Different intracellular polyamine concentrations underlie the difference in the inward rectifier K⁺ currents in atria and ventricles of the guinea-pig heart

Abstract

The outward component of the strong inward rectifier potassium current, I_K1, is significantly larger in ventricles than in atria of the heart, resulting in faster repolarization at the final phase of the action potential in ventricles. However, the underlying mechanism of the difference in I_K1 remains poorly understood. I_K1 channels are composed of subunits from the Kir2 subfamily, and I_K1 amplitude is determined by the voltage-dependent blockade of the channel by the intracellular polyamines spermine and spermidine, and by Mg²⁺. Using a perforated patch-clamp method, which minimizes changes in the intracellular polyamine and Mg²⁺ concentrations, we detected repolarization-induced outward I_K1 transients, which are caused by competition between Mg²⁺ and spermine to block the channel, in ventricular but not in atrial myocytes from guinea-pig heart. The contribution of the Kir2.3 subunit to the I_K1 channel was found to be minor in the guinea-pig heart, because the activation time course of the Kir2.3 currents was ∼10-fold slower than those of I_K1, and the marked external pH sensitivity of the Kir2.3 currents was not found in I_K1. Both the Kir2.1 and Kir2.2 currents recorded from inside-out patches exhibited outward transients similar to those of ventricular I_K1 in the presence of 5–10 μm spermine and 0.6–1.1 mm Mg²⁺, and their amplitudes were diminished by increasing the spermine or spermidine concentrations. The total and free polyamine concentrations in guinea-pig cardiac tissues were higher in atria than ventricles. These results strongly suggest that different intracellular polyamine concentrations are responsible for the difference in atrial and ventricular I_K1 of the guinea-pig heart.

The strong inward rectifier potassium current, I_K1, contributes to the action potential repolarization and determines the resting potential of both cardiac atria and ventricles. However, it is worth noting that the amplitude of the outward component of I_K1 in ventricular myocytes is significantly larger than in atrial myocytes, and their voltage dependences differ, resulting in a significantly different shape of the action potentials (Hume & Uehara, 1985; Giles & Imaizumi, 1988; Koumi et al. 1995).

Cardiac I_K1 channels are thought to be a mixture of homomeric or heteromeric tetramers composed of closely related subunits from the Kir2 subfamily, Kir2.1, 2.2 and 2.3 (Wible et al. 1995; Raab-Graham & Vandenberg, 1998; Zaritsky et al. 2001; Liu et al. 2001; Preisig-Muller et al. 2002; Zobel et al. 2003; Schram et al. 2003; Dhamoon et al. 2004). Because homomeric Kir2.x channels differ with respect to properties such as unit channel conductance (Perier et al. 1994; Takahashi et al. 1994) and the external Ba²⁺ sensitivity (Liu et al. 2001; Schram et al. 2003), differences in the subunits making up the I_K1 channels may account for the observed difference in the atrial and ventricular I_K1 (Wang et al. 1998). However, evidence showing a clear correlation between the differences in the distribution of Kir2.x subunits and differences in atrial and ventricular I_K1 remains lacking. For example, in the canine heart, the expression of the Kir2.3 subunit, which forms a channel with a smaller unit conductance than the Kir2.1 and Kir2.2 channels, is significantly greater in atria than in ventricles and vice versa for the Kir2.1 subunit, but these differences are insufficient to explain the more than 10-fold difference in the I_K1 densities in atria and ventricles (Melnyk et al. 2002). In the sheep heart, the I_K1 channel in the atrium appears to be composed mainly of the Kir2.3 channel and the difference in the atrial and ventricular I_K1 may be explained by the difference in the inward rectification of the Kir2.1 and Kir2.3 channels (Dhamoon et al. 2004). However, the Kir2.3 subunit is not expressed in the guinea-pig atrium (Dhamoon et al. 2004), and the mechanism causing the difference in the atrial and ventricular I_K1 in the guinea-pig heart is not clear.

On the other hand, outward I_K1 amplitude is determined by the intracellular polyamines spermine (SPM) and spermidine (SPD) and by Mg²⁺, which all block the I_K1 channel (Matsuda, 1991b; Nichols & Lopatin, 1997). Eukaryotic cells generally contain high concentrations of SPM and SPD, which play important roles in cell growth and differentiation, though the levels vary among cell types (Pegg & McCann, 1982; Watanabe et al. 1991). Thus, a difference in the concentrations of free polyamines in atrial and ventricular myocytes could also contribute to the difference in I_K1.

In the present study therefore we investigated the respective contributions of Kir2.x subunit composition and intracellular polyamine levels to the observed differences in atrial and ventricular I_K1 in guinea-pig heart.

Methods

All experiments using animals were performed according to the Saga Medical School Guidelines for Animal Experimentation.

Isolation of cardiac myocytes

Guinea-pigs (0.25–0.4 kg) were deeply anaesthetized with an overdose of sodium pentobarbitone (100–150 mg kg⁻¹i.p.), after which the hearts were excised and single myocytes were enzymatically isolated as previously described (Ishihara et al. 2002).

Cell culture and transfection

Mouse fibroblastic L cells and HEK 293T cells were grown as described elsewhere (Ishihara et al. 1996; Ishihara & Ehara, 2004). For whole-cell current recordings, guinea-pig Kir2.1, Kir2.2 or Kir2.3 cDNAs (Liu et al. 2001; kindly provided by Dr R. Preisig-Müller, Marburg University, Germany) were transiently transfected into L cells, which show minimal endogenous currents with a low-Ca²⁺ internal solution (Ishihara et al. 1996). For inside-out patch recordings, mouse Kir2.1 cDNA (Kubo et al. 1993; kindly provided by Dr L.Y. Jan, University of California, San Francisco) and mouse Kir2.2 cDNA (Takahashi et al. 1994; kindly provided by Dr Y. Kurachi, Osaka University, Japan), which we subcloned into pCXN2 (Niwa et al. 1991), were transiently transfected into HEK 293T cells where they were strongly expressed (Ishihara & Ehara, 2004). The Kir2 genes were always cotransfected with the plasmid DNA encoding the green fluorescent protein (pEGFP-N1, Clontech) as previously described (Ishihara & Ehara, 2004) to identify the cells expressing the exogenous genes by their green fluorescence.

Electrophysiology

The methods for recording currents under voltage-clamp condition from the cardiac myocytes using the amphotericin B perforated patch-clamp technique and from inside-out patch membranes from HEK 293T cells expressing Kir2.x subunits were as previously described in detail (Ishihara et al. 2002; Ishihara & Ehara, 2004). Currents were recorded at room temperature (24–26°C).

Solutions

The Tyrode solution contained (mm): NaCl 140, KCl 5.4, MgCl₂ 0.5, CaCl₂ 1.8, NaH₂PO₄ 0.33, glucose 5.5 and Hepes 5 (pH 7.4 with NaOH). The modified Tyrode solution used for recording I_K1 contained 2 μm nicardipine, 5 μm E4031 and 50 μm 293B (kindly provided by Dr H.J. Lang, Aventis Pharma, Germany) to block, respectively, the L-type Ca²⁺ current and the rapidly and slowly activating delayed rectifier K⁺ currents. The external K⁺-free solution was prepared by replacing KCl by equimolar NaCl in modified Tyrode solution. Titration of the external pH to 6.6 or 8.5 was accomplished using Pipes or Taps, respectively. The pipette (internal) solution used for the amphotericin B perforated-patch recordings contained (mm): potassium gluconate 115, KCl 25, NaCl 10, CaCl₂ 1 and Hepes 5 (pH 7.2 with KOH), and that used for conventional whole-cell recordings contained (mm): KCl 30, potassium aspartate 85, EDTA 4, K₂ATP 2, MgCl₂ 5.3, KH₂PO₄ 10 and Hepes 5 (pH 7.2 with KOH). For the inside-out patch experiments, the pipette (external) solution contained (mm): KCl 145, CaCl₂ 1 and Hepes 5 (pH 7.4 with KOH). The bath (cytoplasmic) solutions, which contained the indicated concentrations of SPM, SPD and Mg²⁺, were prepared by adding the appropriate amounts of stock solutions to Mg²⁺-free, polyamine-free cytoplasmic solution containing (mm): KCl 125, EDTA 4, K₂HPO₄ 7.2, KH₂PO₄ 2.8 (pH 7.2 with KOH). SPM and SPD stock solutions (10 mm each in distilled water) were prepared using spermine-4HCl and spermidine-3HCl (Nacalai Tesque, Japan). To prepare cytoplasmic solutions containing 0.6 or 1.1 mm free Mg²⁺, 1 m MgCl₂ solution (Kishida Chemical, Japan) was diluted with the Mg²⁺-free, polyamine-free solution to 5 or 6 mm, respectively (pH re-adjusted to 7.2). Free Mg²⁺ concentrations in these solutions were determined based on measurements of mag-indo-1 (Molecular Probes) fluorescence using a spectrofluorophotometer (RF-5000, Shimadzu, Japan; Yan & Ishihara, 2005).

Measurement of SPM and SPD content in cardiac tissues

After removing the hearts from deeply anaesthetized guinea-pigs, the atria and ventricles were dissected out quickly in ice-cold Tyrode solution and frozen in liquid nitrogen. Polyamine content was then measured essentially according to the method previously described (Igarashi et al. 1986). Briefly, tissues were treated with 5% trichloroacetic acid, and the polyamines in the supernatant were separated on high-performance liquid chromatography. The retention times for SPM and SPD were 15 and 27 min, respectively. The precipitate was used for measurement of protein content. In addition, the cellular components (i.e. DNA, RNA, ATP and phospholipid) were measured in each sample, and the concentration of phosphate in the cellular components and polyamines was calculated using an intracellular water space/protein ratio of 5.5 (μl cell volume) (mg protein)⁻¹ (Watanabe et al. 1991). The polyamine distribution (see Table 2) was calculated using the mean values of the concentrations of the cellular components in respective tissues (see Table 1) according to the equations described elsewhere (Watanabe et al. 1991).

Table 2.

Polyamine distribution in atrium, left ventricle and right ventricle in the guinea-pig heart (μm)

	Atrium		Left ventricle		Right ventricle
	SPM	SPD	SPM	SPD	SPM	SPD
Free	12.2 (2.7)	14.7 (8.4)	10.1 (3.2)	8.5 (9.7)	11.4 (3.1)	6.9 (9.5)
DNA	57.1 (12.5)	16.8 (9.6)	29.4 (9.2)	6.1 (7.0)	47.1 (12.7)	7.1 (9.8)
RNA	344 (75.1)	129 (73.9)	213 (66.9)	56 (64.1)	253 (68.2)	49 (67.3)
Phospholipid	39.7 (8.7)	10.9 (6.3)	46.7 (14.7)	8.9 (10.2)	48.4 (13.0)	6.6 (9.1)
ATP	5.2 (1.1)	3.1 (1.8)	19.3 (6.1)	8.2 (9.4)	11.3 (3.0)	3.4 (4.7)
Total	458.2	174.5	318.5	87.3	371.2	72.8

Polyamine distribution was calculated using total SPM and SPD concentration shown in the table and the binding constant of SPM and SPD to DNA, RNA, phospholipid and ATP under the ionic condition of 2 mm Mg²⁺ and 100 mm K⁺ (Watanabe et al. 1991). Percentage of distribution was shown in parenthesis.

Table 1.

Cellular components of cardiac tissues in the guinea-pig (mm)

	SPM	SPD	Phosphate in DNA	Phosphate in RNA	Phosphate in phospholipid	ATP
Atrium	0.458 ± 0.034	0.175 ± 0.010	14.20 ± 1.83	24.70 ± 1.51	34.20 ± 5.89	0.48 ± 0.04
Left ventricle	0.319 ± 0.027*	0.087 ± 0.012*	8.78 ± 0.73*	18.00 ± 1.07*	48.50 ± 11.5	2.16 ± 0.38*
Right ventricle	0.371 ± 0.013*	0.073 ± 0.017*	12.50 ± 1.59	19.20 ± 0.88*	44.50 ± 8.13	1.12 ± 0.13*

Data are given as mean ±s.d.; n = 5 in each group.

^*P < 0.001 versus atrium.

Statistics

Unless otherwise noted, data are expressed as means ± s.e.m. Error bars are not shown when they were smaller than the symbols. Statistical significance was determined using Student's t test.

Results

Quasi-steady-state I–V relationships of I_K1 in guinea-pig atria and ventricles

We measured the I–V relationships for I_K1 in atrial and ventricular myocytes using the amphotericin B perforated-patch method, which minimizes dilution of intracellular divalent and polyvalent cations with the pipette solution (Fig. 1). I_K1 was defined as the current that disappeared in the absence of external K⁺ (Ishihara et al. 1989). The holding potential (HP) was kept at near the predicted K⁺ equilibrium potential in modified Tyrode solution (E_K, −85.4 mV with 5.4 mm external K⁺ and 150 mm internal K⁺), and test pulses were applied after a prepulse to −43 mV to inactivate the voltage-dependent Na⁺ current and eliminate the time-dependent components of the outward I_K1 (Ishihara & Ehara, 1998). In both cell types, large inward currents flowed at voltages below E_K in modified Tyrode solution, which disappeared in the K⁺-free solution (Fig. 1A). Figure 1B shows the I–V relationships in the presence and absence of external K⁺. The densities of the inward currents that disappeared in the K⁺-free condition were ∼2-fold larger in ventricles than in atria. Figure 1C shows the difference between the outward current amplitudes in the presence and absence of external K⁺. The reversal potential of the difference currents (V_rev) was −80.6 ± 1.1 mV and −83.0 ± 1.0 mV in the atrial (n = 16) and ventricular (n = 10) myocytes, respectively. The I–V relationship for the outward I_K1 in ventricular myocytes exhibited a marked negative slope, whereas that in atrial myocytes did not, as has been shown for various species including the guinea-pig (Hume & Uehara, 1985; Giles & Imaizumi, 1988; Koumi et al. 1995). The peak amplitudes of the outward currents in ventricles (3.03 ± 0.35 A F⁻¹ at −63 mV) were ∼3-fold larger than those in atria (0.86 ± 0.07 A F⁻¹ at −63 mV). This was not only due to a smaller I_K1 density, but reflected a stronger inward rectification of I_K1 in atria than in ventricles (Fig. 1D).

Figure 1. Inward rectification of I_K1 in guinea-pig myocytes recorded using the amphotericin B perforated-patch method.

A, current records from an atrial (a) and a ventricular (b) myocyte in modified Tyrode (control) and external K⁺-free solutions. Test pulses were applied from −123 to +17 mV in 10 mV increments following a prepulse to −43 mV; HP was −83 mV B, I–V relationships in the presence (•) and absence (○) of external K⁺ in atrial (a) and ventricular (b) myocytes. Currents were measured ∼200 ms after the pulse onset. Membrane potentials were compensated for the series resistance error by calculation, and current amplitudes at the test potentials were obtained by interpolation (Ishihara et al. 2002). Cell capacitances were 60.0 ± 5.5 pS for atrial (n = 14) and 153.0 ± 6.2 pS for ventricular (n = 10) myocytes. C, I–V relationships of the external K⁺-sensitive components (I_K1) in atrial (▪) and ventricular (□) myocytes. D, I–V relationships of atrial (▪) and ventricular (□) I_K1 normalized to the inward current amplitudes at −123 mV (I/I₋₁₂₃).

Time dependence of outward I_K1 in atrium and ventricle

Outward I_K1 in guinea-pig ventricular myocytes exhibits a marked time dependence due to competition between intracellular Mg²⁺ and SPM to block the I_K1 channel (Ishihara & Ehara, 1998; Ishihara et al. 2002). Using the perforated-patch technique we examined whether the outward I_K1 in atrial myocytes exhibits a similar time dependence. When the membrane was repolarized to voltages more positive than V_rev following a large depolarizing pulse (+46 mV) applied after a short hyperpolarizing prepulse to −133 mV, the outward currents usually showed a clear transient in ventricular myocytes (30/34 cells), but not in atrial myocytes (Fig. 2A). The outward transients observed in ventricular myocytes disappeared when the depolarizing pulse was applied from a voltage ∼40 mV more positive than V_rev (−44 mV; Fig. 2B). This is because the I_K1 channels are already blocked by SPM at the depolarized HP, which reduces the block by Mg²⁺ of the channel during the depolarizing pulse (Ishihara & Ehara, 1998). The difference between the currents recorded using a hyperpolarizing prepulse (−133 mV) and a depolarizing one (−44 mV) showed even more clearly that the transient components of the outward I_K1 were not present or were negligibly small in atrial myocytes (37 cells; Fig. 2C).

Figure 2. Outward currents during repolarizing pulses in a ventricular (left column) and an atrial (right column) myocyte in amphotericin B perforated-patch recordings.

Repolarizing pulses to −34, −44 and −54 mV were applied following a large depolarizing pulse (+46 mV, 200 ms), which was applied after a 20-ms hyperpolarizing prepulse (−133 mV) (A) or a depolarizing one (−44 mV) (B). Currents shown in C are the difference currents obtained by subtracting the currents in B from those in A.

Contribution of Kir2.3 to I_K1

Guinea-pig cardiac myocytes express the Kir2.1, Kir2.2 and Kir2.3 subunits (Liu et al. 2001), and the difference in I_K1 between atrial and ventricular myocytes may be caused by the difference in the subunits comprising the I_K1 channels. In the experiments summarized in Fig. 3, we transfected Kir2.1, Kir2.2 or Kir2.3 into L cells and compared the resultant whole-cell currents with I_K1. The activation time courses of the inward Kir2.1 and Kir2.2 currents were similar to those of the atrial and ventricular I_K1, whereas that of the Kir2.3 current was markedly slower (Fig. 3A). Figure 3B shows the time constants of the exponentials fitted to the time-dependent phase of the inward currents; when two exponentials were required for the fitting (Ishihara et al. 1996), the time constant of the slower component was plotted. The time constants of the Kir2.3 were ∼10-fold greater than those of the Kir2.1, Kir2.2 or I_K1.

Figure 3. Activation time course and external pH sensitivity of Kir2 currents and I_K1.

A, representative whole-cell current traces at pH 7.4 (left column), 6.6 (middle column) and 8.5 (right column). Test pulses were from −133 to −73 mV in 10 mV increments; HP was −43 mV. B, comparison of the time constants of the activation phase of the inward currents at pH 7.4 for Kir2.1 (•, n = 6), Kir2.2 (▴, n = 3), Kir2.3 (○, n = 4), atrial I_K1 (▪, n = 10) and ventricular I_K1 (□, n = 10). C, effect of external pH. Current amplitudes measured at −121 mV and ∼200 ms for pH 6.6 and 8.5 were normalized to that obtained at pH 7.4 (I/I_7.4) for Kir2.1 (a, n = 3), Kir2.2 (b, n = 3), Kir2.3 (c, n = 5), atrial I_K1 (d, n = 7) and ventricular I_K1 (e, n = 6). *P < 0.01 versus pH 7.4. D, effects of co-expression of Kir2.1 and Kir2.3 on the external pH sensitivity of the currents. Kir2.1 and Kir2.3 cDNAs were cotransfected at the indicated molar ratios. Shown are the I/I_7.4 values for pH 6.6 (•) and pH 8.5 (○); n = 3–9.

We also examined the sensitivity of the currents to the external pH, which is characteristic of Kir2.3 channels (Coulter et al. 1995). The amplitudes of Kir2.3 currents were reduced by ∼53% at acidic pH (6.8) and increased by ∼66% at alkaline pH (8.5), whereas Kir2.1 and Kir2.2 currents and both the atrial and ventricular I_K1 were not noticeably affected (Fig. 3A and C). Moreover, when Kir2.1 and Kir2.3 cDNAs were cotransfected at various molar ratios, the magnitude of the effect caused by acidic/alkaline pH varied with the amount of transfected Kir2.3 cDNA (Fig. 3D), suggesting that the pH sensitivity of the channel increases as the contribution of the Kir2.3 subunit to the heteromeric Kir2.1/Kir2.3 channels increases. Taken together, these findings indicate that Kir2.3 makes little contribution to I_K1 in either atrial or ventricular myocytes of the guinea-pig heart.

Repolarization-induced outward Kir2.1 and Kir2.2 transients

In the following experiments, we examined the outward transients of the Kir2.1 and Kir2.2 channels using inside-out patches. The currents were recorded at an external [K⁺] of ∼150 mm (E_K=∼0 mV) to increase their amplitudes. An increase in the external [K⁺] shifts both the rectification and the transient outward components of I_K1 to positive voltages (Ishihara & Ehara, 1998) because the voltage dependence of the blockades of the strong inward rectifier K⁺ channels caused by internal Mg²⁺ and polyamines apparently depends on the driving force of K⁺ (V−E_K) when external [K⁺] is altered (Matsuda, 1991a; Lopatin & Nichols, 1996). Thus, the voltages during the repolarizing pulse protocol used in Figs 4 and 5 were shifted, compared to those used in Fig. 2, by amounts that roughly equal the shift of E_K. When 0.6 or 1.1 mm Mg²⁺ was present together with 5 μm SPM in the cytoplasmic solution, outward transients similar to those of I_K1 were observed with both Kir2.1 and Kir2.2, which were not observed with SPM alone (Fig. 4A). Figure 4B shows the relationships between the repolarized voltages and the amplitudes of the outward Kir2.1 and Kir2.2 transients obtained with 1.1 mm Mg²⁺ plus 5 μm SPM.

Figure 4. Outward transients of Kir2.1 and Kir2.2 currents in the presence of cytoplasmic SPM and Mg²⁺.

A, outward Kir2.1 (a) and Kir2.2 (b) currents during repolarizing pulses obtained from inside-out patches in the presence of 5 μm SPM (left column), 5 μm SPM + 0.6 mm Mg²⁺ (middle column) or 5 μm SPM + 1.1 mm Mg²⁺ (right column). Currents were recorded at symmetrical [K⁺] of ∼150 mm(E_K=∼0 mV). Repolarizing pulses were applied after 100-ms depolarizing pulses to +80 mV, which followed short hyperpolarizing prepulses to −40 mV. Outward currents during repolarizing steps to +50, +40 and +30 mV are shown for each cytoplasmic condition. Currents recorded under differing cytoplasmic conditions were obtained from the same patch. B and C, amplitudes (B) and time constants of decay (C) of the repolarization-induced outward of Kir2.1 (•, n = 10) and Kir2.2 (○, n = 9) transients obtained in the presence of 5 μm SPM + 1.1 mm Mg²⁺. Current amplitudes were normalized to that at −40 mV under the same cytoplasmic conditions. Decay phases were fitted using a single exponential function (Ishihara & Ehara, 1998; Ishihara et al. 2002).

Figure 5. Effects of cytoplasmic polyamine concentration on the outward Kir2 transients.

A, attenuation of outward Kir2.1 transients elicited by increases in the concentration of SPM present with 1.1 mm Mg²⁺ (a) or the concentration of SPD present with 1.1 mm Mg²⁺+ 5 μm SPM (b). Outward currents during repolarizing pulses to +40 and +30 mV are shown for each cytoplasmic condition. Inward currents at −40 mV are also shown to confirm that the reduction in transient amplitude was not due to current rundown. B, amplitudes of outward Kir2.1 (left column) and Kir2.2 (right column) transients under different cytoplasmic conditions: a, 1.1 mm Mg²⁺+ 5 (•), 10 (○) or 20 (▵) μM SPM; b, 1.1 mm Mg²⁺+ 5 μm SPM (•) +5 (▿) or 30 (□) μm SPD; (n = 3–10). C, steady-state I–V relationships of the Kir2.1 channel in the presence of: 1.1 mm Mg²⁺+ 5 μm SPM (•); 1.1 mm Mg²⁺+ 5 μm SPM + 30 μm SPD (□); 1.1 mm Mg²⁺+ 20 μm SPM (▵). Current amplitudes were measured ∼200 ms after the onset of pulses applied from a HP of +40 mV. Data in A and C were obtained from the same patch. Current amplitudes in B and C were normalized to that at −40 mV obtained under each cytoplasmic condition.

Although the amplitudes of the transient components showed slightly different voltage dependences, they were not markedly different under the same cytoplasmic conditions. For both channels, the rate of decay became faster as the repolarization voltage approached V_rev, which is characteristic of the time-dependence of the outward I_K1 induced by intracellular Mg²⁺(Ishihara & Ehara, 1998; Ishihara et al. 2002), and their time constants were similar (Fig. 4C). Thus, a difference in the contributions of Kir2.1 and Kir2.2 to the I_K1 channel cannot explain the difference in the time-dependence of the outward I_K1 observed in atrial and ventricular myocytes.

On the other hand, the amplitudes of the outward transients declined as the concentration of SPM present with Mg²⁺ was increased, becoming negligibly small with 20 μm SPM (Fig. 5Aa). Addition of SPD to a solution containing Mg²⁺ and SPM also reduced transient amplitudes, which were significantly reduced by adding 5 μm SPD to the solution containing 1.1 mm Mg²⁺ plus 5 μm SPM, and virtually absent in the presence of 30 μm SPD (Fig. 5Ab). The attenuation of the transients was easily reversed by decreasing the SPM or SPD concentrations (data not shown). Similar findings were obtained for the Kir2.1 and Kir2.2 channels (Fig. 5B). Under conditions in which the outward transients became negligibly small, the steady-state amplitudes of the outward currents were reduced, and the negative slope of the I–V relationship was flattened (Fig. 5C).

Polyamine content of cardiac tissues

The above results suggest that the absence of the transient components and the stronger inward rectification of atrial I_K1 are caused by higher concentrations of intracellular polyamines. Consistent with this idea, we found that the concentrations of SPM and SPD were higher in atria than ventricles (Table 1). We also estimated the intracellular free polyamine concentrations (Table 2) by measuring the cellular contents of DNA, RNA, phospholipid and ATP, which all bind polyamines (Table 1). The proportions of free SPM and SPD were, respectively, ∼3% and ∼9% of the total in both atria and ventricles. The differences in the free SPM/SPD concentrations between atria and left/right ventricles were small compared to the total concentrations. However, the free SPM and SPD concentrations in atria (SPM, 12.2 ± 0.88 μm; SPD, 14.7 ± 0.82 μm), were still significantly higher than those in left ventricles (SPM, 10.1 ± 0.87 μm; SPD, 8.5 ± 1.04 μm; P < 0.001 for both) and right ventricles (SPM, 11.4 ± 0.43 μm; SPD, 6.9 ± 1.49 μm; P < 0.05 for SPM and P < 0.001 for SPD).

Discussion

We have shown that in guinea-pig the difference in the time and voltage dependences of atrial and ventricular I_K1 are not explained by differences in the respective contributions of Kir2.1, Kir2.2 and Kir2.3 to the I_K1 channel, but are explained by higher levels of polyamine concentrations in the atrial tissues.

We compared whole-cell I_K1 recorded in atrial and ventricular myocytes using the perforated-patch method, which has an advantage over the conventional method in that the original concentrations of intracellular free polyamines and Mg²⁺, which function as I_K1 channel blockers, are maintained (Ishihara et al. 2002). The ∼2-fold difference in the densities of the inward currents of the atrial and ventricular I_K1 in the guinea-pig heart (Fig. 1) was smaller than the ∼10-fold differences reported for the human and the canine hearts (Koumi et al. 1995; Melnyk et al. 2002). The outward I_K1 amplitude was significantly smaller in atrial myocytes, which reflected not only the smaller I_K1 density but also stronger inward rectification of I_K1 in atria (Fig. 1). We found that the outward I_K1 transients elicited on repolarization in ventricular myocytes were not observed in atrial myocytes using the perforated patch-clamp method (Fig. 2). Because the transient component of I_K1 may flow during the repolarization phase of the action potentials (Ishihara et al. 2002), the difference in the time dependence of the outward I_K1 may contribute to the difference in the atrial and ventricular action potentials. It is noteworthy that the outward I_K1, which was isolated as the external K⁺-sensitive component, was slightly larger in atria than in ventricles at depolarized voltages (>−20 mV; Fig. 1C). This observation was apparently different from the complete rectification of the atrial I_K1 previously found in the guinea-pig (Dhamoon et al. 2004) and in other species (Wang et al. 1998; Melnyk et al. 2002; cf. Koumi et al. 1995). The difference might be due to the difference in the methods utilized to isolate I_K1; as the blockade of I_K1 by external Ba²⁺ is voltage dependent (Standen & Stanfield, 1978; Imoto et al. 1987), isolation of outward I_K1 at the depolarized voltages using 0.4–1 mm Ba²⁺ might have been incomplete in the previous studies. Alternatively, the external K⁺-sensitive components in atria we showed may contain the basal activity of muscarinic K⁺ channels (Soejima & Noma, 1984; Kaibara et al. 1991). If so, the inward rectification of the atrial I_K1 may be stronger than is shown in Fig. 1D, as the inward rectification of both the basal activity and the Ach-induced components of the muscarinic K⁺ channels is much weaker than that of I_K1 (Soejima & Noma, 1984). Discovery of a specific blocker of the I_K1 channel will solve this issue.

Outward transients are observed in whole-cell I_K1 in ventricles in the presence of physiological concentrations of intracellular free Mg²⁺ (0.5–1.0 mm; Murphy, 2000), but not in the absence of Mg²⁺ (Ishihara & Ehara, 1998). In the presence of Mg²⁺, the relative contribution of the blockade made by endogenous SPM is reduced during depolarizing voltage steps due to the blockade by Mg²⁺, and the relief of the Mg²⁺-induced blockade during the following repolarizing voltage steps and the re-block of the channel by SPM generates the outward transients (Ishihara & Ehara, 1998). We recently demonstrated that outward currents of the Kir2.1 channel recorded from inside-out patches in the presence of 5 or 10 μm cytoplasmic SPM exhibit I–V relationships similar to those of the ventricular I_K1 (Ishihara & Ehara, 2004). In the present study, we showed the presence of transient components similar to those of the ventricular I_K1 in outward Kir2.1 and Kir2.2 currents with 0.6 or 1.1 mm Mg²⁺+ 5 or 10 μm SPM (Figs 4 and 5). The amplitudes of the Kir2 transients obtained with 0.6 or 1.1 mm Mg²⁺+ 5 or 10 μm SPM (+5 μm SPD), normalized to the inward current amplitudes at 40 mV more negative than V_rev (Fig. 5B) were close to those of the I_K1 transients (data not shown). Thus, the homomeric Kir2.1 channel can reconstitute the properties of whole-cell I_K1 in guinea-pig ventricles. We have shown that transient components were present in outward Kir2.1 and Kir2.2 currents under the same cytoplasmic conditions. This finding indicated that the difference in polyamine/Mg²⁺ sensitivity of the Kir2.1 and Kir2.2 subunits does not explain the difference in the time dependence of outward atrial and ventricular I_K1. For the guinea-pig cardiac myocytes, there is a controversy about the expression of the Kir2.2 subunit (Liu et al. 2001; Dhamoon et al. 2004). Even if the Kir2.2 subunit contributes to I_K1 channel by forming heteromeric channels with the Kir2.1 subunit (Zaritsky et al. 2001; Preisig-Muller et al. 2002; Schram et al. 2002; Zobel et al. 2003), our results indicate that it is unlikely that the contribution of the Kir2.2 subunit makes the difference in the atrial and ventricular I_K1 of the guinea-pig heart.

For I_K1, it has been difficult to examine the effects of cytoplasmic polyamines and Mg²⁺ on the macroscopic currents using the inside-out patches. In this study, we have shown that the amplitudes of the outward Kir2 transients were larger at the higher Mg²⁺ concentration (Fig. 4) and smaller with higher polyamine concentrations (Fig. 5). This finding supported our view that the outward transients of the ventricular I_K1 are caused by the contribution of the Mg²⁺-dependent blockade of the channel. In our recent study, we proposed that the homomeric Kir2.1 channel exists in two different forms, which show distinct sensitivities to polyamines (Ishihara & Ehara, 2004). The transient outward components are well explained by the competition of Mg²⁺ and SPM for blocking the high-affinity channels (Yan & Ishihara, 2005).

It has been shown that the Kir2.1 and Kir2.3 proteins are expressed in the ventricle but only the Kir2.1 protein is expressed in the atrium of the guinea-pig heart (Dhamoon et al. 2004), and that the Kir2.1 and Kir2.3 subunits could be co-immunoprecipitated from the membranes of the guinea-pig cardiac myocytes (Preisig-Muller et al. 2002). We used Kir2.3's sensitivity to external pH (Coulter et al. 1995) to examine whether the Kir2.3 subunit plays a substantial role in generating the difference in the atrial and ventricular I_K1. We found that neither the atrial nor the ventricular I_K1 was affected by changing the external pH between 6.6 and 8.5, whereas Kir2.3 currents were (Fig. 3). We also verified these findings by varying external pH using only Hepes, as different buffering compounds may affect Kir2 channels differently (Guo & Lu, 2002). This observation indicates that the homomeric Kir2.3 channel is not a major component of the I_K1 channels in the guinea-pig cardiac myocytes. When Kir2.1 and Kir2.3 subunits were co-expressed, the magnitude of the effects caused by acidic/alkaline pH varied with the amount of transfected Kir2.3 cDNA (Fig. 3D). This finding indicates that the pH sensitivity of the channel increases as the contribution of the Kir2.3 subunit to the heteromeric Kir2.1/Kir2.3 channels increases, if the Kir2.1 and Kir2.3 subunits were co-assembled to form channels in our experiment as has been previously shown using a similar expression system (Preisig-Muller et al. 2002). The examination of the time course of the slowest component of the inward current activation, which reflects the relief of the block by SPM of the Kir2.3 and Kir2.1 channels (Lopatin et al. 1995; Ishihara et al. 1996), also revealed a marked difference between the Kir2.3 channel and the I_K1 channels (Fig. 3A and B). We thus concluded that the contribution of the Kir2.3 subunit to the cardiac I_K1 channels is minor in the guinea-pig, and that Kir2.3 does not play a major role in causing the difference between atrial and ventricular I_K1. We did not show whether the Kir2.3 channel can elicit the outward transients using the inside-out patches, because currents were too small to evaluate. However, whole-cell Kir2.3 currents showed repolarization-induced outward transients like whole-cell Kir2.1 and Kir2.2 currents (data not shown).

We demonstrated using the Kir2.1 channel that increasing the cytoplasmic SPM concentration from 5 to 20 μm changed the time dependence and the steady-state I–V relationship of the outward currents from a ‘ventricular type’ to an ‘atrial type’ (Fig. 5). Although addition of 30 μm SPD to a solution containing 5 μm SPM almost eliminated the outward transients, the effect on the steady-state outward current amplitudes was not large enough to mimic the small amplitudes of outward atrial I_K1 (Fig. 5C). These findings suggested that the SPM concentration is higher, and the SPD concentration may be also higher in atrial myocytes. We recently showed that the presence of 10 μm SPM or 30 μm SPD in the cytoplasmic solution reduces the amplitudes of inward Kir2.1 currents by about 5% and 15%, respectively (Ishihara & Ehara, 2004). Although a larger expression of the Kir2 subunits may be a major factor to cause the difference in the I_K1 density (Melnyk et al. 2002; Dhamoon et al. 2004), higher concentrations of cytoplasmic polyamines also explain in part the difference in the density of the inward atrial and ventricular I_K1.

We measured the polyamine contents in atrial, left ventricular and right ventricular tissues of the guinea-pig hearts to examine whether there is evidence supporting our hypothesis (Table 1). The total SPM and SPD concentrations were obtained using the ratio of the intracellular water space to the cellular protein content (5.5 μl (mg protein)⁻¹) measured in lymphocytes (Watanabe et al. 1991). In cardiac cells, a ratio of 4.6 μl (mg protein)⁻¹ is obtained from the cellular water content of ∼3.25 ml (g dry mass)⁻¹ (Aliev et al. 2002) and the protein content that is reportedly ∼70% of the dry mass content (Bünger & Soboll, 1986). This difference in ratio is probably due to the large amount of contractile protein in cardiac myocytes. If so, the total SPM and SPD concentrations may be underestimated slightly. The total concentrations of both SPM and SPD were indeed higher in atrial than ventricular tissues. It was also found that the concentration of SPD was considerably lower than that of SPM in all parts of the cardiac tissues tested. Although this low SPD content differs from the observations in other cell types (Watanabe et al. 1991), it is consistent with our recent functional analysis of the Kir2.1 channel (Ishihara & Ehara, 2004) and with similar results obtained in mouse heart (Mackintosh et al. 2000). The level of the immediate precursor of SPM and SPD, putrescine, was only ∼2% of the SPM in all of the cardiac tissues examined (data not shown), which is a general finding in eukaryotic cells, and suggests that putrescine does not contribute to I_K1 under physiological conditions.

As intracellular polyamines are mostly bound to RNA, DNA, phospholipid and ATP, we also measured the contents of those molecules in the cardiac tissues and estimated the free SPM and SPD concentrations by calculation (Table 2). The proportions of free SPM (∼3%) were smaller than those of free SPD (∼9%), reflecting greater binding constants of SPM to the macromolecules and ATP than those of SPD (Watanabe et al. 1991). Although our calculations assume a homogeneous distribution of the polyamines in the intracellular milieu, which may not the case (Pegg, 1988), the values of the free SPM and SPD concentrations fell in a range of cytoplasmic SPM and SPD concentrations we used to reproduce the properties of the I_K1 in the guinea-pig atrium and ventricle using the Kir2.1 and Kir2.2 channels (Figs 4 and 5). Furthermore, the concentrations of free SPM and, especially, free SPD were calculated to be higher in atrial than in ventricular tissues, which gave support to our hypothesis that the difference in the intracellular polyamine concentrations contributes to the difference in I_K1 between the atrium and ventricle. The total ATP concentrations (Table 1) were somewhat lower than those usually reported (Matsuoka et al. 2004), which could be due to our dissection procedure, but because the fraction bound to ATP is relatively small, the low ATP concentration did not significantly affect the calculated free polyamine concentrations.

Our study strongly suggests that the difference in the concentration of free polyamines in atrial and ventricular myocytes underlies the difference in atrial and ventricular I_K1. However we should like to emphasize that we do not exclude the possibility that a difference in the I_K1 channel subtypes also contributes to that difference, as the distribution of the Kir2.x subunits in cardiac tissues appears to differ among species. Moreover, many proteins associate with Kir2 channels (Leonoudakis et al. 2004), and thus a difference in the distribution of those associating proteins may also play some roles in causing the difference in the atrial and ventricular I_K1.

Finally, it has been shown that the density of I_K1 is significantly reduced in the failing heart and in dilated cardiomyopathy, which predisposes the failing heart to ventricular arrhythmia (Beuckelmann et al. 1993; Kääb et al. 1996). Because levels of Kir2 mRNA are generally not altered in failing hearts, our study suggests that alteration of the polyamine content of ventricular cells to an ‘atrial type’ may accompany cardiac remodelling. Thus, investigations into the regulation of the biosynthesis, degradation and transport (Pegg, 1988; Fukuchi et al. 2004) of cardiac polyamines and their effects on I_K1 will be important future studies.