes, the K + channel opener cromakalim abolished I NS and markedly shortened the action potential (Liu et al. 1990).
As for voltage ranges of Na + currents and of I NS , the beginning of I NS during the ramps (~À58 mV) indicates the participation of I Na3 as in P cells this current started at~À58 mV (Table 8; Rota and Vassalle 2003) and peaked before or by the end of ramps to À42 mV (Rota and Vassalle 2003). In both P and VM cells, I Na3 contrib- uted to I NS prior to I Na1 , but (its slow inactivation being cut off by I Na1 ) presumably little after I Na1 .
In P cells, the slowly inactivating I Na2 can contribute to I NS after the inactivation of I Na1 , as I Na2 has a À40 mV threshold and is largest at a voltage (À30/À20 mV, Vassalle et al. 2007) which is near to the peak of I NS (À26.5 mV with 260 mV sec À1 ramp). I Na2 activates also in the absence of I Na1 (Fig. 5B and E), in agreement with the findings of Bocchi and Vassalle (2008). Indeed, in those P cells in which I Na1 was not present with 260 mV sec À1 ramps, I NS peaked at a potential (À29.2 mV, present results) near the I Na2 peak. Therefore, with respect to VM cells, in P cells I NS : (1) had a smaller voltage range; (2) was larger with faster ramps; and (3) was smaller with slow ramps, with lower V h and in the presence of TTX. These finding indicate a predominant role of the sodium currents in I NS of P cells and of I K1 block in VM cells.

Dual mechanism of I K1 inward rectification and the NS and PS regions in P and VM cells
As for the inward rectification of I K1 channel, two mechanisms have been demonstrated: block by intracellular polyamines (channel gating by spermine and spermidine; Lopatin et al. 1994;see Lopatin and Nichols 2001) and block by Mg 2+ (Matsuda et al. 1987;Vandenberg 1987). The I K1 block by spermine and spermidine is time dependent (Ishihara 1997), whereas the block by Mg 2+ and putrescine is voltage dependent and virtually instantaneous (Ishihara and Ehara 1998).
In guinea pig VM cells, in the absence of internal Mg 2+ , the block by polyamines occurs between À80 and À40 mV whereas, with in the presence of internal Mg 2+ , I K1 block is present also between À40 and 0 mV (Ishihara 1997). Furthermore, after a depolarization larger than 0 mV (which would cause Mg 2+ block), on repolarization to À50 mV there was a sudden transient increase in outward current (see also Shimoni et al. 1992). The amplitude of the outward current was correlated to the degree of Mg 2+ block during the previous depolarization, indicating that the increase in outward current was due to the removal of Mg 2+ block. During a repolarizing ramp, the outward current at À50 mV was substantially larger in the presence than in the absence of internal Mg 2+ , indicating the importance of the removal of Mg 2+ block of I K1 for the repolarization of the action potential (Ishihara 1997;Ishihara and Ehara 1998).
Because I K1 block by polyamines begins at more negative voltage than that by Mg 2+ (Ishihara et al. 1989;Ishihara 1997;Ishihara and Ehara 1998), the block of I K1 ought to be solely due to polyamines up to I K1 peak which with the slowest ramp occurred at À44.2 mV in VM cells (Table 5). Instead, the voltage range of the I K1 block by Mg 2+ (À40 to 0 mV) overlaps the NS region in VM cells.
Time dependence of I K1 block by polyamines at the beginning of depolarizing steps would account for the initial decline of the outward current. On step repolarization from positive potentials toward E K the sudden increase in outward current due to instantaneous relief of Mg 2+ block gradually declined due to time-dependent block by spermine and spermidine (Ishihara 1997). The time constant of the exponential decay of the outward current was 5 msec at À50 mV which is close to the s of~6 msec for the current decline on depolarization from V h À80 tõ À60 mV (Table 4). As for the symmetrical voltage ranges of NS and PS regions, if the voltage-dependent block of I K1 by Mg 2+ initiates near the beginning of I NS , its complete removal at the peak of I PS (ceteris paribus) should occur at a similar potential. Indeed, in VM cells, with the slowest ramp, I NS initiated at À44.2 mV and I PS peaked at À44.4 mV and their amplitudes were similar (218 and 197 nA, respectively). The peak of I NS (full block) was at À3.9 mV and the beginning of I PS was 8.9 mV (initiation of block removal) ( Table 5).
With faster ramps, other factors modified I K1 and I PS peaks. In P cells, the larger I NS with faster ramps implicates additional time-dependent factors such as a greater activation of I Na3 and of I Na2 . In keeping with this conclusion, in P cells with 6.5 mV sec À1 ramp, I NS was much smaller and so was I PS , as the Na + currents would be inactivated during the depolarizing and repolarizing ramp.
That I K1 undergoes a progressively greater inward rectification on depolarization from the V h is also shown by the progressively smaller increase in sustained outward current at the end of 500 msec depolarizing steps (when most of the other currents would be completely or partially inactivated). Actual decline of the sustained current at potentials positive to À50 mV in P and to À40 mV on VM cells reflects the I NS seen during the ramps. The less negative potential at which the sustained current decreased with respect to the beginning of I NS during the ramps might be ascribed to lack of I Na3 contribution to the sustained current.
At more negative potentials, block and unblock of I K1 might contribute to several changes in the I-V relation during ramps of different steepness. Thus, voltage-and time-dependent block by spermine and spermidine on depolarization from V h À80 mV to I K1 peak would account for the gradually smaller increase of outward current during a ramp and for the decreasing slope conductance both in P (Rota and Vassalle 2003;Bocchi and Vassalle 2008;present results) and in VM cells (Fig. 6).
As at less negative potentials the channel block by polyamines is much slower (Ishihara et al. 1989;Ishihara and Ehara 1998), such a block may also contribute to the current changes that occur during the ramps with steeper slopes. A greater lag between faster voltage change and block of I K1 channel might be responsible for the increase in magnitude of I K1 peak with faster ramps. Also, the larger decrease in outward current during faster repolarizations might include a delay in the removal of I K1 block by polyamines.
The fact that during repolarizing 260 mV sec À1 ramps, in P cells I PS was much smaller (À81%) than I NS (Table 6) suggests that I Na3 and I Na2 have a major role in I NS and that the remainder (~19%) is due to I K1 block by Mg 2+ . The removal of I K1 block would account for the I PS beginning being more outward (no Na + currents contribution) than the I NS peak and I PS peak being similar to I NS beginning. In contrast, in VM cells the similarity of I NS and I PS (Table 6) suggests that block and unblock of I K1 channel by Mg 2+ were the major mechanisms underlying I NS and I PS , respectively. This is consistent with Mg 2+ block and unblock being only voltage dependent and with the slowly inactivating I Na2 not contributing to either I NS or I PS in VM cells. During the slowest ramps, there would be little or no contribution by Na + currents to either I NS or I PS , as supported by the similarity of I NS and I PS in VM as well as in P cells (Table 5).
In VM cells, with 260 mV sec À1 ramp the larger difference between I PS peak and I h (Table 6) reflected the almost double I K1 upon which I PS was superimposed. In contrast, the similarity of the I PS peak voltage in VM and in P cells is consistent with the voltage dependence of the removal of I K1 block by Mg 2+ . With the 6.5 mV sec À1 ramp, the much larger I NS and I PS in VM than in P cells support the role of Mg 2+ block and unblock, respectively, as in both tissues the Na + currents would be largely inactivated during both depolarizing and repolarizing ramps. Therefore, the time-and voltage-dependent block of I K1 by polyamines appears to prevail at potentials negative to the I K1 peak whereas the inward rectification of I K1 during I NS is attributable to Mg 2+ block. This conclusion is supported by the findings that in P cells: (1) when I Na3 and I Na2 were present, I NS was larger than in VM cells and more so the faster the ramp; (2) when I Na3 and I Na2 were inactivated (repolarizing ramps) I PS was smaller in P cells; and (3) with the slowest ramp (I Na3 and I Na2 being inactivated during the depolarizing and repolarizing phases), both I NS and I PS were smaller in P cells.

I Na3 in Purkinje and myocardial cells
With the approach of Rota and Vassalle (2003), in P cells I Na3 was mostly studied at potentials negative to I Na1 threshold. With depolarizing ramps (in the absence of I Na1 ), I NS began at À57.7 mV and peaked before or by the end of the ramp at À42 mV. I NS was attributed to the activation of I Na3 , as shown by its threshold, the marked reduction by TTX and lidocaine and its little sensitivity of Cs + and Ba 2+ . In contrast to I Na3 , TTX or lidocaine did not abolish I CaT (Tytgat et al. 1990). In addition, I Na3 was markedly reduced by 70 mmol L À1 [Na + ] o and it was not abolished by 100 lmol L À1 Ni 2+ (M. Rota and M. Vassalle, unpubl. experiments). These findings contribute to rule out that I Na3 might in actuality be I CaT .
Suitably large and slow ramps that did not activate I Na1 initiated I Na3 at about À60 mV followed at about À40 mV by the activation of I Na2 (Rota and Vassalle 2003). In retrospect, these results in P cells separated for the first time the contributions of I Na3 and I Na2 to I NS . That I Na3 is a sodium current also in VM cells is shown by its decrease with less negative V h (depolarizing steps, Table 1) or ramps (Table 9).
A population of slowly inactivating Na + channels has been also reported in giant squid action. These channels are much fewer than the normal Na + channels, activate on depolarization to~À65 mV whereas I Na threshold is À50 mV, activate maximally at À40 mV and undergo a very slow inactivation (Gilly and Armstrong 1984), similar to I Na3 . As for the mechanism by which I Na1 block I Na3 slow inactivation, it appears that fast I Na1 inactivation blocks slow inactivation of Na + channels by charge immobilization (Richmond et al. 1998).
As to Na + channel isoforms involved in I Na3, among several Na + channels isoforms cloned (Na V 1 to Na V 9; cardiac, neuronal, and skeletal), cardiac Na V 1.5 isoform has a low TTX-sensitivity, whereas the neuronal (Na V 1.1, Na V 1.2, Na V 1.3, Na V 1.6) and skeletal muscle (Na V 1.4) isoforms have a high TTX-sensitivity (Zimmer et al. 2002;Haufe et al. 2005a). The neuronal Na v 1.2, Na v 1.3, and Na v 1.6 isoforms are expressed in VM cells and Na v 1.1 and Na v 1.2 (Haufe et al. 2005a) as well as skeletal N aV 1.4 isoform (Qu et al. 2007) have been identified in P cells. Although I Na3 and I Na2 are more sensitive to TTX block than I Na1 , the noncardiac channels isoform involved is not known, as the skeletal muscle Na + channel isoform Na v 1.4 expressed in P cells is also blocked by low concentrations of TTX (Qu et al. 2007).
The neuronal TTX-sensitive Na + channels were found to contribute to peak sodium current by 22% in P cells and by 10% in VM cells (Haufe et al. 2005b) and therefore to the sodium current responsible for the AP upstroke. Our results show that at the I Na1 threshold, I Na3 was a sizeable fraction of the total I Na current. However, because I Na1 was truncated by the saturation of the amplifier consistently in P cells and often in VM, the I Na3 contribution to the total I Na is bound to be overestimated. In addition (as shown by means of double steps), the fast inward component of I Na2 would be expected to contribute to the total Na + influx during the upstroke. I Na3 appears to be the link between DD and upstroke of AP in P cells by being responsible (Rota and Vassalle 2003) for the depolarizing phase of ThV os (the oscillatory potentials near the threshold for the upstroke ;Vassalle 1965;Spiegler and Vassalle 1995;Berg and Vassalle 2000). Successive ThV os increase progressively in size during diastole until the depolarizing phase becomes large enough to attain the threshold for I Na1 (Vassalle 1965;Spiegler and Vassalle 1995;Berg and Vassalle 2000). In P fibers, TTX suppressed the spontaneous discharge by abolishing the upward swing of the late diastolic depolarization, although the fibers could still be electrically driven due to smaller sensitivity of I Na1 to TTX (see Fig. 3 by Vassalle and Scid a shown in Vassalle 1980). ThV os are present also in sino-atrial node (Kim et al. 1997) and are essential for both the initiation and maintenance of spontaneous activity of cardiac pacemakers (Vassalle 2007(Vassalle , 2013.
In VM cells, the voltage gap between the resting potential and the threshold for I Na3 is closed by the depolarization induced by the conducted AP, as of necessity I Na3 threshold is less negative than the resting potential. I Na3 would then speed up the attainment of I Na1 threshold and the induction of the upstroke.
While a 5 mmol L À1 [Na + ] o has been successfully used to compare Na + currents in normal and diseased VM cells (Pu and Boyden 1997;Valdivia et al. 2005), the low [Na + ] o solution markedly reduced I Na1 . This apparently makes it unsuited to study the much smaller I Na3 , as no slowly inactivating I Na3 was apparent in those studies. In our experiments, I Na3 was demonstrated under physiological conditions at potentials negative to I Na1 threshold. Also, less negative V h markedly reduced I Na1 (Table 2) leaving a separately identifiable slowly inactivating I Na3 (Table 1).
Thus, the present results show that I Na3 : (1) was also present in VM cells; (2) had threshold that was negative to that of I Na1 and was less negative than that in P cells; (3) inactivated slowly; (4) could occur at potentials less negative than I Na1 threshold (as shown by preventing I Na1 activation); (5) was less inactivated at a lower V h than I Na1 ; (6) increased in amplitude during depolarizing steps past~À60 mV due to its voltage-dependent activation and during faster depolarizing ramps (greater Na + channel availability); (7) contributed to magnitude of total Na + influx associated with I Na1 ; (8) was absent during repolarizing ramps; (9) was associated with a reversal of pulse currents and with an increase in slope conductance during I NS ; and (10) its slow inactivation was suppressed by I Na1 .
A limitation of our study is represented by the lack of measurements of membrane capacitance (Cm) in the two cell populations analyzed, posing concerns related to the comparison of ionic current amplitudes. In this regard, values of Cm of 121.9 AE 4.8 pF (Rota and Vassalle 2003) and of 125 AE 6 pF (Han et al. 2000) have been reported for canine P cells, whereas Cm values of 113 AE 6 pF (Han et al. 2000) and of 133.4 AE 6 pF (Pu and Boyden 1997) have been reported for canine VM cells. Thus, the average value of the above capacitances for the P cells is 123.4 pF and that for the VM cells is 123.2 pF.
While P cells have no T tubule system, which is present in VM cells, P cells are larger in size than VM cells. Thus, these two features (cell size and T tubule system) tend to offset each other in determining the surface area in the cell populations.
The fast sodium current I Na1 in P and VM cells The measurement of I Na1 magnitude in P and VM cells was hindered by its being truncated by the saturation of the amplifier. However, the finding that with V h À80 mV I Na1 was consistently cut off at À10 nA in P cells but only in 13/18 in VM cells indicates that I Na1 was smaller in VM than in P cells. Furthermore, with less negative V h (Na + channels being partially inactivated), I Na1 was less often truncated and yet in VM cells I Na1 was still smaller than in P cells. Also, the less frequent I Na1 activation during the 260 mV sec À1 ramp in P cells suggests that during this slower ramp I Na1 channels are more susceptible to time-dependent inactivation in P than in VM cells.
Several differences regarding I Na1 in P and VM cells have been reported. Thus, total amplitude and maximal rate of rise of AP upstroke were much larger in Purkinje than in ventricular fibers (Bal ati et al. 1998). Also, neuronal sodium channels contribute less to the peak sodium current in dog ventricular than in Purkinje fibers (Haufe et al. 2005b). A larger I Na1 with a more negative threshold in P cells (present results) would contribute to the faster conduction of P cells with respect to VM cells. Both in nonspontaneous P and VM cells, conducted action potentials would depolarize the membrane to the threshold of I Na3 , which in turn would allow the attainment of I Na1 threshold, as seen during the ramps.

The slowly inactivating I Na2
In VM cells, the absence of a large slowly decaying component of I Na2 comparable to that of P cells accounts for the fact that (see Introduction), in contrast to P cells, the AP duration of ventricular myocytes was very little affected by TTX, local anesthetics, veratridine, and high [Na + ] o , suggesting that there is little or no slow decaying I Na2 in VM cells under physiological conditions. I Na2 is more sensitive to TTX than I Na1 ) and therefore is possibly due to the activation of neuronal or skeletal muscle isoforms. In P cells, the larger contribution of the noncardiac Na + channels (Haufe et al. 2005b) might be due to both I Na3 and I Na2 . In P cells, the fast activation of I Na2 (Bocchi and Vassalle 2008) would contribute to the influx of Na + during the upstroke and its slow inactivation contributes to the longer plateau. A late I Na was found in canine myocardial cells perfused in a K + -free and very low Cl À medium by applying a 2000-msec pulse to À130 mV (to remove steady-state inactivation) prior to depolarizing steps applied at intervals of 30 sec (Zygmunt et al. 2001). Our results show that I Na2 is small or absent under physiological conditions in VM cells, in contrast to P cells.
However, the findings of Zygmunt et al. (2001) could suggest that under abnormal situations remodeling may shift the voltage range of the late I Na activation to less negative values. In this regard, it is of interest that I NaL was increased in cardiac failure and yet no differences were found in isoform NaV1.1, 1.3, 1.5 subunits and in the subunit b1 and b2, leading to the conclusion that I NaL increase was not due to a subunit isoform switching or to an altered b subunit expression (Valdivia et al. 2005).
In normal human ventricular myocytes, there is little or no I Na2 but in myocytes from patient with hypertrophic cardiomyopathy remodeling leads to the appearance of a late Na + current. Ranolazine, a blocker of the late Na + current, had negligible effect on the action potential duration of normal ventricular myocytes, but shortened the longer AP of the myopathic myocytes and reduced the related arrhythmias (Coppini et al. 2013). Therefore, under some pathological conditions remodeling-related I Na2 may induce arrhythmias also in VM cells.
As for role of P cell slowly inactivating I Na2 in cardiac arrhythmias, an increase in I Na2 (but not of I Na1 ) by neurotoxins anthopleurin or ATX II in Purkinje fibers led to the onset of Torsades de pointes. This electrophysiological mechanism appears responsible for congenital and acquired long QT syndromes (LQTS), as abnormally prolonged repolarizing phase of AP leads to early afterdepolarizations (EADs) in Purkinje fibers. In vitro and in vivo, in Purkinje fibers the neurotoxins induced EADs were abolished by concentrations of mexiletene that had little effect on I Na1 (El-Sherif and Turitto 2003). The present characterization of differences between P and VM cells add insights relevant to mechanisms underlying some LQTS syndromes, especially the drug-induced acquired syndromes.

The calcium current in P and VM cells
That I Ca was larger in VM cells was shown by the appearance of a large inward component, which peaked in the +10 to +20 mV range and became more conspicuous after the block of I to by 4-AP. Instead, in P cells, the smaller I Ca appeared as an indentation on I to trace or it was not altogether apparent, even in the presence of I to block by 4-AP.
That the inward component peaking at positive values was due to I Ca was shown by its persistence at low V h and by the abolition by Ni 2+ of both the inward component during the ramps and I Ca during depolarizing steps. In VM cells, I Ca -induced positive shift of the onset of the smaller outward current would contribute to the more positive plateau and the larger I Ca to the greater force of contraction (Lin and Vassalle 1978) with respect to P cells.
As for the identity of Ca 2+ currents, the L-type (I CaL , isoform Ca v 1.2 containing the pore forming a1C subunit) and T-type (I CaT, isoforms Ca v 3.1, Ca v 3.2, and Ca v 3.3) are expressed both in P and M cells, I CaT being more abundantly expressed in P cells and I CaL predominating in VM cells (see Dun and Boyden 2008). These differences suggest that calcium current recorded in VM cells is of the I CaL type, in accordance with fact that it was recorded with less negative V h and at less negative voltage.

The transient outward current I to in P and VM cells
With respect to VM cells, in P cells a larger I to contributed to the much larger outward current between I NS and ramp peaks, consistent with the shift of the plateau to more positive values by 4-AP (Kenyon and Gibbons 1979;Dumaine and Cordeiro 2007).
However, our results show that the total outward current included different components in P and VM cells, as in VM cells a smaller I to was superimposed on a larger I K1 . In VM cells, the smaller I to would induce a smaller phase 1 repolarization and (together with the larger I Ca ) would contribute to the more positive plateau than in P cells (e.g., see Fig. 1 in Bal ati et al. 1998).
The finding that 4-AP abolished the peak I to , but only reduced the sustained current at the end of depolarizing steps and the outward current during latter part of depolarizing ramps agrees with the results shown in other reports (no elimination of sustained current by 4-AP; Kenyon and Gibbons 1979;Cordeiro et al. 1998;Han et al. 2000). Therefore, I to may have a time-independent background component or it may not inactivate completely (Kenyon and Gibbons 1979 Also, there might be the contribution of an unidentified outward current present at positive potentials which is not eliminated by 0 mmol L À1 [K + ] o (Ishihara et al. 1989). Even I K1 block by Cs + and by Ba 2+ did not suppress the increase in outward current at potentials positive to À40 mV, as shown by figures in Shah et al. (1987) and reported by Du and Vassalle (1999). This is consistent with the finding that 42 K efflux at potential positive to À30 mV was not blocked by Cs + (Vereecke et al. 1980). This could be due either to the voltage dependence of Cs + induced block as it is for Ba 2+ (Hirano and Hiraoka 1986) or to a Cs + insensitive outward rectifier. In the experiments of Cordeiro et al. (1998), 0.1 mmol L À1 Ba 2+ reduced the inward I K1 , but did not decrease the large outward current at positive potentials. In P cells (Du and Vassalle 1999) and in VM cells, 2 mmol L À1 Ba 2+ eliminated I K1 peak, but did not eliminate the outward current at positive potentials.
However, this per se does not rule out a contribution of I K1 to the outward current, as Ba 2+ block of I K1 is removed on depolarization and slowly reestablished on repolarization (Hirano and Hiraoka 1986;Imoto et al. 1987;Valenzuela and Vassalle 1991). Therefore, in P cells after I NS peak, the increase in outward current includes I to (which is not blocked by Ba 2+ , e.g., Cordeiro et al. 1998;Du and Vassalle 1999) and a current which is not suppressed by 4-AP, Ba 2+ , Cs + , or 0 mmol L À1 [K + ] o . The delayed rectifier I K could contribute to the outward current, but is blocked by Ba 2+ , albeit to a lesser extent at positive values (Osterrieder et al. 1982). In addition, in P cells I K is rather small (<20 pA, Cordeiro et al. 1998) or undetectable (Kenyon and Gibbons 1979). Indeed, in the presence of 4-AP the outward current did not increase with depolarizing steps from V h À40 mV (Fig. 9).
The increase in peak outward current with steeper ramps in P cells could be due to a larger I to and/or to a lag in I K1 block by polyamines which is much slower at positive potentials (Ishihara et al. 1989). In VM cells, with steeper ramps the failure of the ramp peak current to appreciably increase suggests little change in I to whereas the increase in current between ramp peak and I h indicates a role of I K1 .
On repolarization, with respect to VM cells, in P cells the greater fall in current is bound to reflect the much greater I to activated on depolarization, as 4-AP reduced both I to and the decrease in current on repolarization. The finding that on repolarization the decrease in outward current reached a more negative peak in P cells (~À16 mV) than in VM cells (~+7 mV) suggests that on depolarization the less negative reincrease of the outward current in M cells does not entirely depend on being masked by a larger I Ca .
In P cell, I to induces the large phase 1 repolarization of AP, and keeps the plateau at more negative level (Kenyon and Gibbons 1979). In rabbit P cells, 4-AP slowed phase 1 repolarization and shifted the plateau to positive potentials, while having in VM cells a much smaller effect on phase 1 repolarization and no effect on the plateau (Cordeiro et al. 1998).
In P cells, the persistence of the bulge current during the slow ramps and its elimination by 4-AP (which only reduced the sustained current) suggest that in P cells the bulge current is related to the activation of I to whereas the 4-AP resistant increase in outward current toward the ramp peak might reflect a voltage-dependent increase in the sustained current. Another difference is that in VM cells I to markedly decrease with lower V h whereas in P cells a smaller I to persisted even with V h À40 (Table 11), consistent with the persistence of the bulge current with V h À50 mV (Du and Vassalle 1999).
The greater I to in canine P cells has been also found rabbit P cells versus VM cells (Cordeiro et al. 1998). However, in P cells the larger I to and the smaller I K1 contributions to the total outward current at the plateau may be important in different situations. For example, an increase rate of discharge (tachycardia) may modify I to (due to its slow recovery; e.g., Cordeiro et al. 1998) more than I K1 .
In agreement with the larger I to in P cells, Kv3.4 is more abundant at both the mRNA and protein level in Purkinje fibers than in ventricular myocardium. As the Kv3.4 subunit carries a TEA-sensitive I to outward current, Kv3.4 current may be responsible for the large TEA-sensitive component of I to in canine Purkinje cells (see Schram et al. 2002 for review). Among the differences between I to in P and in VM cells is the different sensitivity to various blockers (e.g., TEA) and a smaller time constant of inactivation in VM cells (Han et al. 2000).
Thus, with respect to the P cells, in VM cells I to : (1) became apparent at more positive potentials during the ramps; (2) did not undergo an enhancement (the "bulge") during depolarizing ramps; (3) was smaller as measured from I to peak either to the end of the step or with respect to I h ; (4) was abolished by 4-AP with a reduced sustained current (as in P cells); (5) had a smaller time constant of inactivation (in agreement with the finding of Han et al. 2000); (6) was associated with smaller slope conductance; (7) during repolarizing ramps, was smaller and had a smaller voltage range; and (8) was inactivated by lower V h .

General conclusions
The present results indicate substantial differences between P and VM cells that bear on the different functions of P (conduction and pacemaker activity) and VM cells (contraction). In both tissues, I K1 is important for a normal resting potential and its inward rectification is important in more than one way. In P cells, the smaller I K1 conductance at the resting potential is essential for pacemaker function in two respects: (1) it keeps the resting potential less negative than E K and therefore it allows the undershoot to the maximum diastolic potential by the pacemaker current I Kdd ; and (2) I K1 inward rectification by polyamines contributes to diastolic depolarization caused by time-dependent decay of I Kdd and therefore to the attainment of threshold for I Na3 . I Na3 initiates I NS and the associated depolarization allows the attainment of I Na1 threshold. Especially in VM cells, I K1 inward rectification caused by Mg 2+ block in the voltage range À40 to 0 mV facilitates the attainment of threshold for Na + currents and contributes (see Shimoni et al. 1992) to maintain the plateau. The large and fast I Na1 cuts short the no longer needed I Na3 and (being large) rapidly depolarizes the membrane and causes the overshoot to positive potentials (fast and large depolarization for rapid conduction). The depolarization induced by I Na1 allows the attainment of the thresholds for other currents in an orderly fashion. I Na2 (by inactivating slowly in P cells) contributes to their long plateau, thereby preventing the reentry of excitation from myocardium. Once the thresholds for I to and I Ca have been reached, at the end of the upstroke I Na1 is rapidly inactivated (thereby preventing a useless Na + influx).
The fast activation of I to eliminates the overshoot which is no longer needed for fast conduction. In the meanwhile, I Ca initiates the events leading to contraction and its inactivation balances the inactivating I to at the plateau over a background of I K1 inward rectification. In VM cells, I Ca shifted the outward current reincrease by 35 mV and therefore would contribute to the more positive plateau. In P cells, the inwardly rectifying I Kdd becomes activated and whatever slowly decaying I Na2 is not inactivated by the end of the plateau becomes deactivated as a function of voltage (Bocchi and Vassalle 2008). During the repolarization, the removal of Mg 2+ block of the K 1 channels leads to I PS and speeds up phase 3 repolarization, which, in turn, removes the polyamines block. In P cells, I Kdd begins to increase at about I PS peak (~À50 mV, Vassalle et al. 1995) leading to undershoot of resting potential, followed diastolic depolarization.
Among the differences between the two tissues, in P cells the more negative thresholds for the Na + currents, the larger I Na1 and the larger overshoot contribute to earlier activation and faster conduction. The larger I to contributes to less negative plateau and the slowly inactivating I Na2 to the longer AP. The larger sodium influx may involve a higher Na + -K + pump activity, which is the major mechanisms underlying overdrive suppression (Vassalle 1977). Thus, the outward current created by Na + -K + pump activity maintains the diastolic depolarization negative to I Na3 threshold, so that P cell spontaneous discharge is suppressed when not needed during sinoatrial rhythm.
In VM cells, the larger resting conductance contributes to set the resting potential near the K + equilibrium potential, as a resting potential lower than E K is needed only for pacemaker activity (see Vassalle 2007Vassalle , 2013. In VM cells, the smaller I Na1 is consistent with the smaller rate of rise of the smaller upstroke (e.g., Bal ati et al. 1998), as the conduction path is shorter (from Purkinje network to epicardium). In VM cells, the larger I Ca contributes to stronger contraction (Lin and Vassalle 1978); the smaller I to together with the positive range of the larger I Ca contributes to the more positive plateau. The lack of slowly inactivating I Na2 contributes to a shorter AP (which in turn regulates the duration of twitch and of diastole) and the greater I K1 and removal of its block during the larger I PS contributes to an earlier and faster phase 3 repolarization.
The definition of these differences is a precondition also for the understanding of deranged function under pathological conditions.